Inspired By Reality

These are some spacecraft designs that are based on reality. So they appear quite outlandish and undramatic looking. In the next page will appear designs that are fictional, but much more breathtaking. Obviously the spacecraft on this page are all NASA style exploration vehicles, they are not very suited for interplanetary combat (well, most of them at least).

For slower-than-light star ships, go here.

Many of these spacecraft have a table of parameters. You can find the meaning of many of them here. Numbers in black are from the documents. Numbers in yellow have been calculated by me using the document numbers, these might be incorrect.

Gary Johnson Mars Mission


I have gotten involved with some friends on the New Mars forums discussing what might be appropriate for very large colonization ships.  This kind of mission demands the delivery of very large payloads.  Doing this effectively requires a reusable ship.  That means you stage off (or jettison) nothing.

It is easy to run a rocket equation-based trade study that assumes a one-stage round trip, that jettisons nothing. Making it carry the same large payload on the return voyage simplifies the analysis, but very likely over-penalizes the design. But at this level of analysis, that really doesn’t matter.

This is basically just a bounding analysis for screening candidate propulsion approaches to a Mars colony ship design. I included nuclear explosion propulsion, nuclear thermal propulsion, ion propulsion, LOX-LH2 cryogenic chemical propulsion, and storable chemical propulsion.
Update 9-13-19:  there is more than one kind of nuclear thermal rocket.  I took a closer look at 6 different nucear thermal rocket approaches,  and in a more nuanced way,  in "A Closer Look At Nuclear Thermal"

Spreadsheet Inputs

The spreadsheet inputs are highlighted yellow. Payload delivered is common to all the designs, and actually arbitrary, but I thought 2000 metric tons might go a long way toward the beginning of a colony.

Inert fractions vary with the propulsion selection. I used data from Ref. 1 to set a realistic guess for the inert fraction, of the nuclear explosion drive. It is very high, reflecting the massive pusher plate, two-stage shock absorption system, and the armored hull.

The Hall effect ion drive is based on existing Busek satellite thrusters already in service, and modified to “burn” iodine, something plentiful, cheap, and storable at low pressure. Getting to an acceptable vehicle acceleration requires a very large thruster array and a nuclear power source in the multi-megawatt range. I just guessed the inert mass fraction that might cover this.

Because of the heavy reactor core and low engine thrust/weight achieved in the old NERVA nuclear thermal rocket development effort, I used twice the typical chemical stage inert fraction as a “good guess” for the nuclear thermal inert mass fraction. There is good data about this engine in Ref. 2.

Both the LOX-LH2 cryogenics chemical propulsion, and the NTO-MMH storable-propellant chemical propulsion, share the same “typical” stage inert mass fraction.

Delta-vees for the Mars trip are for departing and arriving in low Earth orbit to/from a min-energy Hohmann transfer ellipse, plus the corresponding delta-vees for arriving into and departing from low Mars orbit. The same applies to the Ceres transfer, except that the ship just matches Ceres orbital velocity about the sun instead of entering a “low orbit”. This would be typical of many small main belt asteroids.

For those types of propulsion in the order listed above (nuclear explosion, nuclear thermal, ion drive, LOX-LH2 chemical, and storable chemical), my assumed inputs for Isp were 10,000 sec, 1000 sec, 3000 sec, 470 sec, and 330 sec respectively. Vehicle inert mass fractions were 0.50, 0.25, 0.10, 0.05, and 0.05 respectively.

All these dV’s were summed, as required to do the entire mission single-stage. The total orbital delta-vee (dV) to and from Mars is 3.84+1.83+1.83+3.84 = 11.34 km/s. Impulsive-burn options need supply only that summed delta-vee with zero gravity and drag losses. Long-burn ion must supply a lot more than that, due to very large planetary and solar gravity losses.

All but the ion option were considered as "impulsive burn" and Hohmann min energy transfer, with vehicle acceleration exceeding 0.1 gee to enforce that. These used the unfactored sum of orbital dV's to and from Mars (orbit-to-orbit transport) as the mass ratio-effective dV for the rocket equation. The spreadsheet input is factor equal to one.

The ion option must spiral-out and spiral-in at the planetary orbits, and accelerates to midpoint then decelerates to arrival on the transfer trajectory (a patched spiral about the sun). Propulsion is sized for 0.001 gee to ensure that this kind of transfer is feasible. To account for the planetary and solar gravity losses of the resulting months-of-burn, I just doubled the orbital dV sum to 22.68 km/s. For the spreadsheet, this is factor equal to two.

For Ceres, Earth departure and arrival dV is 5.24 km/s. The orbit-matching dV at Ceres (arrival and departure) is just about 3.49 km/s. That round trip sum is 17.46 km/s for all but the ion drive option, unchanged by factor equal to one. Using factor equal to two for ion drive, that mass ratio-effective total is 34.92 km/s.

All 5 designs carried exactly the same 2000 metric tons of dead-head payload, an arbitrary selection perhaps appropriate for a colony-type mission. (I did not look at how to get that payload up to LEO, or down from LMO, that issue would be the same for all the candidates.) This was done for Mars in a spreadsheet worksheet, whose image is Figure 1. All figures are at the end of this article.

Analysis Equations

Sum the round trip delta-vees, and factor the sum for the mass ratio-effective delta-vee required of each propulsion type: required dV = (factor)(sum of all 4 orbital delta vees), where factor = 1 for impulsive propulsion (acceleration exceeding 0.10 gee), and factor = 2 for long-burn ion propulsion (0.001 gee required).

Estimate the effective exhaust velocity from the specific impulse: Vex, km/s = 9.8067 (Isp, s)/1000

Calculate the mass ratio required: MR = exp(dV/Vex), with both velocities in km/s

Calculate the propellant mass fraction: Wp/Wig = 1 – 1/MR

Input an inert mass fraction Win/Wig (must be justified in some way as “realistic”)

Calculate the available payload fraction Wpay/Wig = 1 – Win/Wig – Wp/Wig (must be positive to be even theoretically feasible)

Input the delivered dead-head payload Wpay, metric tons (arbitrary, but should be realistic)

Calculate the ignition mass Wig, metric tons: Wig = Wpay/(Wpay/Wg)

Calculate the inert mass Win, metric tons: Win = Wig*(Win/Wig)

Calculate the propellant mass Wp, metric tons: Wp = Wig*(Wp/Wig)

Calculate the ignition to payload mass ratio: Wig/Wpay = (Wig, m.ton)/(Wpay, m.ton)

Results Obtained

Results for Mars: nuclear explosion drive 5118 metric tons at ignition with ignition/payload 2.56:1 (see Figure 2). Nuclear thermal 30,945 metric tons at ignition with 15.47:1 ignition/payload (see Figure 3). Hall effect ion drive 5516 metric tons at ignition with ignition/payload 2.76 (see Figure 4). LOX-LH2 56,486 metric tons at ignition with ignition/payload 28.24 (see Figure 5). Storable chemical utterly infeasible with a negative payload fraction available (see Figure 6).

The nuclear explosion drive offers the lowest ignition/payload ratio going to Mars at 2.56:1, based on the old 1950's shaped-charge fission device technology. This would be a very tough ship design, probably usable for a century or more, and likely tough enough to aerobrake, reducing the load of bombs in favor of more payload. Its stout hull and huge pusher plate are effective radiation shields.

The ion propulsion offers the next best ignition/payload ratio going to Mars at a very comparable 2.76:1, which to be practical would require its thrusters operating on something cheap, plentiful, and storable-as-a-condensed-phase (at very low pressure), like iodine. This would be a relatively gossamer structure unable to survive aerobraking, and it would likely also have a limited service life. Radiation protection would have to be added.

Two of the others (nuclear thermal and LOX-LH2), while theoretically feasible, are nowhere close in ignition/payload ratio going to Mars. These are unaffordable “Battlestar Galacticas” for any reasonable payload delivery aimed at colonization. And the storable chemicals are just infeasible in any sense of the word for a Mars colonization ship, simply because there is a negative payload fraction available, once propellant fraction has been determined, and with a suitable inert fraction input. It simply cannot do the mission single stage.

I think you can look at the ignition/payload mass ratio to judge whether-or-not a given propulsion system might serve as a practical way to build a colony ship. This value needs to be no more than about 5 or thereabouts, in order not to build an unaffordable “Battlestar Galactica”. This is a “fuzzy” boundary, dependent upon how much you think you can afford.

The same sort of analysis applies to other destinations. You just need an appropriate list of orbit-to-orbit delta-vees, and the same list of realistic guesses for inert fractions.

Results for Ceres: I added a worksheet to the same spreadsheet for a colony-type ship to Ceres, as “typical” of the asteroid belt. Those spreadsheet results are shown in Figure 7. Figures 26 also show the Ceres results (as well as the Mars results).

The only feasible choices for Ceres colony ships were nuclear explosion propulsion and nuclear-powered electric propulsion. It’s the same basic calculation, just with somewhat bigger delta-vees. The nuclear thermal and both chemical options simply had fundamentally-infeasible negative payload fractions available. They simply cannot perform the mission single-stage.

The same general outcome choices obtain for Ceres as for Mars: your nuclear explosion drive ship is quite robust, promising a long service life, while the ion ship is rather flimsy. For this main belt asteroid application, the ignition to payload ratio is also substantially more favorable for the nuclear explosion ship (2.97), vs the ion ship (4.87).


The trend here is clear: the further out you go with a single-stage, round-trip colony ship, the more the ignition/payload ratio is going to favor nuclear explosion propulsion as the more affordable option. Radiation protection needs will also favor the shielding effect of the stout hull required of the nuclear explosion drive. Bigger also favors ease of incorporating spin “gravity”.


#1. George Dyson, “Project Orion – The True Story of the Atomic Spaceship”, Henry Holt, 2002.

#2. David Buden, “Nuclear Thermal Propulsion Systems”, Polaris Books, 2011.


This article takes a closer, more nuanced look at nuclear thermal propulsion for large colonization ships. It still assumes fairly large dead-head payloads, but only carried on the outbound voyage! Propellant is sized to make the outbound and return voyages in one stage (no stage-off or jettisoning of anything along either way, just unload of the dead-head payload at destination). The journey baseline is low Earth orbit to low Mars orbit, and back.

How the ships or the payload get to low Earth orbit is unaddressed. How the payload gets delivered to Mars’s surface from low Mars orbit is unaddressed. How the ships are refueled and reloaded in low Earth orbit is unaddressed. What is addressed here, that is unlike the earlier simpler study, are the separate inert weights associated with the payload section, the propellant tankage section, and the engine-with-its-associated-subsystems. The minimum vehicle acceleration requirement is increased to 0.5 gee, except for one system deemed adequate at 0.33 gee.

The previous closely-related article was “Colonization Ship Study”, dated 9-9-19. It examined the simpler-to-analyze case of carrying the dead-head payload both ways (outbound and return), so that there was one mass ratio and one delta-vee (dV) to cover the round trip. That scope was multiple fundamentally-different forms of propulsion: nuclear explosion drive (or “pulse propulsion”) as it was envisioned in the late 1950’s, nuclear thermal propulsion (as a version of the solid core NERVA for which engine prototypes were tested), Hall effect ion propulsion based off of plentiful, cheap, and solid-phase-but-sublimable iodine, LOX-LH2 chemical rockets, and storable-propellant rockets.

Scope here is only nuclear thermal rocket propulsion, but with the highly-variable tested or envisioned characteristics of six different design approaches. It is these six that are compared in terms of the ratio of initial ignition to dead-head payload weight, using the same maximum-attractive criterion of 5 as in the earlier study. These six approaches and their relative states of technological readiness are:

  1. as-tested NERVA solid core
  2. the best-anticipated solid-core NERVA derivatives that never got built or tested
  3. the particle bed solid core reactor engine (one version of which was “Timberwind”, which got some exploratory testing revealing unresolved problems, but never reached the engine prototype stage)
  4. the so-called “nuclear light bulb” gas core concept (some insufficient feasibility tests)
  5. the open-cycle gas core concept restricted to regenerative cooling, meaning no radiator required (some insufficient feasibility testing)
  6. the open-cycle gas core concept with a large, heavy external waste heat radiator (some insufficient feasibility testing)

To accomplish this investigation, I added an additional worksheet to the colony ships.xlsx spreadsheet file that I used for the earlier study. Unlike the previous study, there are no closed-form ways to get from dead-head payload to a vehicle weight statement. The calculation uses iterative convergence of the propellant tank inert weight, and iterative convergence of engine thrust sizing in terms of the resulting vehicle acceleration gee capability.

For this investigation, the payload section is presumed to be some sort of enclosed hull, with adequate insulation, radiation shielding, and micrometeor protection for a crew built into it, in some unspecified way. The ratio of dead-head payload mass (contained inside) to the loaded payload section mass is a fraction denoted as fpay. The dead-head payload size drives everything in the end, as all results are directly proportional to the dead-head payload input. For this investigation, dead-head payload was arbitrarily set at 100 metric tons, and fpay = 0.8, the same for all six engine types. Thus:

Loaded payload section mass = dead-head payload mass/fpay

Payload section inert mass = loaded payload section mass – dead-head payload mass

The propellant tank section contains the common propellant for all nuclear thermal engine approaches: liquid hydrogen (LH2). This is a harsh cryogen, requiring solar heating control, significant insulation, and some sort of cryocooler to control evaporation. This is simply going to be heavier than the lightest-possible single-wall bare tank. The ratio of propellant mass to loaded tank mass is the fraction ftank. The single value ftank = 0.95 was used for all six engine types. Thus:

Loaded tank section mass = propellant mass/ftank

Tank inert mass = loaded tank mass – propellant mass

Propellant mass must be the sum for two burns at differing dead-head payload. One starts with a guess for tank inert, and iteratively converges it to the result

The engine “section” is the nuclear thermal rocket engine (or engines, for redundancy), complete with turbopumps and control equipment, a radiation shadow shield for the crew up forward, plus any waste heat radiator that may be required (if regenerative cooling alone cannot do the job). This radiator (if present) and the core-plus-engine hardware lead to a characteristic engine thrust/weight ratio T/We, which is dimensionless under the definition that both thrust T and engine weight We (on Earth) are measured in force units. This ratio is different for each engine type, as is the resulting specific impulse. The values I used follow:

TypeIsp, sT/Wedevelopment status
NERVA7253.6as-tested in engine prototypes
derNERVA10005derivative-of-NERVA, estimated on paper
PBR10007particle-bed reactor, based on “Timberwind”“nuclear light bulb” gas core concept, some feasibility
Open GCR250020open-cycle gas core concept limited to regenerative cooling
GCR+rad60000.5open-cycle GCR with heavy waste heat radiator, concept

For this kind of data, the main results used to size the vehicle are the exhaust velocity Vex (km/s), and the engine system inert mass (metric tons). These are:

Vex, km/s = (Isp, s)*9.8067/1000

Engine system inert mass, metric tons = thrust level, KN/(9.8067 * T/We)

For the remaining vehicle characteristics, all the concepts except “GCR+rad” were required to size thrust level such that the vehicle acceleration at the initial ignition mass was at or just above 0.5 gee. This corresponds to about a 15 minute Earth departure burn, definitely short enough to qualify as “impulse”, and not have the orbital dV be factored-up for gravity loss to be mass ratio-effective.

With the data I used, the GCR+rad system could not reach half a gee, but converged fairly well at 0.33 gee. This is less than a 30 minute burn, still short enough to consider as “impulsive” for Earth departure.

Max gee at final burnout weight upon Earth return should be under about 5, but this proved not to be a problem.

It’s a two-level iteration: first set a thrust level, then converge your guess for propellant tank inert weight with the final result of the calculation for tank inert weight. Then check and adjust your thrust level for the right Earth departure gee level. Then converge the tank inert weight again. Repeat the process as needed to get however-close a convergence you deem tolerable (0.1-0.01 ton range).

The orbital dV’s that are required are those for getting from low orbit onto a min-energy Hohmann transfer ellipse. The values used are worst-cases that do not go together; the difference is a nice little “kitty” to cover midcourse corrections. Earth departure = Earth arrival = 3.84 km/s. Mars arrival = Mars departure = 1.83 km/s. These sum to 5.67 km/s outbound in a heavier ship carrying payload, and 5.67 km/s return in a lighter ship with no payload and already having burned off some propellant on the outbound voyage.

Factored for losses, these dV figures become the mass ratio-effective dV’s for design purposes. Those and the Vex for each engine type give you the mass ratio MR for each engine type, one for outbound, the other for return.

MR = exp(sum dV/Vex) with both velocities in km/s, and the sum dV for outbound or return

You start the calculation with the return voyage by summing up the inerts (payload section inert + tank inert + engine inert), plus zero dead-head payload, as the burnout mass at Earth arrival. This starts with a best guess for inert tank mass, as well as for installed engine thrust level. Apply the appropriate mass ratio to get Mars departure ignition mass. The difference in ignition vs burnout mass is the propellant expended for the two burns of the return voyage.

The next step is the outbound voyage. The Mars departure ignition mass, plus the dead-head payload mass, is the Mars arrival burnout mass. Apply the appropriate mass ratio to get the Earth departure ignition mass. The difference in ignition vs burnout mass is the propellant expended for the two burns of the outbound voyage to Mars.

The sum of the two propellant quantities is the total propellant for the round trip. Divide this total propellant by ftank to find the total loaded tank mass. The difference between loaded total tank mass and total propellant mass is the inert tank mass. This resulting inert tank mass is what your guess for tank inert mass must converge to! The best next guess is close to the last result.

Thrust divided by Earth weight is the vehicle acceleration gee estimate. This is done at each of the 4 vehicle masses: Earth departure ignition, Mars arrival burnout, Mars departure ignition, and Earth arrival burnout. Two of these are of real interest: Earth departure ignition (min gees), and Earth arrival burnout (max gees). The other two conditions fall in-between. You must adjust your installed thrust level to achieve min gees. Then iterate to convergence again on tank inert mass.

Max gees at Earth arrival burnout did not prove to be a problem, but should fall under 5 gees for the most tolerable results. Be sure you check for that outcome.

The last calculation sets up weight statements and estimated dV performance for the six propulsion types, using the data already calculated. The initial part of the weight statement is the vehicle buildup from payload and inert items to Earth departure ignition mass. Subtracting the total outbound propellant gives the Mars arrival burnout mass. Their ratio produces an outbound summed dV for both burns, to be calculated for each type (for comparison to the initial summed requirement).

That Mars arrival burnout mass, less the dead-head payload, is the Mars departure ignition mass. Subtracting the return voyage propellant produces the Earth arrival burnout mass. Their ratio produces a return summed dV for the two burns, done for for each propulsion type (for comparison to that summed requirement).

The deviations of these weight statement dV’s from the required values reflect just how closely you converged your tank inert weights. These should be only trivially off (by under 0.001 km/s = 1 m/s). If you see bigger errors, you didn’t converge your tank inert masses closely enough. The effect of being “off” on min gee (as set by installed thrust level) is small, when compared to the effect of being “off” on guessed tank inert mass.

At the very bottom of the weight statements are the vehicle payload fractions, in both definitions. One is the conventional definition: dead-head payload mass / Earth departure ignition mass. You probably should not consider anything under 0.2 for a practical colonization ship design. Its inverse is Earth departure ignition mass / dead-head payload mass. In that definition, you probably should not consider anything over 5 for a practical colonization ship design.

This limit (in either form) is inherently a very fuzzy judgement call. But, if dead-head payload mass is too small compared to Earth departure ignition mass, the resulting design will be inherently very expensive to build and to operate, just like with ocean-going transport when the cargo mass is small compared to the tonnage of the ship.

What I got for this study is given in Figures 1 and 2, a two-part image of the completed spreadsheet worksheet page. Of the six propulsion types, four look reasonably-to-very attractive. These are the derivative of NERVA, some form of PBR, and the two gas core concepts that do not require a huge waste heat radiator. The as-tested NERVA falls short because its engine thrust/weight is too low and the resulting large engine inerts drive the vehicle inerts, constrained by the large thrust level to achieve min acceleration gees. The gas core with radiator falls short because of the gigantic, heavy radiator.

Near-term, the higher Isp and engine thrust/weight of the derivative NERVA could be realized in a few short years, to an engine prototype ready for flight test. The PBR concept would take a few more years than that, since no prototype engines were ever ground tested, and some fundamental problems identified in testing of “Timberwind” components remain unresolved. The gas core concepts would require several-to-many years to reach a flight-testable prototype, since only very sparse lab-type feasibility demonstrations were ever done; plus, there is no guarantee of eventual success, either.

My own recommendation would be to base an initial design around the derivative NERVA as lowest-risk option of acceptable benefit, and plan on replacing it later with one of the non-radiator gas core designs, should that development prove successful.

Figure 3 sketches a ship design concept based on the derivative of NERVA, figured at 100 metric tons of dead-head payload delivered to Mars. Volume of LH2 and a guess for tank L/D set the tank dimensions. Everything else scales one way or another from that, as a first approximation. Everything about the weight statement and thrust level sizing is proportional to dead-head payload size. Dimensions would scale as the cube root of mass, provided that L/D ratios are preserved.

This vehicle rough-out delivers the same design dead-head payload to Mars as the proposed Spacex “Starship” design. The differences are several: this vehicle never lands on Mars (delivery to the surface is by unspecified other means), this vehicle must make a full Mars arrival burn into low orbit (“Starship” only makes a final touchdown burn after an aerobraking direct entry), and this vehicle returns all the way to low Earth orbit for reuse, unrefueled. It never needs to survive any sort of atmospheric entry.

This design makes the round trip single-stage unrefueled. The Spacex “Starship” is entirely one-way only, unless and until it can be refueled on the surface of Mars from local resources.

There is enough payload section volume to support a crew of up to 15, at about 300 cubic meters per person, in addition to the volume occupied by the dead-head payload, at a payload specific gravity averaging only 0.3.

This result says a Mars colonization ship able to carry 100 metric tons of dead-head payload one-way to Mars, and return to Earth with no payload, all one-stage, is not that large an item. It is not large enough to spin for artificial gravity like a rifle bullet, but it is large enough to spin end-over-end (like a baton) for artificial gravity. At about 3.24 rpm, there is about one full gee available in the payload section. That spin rate is tolerable to untrained, unacclimatized people, for long-term exposure.

The insulation and meteor shielding is about a meter thick on the payload section, meaning it can double as radiation protection. If those layers of fabric average 0.20 effective bulk specific gravity, that is some 20 g/ shielding mass, adequate for solar flare events, and offering some reduction of galactic cosmic radiation. The insulation and tank shell thickness of the propellant tank section was assumed to be 0.1 m. Engine section length was just a guess.

Key to this design as-sized is carriage of dead-head payload to Mars, but not from Mars. The return dead-head payload must be zero! If not, the propellant tank section must be significantly larger, to the detriment of the payload fraction criteria. Any crew and their life support must come out of that dead-head payload allowance (meaning near-zero crew on the return voyage).

These results look more favorable than the otherwise-comparable nuclear thermal option in the earlier study. That is precisely because dead-head payload is only carried one-way in this study, and it was carried both ways in the earlier study. That is one huge effect. But the trend from the earlier study applies here as well: if we design for a farther destination than Mars, the design won’t look so attractive in terms of the payload fraction criteria.

The restriction of zero dead-head payload on the return voyage is not as constraining as it first sounds, when one considers the goal is building a colony with these payloads. During that process there is little-or-nothing to ship home to Earth, except information, which is better sent electronically. Later, when an operating economy results in two-way trade, one will need commerce ships, not colonization ships. But, by the time that need arises, significantly-better propulsion technology should have become available.

Brief Result Summary: The best near-term option of the six nuclear thermal approaches, for a Mars colonization ship design, is the derivative-NERVA nuclear thermal propulsion approach (Isp ~ 1000 s and engine T/W ~ 5). For 100 metric tons dead head payload, the initial ignition mass is about 500 metric tons. That means for 1000 metric tons dead-head payload, the sized ship will initially mass about 5000 metric tons. For 2000 tons payload, the ship will be around 10,000 tons, etc. This is restricted to orbit-to-orbit operation, and to no dead-head payload on the return voyage. Even the small 100-ton payload size is large enough to spin end-over-end for artificial gravity at near 1 gee and an easily-tolerated spin rate. The payload section insulated design (if a meter of fabric layers) also inherently provides a fair amount of radiation protection.


This year has been the 50th anniversary of the first man on the moon. That was the culmination of the space race between the US and Soviet Russia. That accomplishment was a whole lot more about “flags and footprints” and experimental flight test, than it was about science or real exploration.

This article builds upon some earlier articles posted upon this site.It presents the latest version of my Mars mission outline plan,with an enlarged manned transport,and the latest sizing of 1-stage 2-way reusable chemical landers.These earlier articles are as follows:

Why We Should Go Back (And Farther Still)

Is there anything worthwhile to accomplish out there?Yes,definitely!

In the longer term,there are those future off-world settlements and the associated future economies.I cannot tell you the details of how this might benefit us,because it has yet to be done.But it has always proven beneficial in prior centuries here on Earth.

In the shorter term,there are the possibilities of space resource businesses,and of planetary protection against rogue asteroid and comet impacts.That second item is the most important of all:there is simply no better reason for continuing both unmanned and manned space programs than finding ways to protect the folks back home!

It’s not about winning some race,and it’s not so very much about doing pure science just for the sake of knowledge.It’s about real exploration of the unknown,something hard-wired into humans.In centuries past,this was exploration of the unknown parts of the Earth.Now it is about space and the deep ocean floor.This article is concerned with the real exploration of space.

“Exploration” is a really an emotionally-loaded code word,something most people do not think about.What it truly means is you go there to find out “what all is there” (resources,including those you don’t at first recognize),and “where exactly it is” (how hard to obtain,as well as how much is there).Then you have to stay a while to figure out how to use what you found, in order to cope with living in the local environment. All of that is part of “real exploration”.

Unless you do that correctly,there is no real possibility of future settlements and the associated future economies,or any of the benefits that would ultimately derive therefrom!There is no way to accomplish much of anything else,except just the “flags and footprints” act of going there and returning (which is the bulk of what Apollo itself really accomplished at the moon).

Those who “get there first” do tend to do a little better in the long run,in terms of those benefits,provided that they do it “right” when they go.That is one crucial lesson from history.

My Suggestions for the Near Term

Establish a continuous human presence on the moon,the first item.Start small and expand it slowly over time. Do the lunar “exploration” thing right,this time.

Send humans to Mars as the fulfillment of a dream centuries old,probably the second item.When we go,do the “exploration” thing right,from the very first landing.Further,it starts long before the first item (going to the moon) is “done” in any sense of that word.

But,any vehicle capable of taking crews to Mars can also take a crew to near-Earth asteroids and comets.Visit those asteroids and comets and properly explore them,in order to learn how to defend against their impacting Earth,as well as “ground truth” for how to really do space mining.

That’s the third item,but it is just as easily done,and at least as important,as going to Mars. 

Maybe we do them at pretty much the same time.

Ethically and Responsibly Addressing Known Risks For Spaceflight

We are ethically-bound to address the known risks of manned spaceflight as best we can.There is a whole long list of safety risks associated with any sort of manned spaceflight.Three come to mind as the most truly credible risks:(1) reliability of,and escape from,spacecraft and booster rockets,(2) microgravity diseases,and (3) exposure to space radiation.

The first one has cost us three American crews totaling 17 people dead (Apollo 1,shuttle Challenger,and shuttle Columbia). Each caused a year-or-more stand-down,and very expensive investigations,plus very expensive changes.

The two shuttle losses were ultimately caused by bad management decisions valuing cost or schedule above safety.Apollo 1 was about a really-poor basic management attitude (“good enough for government work”) combined with technical ignorance,because we had never done this sort of thing before.

Those outcomes and their actual causes are why I claim “there is nothing as expensive as a dead crew,especially one dead from a bad management decision”.Bear in mind that those expenses are both economic and political (which includes public opinion as well as DC politics).

Making spaceflight more safe, from a reliability and escape standpoint, is now also something we already know how to address!This takes careful design allowing for failure modes,redundant systems, and copious verification testing.Mitigation efforts will never be perfect,but they can be quite good. Ethics requires that you treat this as a required constraint upon your designs.

It means you always provide “a way out” for your crew at every step of the mission.It really is as simple (and as hard to do) as that!This very seriously constrains your overall mission architecture,as well as your detailed space vehicle designs.

The other two have been long studied in low Earth orbit,where microgravity exposure is inherent in everything we have done there,and radiation exposure is somewhat more than on Earth’s surface,but less than outside the Van Allen radiation belts (and far less than inside the belts themselves).

Microgravity Diseases

Microgravity has proven to affect the human body in a variety of expected, and unexpected, ways.The longer one is exposed,the worse the various diseases become.Beyond the bone decalcification and muscle-weakening that we have long expected,there are also degradations of the heart and circulatory system,degradation of vision from eye geometry changes due to the fluid pressure redistribution,immune system degradations that we have yet to understand,and most recently genetic changes whose meanings are still a total mystery. No doubt more will be discovered,as that has been the trend.

The longer exposed,the longer it takes to recover upon returning home,with full recovery actually still in doubt for some of the effects,despite diet,drugs,and exercise. The practical time limit seems to be only a bit more than a year.For that very reason,usual practices on the International Space Station (ISS) call for 6 months to a year’s exposure at most,with 6 months the preferred limit.

We do know that something near one full Earth gravity (one “gee”) is therapeutic,precisely because that is what we evolved in.So,until we know better,any artificial spin gravity schemes need to supply very near one gee,in order to obtain the full Earthly benefits that we already know will work.

Destinations outside of Earth-moon space are very much further away than the moon:one-way travel times range from near 6 months to multiple years.This is pretty much outside the preferred limit of microgravity exposure that we have already established on ISS.

Mars is 6-to-9 months away one-way,and we do not know how therapeutic its lower gravity (38%) really is for the rigors of the return voyage.Other destinations are further away still,and all those we can actually land upon, are even lower gravity than Mars.That situation says quite clearly that we need to provide artificial gravity (no matter how inconvenient that might otherwise be !!!!) at something near one gee (until we actually know better !!) during these one-way transits to-and-from,in order to best preserve the health of the crews.

Ethically,you simply cannot argue with that conclusion,no matter how inconvenient for design purposes,or for total mission cost purposes.That is the only “box on thinking” applied here.

Supplying Artificial Gravity

There is as yet no such thing as “Star Trek”-type artificial gravity.The only physics we have to serve that purpose is “centrifugal force”.You must spin the vehicle,to generate “centrifugal force” as an equivalent to gravity.If the spin rate is low,then Coriolis forces (something everyone has experienced on a merry-go-round) become less important,and so fewer folks can tell the difference between this and real gravity,and there are fewer problems with disrupting the balance organs in the ear.

The physics of spin say that the acceleration you feel is proportional to the radius of spin and to the square of the spin rate.The actual physics equation says

a = R w2

where a is the acceleration,R the spin radius,and w the spin rate

Another form expressed in gees,and not absolute acceleration units, is

gees = 1.00 * [(R, m) / (55.89 m)] [(N, rpm) / (4 rpm)]2

Earthly experience with spin rates says that normal untrained and unacclimatized people can tolerate 3 to 4 rpm immediately,and for long-term exposures,without getting motion sick.People extensively trained might (or might not) tolerate higher spin rates in the 8-12 rpm class, without getting motion sick from long exposures.Still-higher spin rates (16+ rpm) induce blood pressure gradients head-to-toe in a standing individual, that are just unacceptable for long term exposures.Stand up,and you faint.

3-dimensional objects typically have 3 axes.About these axes these objects have properties called “mass moment of inertia” that relates to spin dynamics.Usually,higher moment of inertia correlates with a larger dimension along some axis perpendicular to the actual spin axis.These are typically proportional to mass,but proportional to the square of its distance from the center of gravity.

There are two (and only two !!) stable spin modes for most objects:  about the axis for highest moment of inertia (longest dimension),  and about the axis for lowest moment of inertia (shortest dimension).  The first case is exemplified by a baton twirler’s spinning baton,  and the second case is exemplified by a spinning bullet or artillery shell.  There are no other stable modes of spin.See Figure 1.

Clearly,  building a “spinning rifle bullet” 112 m in diameter at 4 rpm for one full gee at its outer girth is not so very feasible:  this is just too big to afford at this time in history.  But spinning a smaller-diameter “something” that is 112 m long,  end-over-end at 4 rpm,  for 1 gee at each end,  would indeed be a feasible thing to attempt!  That says select the baton-spin mode for practical designs.

We already know a lot about the transient dynamics of spinning rigid objects,  something important for spin-up and spin-down,  as well as for applying any thrust while spinning.  There would be no fundamental engineering development work to design a long,  narrow spacecraft that spins end-over-end for artificial gravity.  There would only be proving-out the specific design in tests before we use it.

The most-often-proposed alternative is a cable-connected structure,  because it is conceptually easy to reel-out long cables between two small objects.  Cable-connected transient dynamics for spin-up and spin-down,  and especially for applying thrust while spinning,  are incredibly complex and still not very well-known.  “You cannot push on a string”,  that is the complication!  So there is a huge fundamental engineering development effort needed,  beyond just proving-out the actual design to be used.

What this really says is that the preferred near-term spacecraft design is a long and rigid,  more-or-less cylindrical shape,  to be spun end-over-end,  baton-style.  This will generate varying artificial gee from a maximum near the ends,  to zero at the spin center. 

We know that microgravity vs gravity has no impact while prone sleeping,  or else Earthly bed rest studies would not be a decent surrogate for some of the in-space microgravity effects.  That means you can put the sleeping quarters in the low gravity section of the spacecraft near the spin center,  and just put the daily workstations in the full-gravity sections of the spacecraft near the ends.  See Figure 2.

Radiation Hazards

There are basically three radiation hazards to worry about:  galactic cosmic rays (GCR),  solar flare events (SFE),  and the Van Allen radiation belts about the Earth (or similar belts around some of the outer planets).  All three hazards are atomic or subatomic particles,  just at different speeds and quantities.  The threats they pose are location-dependent.

GCR is a very slow drizzle of really-high-speed particles,  moving at a large fraction of the speed of light.  Particles that energetic are very difficult to shield against,  because they penetrate deeply into shielding material,  and quite often create “secondary showers” of other harmful radiation when they strike the atoms in the shield material.  If the shielding atoms are low atomic weight,  the secondary shower effect is greatly reduced.

GCR comes from outside the solar system.  Its quantity is affected by the solar wind,  in turn affected by the sun’s sunspot cycle,  which is about 11 years long.  The solar wind is stronger when sunspots are active,  making GCR lower in the vicinity of the Earth-moon system at that time.

From NASA’s radiation effects website (ref. 1) I obtained these values that apply in the general vicinity of the Earth-moon system.  GCR maximizes at about 60 REM per year when the sun is quiet,  and minimizes at about 24 REM per year,   when sunspots are most active.  To “calibrate” the effects of what may be unfamiliar units of radiation,  the natural Earthly background radiation is about 0.3 REM per year (and up to 10 times higher in some locations),  and a lethal dose would be 300 to 500 REM accumulated in a “short time”,  meaning hours to a week or so.  (Just for information,  1 Sievert is 100 REM.)

The NASA astronaut exposure standards are set at about twice the levels allowed for Earthly nuclear workers.  Those NASA standards are no more than 50 REM per year,  no more than 25 REM in any one month,  and a career limit that varies with age and gender,  but maxes-out at no more than 400 REM accumulated over an entire career.  These career limits are predicated upon a single-handful percentage increase in the likelihood of late-in-life cancer.  

Clearly,  with a very modest shielding effect (to reduce worst-case 60 REM to an acceptable 50 REM annual),  GCR is not the “killer” it is often portrayed to be.

SFE (solar flare events) are different.  They are much lower-speed particles,  much easier to shield,  but there is an incredibly-huge flood of them,  when these events happen.  They come in very-directional bursts from the sun,  at rather erratic intervals.  There are usually more of them during times of active sunspots,  but they can indeed happen when the sun is quiet.  They come at irregular intervals measured in durations of “several months apart”.

The intensity of a burst can vary wildly from only tens of REM received over a few hours,  to tens of thousands of REM received over a few hoursThe median dose would be multiple thousands of REM over a few hours.  Obviously,  for unshielded persons,  the great bulk of events like this (those over about 300-500 REM) would be fatal doses,  and it is an ugly,  irreversible,  and miserable death.  There was a massively-fatal-level event in 1972 between the last two Apollo missions to the moon,  and a low-intensity (non-fatal) event during one Apollo mission to the moon.

We had chosen to ignore this SFE threat during Apollo because the short duration of the missions (at most 2 weeks) was small,  compared to the typical interval (several months) between events.  But,  had a large event hit during an Apollo mission,  the crew would have died in space in a matter of hours.  As it turns out,  this actual record shows that Apollo’s “ignoring-the-risk-as-low-probability”-assumption was not a good assumption to make!  That’s 20-20 hindsight,  but it is still a crucial lesson to learn!

For an extended or permanent return to the moon,  or going elsewhere,  radiation shielding is obviously imperative!  On Earth,  we are protected from these SFE’s (and the GCR) by both the Earth’s magnetic field and its atmosphere.  These are a very real threat anywhere outside the Earth’s magnetic field!  In low Earth orbit,  we are protected only by the magnetic field,  and the background exposure there is higher than down on Earth,  but still much less than beyond the magnetic field.

The Van Allen belts are concentrated regions of these same radiation particles trapped in the Earth’s magnetic field.  The intensity is lethal on a scale of days-to-weeks,  but tolerable on a scale of hours-to-a-day-or-so.  The inner boundary is not sharp,  but this is generally considered to become a problem at about 900 miles orbit altitude,  and extending many tens of thousands of miles out from the Earth. 

The exception is the “South Atlantic Anomaly”,  where the inner side of the Van Allen belt dips down locally to low Earth orbit altitude (100-300 miles).  Satellites and spacecraft in high-inclination orbits inherently pass through the South Atlantic Anomaly every several orbits.  The ISS does indeed encounter this threat,  it being short “flashes” of exposure that accumulate over time,  but these still fall well within the astronaut exposure standards (no more than 50 REM annually,  no more than 25 REM in any one month).  Their main effect is accumulation toward career limits.

Spacecraft traveling to the moon or elsewhere must transit the Van Allen belts.  Because of the potential for lethal exposure if you linger within them,  such transits must be made quickly!  Apollo did this correctly,  transiting within only several hours.  Given the state of today’s electric propulsion technology,  this rules out using electric propulsion for people to leave Earth orbit for the moon or elsewhere,  because the spiral-out time is measured in multiple months.  That would quickly accumulate to a lethal exposure,  even with some shielding.  

Passive Shielding

The same NASA radiation site has data regarding the shielding effects of typically-considered materials.  Those are hydrogen,  water,  and aluminum.  Mass of shielding above a unit exposed area turns out to be the “correlating variable”,  and 15-20 g/cm2 seems to be enough to generally address the worst SFE. 

Hydrogen has the lowest density,  requiring the thickest layering,  but also has the least secondary shower potential,  when used against GCR.  211 to 282 cm of liquid hydrogen suffices. 

15-20 cm of water is 15-20 gm/cm2,  same shielding effect as a really thick layer of hydrogen.  Water molecules are still light enough not to have much secondary shower risk. 

Aluminum would be the thinnest layer,  but with the greater secondary shower effect.  However,  of the practical metals,  its atoms are the lightest,  and this secondary shower effect is deemed tolerable with it.  6-8 cm thick aluminum plate would be required.  That is quite out-of-line with current spacecraft hull design practices:  something nearer a millimeter.

Other materials based on polymers,  and even most rocket propellants,  are light-enough atoms to be effective shielding with a low secondary shower risk,  yet with densities roughly in the same ballpark as water,  for a thinner layer thickness.  So,  any of these could be practical shielding materials!

Because weight is critical,  what you have to do is not simply add shielding weight to your design,  but instead rearrange the distribution of masses you already otherwise need,  so that they can also serve as radiation shielding.  You will need meteoroid shielding and thermal insulation,  and any manned craft will have water and wastewater on board,  as part of the life support system.  All spacecraft will need propellant for the next (and subsequent) burns.  You use a combination of these,  acting together.

The real suggestion here is to use water,  wastewater,  and next-burn propellant tankage as shadow shields,  in addition to the meteoroid protection and thermal insulation materials that the manned modules require anyway.  It doesn’t take much of this at all to cut the worst-case 60 REM/year GCR to under 50 REM/year.  It takes only a little more to cut worst-case SFE to safe short-term exposure levels. 

If you cannot protect the whole manned interior,  then the flight control station becomes first priority,  so that maneuvers can be flown,  regardless of the solar weather.  Second priority would be the sleeping quarters,  to reduce round-the-clock GCR exposure further.  These seriously constrain spacecraft design.

See Figure 3 for one possible way to do this,  in an orbit-to-orbit transport design concept.  This would also be a baton-spin vehicle for artificial gravity during the long transit.  Plus,  the habitation (“hab”) design requires a lot of interior space for the mental health of the crew,  something else we know is critical.  Somewhere between 100 and 200 cubic meters per person is needed as a minimum,  and at least some of it must be reconfigurable as desired by the crew. 

Spin-up is likely by electrically-powered flywheels in the center module.  The vehicle is spun-up after departure,  and de-spun before arrival.  If a mid-course correction is needed,  the vehicle could be de-spun for that,  and spun back up for remainder of the transit.  

Note in the figure how the arrival propellant and the water and wastewater tankage has been arranged around the manned core to provide extra shadow shielding,  for really effective radiation protection.  The manned core modules are presumed insulated by polymeric layers that also serve as meteor shielding (while adding to the radiation protection,  without being driven by that issue).  The pressure shell on the inside of this insulation should be unobstructed by mounted equipment,  so that easy and rapid access for patching of holes is possible. There is not time to move stuff when a compartment is depressurizing!  Ethics!

At departure,  the vehicle can be propelled by a different propellant and engine choice,  since departure is a short event.  The arrival propellant is likely a storable to prevent evaporation losses.  Storable return propellant tankage sets can be sent ahead unmanned,  for docking in orbit at the destination.  

There is an emergency return capsule (actually two capsules) mounted at the center module,  each one enough for the entire crew.  (“Bailout” at Mars presumes a rescue capability already exists there,  so we need redundant engines instead.) Emergency bailout,  upon a failed burn for returning to Earth orbit,  is the main function of this capsule.  Routinely,  it could return a crew to Earth from the spaceship,  once it is parked safely in Earth orbit.

This kind of orbit-to-orbit transport design could serve to take men to Mars or to the near-Earth asteroids and comets.  For Mars,  the lander craft could be sent ahead unmanned to Mars orbit,  and none are needed to visit asteroids.  But you cannot send return propellant ahead on an asteroid mission.

By refueling and re-supplying in Earth orbit,  such a manned hab design could easily be used for multiple missions,  once built.  Care must be taken in its design and material selection to support many thousands of cycles of use.  Thus the craft could safely serve for a century or more,  updated with better propellants and engines as the years go by.

There I went and wrote a basic “how-to” document for practical and ethical interplanetary spaceship design!

These first few sections so far have been reprised (with edits) from “Just Mooning Around”,posted 7-14-19.Everything that follows is new.

Mars Mission Outline 2019:Overall

The new version uses a larger orbit-to-orbit transport,and recovers the solar-electric tugs that preposition unmanned assets at Mars for the manned mission (2016 did not).It uses similar (but larger) landers as the 2016 version,and it still jettisons the Earth departure stage without recovery.

That last could be addressed by fitting the departure stage with a second propulsion system, possibly electric,and putting it into a 2-year-period orbit after stage-off.Then it could be captured into Earth orbit for reuse.That recovery possibility is beyond scope here in the 2019 version. Consider it to be a “future update”.

Main point here:if one does spin gravity in a baton-spin mode,the resulting transit vehicle is ill-adapted for a direct entry at Mars,or a direct entry at Earth.Such a design is far better-adapted as an orbit-to-orbit transport,with any Mars lander function relegated to a separate vehicle,sent separately.Long-life reusability also points toward an orbit-to-orbit transport design,free of entry heat shield requirements.It means we base our exploration forays onto the surface of Mars from low Mars orbit.

The resulting mission architecture requires that both the landers and the Earth return propellant get sent ahead unmanned to parking orbit about Mars,with the manned orbit-to-orbit transport arriving afterward,and rendezvousing in Mars orbit with those items.This powerful concept is not unlike the Lunar Orbit Rendezvous architecture that made it possible to mount each Apollo landing mission with only one Saturn 5 booster.See Figure 4 for the overall mission architecture.

The landers themselves are envisioned as one-stage reusable articles that make multiple flights,  based out of low Mars orbit.  Sending 3 landers ahead with their propellant supply allows one lander to make a landing with only part of the human crew,  with a second lander in reserve as a rescue craft.  Thus,  there is a “way out” even during the landings,  unlike with Apollo! 

Because of storability concerns,  the wisest choice is that the lander propellant and engine design be the same as the transport propellant and engine design.  This maximizes the interchangeability of engine hardware and propellant supplies,  in the event that there are mishaps from which to recover,  without aid from Earth.  It also simplifies the overall design and hardware development and prove-out.

The presence of a third lander allows one lander to become unserviceable,  while still maintaining the reserve rescue lander capability,  without which landings so far from Earth become too risky to ethically attempt.  This is shown in Figure 5,  including the velocity requirements for the lander design. 

The initially-sized version of the lander design concept was used in the 2016 posting,  and came from one of the options explored in another posting titled “Reusable Chemical Mars Landing Boats Are Feasible”,  dated 31 August 2013.  These landers are resized somewhat for this posting.

Note that for a rescue possibility to exist,  some of the crew must stay in the transport in low Mars orbit,  while others descend to the surface in a lander.  Because we do not know how therapeutic Mars’s 0.38 gee gravity might be for the surface crew,  I suggest we spin the transport for artificial gravity while it is in orbit,  de-spinning for lander departures and arrivals.  Thus everybody stays fully healthy no matter what,  while we alternate crews on the surface.

Now,  overall,  it is worst-case 9 months to and from Mars,  and in any case,  13 months at Mars waiting for the orbital “window” to open for the voyage home.  That last is simply inherent from the choice of min-energy Hohmann transfer orbits.  That leaves a long time for the crew to explore on Mars.  That plus the possibility that the initial landing site might not prove to be desirable,  makes it wise to plan for multiple landings,  at possibly-multiple sites. 

Basing exploration forays from low Mars orbit is what makes multiple landings at multiple sites possible at allNo other mission architecture can provide this capability.

It is that orbit-based architecture allowing for multiple landings which lets us alternate roles for the crew,  so that all of them get to spend time on the surface of Mars (unlike what was possible with Apollo).  With a mission crew of 6,  that means we could send down alternating crews of 3 in the lander,  while the other 3 do science from orbit and provide the critical watchdog rescue capability with the other two landers (two for the reliability of redundancy).  It is already known that odd numbered crews fare better in hazardous situations,  there being no possibility of the stalemate of ties,  in decision-making.

Given the existence of the rescue capability from low Mars orbit,  we can address lander reliability in two ways,  thus increasing the odds of success,  and also the odds of still saving the lander crew,  if things go seriously wrong.  (We are ethically bound to do this!)  First,  the lander must use redundant engines,  so that if one fails,  the remaining engine (or engines) can still perform the mission.  

Second,  the crew piloting cabin could be rigged as an abort-to-surface (or abort-to-orbit) capsule,  in the event that too many redundant engines fail,  or that there is some overall catastrophic failure of the lander.

The minimum number of landings is two,  one for each half of the crew.  Allowing some time for reconnaissance-from-orbit prior to the first attempt,  and for preparations for returning to earth,  we can plan on 12 months total for the landings,  splitting the remaining month between those other two needs in orbit about Mars.  That does cover up to two possible landing sites in the one voyage to Mars!

The surface crew will live inside the lander on the surface.  That means it must carry them,  their exploration gear,  and up to 6 months of life support supplies,  on each trip.  More exploration gear could be carried to the surface if we shorten the stay for each lander. 

If four trips will be made,  that’s 3 months each (not 6),  and one can trade away life support supplies for extra exploration gear carried down.  That could cover up to four possible landing sites in the one trip to Mars,  and each crew of 3 making 2 trips,  all with the same overall resources sent to Mars,  excepting the total lander propellant supply.

Continuing that logic,  if 6 trips are planned,  that’s 2 months each,  each crew of 3 making 3 trips,  and a higher weight of exploration gear relative to life support supplies.  That’s up to 6 separate sites that could be explored in the one voyage to Mars!  Or,  12 trips of 1 month each,  which is up to 12 sites explored.  Since the lander propellant is sent ahead by SEP,  it is rather easy to afford such a capability. 

The biggest mass ratio-effective burn for the lander is the ascent burn,  which can be at significantly-reduced payload,  since wastes can be left on the surface along with some exploration gear,  while the weight of a plethora of samples is far less than the weight of gear and supplies during the less-demanding descent.   That makes the overall 5.22 km/s delta vee far more affordable with an overall realistic mass ratio and storable propellant specific impulse (Isp).  

Those considerations very dramatically impact and constrain the design of the lander.

Sending Assets Ahead Unmanned

The unmanned transfers can be done more efficiently (lower total mass to be launched) with solar electric propulsion (SEP).  The manned transport uses short-burn chemical rocket propulsion to avoid long spiral-out/spiral-in times.  (An SEP-based transport would give the crew a lethal radiation dose spiraling-out through the Van Allen belts on departure from Earth,  and again spiraling-in through the belts on return to Earth.)  At least approximately 0.1 gee vehicle acceleration is required to qualify as a gravity loss-free “short burn”. 

This prepositioning of assets at Mars using SEP was also a part of my 2016 Mars mission posting.  The differences here are that I recover the SEP “tugs” for reuse on future missions,  and that I use a larger “hab” for the orbit-to-orbit transport. 

Earth Departure of Manned Transport

The Earth departure can be done with higher-performing LOX-LH2 tankage and engines on one end,  that are staged off after the burn.  To recover these,  a higher aphelion orbit with a 2 year period is required,  plus some sort of propulsion to return to Earth orbit.  This could be electric,  or some storable propellant rockets.  (Expecting LOX-LH2 cryogens not to evaporate over a 2 year period is just nonsense!)  I did not include that here,  but it is required for more reusability.  That’s a future growth item.

Velocity Requirements for the Mission

The orbital mechanics of min-energy Hohmann transfer determine the minimum velocity requirements for the manned (and unmanned) vehicles,  as well as the one-way travel time.  Shorter flights require more energy,  which is more propellant and tankage that must be sent to low Earth orbit and assembled. 

The basic velocity requirements for the manned orbital transport are shown in Figure 6.  These take the form of unfactored orbital mechanics values serving as the mass ratio-effective values for vehicle design.  This is allowable because all these chemical rocket propulsion burns are “short” and exoatmospheric.  The resulting mass-ratio-effective design values are given in Figure 7.

For only Mars arrival with the manned transport,  there is a need for a rendezvous propellant allowance.  It is necessary to adjust orbital position to coincide with the assets sent ahead.  As a wild guess,  add another 0.2 km/s delta vee to the value shown in Figure 7 as the mass ratio-effective value for design.  

For the assets sent ahead with SEP,  design velocity requirements are much more problematic.  There are no drag losses,  but the gravity losses are huge,  since the burns are months long!  For a rough rule-of-thumb estimate,  just use twice the values in Figure 7.  That is what I did here. 

Propulsion Estimates

No particular existing chemical rocket engine’s characteristics were used.  Ballistic estimates were made “from scratch” using shortcut methods.  For both the transport and Earth-departure engines,  it was assumed that no gas used to drive pumps was dumped overboard,  meaning 100% of the hot gas generated went through the propulsion nozzle.  This requires an efficient engine operating cycle. 

Estimates were made from 1000-psia data for chamber characteristic velocity and gas specific heat ratio,  using standard ideal-gas compressible flow methods to develop vacuum thrust coefficient (to include the effects of a nozzle kinetic energy efficiency reflecting streamline divergence).  The c* and r “constants” vary with chamber pressure in a way that conforms to empirical ballistic methods I have long used successfully.

This gets us to specific impulse (and thus effective exhaust velocity) for vehicle mass ratio determinations with the rocket equation dV = Vex ln(Wig/Wbo).  The actual design thrust level is driven by vehicle mass and the min 0.1 gee acceleration requirement,  which sizes throat (and exit areas) via the thrust/throat area/thrust coefficient equation F = CF Pc At.  That leads to real engine dimensions.  For not-quite-the-highest-tech in engine design technology,  a good “wild guess” for engine weight would be thrust/50,  both in force units,  figured at 1 gee Earth gravity for the weight. 

Assuming redundant engines for safety and reliability,  these rockets won’t be simultaneously run at full thrust.  For vacuum-only operation,  there is no need for really high chamber pressure,  and there is no need to worry about backpressure-induced separation effects,  because there isn’t any backpressure.  6-7 mbar on Mars is also effectively no backpressure at all,  so the lander engines can be the same vacuum design as the transport engines.

Reflecting those considerations,  I assumed 1000 psia at max thrust,  typical operation at 500 psia,  and min throttled-down pressure 200 psia.  Others may disagree,  but that is what I did.  The higher the Pc,  the higher the c*,  and thus the higher the Isp.  But so also the higher is the weight of the engine.

The data I got for the NTO-MMH storable transit engines are given in Figure 8.  The data I got for the LOX-LH2 Earth departure engines are given in Figure 9.  For both I assumed an expansion bell equivalent to a constant 15 degree half-angle conical bell,  leading to a kinetic energy efficiency of 0.983 for the nozzle efficiency.  Any real-world curved bell will have an average half angle not far at all from that value;  it will be slightly shorter than the equivalent conical bell,  and just about the same efficiency. 

The solar electric propulsion is more problematical in its characteristics,  it being currently available only in small sizes,  with scaleup efforts underway at both Ad Astra and NASA.  What is important for vehicle design purposes would be thrust/weight for the actual electric thruster equipment,  its operating specific impulse,  its electric power/thrust requirement,  and the type and phase of its propellant (liquids or solids are easier to store at lower total tankage weight than gases).   

Add to that the producible electric power/panel area,  the weight/panel area,  and miscellaneous equipment weight (if any),  for the solar power supply equipment,  and for autonomous robotic vehicle guidance.  The size of the thruster’s thrust relative to the full vehicle weight should probably fall near what the current small thrusters on satellites use:  something near or above 0.001 gee.

Here are the values for the putative system I “chose”,  it being something that does not yet exist,  but likely could be made to exist near-term.  Bear in mind the available solar power at Mars is half that at Earth (Mars actually sizes the panels).  The value shown for electric power/area of solar panel is for near-Earth space,  turned to face the sun directly. This data represents a Hall-effect device on iodine.

The solar photovoltaic power per unit area was estimated as the solar constant at Earth (in space 1353 W/m2) multiplied by a 20% conversion efficiency of sunlight power to electric power. That represents a high-tech space-industry type of solar cell.  The weight was estimated from reported data for the Alta Devices Alta 5x1 2J and Alta 5x1 1J satellite solar panel devices.  The miscellaneous equipment is not structure,  that is in the weight/area figure for the panels.  It is the mass of the autonomous guidance equipment,  including things like star trackers,  computers,  communications,  and accelerometers. 

Space Hab for the Crew:  Characteristics

I based these guesses off the Bigelow Aerospace B-330 space station module design as seen on the internet (ref. 2).  This is the big commercial product,  not the simple,  small BEAM unit attached to ISS for testing and evaluation by NASA.  These are nominally 15.7 m long and 20 metric tons.  They are somewhat inflatable,  and feature a core equipment and framing structure around which the inflated hull is unobstructed.   There is a meter of layers of micrometeoroid shield that also serve as thermal insulation,  and as low-molecular-weight radiation shielding.   Each module contains some 330 cubic meters of interior space. The hard core protrudes on one end,  providing a place for solar panels. 

The modules of the orbit-to-orbit transport cannot be exactly these B-330 modules,  but can be something rather similar!  Docking multiple modules end-to-end creates the baton-shaped vehicle this mission design needs.  The modules must have external features of some type that allow tankage to be attached around the outer periphery,  and internal fold-out decks as part of the core.  The center module must be very stout,  and contain big electrically-driven flywheels for vehicle spin-up and spin-down,  plus places to dock space capsules.

It would seem wiser to put big solar panels on the center module,  with the docked capsules,  and the flywheels inside,  where spin forces are zero-to-minimal.  It is likely to be hard shell,  not an inflatable,  for strength.  That module is also likely to be quite heavy.  As a wild guess,  call it 16 m long and 40 tons.  The others can be nominal 16 m long,  and nearer 20 tons,  reflecting inflatable pressure shell along almost the entire length,  plus the features for attaching external tankage.   Call the internal volume 350 m3 each as a best guess,  excluding what the hard core occupies.

Counting the center module,  some 7 modules each 16 m long docked end-to-end is 112 m long,  for 1 full gee at each end if spun at only 4 rpm.  That basic structure would total 160 metric tons,  using the guesses in the previous paragraph.  To that one must add masses for crew and 2 space suits each,  their personal effects,  and personal equipment (call it 0.5 metric ton per person as a guess),  and for fully-expendable supplies of food,  water,  and oxygen (call it 0.75 metric tons per person per month,  knowing that these are just “reasonable guesses”).  Crew and supplies must fit within the vehicle,  which has (for the 6 modules not filled with flywheels and heavy equipment) some 2100 m3 volume. 

If one assumes half the volume is packed supplies,  and also a crew of 6,  that leaves some 175 m3 per person as living space available.  That’s about like 3 large living rooms in a typical middle-class house.  That seems adequate at first glance,  if it is well distributed,  and some part of it is reconfigurable at some level.  

The crew weight allowance is 3 metric tons,  and the packed supplies mass is about 4.5 tons per mission month.  If the mission is 31 months long (9 months transit,  13 months at Mars,  9 months return),  that’s about 140 tons of supplies,  with no margin for error.  So call it a nominal 150 tons.  This presumes no recycling or growing-of-food in space or on Mars.  It’s a worst-case deal,  but we can do this “right now”.   

So,  the empty hab section is estimated at 160 tons.  It gets loaded with about 150 tons of supplies,  allowing for 7.5% safety factor on supply mass,  and loaded with about 3 tons of crew with their suits,  equipment,  and personal effects.  Fully loaded,  that’s 313 tons.  That would be crew of 6,  and supplies for a 31 month mission plus a small margin.  See Figure 10.  Figure 11 shows an image of the spreadsheet where these numbers were calculated.  Yellow highlighting denotes inputs.  Some selected outputs are highlighted blue. 

Assumed depleted at a constant rate,  the supplies total 150 tons at departure,  109.5 tons at Mars arrival,  51.0 tons at Mars departure,  and not-zero at 10.5 tons at Earth arrival,  assuming the safety margin is not consumed.  This presumes wastes are dumped overboard with no recycling at all!   This dumping reduces the effective mass of the hab section,  at each mission segment,  a benefit to propellant required. 

We can already do somewhat better than that with recycled water,  but this is a worst case estimate!  Yet this open-cycle assumption gets the smaller propellant supply for return to Earth.  “Efficiency” is not always beneficial:  that is too often presumed erroneously!  Jettisoned mass reduces next-burn propellant requirements.  That’s just physics you cannot fight!

Sizing the Manned Transport and Its Return Propellant

The fundamental notion for sizing propellant supplies for the four events (Earth departure,  Mars arrival,  Mars departure,  and Earth arrival) is that the mass of the loaded,  crewed hab,  plus the mass of all propellant tankage,  plus the mass of the engines,  is the ignition mass.  That minus the mass of propellant burned from that tankage is the burnout mass.  That produces a mass ratio for the burn,  and the delta-vee it will produce,  which must meet or exceed the requirement for that burn.  This is subject to the constraint that we want 0.1 gee or thereabouts as a min vehicle acceleration at each burn.

To do this,  one must estimate the ratio of propellant to loaded tank mass for the added tankage.  This has to reflect a long,  slim tank geometry for docking multiple tanks around the periphery of the hab,  and it must account for the mass of the docking structures needed to achieve that result.  As a guess,  I am assuming that the empty tank inert mass (with all those fittings) is 5% of the loaded tank mass,  so that the contained propellant is 95% of the loaded tank mass. 

To that end,  I used a series of calculation blocks in a spreadsheet worksheet to run the calculations.  Again,  inputs are highlighted yellow,  and significant outputs are highlighted blue.  Figures 12,  13,  and 14 show the results. 

Bear in mind that the loaded tank mass for the Mars and Earth arrival burns must be part of the “payload” for the Earth and Mars departure burns,  respectively.  They are unique in this way.  That means the dead-head payload is the appropriate hab mass plus the mass of the next burn’s loaded tanks.  The current burn’s tanks must push this (plus the added engine mass) to the required delta-vee for that burn. 

Added engine mass is handled by an iteratively-applied tankage scale-up factor just slightly over unity.

As it turns out,  finding the propellant tankage mass to push the hab to the required delta-vee is not an excruciating iterative process.  You first find the mass ratio MR that is required from the required mass ratio-effective delta-vee,  and the propulsion’s effective exhaust velocity,  by the rocket equation.  Ignoring the mass of the engines themselves,  it turns out to be closed-form to find the loaded tankage mass Wtf from that mass ratio,  and the total “dead head” mass to be pushed in that burn. 

For both departures,  the “dead head” mass is the appropriate loaded hab mass plus the loaded mass of the corresponding arrival tankage.  For both arrivals,  the “dead head” mass is just the loaded hab mass.  This can be corrected at the 1 or 2% level for total engine mass later,  to ensure fully meeting the delta-vee requirements,  simply by scaling up the loaded tank mass Wtf with a factor applied iteratively until delta-vee produced meets the requirement.   

Wtf = Wdead (MR – 1)/(1 – MR f)  where f = Wt/Wtf and Wt is dry tank mass

That’s the orbital transport rough-out design for Mars.  It can get there to low Mars orbit from low Earth orbit where it was assembled.  It can rendezvous with its Earth return propellant,  the Mars landers,  and the Mars lander propellant supply,  all three of which were sent ahead by electric propulsion.   The nonrecoverable items are the Earth departure stage and the empty Mars departure tanks.  The empty Mars arrival tanks are left in Mars orbit.  Everything else about this design is recovered in low Earth orbit.

Note that this ship is 1413 metric tons,  as assembled and loaded in low Earth orbit,  ready to go to low Mars orbit.  Its use requires that some 997.26 metric tons of loaded propellant tanks be sent ahead to Mars for the return propellant supply.  In order to actually make landings on Mars as staged out of low Mars orbit,  the landers and their propellant supply must also be sent ahead to low Mars orbit. 

With much bigger propellant tankage,  this same design could take men to a near-Earth asteroid.  For such missions,  landers are not needed,  and there is no practical opportunity to pre-position return propellant,  except many years ahead.  Those missions are far more difficult.  Analysis of one is not attempted in this posting.

Sizing the Lander and Lander Propellant Supply

The lander payload is its crew,  their suits and personal equipment,  plus an amount of life support supplies that depends upon how long the crew will live in the lander on the surface,  each landing.  The de-orbit burn for a surface-grazing ellipse is a trivial 50 m/s delta-vee.  Most of the deceleration is aerodynamic drag,  effectively terminating at end-of-hypersonics at Mach 3,  just about 1 km/s velocity,  but at a low altitude because of the high ballistic coefficient.  That altitude is only about 5 km

From there,  deceleration is by retropropulsion alone,  with a large “kitty” to cover hover and divert requirements.  Assuming 1 km/s velocity at 5 km altitude,  along a straight slant trajectory at 45 degrees,  the average deceleration level required is 70 m/s2,  or 7.211 gees,  which with the lander mass,  sets the required engine thrust level for landing.  That is a rough ride,  about twice the rigors of return from low Earth orbit,  and justifying all by itself the maintenance of full crew health with artificial spin gravity!

The lander is a one-stage reusable “landing boat” intended to make multiple flights,  each fueled from a propellant supply sent with it to low Mars orbit.  Factored,  the mass ratio-effective descent delta-vee is just about 1.5 km/s.  Propellant is storable NTO-MMH,  to preclude evaporation losses and massive energy requirements to prevent freezing or boiling.  The ascent must account for small but non-zero gravity and drag losses (about 2% of velocity),  and a “kitty” for rendezvous maneuvers.  That mass-ratio-effective delta vee is just about 3.62 km/s. 

The payload requirements for crew,  equipment,  and supplies as a function of surface duration are given in Figure 15,  along with a crude estimate of the “larger-than-minimum” vehicle inert weight fraction that is appropriate to the necessary structural robustness,  and to the equipment required to function as a reusable entry-capable vehicle,  and as a surface habitat.  Conceptually,  the lander is sketched in Figure 16.  Some of its backshell panels double as cargo load/unload ramps.  Most of the cargo space can be isolated and pressurized as living space,  once unloaded.  The piloting cabin is the abort capsule,  something somewhat similar to a crew Dragon from Spacex.  This thing is NOT a minimalist lander the way the Apollo LM was.  

The ascent payload is smaller,  since most (but not all) the supplies are used up (and wastes left behind) at ascent liftoff.  There is a generous allowance for Mars samples to be returned to the orbital transport. This has to be taken into account in calculating the actual vehicle masses,  since the two delta-vees are handled at two different payload fractions,  in the one vehicle design.  That process is inherently iterative,  as shown by the data given in Figure 17.  

In order to determine these numbers,  one guess a value for the max lander mass,  which is ignition-at-descent (Wig-des).  The inert fraction times this gives the vehicle inert mass Win.  The ascent and descent payloads are determined vs mission surface duration separately.  The mass ratios already determined are used to estimate propellant masses. 

The ascent propellant mass Wp-asc is determined first as (MR-asc – 1)(Wpay-asc + Win),  then the descent propellant mass Wp-des as (MR-des – 1)(Wpay-des + Win + Wp-asc),  treating the ascent propellant as part of the effective “payload” during descent.  The descent payload plus both propellant masses plus inert mass sum to the result for descent ignition mass. 

The input guess for descent ignition mass is then adjusted iteratively,  until it converges to the result for descent ignition mass.  This is done by simple trial and error in the spreadsheet.  There is such a result computed for each of 4 possible surface durations that divide evenly into the 12 months available.  These results are then the inputs for a characterization of the lander sizing as a function of design surface duration. 

For the selected 2-month duration (corresponding to 6 total lander flights),  those results are given in Figure 18.  These show ascent and descent weight statements,  confirmation of delta-vee capability,  and characterization of vehicle mass fractions,  plus the propellant supply required to cover the appropriate number of flights.  Similar tables exist in the spreadsheet for the other 3 durations,  but those are not shown here.   

Figures 19 and 20 show the trade-off of vehicle sizes and propellant supply sizes versus surface duration options.  The selected design is near the “knee” in the curve of number-of-flights vs surface duration,  at 2 month duration for 6 flights.  For shorter duration,  the required propellant supply is significantly larger.  For longer duration,  the required propellant supply is smaller,  but not so significantly smaller. 

The lander size itself is significantly affected by the design surface duration,  being larger at longer duration.  The 2-month duration selected limits this affect,  without so significantly penalizing the payload fraction (which ranges from about 2 to about 3%).  The selected 2-month duration is also near the “knee” in that curve.  Longer durations do not improve this as much as was gained going from 1 month-12 flights to the selected 2 month-6 flights option. 

For this selected design (6 two-month surface stays),  three landers fueled and loaded with supplies,  less crew,  suits,  and personal equipment,  each massing 376.5 metric tons,  must be sent to Mars along with some 1764 tons of propellant to support all 6 flights.  If 95% of the tank weight is propellant,  the mass of loaded tankage to be sent is some 1856.8 metric tons.  If sent as tanks docked to each of the 3 landers,  that’s a 376.5 ton lander plus 619 tons of loaded propellant tanks. 

The “smart” thing to do from a reliability / self-rescue standpoint is to send the transport return propellant with those same three landers,  so that if one is lost,  the transport can still return safely by drawing the shortfall instead from the remaining lander supplies.  That return propellant was determined above to be 997.26 metric tons of loaded tanks.  Divided by 3,  that’s an additional 332.4 metric tons of Earth return propellant tankage sent to Mars with each lander. 

That makes each lander plus propellant tanks a 1327.9 metric ton item to be moved by solar electric propulsion from low Earth orbit one-way to low Mars orbit.  Each such is thus quite comparable to the departure mass of the manned orbital transport.  That would not be true at the other surface durations. 

There are 6 landings to be made,  and three such landers sent to Mars.  Distributed evenly,  that is two flights per lander minimum,  and 6 maximum.  Bear in mind that only one lander is sent to the surface at a time,  carrying a crew of 3,  while the other three crew do science in orbit,  while acting as the safety rescue “watchdog”,  with at least one functional lander,  even if the other one fails.  The worst case is that all 6 flights are made with one lander.  Thus the lander design must allow for at least 6 flights per vehicle,  justifying in part the higher inert mass fraction used in this design rough-out.

Landers get left in low Mars orbit at mission’s end,  when the transport departs for Earth.  Subsequent missions might utilize these assets,  and reduce the sent mass to only more lander propellant.  That possibility argues for much more than 6 flights per vehicle,  in turn a really good argument for the very robust inert mass fraction of 20% used here. Alternatively,  they could be landed robotically.

Common Engine Design for Transport and Lander?

The lander mass is 378 metric tons at ignition,  and 241 at touchdown,  as just determined above.  The average is 309.5 metric tons.  Also as determined above,  the average deceleration required is 70 m/s2.  That translates to 21,665 KN of retropropulsion thrust required to safely land (nominally 22,000 KN).  This is totaled for multiple engines.  Less may be used for ascent,  as such high gee capability is not required for that.  Something nearer 2 gees at ascent ignition mass 236.3 metric tons (4726 KN thrust) is more appropriate.

As described above,  something near 1170 KN thrust from multiple engines is the minimum required for the orbit-to-orbit transport.   This was set by the min 0.1 vehicle gee capability at max vehicle mass,  and still resulted in only large fractional gee capability at min vehicle mass.  This thrust level selection could be doubled or tripled (or more) with relative impunity.  

A worksheet page was set up in the spreadsheet to explore how this could be done,  in two steps.  The results are shown in Figure 21,  which indicate the possibility of using some number of 3600 KN max thrust engines,  throttleable from 20 to 100%.  In the first step,  I input factored thrust requirements,  plus a number of engines,  and a max number of inoperative engines. 

The thrust requirement for the lander descent is based on slowing the average descent mass (as a constant) from 1 km/s to zero,  in a slant path length of 7.1 km,  using the oversimplified kinematic equation V2 = 2 a s.  This is a very high-gee descent!  Reducing that requires not just supersonic retropropulsion,  but hypersonic retropropulsion (starting retropropulsion earlier in the entry sequence).  It is an inevitable consequence of the high ballistic coefficient producing very low altitudes (on Mars) for end-of-hypersonic deceleration.  This is an area for further design work!

The thrust requirement for the lander ascent is its Earth weight,  factored-up just slightly,  to accommodate flight tests on Earth.  That’s “overkill” for Mars with its lower gravity.

The thrust requirement for the orbital transport is based on its Mars departure mass (largest of the masses under storable propulsion) and a min 0.1 gee vehicle acceleration requirement.  This is arbitrarily factored-up by 3 to achieve commonality,  without exceeding max gees ~ 2 at Earth arrival.

That initial result indicated that something like 3600 KN max thrust per engine would be suitable,  with 9 engines in the lander operating at part throttle in descent,  and 4 engines operating at part throttle in ascent,  able to lose up to 3 engines either way,  and still function within limits.  This was explored further,  looking at vehicle gees and engine throttle percentages,  in the second step. 

Up to 3 of these lander engines could cease operation during ascent or descent.  The remainder could supply adequate thrust at 100% throttle or less,  without waiting for lightoff of any inactive engines. That’s an important safety consideration,  which ethics demands!  Two of these same engines would be adequate to push the orbital transport at part throttle,  with only one operating engine still able to supply much more than the demanded minimum thrust.

In all cases,  engines operate between 20 and 100% throttle setting,  and appropriate gee limits are not exceeded.  Min transport vehicle gee requirement (0.1 gee) is exceeded. 

For descent,  the lander retropulsion operates between about 6 and about 9 gees.  This event is only about 14-15 seconds long!!!  “At the last second” to actually land,  some 8 of the 9 engines must be shut down to reduce thrust to nearer Mars weight of the lander (about 749 KN to 872 KN,  depending upon how much propellant was burned) at touchdown,  with the remaining active engine operating at about 21-24% thrust setting.  This single-engine point is the riskiest aspect of the landing,  but it is mitigated by the facts that (1) this engine is already operating,  and (2) it need only continue to operate at reduced thrust for a second or two.

On ascent with a reduced number of engines,  this is 1.2 to 3.6 gees for the lander at full thrust,  far more than is needed to depart against Mars gravity (only 0.38 gee).  Active throttling reduces that some.

The transport operates between 0.3 and 1.8 gees during the return to Earth.  This exceeds the min acceleration requirement,  but not the maximum.  A 3600 KN engine design for this NTO-MMH common engine would resemble the notional sketch in Figure 22. 

If the Earth departure stage at 1350 KN uses 5 engines,  each would be approximately 1350 KN max thrust capability operating at 20% thrust.  Up to 4 could be non-functional,  and still easily meet the overall min departure thrust requirement,  without exceeding 100% throttle.  Higher vehicle acceleration than 0.1 gee is easily obtained,  but even with all 5 engines at full thrust,  it is still only fractional gee.  Such a 1350 KN LOX-LH2 engine would resemble the notional sketch in Figure 23.

Sizing the SEP for the Unmanned Assets Sent Ahead

This item is the most speculative,  because (1) it uses the most assumed data,  and (2) this kind of solar electric propulsion has yet to be scaled up to such sizes to push masses this large.  To cover the gravity losses (both planetary and solar),  I simply doubled the required orbital delta-vee data. 

I simply assumed the average characteristics of small Hall effect thrusters operating on iodine could be scaled way up by simple clustering,  at the same thrust/weight and thrust/power ratios.  And,  I just assumed the characteristics of satellite-sized solar panels could be scaled up to the low-hundred kilowatt range at the same power/area and weight/area ratios.

My approach was a self-contained solar-electric propulsion (SEP) “tug”,  that incorporates the clustered thruster unit,  the solar panels to power it,  sized for reduced sunlight at Mars,  a robot guidance package,  and a low-pressure “tank” to contain the easily-sublimated  and inexpensive iodine propellant.  I used published data for two Busek Hall-effect thrusters,  and for a couple of Alta Devices satellite solar panels,  for these estimates. 

This SEP “tug” is coupled to a dead-head payload for the trip from Earth orbit to Mars orbit,  using all of its 120 clustered SEP thrusters to achieve a milli-gee of vehicle acceleration capability at Earth departure.  That payload is one (of the three) Mars landers (fully fueled and supplied),  plus a 1/3 share of the total lander propellant supply,  and plus a 1/3 share of the manned orbital transport’s Earth return propellant supply.  This dead head payload is over 1300 metric tons.

For the return trip (these “tugs” are fully reusable),  there is no dead-head payload,  only the “tug” and its iodine tank,  still containing just enough iodine propellant to get home.  During the trip home,  only one SEP thruster in the cluster need be used to achieve near a milli-gee of vehicle acceleration at Mars departure.  That leaves many “spares in case the one fails”,  insuring utter reliability.  (Outbound,  the cluster is large enough that the loss of a few thrusters is no significant percentage loss of thrust.)

The size of one such thruster (200 mN,  mN meaning milli-Newtons) falls within the range of thrusters produced today.  This produces adequate acceleration of the unladen vehicle.  The scaleup is by clustering,  not by increasing the size of the thrust in such a device.  The clustering-together of 120 of these units produces some 24,000 mN,  needed to move the laden vehicle at adequate acceleration. 

The resulting SEP “tug” design is depicted in the sketch of Figure 24.  I used a big two-stage spreadsheet worksheet to iteratively size this “tug” system,  examining the 4 “burns” individually.  The second stage of this process fully defines the characteristics of the “tug” and its estimated performance.  This is the tabular data in the partial spreadsheet image shown in Figure 25. 

Hopefully,  this rough-sizing is “overkill”,  due to my just-assumed doubling of the orbital delta-vee requirements.  The intent here is to slowly spiral-out of low Earth orbit to escape,  and continue an accelerating spiral about the sun to an appropriate midpoint,  then use a decelerating spiral about the sun toward capture at Mars.  From there,  it follows a decelerating spiral-in to low Mars orbit.  The return uses the same spiraling processes,  just unladen of dead-head payload,  and at far-lower thrust and propellant requirements. 

Sizing the Earth Departure Stage

Of all the items analyzed,  this is the easiest and most straightforward,  because there is one and only one burn (the Earth departure burn).  Then this stage is jettisoned.  The stage layout concept and sized data were already determined as part of the orbital transport propulsion sizing above.  These data were given as part of Figures 12,  13,  and 14 above,  plus part of the common engine discussion just above,  with sized engine dimensions in Figure 23.

Just to summarize,  the departure stage has 5 LOX-LH2 engines each designed for 1350 KN thrust,  weighing an estimated total of 5.139 metric tons.  The stage comprises LOX and LH2 tankage whose combined dry weight is 41.906 metric tons.  The total propellant load is some 796.210 metric tons.  Thus the loaded stage itself is some 843.255 metric tons.

This stage pushes a fully loaded and crewed hab plus Mars arrival propellant tankage that totals some 569.810 metric tons of dead-head payload.  Total orbital transport vehicle mass,  at Earth departure ignition,  is thus some 1413.065 metric tons.  This was shown in Figure 14 above,  including weight statements and performance.

Not considered here is reuse of the Earth departure stage.  Its engine sizing would be fine,  but it needs larger tanks and propellant to accomplish 2 burns.  The first is to put the orbital transport onto a Hohmann transfer ellipse trajectory. 

After releasing the transport,  it burns a second time to enter an ellipse about the sun with an exactly two-year period.  That way the Earth is there when it reaches perihelion,  thus making recovery feasible at all.

It is just not reasonable to expect that cryogens like LOX and especially LH2 will not completely evaporate away over a 2 year interval.  Therefore,  the reusable form of the stage must also incorporate a second propulsion system storable over long periods.  This added propulsion provides the delta-vee to return to Earth orbit from the 2-year solar orbit perihelion conditions.  

Being unmanned,  there is no reason this second propulsion system could not be solar-electric using iodine.  The stage then executes a spiral-in to low Earth orbit after capture.  The alternative is storable propellants like the NTO-MMH. 

Being out of scope here at this time,  these designs have not been explored.  Consider that as a future upgrade.

Totaling Up the Mission and Its Launch Requirements

This mission to Mars requires a fleet of 4 vehicles to be sent from Earth orbit to Mars orbit.  One of these (the manned vehicle) returns to Earth.  The other three are unmanned assets sent ahead earlier by electric propulsion,  for the crew to utilize when they arrive by conventional rocket propulsion. 

The three unmanned vehicles are identical,  comprising a dead-head payload and a reusable solar-electric “tug” that returns to Earth for reuse,  after delivery of the dead-head payload into orbit at Mars.  

That dead-head payload payload is the same for each of these vehicles:  an uncrewed but loaded and fueled reusable Mars landing boat,  plus 1/3 of the total Mars lander propellant supply,  plus 1/3 of the crewed vehicle’s Earth return propellant supply.  That dead-head payload is 1327.9 metric tons for each of these 3 vehicles.

Each of these three unmanned vehicles totals some 2413.5 metric tons as assembled in Earth orbit,  that being the dead-head payload plus the fueled SEP “tug”.

The crewed vehicle (the orbit-to-orbit transport) comprises the crewed and loaded hab section,  plus the loaded Mars arrival propellant tankage,  plus the expendable Earth departure stage that uses cryogenic propellants.  (All the other rocket propulsion uses the same storable propellants,  and the SEP “tugs” use sublimable iodine to keep the iodine “tank” weight down.)  Ready to depart Earth orbit,  the transport and departure stage total some 1413.065 metric tons. 

The grand total that must be assembled in orbit for the fleet of 4 ships is some 8653.6 metric tons.  For that,  you get 6 landings at up to 6 different places on Mars,  all in the one manned trip to Mars.  That’s 1442.3 tons to support each of the 6 landings,  essentially.  These are 2-month max stays at each landing site. You get all this,  plus a “way out” or a self-rescue capability built into the mission at every step,  plus a fully-healthy crew with radiation shielding and artificial gravity during the transits,  and in low Mars orbit. That’s a lot of benefit for the cost.

Getting Landers To Low Earth Orbit

The selected lander design is just about 378 metric tons,  crewed,  loaded and fueled.  Less crew (and their suits and gear),  that’s just about 376.5 metric tons.  Just about 294 tons of that lander weight is propellant.  So,  a loaded,  crewless,  empty-of-propellant lander is just about 82.5 metric tons.  Remove the supplies,  but leave the surface equipment and rover aboard,  and this is about 77 tons.  Completely unladen,  the lander is about 75.6 tons. 

I looked at SLS (150 metric tons to LEO,  guessing $1,000M per launch),  Spacex’s “Starship” (100 metric tons to LEO,  guessing $150M per launch),  Spacex’s Falcon-Heavy (63 metric tons to LEO flown expendably,  about $85M per launch),  ULA’s Atlas-V (20 metric tons to LEO at about $85M per launch),  and Spacex’s Falcon-9 (20 metric tons to LEO flown expendably,  and $63M per launch). 

The loaded unfueled lander mass of 75.6 metric tons is out of reach of Falcon Heavy,  much less Atlas V or Falcon 9,  even if an 8-meter payload diameter could be flown on any of them.  NASA’s SLS might possibly launch it dry of propellant,  maybe even two of them at once,  although it has yet to fly.  That would be 2 or 3 flights of SLS at $2-3B to put 3 landers into orbit,  unladen of propellant.  It would be 3 flights of “Starship” at $450M total.  The most cost-effective of those two options is “Starship”.  3 “Starships” deliver 3 landers loaded but unfueled. 

At 294 tons of propellant per lander,  and 100 tons per “Starship”,  some 9 “Starship” tanker flights would be required to fuel them fully up.  At 150 tons per SLS,  some 6 SLS flights would be required to fuel them up fully.  At about 60 tons per flight,  some 5 Falcon Heavy flights could be those tankers per lander,   for some 15 Falcon-Heavy flights to fuel the 3 landers up.  At 20 tons per flight,  it would require some 45 flights of Falcon-9 or Atlas-V to fuel the 3 landers in orbit.  The most cost-effective way to deliver these bulk liquid propellant supplies turns out to be 9 “Starship” flights,  with 15 Falcon-Heavy flights a rather close second.  If “Starship”,  the transfer crew need not be sent up separately.

Getting Earth Return and Lander Propellant Supplies to LEO and Docked

Remember,  we must send to Mars each lander loaded and fueled,  plus 1/3 of its Mars landing propellant supply,  plus 1/3 of the transport’s Earth return propellant supply.  These propellant supplies are pre-loaded tanks.  They are 1764.1 tons for the lander operations,  541.3 tons for the transport’s Mars departure,  and 455.9 tons for the transport’s Earth arrival.  That totals some 2761.3 metric tons of propellant,  which must be in tanks,  at about 95% propellant and 5% tank inert.

Unconstrained by other considerations,  I chose to break this up into nominal 60-ton loaded tanks.  The lander supply is 31 of these,  the Mars departure supply is 10 of these,  and the Earth arrival supply is 8 of these.  That’s a total of some 49 tanks to deliver to LEO,  at 60 metric tons each.  The most cost-effective way to do this was 49 flights of Falcon-Heavy,  flown expendably.

We will need a docking crew on-orbit for about a week max to assemble the docked cluster for each of the landers.  This can be a crew of 2 to 4 in a Crew Dragon atop a Falcon-9.  This probably will not happen in parallel for the 3 landers,  but serially.  So plan on 3 manned Falcon-9 launches to support these assemblies.

               Getting the Transport to LEO,  Loaded,  and Assembled

The orbit-to-orbit transport goes up as separate modules (without supplies) to be docked in orbit.  There are six 20-ton modules and one 40-ton center modules,  complete with solar wings that must unfold.  All the listed boosters could launch the 20-ton modules,  only Falcon-Heavy,  “Starship”,  or SLS could launch the 40-ton module.  The most cost-effective means was a tie:  2 flights of “Starship” or 3 flights (expendable) of Falcon-Heavy deliver these 7 modules to LEO. 

There is about 150 tons of supplies,  crew suits,  and crew personal equipment to deliver to the transport and load inside (152 exactly,  per these admittedly-uncertain estimates).   This is separable into lots deliverable by any of the boosters listed.  From a cost-effectiveness viewpoint,  this was another tie:  2 flights of “Starship”,  or 3 expendable flights of Falcon-Heavy. 

This is going to require a temporary docking and loading crew of perhaps 4 to 6 astronauts for a week or so in orbit.  If we send them up in two Crew Dragon capsules atop Falcon-9 boosters,  they can come home in one,  and leave the other Crew Dragon docked to the transport as one of its emergency return escape craft.  Add 2 Falcon-9 flights for the transport assembly crew unless “Starship” is used instead.

Getting the SEP “Tugs” to LEO and Fueled

The SEP “tug” hardware,  empty of the solid iodine fuel,  are not heavy at all.  This crude estimate says they are 14.42 tons each,  and there are 3 of them. That includes the folded solar panels,  the big thruster array,  the guidance package,  and the empty tank which doubles as the vehicle core structure,  about which dead-head payload gets docked.

Any of the listed boosters can get an empty tug to LEO.  The most cost-effective means is 3 Falcon-9 launches,  possibly flown recoverable,  but the expendable price was used here.

The iodine thruster fuel is a sublimable solid,  which can be sent up in portions that fit the various boosters,  determining the number of flights.  For the three tugs together,  we need 3213.54 metric tons of iodine sent to LEO.  (Most of this,  by far,  gets used sending payload to Mars.  Only a few tons with only 1 thruster firing is needed to return to Earth.)

Any of the listed boosters can do this job.  The most cost-effective means is by “Starship”,  with Falcon-Heavy a close second.  That would be 33 “Starship” flights,  or 54 Falcon-Heavy flights flown expendably. 

It will take a crew of 4-6 astronauts to load the iodine fuel and unfold the solar arrays,  plus some checkout.  We probably do not do all 3 vehicles in parallel,  but serially.  If by “Starship”,  that vehicle can carry the crew.  If by Falcon-Heavy,  a separate Falcon-9 launch is needed to send this crew up for a week or two in orbit as the payloads arrive,  which is a huge Falcon-Heavy flight rate!  “Starship” with payload and loading crew aboard is thus the preferred way,  by far.

Getting the Earth Departure Stage to LEO and Fueled

This is assumed an empty stage delivered as one piece of hardware at 47 metric tons,  plus 796.2 metric tons of LOX-LH2 propellants delivered as bulk liquid.  Bulk liquids can be delivered in multiple payloads by any of the listed boosters,  but requires special tankage and a human crew to do the transfers. 

The most cost effective way to deliver the empty stage is by a single Falcon-Heavy,  possibly flown recoverably,  but priced expendably for this analysis.

The most cost-effective means to deliver bulk propellant is 8 “Starship” flights,  followed fairly closely by 14 Falcon-Heavy flights.  These require crews,  which can be aboard the “Starship” flights.  They would have to come up in some 14 Falcon-9 launches with Crew Dragon if Falcon-Heavies were the propellant ferries.  By far,  the preferred approach is 8 crewed “Starship” flights.

Getting the Crew Onto the Transport for the Mission

The Mars mission crew is only 6 people.  This is one Falcon-9 Crew Dragon flight to send them up.  Their Crew Dragon docks with the transport to be its second (and redundant) emergency escape capsule. If not covered earlier,  make this 2 flights so there are two Crew Dragons as escape capsules.

Totaling Up Mission Launch Requirements & Guessing Costs

I totaled-up the launch costs for this mission.  On the assumption that launch costs are 20% of overall program costs,  that puts this mission in a rather modest cost category,  despite the large tonnages.  That is precisely because it does NOT use SLS to launch anything,  at a billion dollars per flight (if not more)!  See Figure 26 for a summary of the launch requirements and costs.  The basis for comparison is the infamous “90 Day Report”,  based on mounting essentially “Apollo-on-steroids-plus” as executed by the long-favored contractors,  to send a crew of 4-to-6 to one site on Mars,  in the one trip. 

Totaling Up What the Mission Accomplishes

This makes the comparison to the “90 Day Report” even more stark.This mission as planned has a “way out” or a self-rescue capability at every step,plus inherently designed-in artificial gravity and radiation protection (to include solar flare events). The likelihood of this crew returning alive and healthy is actually quite high.In comparison,with the “90 Day Report” mission,that likelihood is rather low,because it does not offer those characteristics.

What this mission accomplishes is up to 6 different sites explored in the one manned trip to Mars.With the “90 Day Report” mission design,only one site gets explored.

This mission leaves considerable usable assets at Mars for future missions to utilize.That would include the reusable landers,either in low Mars orbit,or on the surface if landed robotically.Plus, there might be some leftover propellant,probably in Mars orbit.The “90 Day Report” mission leaves few (if any) usable assets on Mars for future missions to utilize:maybe a surface habitat structure and a rover or two,and possibly a nuclear power supply item.

See Figure 27 for a listing of what this mission accomplishes,compared to that of the “90 Day Report”.

“Bang-for-the-Buck” Discussion

The first gross indicator is program cost for the one trip to Mars,  divided by the number of sites explored while the mission is there.  For my mission design,  cost per site ranges from $11.7B/site to at most $70.3B/site,  depending upon whether the minimum 1 or maximum 6 sites get explored.  That is factor 6.4 to 38.5 times better cost per site than that of the “90 Day Report”. 

The second gross indicator is the likelihood of getting the crew back alive and healthy.  Because of the features demanded by ethics,  and designed-in from the start,  this mission plan can truthfully claim a high likelihood of accomplishing this.   The “90 Day Report” mission plan cannot truthfully claim that. 

For one thing,  there is no rescue for a crew stranded on Mars.  For another,  there is a high likelihood of a solar flare event during a 31 month mission,  and almost zero chance of surviving that event with no radiation shelter.  And yet another:  there are two 9-month transits in zero-gee,  separated by a 13 month stay on 38% gee Mars,  with undetermined therapeutic effect,  if any.  Should an emergency free return at Earth arrival be required,  that is a high-gee event (likely 10+ gees).  A crew weakened by microgravity diseases is unlikely to survive this.

Now remember,  spaceflight history clearly demonstrates that there is nothing as expensive (economically and politically) as a dead crew.  Especially one dead from a bad management decision.  My mission design raises crew survival probability,  the “90 Day Report” mission design does not;  that survival probability is quite low,  if one is truthful about it.

In order to get both high “bang for the buck” and a high likelihood of getting a crew back healthy,  I had to think way outside the usual boxes.  One of those boxes is “nothing can look much different than what we already did during Apollo,  shuttle,  and ISS”.  Another is “no mission can be affordable if there must be a high tonnage launched”.  A third is “you simply must do direct entry at Mars to save launched tonnage”.  A fourth is “you must use SLS no matter what in order to launch this mission”.

All proved to be false constraints on thinking.  The only one that is true is the one I used:  crew survivability above all,  driven by basic ethics.  In a nutshell:  “provide a way out or a self-rescue capability at every single step”.  That drove me to orbital-based exploration and a manned orbit-to-orbit transport design.

The main possible weakness of my mission design is the low payload fraction of my one-stage reusable landers:  around 2%.  A one-shot two-stage design would have a far higher payload fraction (perhaps 5-6% if you include the safety-required abort capsule,  only higher if you fail this safety requirement),  resulting in a smaller mass sent to Mars for each lander.  But I would have to send more of them (8) to maintain a rescue capability and a spare,  and still visit as many as 6 sites.  This I leave to others to explore.

Final Comments

In terms of both cost and safety,  the comparison of this mission plan to that of the “90 Day Report” demonstrates the unattractiveness of the usual way NASA did things in the past.  There is far more “bang for the buck” and an enormously-higher probability of getting the crew back alive and healthy in my plan.   Not only that,  my program cost is far,  far lower.

The astute reader will observe that I have selected a lot of Spacex hardware as the most cost-effective means to launch and assemble this mission.  That begs a comparison to the Spacex plan just to send multiple “Starships” to Mars by direct entry from the interplanetary trajectory.  According to the presentations released,  that would be about 6 “Starships” initially landed on Mars,  with probably one or at most two of them eventually returning to Earth,  if the local propellant production works,  and it can supply them fully and quickly enough. 

It is as yet unclear whether 5 or 6 “Starship” tanker flights are required for refueling each interplanetary “Starship” in LEO for the journey to Mars.  So somewhere between 36 and 42 total “Starship” launches are required to support their mission.  Using $150M per launch,  and launch costs equal 20% of program cost,  that’s $5.4-6.3B launch cost,  and $27.0-31.5B program cost,  to put their mission onto Mars. 

That program cost scaleup is real,  even for them,  because they are counting on others to supply the local propellant production hardware,  local rover vehicle capabilities,  and local life support capabilities (cannot live in the landed “Starship” forever !!),  not to mention local electric power.  They have their hands full just developing the “”Starship” vehicle.

That’s comparable to my costs,  and (like me) way below the costs in the “90 Day Report”.  The differences are many,  however.  They explore only 1 site,  period.  If the local propellant production fails to meet expectations,  nobody comes home.  They say they will supply radiation sheltering,  but not artificial gravity.  They are counting on Mars’s 0.38 gee being “therapeutic enough”,  when in point of fact,  nobody yet knows that to be true. My mission takes none of those risks and explores up to 6 sites.

There is no aborting or bailing-out during the “Starship” direct entry at Mars.   There is no aborting or bailing out during the landing and touchdown.  They have yet to address soil bearing loads vs landing pad size for Mars,  or rough field landing hazards such as slope,  local roughness   and big boulders.  There is no bailout or abort during the return ascent.  There is no bailout or abort for the direct entry at Earth return.  There is no bailout or abort during the Earth landing and touchdown.    A failure during any one of these events is inevitably a loss of the vehicle and everybody aboard.  My mission takes none of those risks.

Yeah,  you can save the money using “Starship” as the transit vehicle (by about a factor of 2-3 over my plan).  But you are also very much more likely to kill one of your crews if you do (which also very likely would put a stop to the ongoing mission). 

Ethics-driven spaceflight design “from the get-go” seems the more prudent course,  especially when you consider the consequences of killing a crew. 


#1. NASA radiation website, titled Spaceflight Radiation Health Program at JSC (no cited reference newer than 1992).

#2. From Bigelow Aerospace website  as of 3-7-17

GCNR Spacecraft

RocketCat sez

You want an atomic rocket? I'll give you an atomic rocket!

Yeah, yeah, this ain't an over-the-top torchship like an Orion Drive ship much less Zubrin's outrageous Nuclear Salt Water Drive. But it is a good working-man's atomic rocket that has the horsepower to Get The Job Done. Orion drives are for battleships, this one is a space trucker hauling cargo.

Bloated chemical drives can barely do a Mars mission in two years, this little atomic number can smoke the mission in 80 days flat! I know that saying the exhaust is radioactive is putting it mildly, but nobody is near enough to it to be harmed (well, except for the poor working-class slobs who are the ship's crew).

Bottom line:

  • It is undisputably an Atomic Rocket

  • It has both high thrust and high specific impulse, approaching torchship levels

  • The design does a clever end-run around the "melting reactor" problem with a solution both elegant and brute force

  • 80 day round-trip to Mars, man! How cool is that?

GCNR Spacecraft
Specific Impulse2,500 to 6,500 s
Exhaust Velocity24,500 to
63,800 m/s
Mass Flow0.8 to 6.7 kg/s
Thrust20,000 to
430,000 N
Fixed Thrust224,000 N
Thrust Power0.25 to 13.7 GW
Initial Accel0.01 to 0.05g
GCNR Spacecraft
Mars Courier
80 days
Wet Mass950,000 kg
Dry Mass290,000 kg
Mass Ratio3.28
Thrust150,000 N
Initial Accel0.016 g
Specific Impulse5,500 s
Exhaust Velocity53,955 m/s
ΔV64,100 m/s
H / 235U Ratio200:1
235U Fuel3,300 kg
660,000 kg

Data from Gas Core Rocket Reactors - A New Look.

This little hot-rod can do a round-trip mission to Mars in 80 days flat! That's only 2.7 months. Using Hohmann trajectories a round-trip Mars mission will take 32.3 months (2.7 years) when you take into account the wait for the Mars-Terra launch window to open.

The report starts off with the common complaint that most rocket propulsion is either high-thrust + low-specific-impulse or vice versa. The problem being that rocket designers want a high-thrust + high-specific-impulse engine. In other words they want a torchship.

The closest thing they can find that is actually feasible is a Gas-Core Nuclear Thermal Rocket. Open-cycle of course, closed-cycle has only half the exhaust velocity. So what if it spews still-fissioning uranium in an exhaust plume of glowing radioactive death?

The report examines the GCNTR's performance to see if it is a torch drive. It comes pretty close, actually.

The higher the specific impulse / exhaust velocity, the more waste heat the engine is going to deal with. They figure that a GCNTR can control waste heat with standard garden-variety regenerative cooling like any chemical rocket, but only up to a maximum of 3,000 seconds of specific impulse. Past that you are forced to install a dedicated heat radiator to prevent the engine from vaporizing. Otherwise the engine vaporizes, your spacecraft has no engine, and perhaps centuries from now your ship will come close enough so that space archaeologists can recover your mummified remains.

As everybody knows, thermal rockets use a heat source to heat the propellant (usually hydrogen) so that its frantic jetting through the exhaust nozzle creates thrust. Solid-core nuclear thermal rockets (NTR) use solid nuclear reactors. They are limited to a specific impulse (Isp) of about 825 seconds, since that corresponds to a propellant temperature of about 2,500 K. Any higher specific impulse raises the temperature high enough that the reactor starts to melt. And nobody likes an impromptu impression of the China Syndrome. If you want an Isp of 5,000 seconds you are talking about a propellant temperature of 22,000 K!

Also as everyone knows the gas-core NTR concept is the result of clever engineers thinking outside of the box and asking the question what if the reactor was already vaporized?

Instead of solid nuclear fuel elements it uses a super hot ball of uranium vapor which is dense enough and surrounded with enough moderator (neutron reflector) that it still undergoes nuclear fission. The fission produces huges amounts of thermal radiation, which heats the hydrogen propellant. The fissioning uranium is like a nuclear "sun" in the center of the engine. The reaction chamber directs a flow of propellant around the sun to be heated.

Since this is using the concentrated energy of fission there is no real limit to the thermal energy generated (think nuclear weapons). Unfortunately there is a limit to the hydrogen propellant's ability to absorb heat. Any heat that the hydrogen fails to sop up will hit the engine walls. If this unabsorbed heat is more than the heat radiator can cope with, bye-bye engine. This puts the upper limit on the engine's Isp capability.

Cavity Linergraphite +
5% niobium
Moderatorberyllium oxide
5.07×107 to
20.34×107 N/m2
10% by weight
0.46 m
Cavity Liner
0.0063 m
Engine Cavity
2.44 m
Plasma Dia
1.8 m
Plasma Vol
3.04 m3
Critical Mass
21 kg
Engine Mass
(including 235U)
40,000 to
210,000 kg

The engine is spherical. The outer layer is the pressure vessel (since both the propellant and uranium gas needs lots of pressure to make this thing work), a layer of beryllium oxide (BeO) moderator (a neutron reflector to help the uranium undergo nuclear fission), and an inner porous slotted cavity liner that injects the cold propellant to be heated. In the center is the furious blue-hot atomic vortex of uranium plasma.

Sadly, this structure does suffer from waste heat:

[1] a bit under 0.5% of the reactor power gets to the slotted cavity liner from thermal radiation emitted by the hot propellant. Which is a problem but not a major one. Most of the thermal radiation is soaked up and removed by the propellant.

[2] A whopping 7% of the reactor power hits all three layers of the engine, because part of the fission output is in the form of gamma-rays and neutrons, instead of useful thermal radiation. Hydrogen propellant does not do zippity-doo-dah to soak up gammas and neutrons, all of it sails right through the propellant to hit the engine structure. Deep inside the engine structure, gamma-rays and neutrons are more penetrating than x-rays.

This waste heat is managed by the engine heat radiator (and a bit managed by regenerative cooling, about as effectively as a 3-year-old helping Daddy wash the car). Most of the engine is the beryllium oxide moderator. It is designed to operate at 1,400 K, which is below the 1,700 K melting point of the BeO but above the 1,100 K radiator temperature (otherwise the radiator will refuse to remove the heat).

The hydrogen propellant is pumped into the engine at about 5.07×107 to 20.34×107 newtons per square meter (which is why the engine needs a pressure vessel).

As it turns out hydrogen propellant is transparent, which means it is lousy at absorbing thermal radiation. That's not good. To remedy this sad state of affairs, it is "seeded" by adding tiny metal bits about the size of particles of smoke, about 5% to 10% seeding material by weight. This is done right before the propellant exits the porous cavity liner into the flood of heat from the nuclear vortex. The seeding absorbs all the thermal radiation and passes the heat to the propellant by conduction. The seeding material will be something like graphite, tungsten, or non-fissionable uranium 238.

Around the exhaust nozzle the seeding concentration will have to be increased to 20% to protect the nozzle from propellant heat. The cold 20% seeded hydrogen will reduce the specific impulse a bit but it has to be done.

The porous cavity liner (in some as yet to be defined manner) magically sets up flow patterns so that the propellant flows around the hot uranium and exits via the exhaust nozzle. Meanwhile miraculously the uranium is trapped in a stagnant cavity in the center so hideously radioactive fissioning uranium does not escape through said exhaust nozzle. Uranium escape not only exposes the crew to deadly radiation, it is also a criminal waste of uranium (that is, it lets get away uranium that is not contributing to the engine's thrust).

The interior of the engine (cavity diameter) is 2.44 meters in diameter (7.61 cubic meters), and the incandescent ball of violently fissioning uranium is planned to have a diameter of 1.80 meters and a volume of 3.04 cubic meters. This gives a fuel-to-cavity radius ratio of 0.74. The idea is for the uranium sphere to be 40% of the volume of the entire chamber. However since hydrogen propellant is going to diffuse into the atomic vortex, the uranium sphere might be up to 50% hydrogen. This means the effective volume of pure uranium will be closer to 20% to 30% of the entire chamber.

The uranium can be injected by pushing a very thin rod of solid uranium into the chamber. The uranium penetrates the BeO moderator inside a tunnel lined with a cadmium oxide neutron poison, because otherwise there would be a nuclear explosion once the uranium was surrounded by BeO. This is a bad thing. The engine was designed to have the nuclear reaction happen in the core of the chamber, not in the walls.

As the uranium rod enters the chamber, the heat of the fission ball vaporizes the rod so the fresh uranium atoms can join the party.

A problem is how to get the process started. At startup, there ain't no ball of fissioning uranium to heat up the rod. The report says that the engine will have to be started by first blowing in some hydrogen and somehow injecting some powered uranium metal into the stagnant cavity until it reaches critical mass. Sounds tricky to me.

Figures 2a through 2c above are for a reactor of the following characteristics:

  • Spherical geometry
  • Uranium-235 fuel
  • Beryllium-oxide (BeO) moderator
  • Fuel-to-cavity radius ratio 0.67
  • Cavity liner thickness 0.63 cm
  • Cavity liner graphite + 5% niobium

Figure 2a shows that the 235U critical mass ranges from 10 to 35 kilograms for the cavity diameters and moderator thicknesses considered (all the curved lines are more or less above the 10 kg line and below the 35 kg line). Now for a given cavity diameter, you can reduce the critical mass required by adding more BeO neutron reflector. This means the pressure inside the engine can be lowered, which means the mass of the pressure shell can be lowered. Alas the increased penalty mass of the BeO moderator more than offsets the mass saving on the pressure shell.

(Elsewhere the report notes an optimal thickness of BeO to be 0.46 m, and a cavity diameter of 2.44 m. Eyeballing the graph implies a critial mass of 21 kg of 235U.)

Figure 2b shows that if the BeO moderator thickness is fixed, increasing the cavity diameter will decrease the critical density (the curved line will be closer to the bottom of the graph). Not shown in the table is the unfortunate fact that increasing the cavity diameter also has the side effect of increasing the total BeO weight.

(Elsewhere the report notes an optimal thickness of BeO to be 0.46 m, and a cavity diameter of 2.44 m. Eyeballing the graph implies a critial density of 18 kg/m3. If the uranium plasma ball has a volume of 3.04 m3, at that density it will contain about 55 kg of uranium, which is more than the 21 kg (from eyeball value above) it needs for criticality. However, since propellant seepage will make the sphere about 50% hydrogen, this means it will have about half of 55 kg. Which is a reasonably close eyeball value to a second eyeball value. I'm just playing number games with the graphs, do not put too much credence to these speculations on my part.)

Figure 2c shows that there is an optimum BeO moderator thickness which gives a minimum critical density for a given BeO moderator weight.

Why is there an optimum BeO moderator thickness?

If the BeO is too thin there is excessive neutron leakage (the purpose of the BeO moderator is to reflect escaping neutrons back into the fissioning uranium, basically kicking the out-of-bounds neutrons back into play). Excessive neutron leakage means the blasted cavity diameter will have to be extremely large to avoid very high critical densities.

If the BeO is too thick, the total BeO weight becomes very large. Even though you can get away with smaller cavity diameters without the heartbreak of very high critical densities.

Figure 2c is telling you that the optimum BeO thickness is 0.46 meters (for a reactor of the specified characteristics). 2c goes on to tell you that above a moderator weight of 40,000 kg larger cavities only give a slight reduction in the critical density (the curved lines are almost horizontal).

So all the engine weight estimates below are assuming a BeO thickness of 0.46 meters.

Experiments show that an effective fuel volume is about 20% to 30% of the cavity volume, for a uranium flow rate less than 1% of the hydrogen flow rate.

The paper assumes the engine can accelerate at about 0.01 to 0.05g (0.098 to 0.491 m/s)

The idea is to get the maximum thermal radiation from the fissioning atomic fireball into the cold hydrogen propellant, and the minimum thermal energy escaping the hot hydrogen propellant (which reduces the specific impulse and scorches the heck out of the cavity wall).

Figure 5b shows experimental data for tungsten-seeded hot hydrogen. It says that adding just a few percent by weight of tungsten will increase the thermal absorption cross section to between 2,000 to 100,000 square centimeters per gram. The figure also shows the thermal absorption increase at elevated pressure, which is a good thing since the engine is a high-pressure rig.

These cross sections are high enough to protect the cavity wall from damage for Isp from 4,000 to 7,000 seconds.

Figure 6 is the straight dope on the gas-core NTR engine parameters. The critical density of uranium given cavity size and moderator is as per figure 2. Thermal absorption of seeded hydrogen is as per figure 5. Heat tranfer analysis is used to determine maximum specific impulse that will keep heat load on cavity wall below 1,000 K. Engine pressure is whatever is required to have a critical mass of uranium.

The engine weight is assumed to be the sum of the three major components: BeO Moderator, Pressure Shell, and Heat Radiator. Plus 4,000 kg or less for the uranium fuel.

Pressure Shell assumes a strength-to-density value of 1.7×105 N-meters/kg.

Heat Radiator assumes a unit weight of 140 kilogrmas per megawatt of radiated power. Heat depostion rate is assumed to be 7% of reactor power. Heat radiator operates at 1,100 K (instead of 945 K), which reduces the required radiator surface area by a factor of 2. This kind of radiator more than doubled the specific impulse without adding enough weight to offset the gain. Future radiator designs with even lower unit weights would give even more specific impulse gains.

  • Spherical geometry
  • Uranium-235 fuel
  • Beryllium-oxide (BeO) moderator
  • Beryllium-oxide (BeO) thickness 0.46 meters
  • Fuel-to-cavity radius ratio 0.67
  • Fuel volume 30% of cavity volume
  • Uranium loss rate is 1% or less of hydrogen flow rate
  • Cavity liner thickness 0.63 cm
  • Cavity liner graphite + 5% niobium

In figure 6, the abscissa for both charts is engine thrust. The charts are for thrust levels from 20,000 to over 400,000 Newtons.

The ordinate of the upper chart (Fig 6A) is specific impulse, engine weight for lower chart (Fig 6B). Specific impulse ranges from 2,500 to 6,500 seconds. Engine weight ranges from 40,000 to 210,000 kg.

The curved lines are Engine Pressure, for ranges between 0.5×108 to 2.0×108 N/m2. Note in Fig 6A the three curves are labeled "Low", "Nominal", and "High". These labels are used in the Mission Chart below.

A higher engine pressure allows higher specific impulse because higher pressure makes the hydrogen propellant more opaque. But higher pressure also makes the engine heavier.

Higher thrust increases the specific impulse because there is more propellant flow to cool the cavity wall (note this is the opposite of what occurs when shifting gears). But this also makes the engine heavier.

The two reason above are why it is impossible to chose the "best" engine. What you have to do is specify a specific mission in order to have enough determining factors to figure which engine would be best.

The spacecraft is composed of a gas-core engine (with heat radiator and uranium fuel), a command module, payload, various jettisonable liquid hydrogen propellant tanks , and interconnecting structure.

The engine provides four burns:

  1. Terra orbit escape/target planet trajectory insertion
  2. Target planet orbital capture
  3. Target planet orbit escape/Terra trajectory insertion
  4. Terra orbital capture

After each burn the associated empty propellant tanks are jettisoned, except for the last burn. This is because the command module is attached to the last tank, and the crew would object strongly to being cast off into deep space. The command module also relies upon the hydrogen in the last tank for extra engine-radiation shielding.

Initial Mass In Orbit
Command Module50,000 kg
Payload to Planet150,000 kg Science/Exploration
0 kg Courier
Expendables50 kg/day
Propellant Tankage20% of hydrogen mass
Interstage Structure2% of transmitted load
Thrust Frame5% of thrust
Gas-core Engineas per Figure 6, including uranium storage and supply
Parking orbits600 km circular at Terra
high ellipse at target planet
Propulsive Effortideal ΔV from ref. 19
gravity-loss corrections Cg from ref. 20
Propellant Fraction1 - exp(-((ΔVi * Cg) / (Isp * g0)))
Ref. 19. Fishbach, L. H., Giventer, L. L., and Willis, E. A., Jr., "Approximate Trajectory Data for Missions to the Major Planets," TN D-6141, 1971, NASA, Cleveland, Ohio.
Ref. 20. Willis , E. A., Jr., "Finite Thrust Escape from and Capture into Circular Elliptic Orbits," TN D-3606, 1966, NASA, Cleveland, Ohio.

Propellant Fraction equation comes from combining these four equations into one big equation:

Pf = 1 - (1/R)
R = ev/Ve)
1/ex = e-x
Ve = Isp * g0


Cg = gravity-loss corrections Cg from ref. 20
Δv = ship's total deltaV capability (m/s)
ex = antilog base e or inverse of natural logarithm of x.
g0 = acceleration due to gravity = 9.81 (m/s2)
Isp = specific impulse (seconds)
Pf = propellant fraction, that is, percent of total rocket mass M that is propellant: 1.0 = 100% , 0.25 = 25%, etc.
R = mass ratio (dimensionless number)
Ve = exhaust velocity (m/s)

In the charts below

  • SCNR: Solid-Core Nuclear Rocket (an old-fashioned NERVA)
  • REGEN GCNR: Gas-Core Nuclear Rocket cooled with Regenerative Cooling (choked down to avoid need for heat radiators)
  • RAD GCNR: Gas-Core Nuclear Rocket cooled with Heat Radiator (uses heat radiators so it can run full-bore)
  • FUSION: Fusion Rocket (for comparison purposes)
  • SCIENCE/EXPLORATION: A mission where you bring along tons of scientific payload, and stay on Mars for 40 days to do some science.
  • COURIER: A mission with no payload just a Very Important Person. And no staytime on Mars, just a quick unloading/loading and immediate return to Terra.

In many of the charts Initial Mass in Earth Orbit (IMEO) is used to measure efficiency. The lower the IMEO value, the more efficient. Usually because it means lower propellant requirments, and may allow more payload.

Figure 6 shows that the radiator-cooled gas-core nuclear rocket becomes more efficient (higher Isp and lower specific weight) as the thrust level is raised. So the GCNR is best for missions with large payloads and/or big thrust-to-weight requirements. The missions depicted in the charts below were chosen with this in mind.

This chart shows the effect of changing the duration of the mission on the Initial Mass in Earth Orbit (IMEO). You want IMEO to be as low as possible. The shorter the mission duration, the more propellant you have to pack to increase ΔV, so the higher IMEO becomes. Obviously you can lower IMEO by increasing the mission time, but who wants to spend years on a Mars mission?

The scientific missions assume a 40 day stay on Mars to do science stuff.

The patheticaly weak SCNR (NERVA style solid-core nuclear rocket, shown with yellow curved line) has minimum mass at around 500 days and 1.5×106 kg IMEO (very roughly). This wimp ain't gonna manage a trip time below 400 days, not with a practical IMEO it isn't.

The first gas-core nuclear rocket (green curved line) show an immediate performance improvement. This is the gas-core with no heat radiator, deliberately throttled down so it can make do with mere regenerative cooling. If it is given the SCNR's 1.5×106 kg IMEO, it can do the mission in half the time, only 250 days. Its lowest IMEO is about 0.7×106 kg (700 metric tons) with a mission time around 480 days.

But the other gas-core rocket is even more powerful.

The gas-core nuclear rocket with a heat radiator (blue curved line) lowest IMEO is 0.4×106 kg (450 metric tons). This is only twice the payload (150 tonne payload + 300 tonnes = 450 tonnes). If it is loaded at a IMEO of 0.7×106 kg (the regenerative GCNR's minimum) it will do the mission in 250 days flat instead of 480 days.

With performance this high, the 40 day stay on Mars becomes an appreciable fraction of the total mission time. However low transit times mean high ΔVs and high propellant fractions.

So we now present "courier mode." This has a zero day stay on Mars, instead it immediately turns around to return to Terra. No payload either, except for something way under 1 metric ton (like a Very Important Person or a box of serum to treat the Martian Anthrax-Leprosy Pi epidemic.). The entire mission is nothing but Terra/Mars transits.

A gas-core rocket with radiator on a courier mode mission (hot pink curved line) has truly jaw-dropping performance. It can do an entire mission in only 80 days!

Just for comparison sake, the report includes a fusion rocket with typical high specific impulse but miniscule thrust (orange curved line). The fusion ship has a power plant specific mass ("alpha" or "α") at a very advanced 1 kg/kW. It has extremely low IMEO's if the mission time is greater than 250 days. But below that mission time the fusion ship's performance is lackluster. This is because the fusion drive is low thrust and is power-limited. In order to accelerate up to cruising speed in sometime less than a decade it has to increase its thrust at the expense of the specific impulse. Which sends its IMEO skyrocketing.

Unlike the fusion drive, the radiator-cooled gas core nuclear rocket is not power-limited, it is specific impulse limited (as shown in Figure 6A, see how it rapidly reaches a plateau?). This means if it trades thrust for specific impulse, it isn't reducing the specific impulse very much at all. It can crank up the thrust so it gets up to cruising speed in only two or three days. Then it can drop down to high specific impulse fuel economy gear for the rest of the 80 day mission, at a vast savings in IMEO.

Actually one can calculate the functional equivalent of α for the gas-core drive by using Figure 6. Thrust power is:

Fp = (F * Isp * 9.81) / 2

where Fp is thrust power in watts and F is thrust in Newtons. Divide Fp by the engine weight We' to get the engine α. When you do that with Figure 6, all the engines have an α in the range of 0.01 to 0.1 kg/kW, which makes the fusion drive look like a hippo.

Since these rockets were designed to be reusable, it is important to look into the difficulty of refurbishing one for a new mission.

Insipid solid-core nuclear rockets are woefully weak, but at least their nuclear fuel elements don't go anywhere. They stay safe inside the reactor ready for the next trip. Gas-core on the other hand have the drawback that the nuclear fuel elements eventually spew out the exhaust nozzle. The gas-core rocket's uranium requirement for one mission may be considerably less than the solid-core. Unfortunately the solid-core can re-use its uranium several times before more has to be added, while the gas-core has to restore its entire supply with each mission.

In figure 9 the H/U numbers are Hydrogen-Uranium flow ratios. So for instance, a rocket with a H/U of 200 will expend 200 units of hydrogen propellant for each single unit of uranium. The green SCNR curved line has no H/U number, it is a solid core rocket so zero units of uranium are expended regardless of the hydrogen flow (unless there is a catastrophic engine malfunction).

The family of yellow lines of the scientific/exploration missions show several flow ratios. There is only one flow ratio for the courier mission (200), the one in orange.

Since these are ratios you can take the uranium fuel requirement, multiply by the flow ratio, and thus calculate the hydrogen propellant requirement. For example, the 80 day Mars courier mission requires 3,350 kilograms of weapons-grade uranium-235 (98% enrichment) at a H/U of 200. Therefore the hydrogen propellant requirement is 3,350×200 = 670,000 kg.

Due to the fact that solid-core rockets can re-use their uranium a few times, a gas-core needs a H/R of 200 or more to have a lower uranium fuel bill. In 1971 (when the report was written) uranium fuel was roughly $10,000US/kg. Which means the 150-day Mars courier mission, needing 1000 kg of the hot stuff, has a uranium bill of about ten million dollars.

Not that the hydrogen propellant is exactly cheap, mind you. The element is inexpensive but shipping it from the ground into LEO can make the price tag for the 200,000 kg of propellant somewhere between $44,000,000 and $440,000,000US. This is why space fans are so keen on things like space elevators and in-situ resource utilization, to reduce these outrageous costs.

The preceeding charts assume that the spacecraft uses the optimum thrust level given the mission time and engine. This is shown in figure 10.

If low IMEO missions are desired, the thrust should be within the range of 70,000 to 90,000 Newtons (green area, favoring the right side of each curve). For low mission durations ("fast" missions) the thrust should be within the range of 112,000 to 224,000 Newtons (gold area, favoring the left side of each curve).

This chart shows the effect of using a fixed, non-optimum thrust levels. Since both lines are virtually horizontal the chart is saying there is very effect at all. Over huge ranges of thrust the IMEO doesn't really change.

If you needed a fixed thrust spacecraft that can do both missions, 150,000 Newtons is a good compromise.

But not so fast on choosing 150,000 Newtons.

Remember how shifting gears to increase the thrust imposes a penalty on specific impulse? Well, gas-core rockets with heat radiators laugh at your puny Isp penalties (the technical phrase is "relatively insensitive to Isp penalties").

In the chart, look at the area between "Low" and "Nominal". Notice how the 112,000 Newton curve is far more steep than the 224,000 Newton curve. True a gas-core is relatively insensitive to Isp penalties, but the 112,000 N engine is the more senstive of the two. Lower its Isp and the IMEO penalty mass shoots up to ugly levels.

In light of this information, a fixed thrust spacecraft that can do both missions was given a compromise of 224,000 Newtons.

The paper decided to look beyond Mars to see how the gas-core rocket would handle outer solar system missions. These all use the 224,000 Newton engine.

The science/exploration missions have a 200 day stay time, courier is still 0 day stay time. The chart shows a family of missions for each planet of gradually increasing mission durations, with the first being the courier mission (obviously). The actual feasible missions only occur at 12 to 13 month intervals, so they are marked with squares or circles. There are no missions on the connecting lines, those are just to group the planets and to indicate trends.

The Jupiter courier mission is 1.67 years (600 days) round trip and only requires an IMEO of 1.3×106 kg. The very next mission is a scientific/exploration mission with a 2.75 (1000 day) round trip and an IMEO under 106 kg. This is almost as efficient as the Mars mission.

The Saturn mission IMEOs are almost as good. Of course the trip times are about a year longer (400 days) than the Jupiter missions.

The IMEOs for the Neptune and Uranus missions are very discouraging. This probably means they are better performed with a nuclear-electric, a fusion drive, or other propulsion with a much higher Isp.

GCNR Liberty Ship

RocketCat sez

Ho, ho! This brute kicks butt and takes names! You want to boost massive amounts of payload into orbit? Freaking monster rocket has eight times the payload of a Saturn V rocket. It can haul three entire International Space Stations into LEO all at once!

But to do this it packs seven honest-to-Heinlein nuclear lightbulb engines! The only rocket that could come close to this beast is a full blown Orion drive rising on a stream of nuclear explosions at about one Hertz.

Liberty Ship
ΔV15,000 m/s
Specific Power350 kW/kg
(350,430 W/kg)
Thrust Power560 gigawatts
Specific Impulse3060 s
Exhaust Velocity30,000 m/s
Wet Mass2,700,000 kg
Dry Mass1,600,000 kg
Mass Ratio1.6875
Mass Flow1246 kg/s
Thrust37,380,000 newtons
Initial Acceleration1.4 g
Payload900,000 kg
Length105 m
Diameter20 m wide

Anthony Tate has an interesting solution to the heavy lift problem, lofting massive payloads from the surface of Terra into low Earth orbit. In his essay, he says that if we can grow up and stop panicking when we hear the N-word a reusable closed-cycle gas-core nuclear thermal rocket can boost huge amounts of payload into orbit. He calls it a "Liberty Ship." His design has a cluster of seven nuclear engines, with 1,200,000 pounds of thrust (5,340,000 newtons) each, from a thermal output of approximately 80 gigawatts. Exhaust velocity of 30,000 meters per second, which is a specific impulse of about 3060 seconds. Thrust to weight ratio of 10. Engine with safety systems, fuel storage, etc. masses 120,000 pounds or 60 short tons (54 metric tons ).

Using a Saturn V rocket as a template, the Liberty Ship has a wet mass of six million pounds (2,700,000 kilograms). Mr. Tate designs a delta V of 15 km/s, so it can has powered descent. It can take off and land. This implies a propellant mass of 2,400,000 pounds (1,100,000 kilograms). Using liquid hydrogen as propellant, this will make the propellant volume 15,200 cubic meters, since hydrogen is inconveniently non-dense. Say 20 meters in diameter and 55 meters long. It will be plump compared to a Saturn V.

Design height of 105 meters: 15 meters to the engines, 55 meters for the hydrogen tank, 5 meters for shielding and crew space, and a modular cargo area which is 30 meters high and 20 meters in diameter (enough cargo space for a good sized office building).

A Saturn V has a dry mass of 414,000 pounds (188,000 kilograms).

The Liberty Ship has seven engines at 120,000 pounds each, for a total of 840,000 pounds. Mr. Tate splurges and gives it a structural mass of 760,000 pounds, so it has plenty of surplus strength and redundancy. Add 2,400,000 pounds for reaction mass, and the Liberty Ship has a non-payload wet mass of 4,000,000 pounds.

Since it is scaled as a Saturn V, it is intended to have a total mass of 6,000,000 pounds. Subtract the 4,000,000 pound non-payload wet mass, and we discover that this brute can boost into low earth orbit a payload of Two Million Pounds. Great galloping galaxies! That's about 1000 metric tons, or eight times the boost of the Saturn V.

The Space Shuttle can only boost about 25 metric tons into LEO. The Liberty Ship could carry three International Space Stations into orbit in one trip.

Having said all this, it is important to keep in mind that a closed-cycle gas-core nuclear thermal rocket is a hideously difficult engineering feat, and we are nowhere near possessing the abilty to make one. An open-cycle gas-core rocket is much easier, but there is no way it would be allowed as a surface to orbit vehicle. Spray charges of fissioning radioactive plutonium death out the exhaust nozzle at fifty kilometers per second? That's not a lift off rocket, that's a weapon of mass destruction.

There is an interesting analysis of the Liberty Ship on Next Big Future.

Hariven-class Free Trader

This is not actually "real", but the science is admirably hard.


     “It’s a steel box.”
     “It’s a fully functional – well, mostly functional, but all primary systems are functional – Hariven-class free trader. Just what you want when you’re starting out in this business.”
     “It’s a steel box with a plasma torch welded on the back.”
     “And a generous cargo capacity for its displacement, regenerative life support, ah – adequate crew quarters and food vats, and docking room for a single surface-orbit shuttle.”
     “And it’s –”
     “– a steel box, yes. If you wanted to pay for stylish, would you be shopping for starships in a wreckyard?”
     “Show me the contract again.”

– overheard in Kathar orbit, Cilmínár system


So, I got a request from a reader for a few specs on the Hariven-class free trader. Well, why not?

(Sadly, they were imagining something like Vaughan Ling’s Planetes-inspired debris collector with comparable dimensions, capacity, etc. Sorry to say it, but that ship? Had some style. The Hariven? Really doesn’t.)


Operated by: Desperate free traders, just starting-out bands on tour, your sketchy brother, refugees, space hobos, and anyone else who can’t afford a better ship.
Basic freighter.
Under open-source license; produced by multiple manufacturers, most of whom would prefer not to admit it, along with various backyard fab shops.

(And when I say “desperate free trader”, I don’t mean, say, the people who fly around in a Firefly-class in Firefly. Those people, in this verse, own something like a Kalantha-class. This is down from there at the true ass end of space travel.)

Length: 46m, of which 30m is the hold.
8m (not including radiators)

Gravity-well capable: No.

Personnel: 3, as follows:

Flight Commander
Flight Director
Flight Engineer

(This assumes you’re following the typical regulations which require – since the Hariven has no AI, and only dumb automation – that at least one qualified person be on watch at all times, hence a minimum of three. In practice, a Hariven can be flown by one and very often is, if they don’t mind violating the rules of navigation of every halfway sane polity in space.)

Drive (typical; may vary from build to build): Nucleodyne Thrust Applications “Putt-Putt” fusion pulse drive.
Deuterium pellets. (dirty D-D fusion)
Cruising (sustainable) thrust:
0.6 standard gravities (0.56 g)
Peak (unsustainable) thrust:
1.2 standard gravities (1.12 g)
Delta-v reserve:
(Not yet calculated, but limited; if you’re flying a Hariven, you ain’t going brachy unless you devote a lot of your hold space to extra tanks. Be prepared to spend much of your voyage time on the float.)
Maximum velocity:
0.02 c (based on particle shielding)


Not supplied as standard, but buy some. You’re gonna need ’em.


Orbital Positioning System sensors
Inertial tracking platform
Passive EM array
Short-range collision-avoidance and docking radar



Other systems:

Omnidirectional radio transceiver
Communications laser
Whipple shield (habitable area only)
Mechanical regenerative life support (atmosphere/water only)
Algi-prote vat
2 x information furnace data systems
Sodium droplet radiators

Small craft:

Not supplied as standard, but a common as-supplied variant adds a partition to convert part of the forward hold into a bay with docking clamps suitable for many surface-to-orbit vehicles.


It’s a classic tail-lander layout of the crudest form: a 30m steel box welded on top of an 8m steel cylinder welded on top of a cheap fusion pulse drive, the latter two surrounded by pellet containers. It couldn’t look more brutalist/functional if it tried. At least most Hariven owners try to give it a bright paint job.

The hold is up front, a big steel box roughly the size of eight standard shipping containers. (Indeed, sometimes it’s made from eight standard shipping containers.) Putting it right for’ard has the advantage of simplifying construction greatly – all the machinery is at one end – and giving Hariven captains the assurance that if they ram their junker into anything accidentally, at least there’s 30m of other stuff between them and whatever they hit.

The hold opens up along its entire length on the port side to permit access. Responsible captains who convert their Hariven for passenger transport (the aforementioned touring bands, refugees, and space hobos, for example) by attaching deck partitions inside the hold and adding canned air have these welded shut. Less responsible captains simply pray for a lack of wiring faults.

The habitable section (the cylinder at the back) is wrapped in auxiliary engineering machinery and fuel storage, to the point that it’s only 4m in internal diameter. (If you need to fiddle with most of the engineering systems, you’re going to need a drone, or to take a walk outside.) It’s divided into four decks, from the bow down:

The bridge, which shares space with most of the avionics;

A small living area, which contains the food vat, a tiny galley, the inner door of the airlock, and any luxuries you see fit to squeeze in there. Like chairs;

The crew quarters, which means four vertically-mounted sleep pods, and maybe room for another luxury or two if they’re small;

And a tiny workshop, for any repairs that need doing.

That all sits right on top of the shadow shield and the business end of the drive. If you need to adjust anything below that – well, hope you brought a drone.

But enough of this. You buy this ship, treat her proper, she’ll be with you the rest of your life.

Ain’t sayin’ how long that’ll be, mind.

Ru said:

     Not that long ago, I spent quite some time running the numbers on fusion-pulse torch drives, and working out various performance figures and limitations. It was all quite informative and interesting. You haven’t given nearly enough information about the ship (eg. dry mass) or the engine (exhaust velocity ranges, reaction mass) for me to hazard a guess at its parameters, so it’ll be interesting to see what else you have for us…
     There’s a lot of steel mention in its design. Sounds pretty heavy. Also, metal shells plus charge particle radiation equals bremsstrahlung delight (meaning crew will be constantly irradiated by deadly x-rays) (and it makes for a poor neutron shield, which this sort of drive badly needs). Carbon is probably easier to come by, and much easier to push around.
     You don’t mention reaction mass, but with those performance figures you won’t be using pure fusion for peak acceleration. Presumably the drive expends additional deuterium for that purpose (though lithium might be a better choice).
     You’re using pure deuterium fusion, but that’s a terrible choice for spacecraft fuel, really. For flights much less than the half-life of tritium, D-T offers easier ignition, lower neutron flux and more charged particles to thrust against. D-3He would be the fuel of choice, but if you want stable and conveniently mineable fuel p-6Li or p-11B would be a much better choice than pure deuterium.
     Delta-V reserves for even a fairly conservative fusion spacecraft design are pretty generous. You might not be tooling around at cruising speed for long, but it should be able to sustain a centigee for weeks (or even months if the drive is good enough) at a time with a 3 or 4:1 wet to dry mass ratio.
     (on reflection, I am of course wrong that D-T offers a lower neutron flux or a higher proportion of charged reaction products than D-D, but it does offer a significantly lower x-ray output and a higher exhaust velocity)

Alistair Young said:

     Ah, but you’ve got to bear in mind the target market, and therefore the design paradigm. If this were higher up the scale of starships, it’d have all the fancy carbon-composite hulls, high-efficiency fuel blends, etc., etc., one could possibly desire.
     It’s steel, though, because it’s designed to be repaired – and in some cases, even built – by a monkey with a wrench, a backyard welding kit, and duct tape, not by professional yard dogs with all the nanowhatsits in the catalog.
     (Same reason the neutron protection is a slug of paraffin in the lower hull space rather than proper formed HICAP.)
     Likewise, it uses D pellets to power an old-style fusion pulse drive rather than D-He3 slush to power a new-style fusion torch because that drive needs much less maintenance, any backplanet schmuck can separate deuterium from water, and the calibration is rough enough that in a pinch, you can stuff just about anything that’ll fuse in there and it’ll mostly work for a while.
     (Basically, you want to picture the spacegoing equivalent of the beat-to-hell jalopy that’s been driven around the rainforest for forty years, being fixed with banana peels and duct tape and occasionally run on rough home-cooked rum when gas was short. It’s a sh*tbox, but it’s a sh*tbox that’s hard to kill by design.)


HELIOS Stage One
Thrust12,000,000 newtons
Wet Mass700 metric tons
not including
Stage 2
Dry Mass32 metric tons
Body Diameter6 meters
Wingspan27 meters
HELIOS Stage Two
ΔV21,000 m/s
Specific Power57 MW/kg
(566,100 W/kg)
Thrust Power3.8 gigawatts
PropulsionSolid Core NTR
Thrust981,000 newtons
Exhaust Velocity7,800 m/s
Reactor Power2,600 MW
Wet Mass100 metric tons
Payload6.8 metric tons

HELIOS stands for Heteropowered Earth-Launched Inter-Orbital Spacecraft. Unfortunately "HELIOS" became a catch-all term for quite a few post-Saturn studies around 1963. This entry is about the 1959 version from Krafft Ehricke at Convair.

As you should recall, when dealing with a radioactive propulsion system the three anti-radiation protection methods are Time, Distance, and Shielding. A rocket cannot shorten the time, a burn for specific amount of delta V takes as long as it takes. Most designs use shielding, even though the regrettable density of shielding savagely cuts into payload mass.

But some designers wondered if distance could be substituted. The advantage is that distance has no mass. The disadvantage is it makes the spacecraft design quite unwieldy. You'd have to either put the propulsion system far behind the habitat module on a long boom, or more alarmingly have the propulsion system in front with the habitat module trailing on a cable. In theory the exhaust plume is not radioactive, so again in theory the habitat module can survive being hosed like that. The propulsion exhaust is poorly collimated so it is not like a spacecraft weapon is being directly aimed at the hab module.

There is no way this design would work as a warship. It would be like trying to run through a maze while carrying a ladder. If you made too tight a turn the tow cable will be subject to the "crack-the-whip" effect, the cable will snap, and the hab module will be shot into deep space like a stone from a shepherd's sling.

The break-even point is where the mass of the boom or cable is equal to the mass of the shadow shield. Past that point it is much less trouble just to use a standard shadow shield and deal with the mass.

This is the Waterskiing school of spacecraft design.

Dr. Ehricke design was two-staged. It has a liftoff mass of 800 metric tons, a diameter of 6 meters (omitting the delta wings) and a length of 60 meters.

The first stage was chemical powered since even in 1959 they knew nobody was going to allow a nuclear propulsion system to lift off from the ground. The lower stage has a delta wing, and will glide back to base after stage separation to be reused on future missions. The lower stage has a diameter of 6 meters, and a wingspan of 27 meters. Wet mass of 700 metric tons, dry mass of 32 metric tons, twin chemical engines with a combined thrust of 12,000,000 newtons. The first stage pilot rides in a little red break-away rocket in case the first stage has an accident. In which case it will just be too bad about the crew riding next to the nuclear reactor.

The first stage separates from the second at an altitude of about 50 kilometers when the velocity reaches 4.5 km/s. The corrugated coupler that held the two stages together falls away.

The second stage will use retrorockets to lower the habitat module on cables about 300 meters below the nuclear stage, then let'er rip. The second stage has a wet mass of 100 metric tons, the nuclear reactor has a power of 2,600 Megawatts, and a thrust of 981,000 newtons. Initial acceleration is 1 g.

When it comes to Lunar landing, the habitat module touches down, then the nuclear stage move down and sideways so it stays 300 meters away as it lands. HELIOS can deliver about 6.8 metric tons of payload to the Lunar surface, and stil carry enough propellant to make it back to LEO.

Dr. Ehricke does not give details above the return trip, but it would need to involve some sort of ferry rocket to retrieve the crew from Terra orbit. There is no way anybody would allow that radioactive doom rocket to actually land. Even if it could carry enough propellant. Dr. Ehricke Convair Space Shuttle would do nicely to retrieve the crew.

Nowadays most experts agree that a 300 meter separation from a 2,600 MW reactor is totally inadequate to protect the astronauts from a horrible radioactive death. I've heard estimates of a minimum 1,000 meter separation from a 1 MW reactor. For 2,600 MW you'd want a separation more like 14,000 meters, which probably has more mass than a conventional radiation shadow shield.


Cole Nuclear Pulse
Cole Mod I
Chamber Dia40 m
Chamber Mass454,000 kg
Chamber Wall1.27 cm
Height91 m
67 m
ΔV7,900 m/s
Pulse Rate1 per sec
Pulse Yield0.01 kt
Num Pulse2,400
Water Propel
per pulse
389 kg
Thrust @
1 pulse/sec
3,560,000 N
Isp931 sec
Exhaust Vel9,100 m/s
T/W @
1 pulse/sec
Wet Mass1,611,502 kg
934,400 kg
Dry Mass677,102 kg
Payload159,000 kg
Inert Mass518,102 kg
System Mass
454,000 kg
Structural Mass64,102 kg
Cole Mod II
Chamber Mass90,720 kg
Isp1,150 sec
Exhaust Vel11,280 m/s
Num Pulse5,800
Payload1,325,000 kg
Wet Mass3,048,000 kg
Cole Mod IIa
(x10 scaleup of II)
Chamber Dia86 m
Pulse Yield0.10 kt
Isp1,350 sec
Cole Nuclear Pulse Jet
Pulse Rate2 per sec
Thrust @
2 pulse/sec
42,970,000 N

This is mostly from Aviation Projects Review volume 1 number 3 which has more details than given here. Additional material from Helios pulsed nuclear propulsion concept (1965) which discusses the Lawrence Radiation Laboratory (LRL) Helios.

Again, there were several spacecraft designs that all wanted to use the name "Helios", which is confusing. Almost as many as the designs who all want to use the name "Orion."

This Helios is closely related to the Project Orion designs, in as much as they both used tiny nuclear bombs as propulsion. Sadly the Helios concept had some fundamental design problems that it never overcame.

The basic idea was created by visionary Dandridge Cole who was then working at the Martin corporation. Mr. Cole was unaware of the nuclear-shaped-charge innovation, so he thought the Project Orion design was wasting 90% of the bomb energy. He figured he could do better than that. The more you surround the bomb, the less energy you will waste. Since most material objects fare poorly when hit by a nuclear blast, Mr. Cole used three strategies:

  • The reaction chamber surrounding the bomb was given a huge radius. This spreads the ravening energy of the blast over more chamber wall area, so each square meter of wall has to deal with a smaller portion of the total blast. Keeping in mind that when he said "huge", he wasn't fooling. The first design had a reaction chamber diameter of a whopping 40 meters (130 feet).
  • The bombs were much weaker than the Project Orion pulse units, so the total blast was less. Project Orion units were 1 kiloton, Helios units were 0.01 kiloton, or one hundred times weaker.
  • 390 kilograms of water propellant was injected into the chamber prior to each bomb. The pious hope was that the water would soak up the blast and go shooting out the exhaust nozzle at high velocity, instead of the chamber walls. Hopefully the water would also cool off the chamber walls so they wouldn't melt.

The Cole model I had engine performance that can be charitably described as "disappointing". Specific impulse was 931 seconds, which is in the upper range of conventional solid core nuclear thermal rockets. At one pulse per second the engine had a thrust-to-weight ratio of only 0.25, enough to land on Luna but not enough to lift-off from Terra. By comparison small first generation Project Orion ships were expected to have a specific impulse of 2,500 seconds and a thrust-to-weight ratio of at least 4.0.

One little inconvenient detail that Mr. Cole glosses over is the problem with tiny nuclear bombs. You see, fission reactions have that tiresome "critical" mass requirement. Meaning that if you use less than the critical mass there will be no fission chain reaction. The problem is that a critical mass of uranium-235 or plutonium will ordinarily make a much bigger bang than 0.01 kiloton. Damping the bomb down to 0.01 kiloton means that most of the uranium or plutonium does not enter the reaction. Instead they are merely volatilized into glowing radioactive vapors of death and spread to the four winds at high velocity. This makes it difficult to get permission to launch this monster from Terra's surface.

Even ignoring the radioactive contamination the inefficient use of fissionables is unconscionable. Weapons-grade uranium and plutonium are monstrously expensive, and this design will use tons of the stuff.

A more ambitious (and utterly insane) version was Cole's nuclear pulse jet. This would be a titanic airbreathing version that utilizes Terra's atmosphere as propellant until the ship climbs into space. The radioactive fallout would be only slightly less horrific than that from Project Pluto. The difference was that Pluto was supposed to be a weapon.

Cole and the Martin corp stopped working on the concept in the early 1960s, because of the lack of interest on the part of the USAF, NASA, and Martin higher management. There were a few amusing "artist conceptions" of the concept created by the advertising department of other aerospace companies that wanted to appear new and trendy.



American Bosch Arma Corporation

This is the Atomic Pulse Rocket, a pot-bellied space ship nearly the size of the Empire State Building, propelled by a series of atomic blasts.

The enormous rocket (weighing 75,000 tons fully loaded) is designed to leave Earth with a thrust of 100,000 tons. Altogether a thousand atomic blasts—each equal to 1,000 tons of TNT—are fired from a low velocity gun into a heavy steel rocket engine at a rate of one per second until the vehicle leaves Earth's atmosphere. Then steam and vaporized steel maintain the thrust. After transit speed is reached, and the propulsion system shut off, power is provided by solar batteries plating the wing and body surfaces.

Inside the rocket. living quarters are situated in the rim of a pressurized wheel-like cabin which revolves to provide artificial gravity. Radio and radar antennae revolve with it. Tubular hydroponic "gardens" on either side of the rim grow algae to produce oxygen and high protein food.

The Atomic Pulse Rocket could transport payload to the Moon at $6.74 per lb., less than one quarter the prevailing air freight charges over equivalent distance. A similar project is past the pilot-study stage in the Defense Department

(ed note: This is vaguely based on the Cole study, but is more public relations than a real engineering design study)

Helios Nuclear Pulse
Wet Mass680,000 kg
Payload Mass91,000 kg
ΔV18,000 m/s
Chamber Dia9.2 n
per pulse
100 kg
Pulse Rate1 per 10 sec
Pulse Unit
32 kg

Helios Nuclear Pulse

In 1963 the Lawrence Radiation Laboratory started working on their own version under the name of Project Helios. This was for a crewed mission to Mars. Mass in low Earth orbit (IMLEO) was to be 680,000 kg, delta-V of 18,000 m/s, delivering a 45,000 kg Mars lander into Mars orbit (total payload 91,000 kg).

The reaction chamber would have a diameter of 9.2 meters (radius 4.6 m); into which would be introduced 100 kg or so of hydrogen propellant, a small nuclear explosive charge, and a sacrificial positioning framework to hold the nuke at the center. This will be added with each detonation, at intervals of 10 seconds or longer. Of the hydrogen propellant, nuke, and framework mass; the fraction that is hydrogen propellant is called χ.

The nuclear pulse units were one meter in diameter. The core is a 2 kg sphere of weapons-grade uranium. It is coated by 5 kg of high density chemical explosive, and the entire clanking mess is jacketed by 15.7 kg of low density chemical explosive. The nuclear explosive yield is a miniscule 0.0051 kilotons (5.1 tons).


The nozzle sticking out of the chamber is conical with a 20° half-angle. The mass of the nozzle is approximately:

MN: mass of nozzle
k: a constant, report does not specify its value
ϵ: area expansion ratio of the nozzle
p0: initial pressure within the chamber
rt: radius of the nozzle throat
(ρ/σ)N: weight/strength ratio for nozzle material

Pressure Vessel

The minimum mass of a spherical pressure vessel that can withstand a steady internal pressure p without exploding into a zillion pieces is:

A factor of 4 is then included because the engine is NOT subject to a steady pressure, the pressure pulsates. Then an additional safey factor of 2 is added. So the equation becomes:

Ms: mass of pressure vessel
V: cavity volume in the shell
ρ: density of shell material
σs: shell material yield stress
p: steady pressure
p0: initial pressure within the chamber


The analysis used the payload mass MF to "hide a multitude of sins." It includes the mass of the nozzle throat valve, shock absorbers, shadow shields, life support, observational equipment, Mars excursion vehicle, and Terra atmospheric reentry vehicle. They figure that the sum of the nozzle throat valve, shock absorbers, and shadow shields will come to a total of less than 9,100 kg.


The analysis assumes that the liquid hydrogen propellant will require an additional 8% of propellant mass for tanks, insulation, and boil-off. The ratio of hydrogen tankage mass to useful hydrogen mass is α.

Nuclear Charges

Each nuclear charge and the sacrificial positioning framework is assumed to have a combined mass of 32 kilograms. There will be an additional 2.3 kg per unit for storage and handling in the pulse unit magazine. The ratio of the charge storage/handling mass to the total mass of the charges is β

Total Mass of Vehicle, Propellant, and Nuclear Charges

Remember that each pulse start with the pressure chamber containing hydrogen propellant, a nuclear pulse unit, and a sacrificial framework holding the nuke at the chamber center. The nuke and the framework will be volatilized in the explosion, and the volatilized gas plus the propellant will be heated and sent out the exhaust nozzle. Of the combined mass, the fraction that is hydrogen propellant is called χ.

δMH: mass of hydrogen propellant
δMc: mass of charge debris: volatilized nuclear charge and sacrificial framework
χ: propellant fraction

If it takes N pulses total to accelerate the vehicle to the mission delta-V, then the total amount of effluent mass is:

Remember that the ratio of hydrogen tankage mass to useful hydrogen mass is α and the ratio of the charge storage/handling mass to the total mass of the charges is β

Total Initial Vehicle Mass

Using the equation to determine mass ratio (μ) from delta-V and specific impulse (or exhaust velocity) we can make an equation that will spit out the total vehicle mass (M0) given the mission delta-V (ΔV) and engine specific impulse (Isp)

Combining the effluent mass equation and the total vehicle mass equation we can create three new equations:

Vehicle "Cost"

The cost of the vehicle is assumed to be $91 per kilogram (cost of delivering vehicle components into LEO) plus $50,000 per nuclear charge. Both in 1960 US dollars.

Above graph is number of pulse units (N) vs plenum chamber radius (r). Superimposed on top is a grid of chamber pressure (p) vs propellant-to-total-chamber-contents fraction (χ).

Plotted are the family of curves for vehicle cost COST (109$) in units of billion of 1960 US dollars.

For this chart the constants are:
  • Payload Mass (MF) = 9,100 kg
  • Nozzle expansion ratio (ϵ) = 200
  • Chamber temperature (T) = 6000 K
  • Delta-V (ΔV) = 18,000 m/s

The cost curves close on the left because the mass of the chamber increases rapidly with pressure, due to the thick-shell correction.

The cost curves close on the right because the enthalpy and specific impulse decrease with decreasing pressure for a fixed expansion ratio and initial temperature.


      As was mentioned, the HELlOS concept is by no means new. In its simplest form HELlOS is a large cavity, or pressure vessel, filled with high-temperature and pressure gas. This gas is obtained by introducing it cold to the chamber and then adding a large amount of energy. A plug in the throat of a nozzle attached to the vessel is removed and the gas flows out the nozzle, producing thrust.

     What we feel is reasonable to include on the rocket for which HELlOS is the engine is shown in Figure 1, which may or may not bear any relation to reality. Forward of the engine is shown a warehouse for the energy sources. Next are propellant tanks, biological compartments, re-entry vehicles, mission vehicles, and what-not. Since most of the thrust is produced by the nozzle, this has to be a fairly substantial structure. Shown in Figure 1 are superstructures surrounding the nozzle which would be dead, wasted weight to be carried along.

     Figure 2 is a rather exotic device by which we bring this dead weight back to life. We use the nozzle itself as part of the structure. The biological comparimetitS, mission vehicles, and propellant storage tanks are carried on the edge of the nozzle. The skirt on the end of the nozzle was placed there by the artist to serve as a warehouse for the nuclear devices. The nozzle expansion ratios we talk of are on the order of 100 in area.

     In all of the schemes that have been discussed so far in the meeting, there are two questions that must be answered. First, is it possible to build the machine, and the second, is it worth building? I believe that we can categorically state that it is possible to build this engine given enough mass. Naturally, our primary concern is seeing just how light we can build the engine—and still make it work.

     As was mentioned, the idea has been discarded several times previously. The reasons for this were two-fold. First, until recently, reliable steels had yield strengths below 100,000 psi, and second, nuclear device technology dictated that reproducible yields were in excess of 20 tons of H.E. equivalent energy, and that in order to make efficient use of the nuclear fuel the yields had to be in the kiloton range.

     How these two figures tend to fix the weight of the pressure vessel is explained as follows. If we assume that the nuclear energy is transformed into. the internal energy of a perfect, γ-law gas contained within an elastic, thin-shelled, spherical container, then it is quite simple to show that the total weight of the pressure vessel is a linear function only of the specific internal energy of the gas. A few years ago it was estimated that the constant of proportionality between shell weight and nuclear yield was 100 tons. of pressure shell per ton of yield. This, coupled with a minimum of 20 tons of yield (8.37×1010 J), indicated excessive engine weights. Recent advances in materials technology have led us to believe that a value of 20 tons/ton is realizable. This will be discussed in a bit more detail later. Additionally it is felt that it is not unreasonable to consider nuclear yields down to around 2 tons of H.E. equivalent (8.37×109 J).

     Soon after the energy of the nuclear device is released the vessel will be subjected to nuclear radiation, recurrent shock waves, high pressures, thermal radiation, and energy transfer by conduction and convection from high temperature gas. In order to arrive at quantitative estimates of the internal environment of the shell we ran a series of problems on one of the existing digital computer codes at LRL. The code used was designed for the weapons program, and the equations-of-state are known to be incorrect in our region of interest, although not grossly so. The greatest inaccuracy is in the computation of radiation transport. However, we expect that the results are not grossly in disagreement with nature.

     Figure 3 shows the computed pressure on the inside of the vessel wall as a function of time. The first spike is the initial shock wave that hits the wall and reflects back inwards to the center of the cavity and reverberates repeatedly between the wall and the center. There may well be structural problems created by the shock waves traveling through the material of the pressure vessel. We are just beginning an experimental program to determine the internal behavior of metals under repeated, low-intensity shocks.

     Figure 4 indicates the response of the shell to this sort of a pressure history. The wave labeled "b" represents response of the shell to a step in pressure to the final steady-state pressure. The curve labeled "a" is the response of the shell to the same pressure step with a short, square-wave pulse near time t = 0. We have shown that the pressure profile of the previous slide yields about this same response with minor corrections for the later reverberations.

     We expect the pressure in the vessel to start dropping noticeably in a few milliseconds. From Figure 3 we see that the gas is not too far from hydrodynamic equilibrium at times of this order. This exhaust-time scale is varied by the area of the exhaust port. At present the nozzle throat is fixed such that about 90% of the gas will be exhausted in around a tenth of a second. From Figure 4 we see that there is an overshoot of the wall displacement from what would be expected from a step in the pressure. It is something like two and a half times the mean pressure expansion. The weight/yield constant of 20 tons/ton quoted earlier comes from increasing the weight of the pressure shell by a factor of four over that which would just contain the steady-state pressure without yielding. This slide indicates that we only need a factor of 2.5; however, this would leave no safety margin. The factor of four is used throughout all of our performance and mission analyses.

     Perhaps our most serious and least understood problem is the energy transfer from the gas to the vessel. Very crude calculations indicate that there will be something like 100 cal/cm2 deposited on the inside of the wall of the pressure vessel. About 10% of this comes from the initial fire ball, and will be deposited within the first milliseconds. We need very much to have a physical understanding of the processes that go on here. We are working on this from a purely theoretical standpoint at the moment. The high-energy deposition rate from the fireball may necessitate the use of some kind of ablative coating on the inside of the wall.

     At any rate, we believe that there will be between 3 and 4% of the total nuclear energy, of the total energy released, deposited in the metal itself. Perhaps half a percent of this will be in the form of dynamic ringing of the vessel.

     This 3 to 4% of the energy that gets into the wall will account for something like a 15 to 20°K rise in the temperature of the wall per pulse. Sooner or later, it will be necessary to remove the heat from the wall as demonstrated in Figure 5. Here is plotted the effective strength as a function of temperature and pressure. The q is that "safety" factor four; omega (ω) is the tons/ton of yield which, it will be remembered, was claimed to be completely constant, dependent only on the energy contained within the pressure vessel. This is true only for a perfect γ-law gas which hydrogen is not, at temperatures in the dissociation region. The formula for omega includes a factor which is the ratio of the PV energy of the gas to its internal energy. The fact that this ratio is a complicated function of both P and E generates the three different curves shown. From Figure 5 we see that the weight of the pressure shell needed to contain this energy is fairly flat as a function of temperature up to around 700°K, then it starts rising rapidly. In order to keep the weight to a minimum, we must keep the temperature as low as possible. Since energy cannot be radiated to space from a body at 700°K rapidly enough to be commensurate with the pulse rates we contemplate, it seems that the only reasonable place to deposit this energy is in the propellant itself. Even there, energy may not be removed from the wall unless there is a temperature difference maintained between the propellant and the wall. It will take several pulses to get up to a reasonable temperature difference, and we feel that we can maintain this difference by propellant flowing through the pressure shell wall. Naturally, if we were to expel the gas faster, we wouldn't have as much of a heating problem; but this would say that there would be more peak thrust from the engine which would demand that the nozzle and shock absorbers be increased in strength and weight. There would thus appear to be some optimum peak thrust dictated by the cooling requirements of the pressure shell and the mass requirements of the nozzle.

     As the gas expands out the pressure vessel, it follows the curve given in Figure 6, if an infinite nozzle, complete recombination, and 100% isentropic efficiency are assumed. We have performed some hand calculations to determine the effect of a fixed expansion ratio nozzle. We found that this actually tends to sharpen up the shoulder of the curve. Naturally, it drops the whole Isp curve; but the Isp does hold up flatter a little longer before it starts dropping off. We must also include the effects of recombination and the effects of a fixed amount of contaminant from the energy source, perhaps 70 lb, in the propellant. This also has to be included in the hydrodynamic calculations. We must know what effect the contaminants have on the recombination rate. We suspect that it will increase the rate by catalysis. This is something that we have not done much work on and will have to look into.

     The basic question in our minds is the value of building the engine—as it seems possible to do so. The answer to this entails a certain amount of mission analysis. Very preliminary, but indicative results are given in Figure 7. It has been stated that mission analyses are to be avoided for this conference; however, we feel that they are necessary for an understanding of the problems of HELlOS.

     We have assumed for this a payload of 50 tons. The equilibrium temperature of the gas would be 6000°K and the pressure 100 atmospheres. We have assumed a rather optimistic Martian mission ΔV of 60,000 feet/sec (18,000 m/sec) for our analysis.

     Included in the analysis are analytic, dissociating-gas, thermodynamic properties for hydrogen, the rocket equation, material properties of steel for the pressure shell, and fairly optimistic estimates of tankage and nozzle weight. The thin-shell approximation was used to obtain the vessel radius and mass. Exclusive of specific impulse, the quantities obtained from this analysis are, in part, listed in Figure 7. It turns out that there are three independent variables which must be specified to completely determine a given system; and these we chose to be temperature, pressure, and specific. impulse. One might wonder how specific impulse could be an independent variable here. This obtains because we fix the amount of energy-source debris we have in the propellant at 70 lb. The rest of the gas is hydrogen. The debris is treated as a perfect, γ = 5/3 gas of molecular weight 12. Figure 7 is a typical page of computer output and shows, for example, that if we want 1800 sec specific impulse from the given conditions, the mass fraction of hydrogen would then have to be 96%. Now with 70 lb of debris, the mass of gas in the cavity is quite large. The vessel radius would be something like 30 ft and the yield around 40 tons, which means that the pressure vessel would have to be quite heavy. The number of pulses for this particular case is around 2000. This is a fairly low nuniber, because of the high specific impulse and the large amount of propellant exhausted per pulse. The total mass of the rocket is quite large because of the large engine weight.

     Now at the other extreme in Isp we see that we need only a small yield and a low weight percentage of hydrogen in the bottle. However, the low mass expelled per pulse and low Isp demands many such pulses for a given mission ΔV. In this case a large total rocket mass is produced from the rocket equation with low Isp rather than a high engine weight, as in the high Isp case. As shown, between these extremes there is a minimum in total weight.

     We have devised a figure-of-merit for a given set of parameters which we guardedly call the mission cost. It is expected that the amount of fissionable fuel is around 1 kg per pulse. We take $50,000 to be the cost per pulse. NASA indicates that $200 per pound is not an unreasonable freight charge to place the total mass in orbit. Our figure-of-merit is the sum of these two. We find that the minimum in mission cost occurs not too far from the minimum in total mass. Thus, for these particular parameters the least expensive mission would start with around 700 tons in orbit, have average Isp a bit above 1500 sec, and use between 4000 and 5000 pulses.

Question: And the Isp is at equilibrium?

Answer: It is the equilibrium value multiplied by some number less than one to account for the effects of a finite nozzle, finite recombination, and pulsing. We have used 0.8 for the results given in Figure 7.

Answer (Cooper): The point is that if the energy of the bomb were added to pure hydrogen, the temperature would be over 10,000°K. However, the energy is divided between the hydrogen and the bomb material which has a molecular weight of 10 or 12. This extra material thus lowers the temperature and raises the average molecular weight resulting in a lower equilibrium specific impulse

     One good feature of this machine is that it can be tested underground. In this connection one can conceive of many good physics experiments that can be carried out with a pressure vessel which could contain a nuclear shot.

Question: What is the time between shots?

Answer: There is no a priori fixed time between pulses. This depends upon the cooling rate of the vessel and the complexity of introducing the next device. We want to make the pulse rate as fast as possible. Typically, in the absence of shock absorbers, the acceleration is about a 2 g pulse lasting for a tenth of a second. Now, if you wait 10 sec between pulses the average thrust is then 0.02 g. We would like as high an average thrust as possible.

Question: Would you assume that your charges would need no shielding for radiation?

Answer: That's right. We have so far.

Question: Do you think this is realistic?

Answer: Perhaps not.

Question: You mentioned 1 kg of nuclear fuel. Is that a critical mass in this case?

Answer: Well, it depends upon how you build these devices. I can't go into that deeply, but I think 1 kg is about as low as you're going to go.

Answer (Cooper): I think that the point is, that it's kilogramish and that it's not 2 to 500 grams. A factor of two one way or the other is reasonable if the fuel is plutonium and somewhat larger, naturally, if you use uranium as the fuel.

Question: Does some of the propellant go through the wall? You can see that the dynamics of the shell would be changed when you have some flow of the coolant through the wall.

Answer: We have not yet considered the dynamic response of any structure other than a thin spherical shell. We would like, if possible, to spray the inside or have flow holes in the pressure vessel, which is going to increase the weight. Practically any modification of this sort is going to increase the weight of the pressure vessel. We should also be worried about the thermal stresses on the inside. wall. This is something that we have not looked at closely and may well be one of the more serious problems.

Question: What do you think will be the total mission cost?

Answer: The last slide will give that to us.

Answer (Cooper): These are based on something like 50 tons round trip through 60,000 ft/sec, and the cost generally turns out to be divided about equally between the fuel and the weight that you have to put into orbit and the minimum cost is about the minimum weight. So, it's about $148,000,000 to put the thing into orbit, and it's something like a $140 to $200 million for the charges.

     The total cost is thus around $500 million (in 1965 dollars. About $4 billion in 2020 dollars).

Question: How much hydrogen do you use per pulse?

Answer: Each case is different. Shown in this slide is the percent of hydrogen necessary to produce a given Isp· As you will remember, we fix the weight of non-hydrogen additive at 70 lb. From this you can easily get the total mass of gas expelled per pulse. The numbers presented here are merely indicative of our general analysis.

Question: What's the average thrust to weight ratio for the engine alone?

Answer (Cooper): Well, he really answered the question before. It depends on the rate at which you fire; and in particular, it doesn't matter. I think they want to fire at such a rate that they have better than 0.1 g, on the average. We made some computations on ground take off, and then it appeared to be quite a problem. It was necessary to fire one per second or faster.

Answer: Well, if you wanted to take off, there are two problems. The first is a political one, and the second is how fast the engine can be pulsed.

Question: Was the mass number for the pressure shell based on the properties of maraging steel?

Answer: That's right. It's just about the best we can do. We don't know if the yield strength we assumed is an over or under estimate. We are just getting into the experimental study of the problem.

Answer (Cooper): Right. Oh, incidentally, let me mention that there's an unclassified report that Bob Fox did on this, in I guess about 1957, which has almost all of this material.

R. H. Fox, "A Study of the Nuclear Gaseous Reactor Rocket,". UCRL-4996, Oct. 31, 1957 (Uncl.)

From HELIOS (page 71) by Theodore Stubbs, Lawrence Radiation Laboratory (1965)

1. Improvements in Computations

     The performance code described in AFN 32 and 45 has undergone two additional, important improvements. For the most part, the code employs tables of real gas properties to determine the thermodynamics of the propellant gas. However, in three crucial steps of the performance calculation it employs an approximation which assumes the gas to behave like an ideal, γ-law gas.

     Our approach to approximating real gas behavior for each of these steps was to keep the functional form of the equations used in each step unchanged. We then used three different γ's which were functions of To , po, χ, and ϵ and forced a fit to the real behavior of the gas.

     The second major improvement was to include the temperature dependence of the yield strength of the pressure vessel material.

2. Results

     Figure 2 is a plot of the number of pulses vs the radius of the cavity with grid of po vs χ (the mass fraction of hydrogen in the gas) overlaid. We note a marked difference in the optimum value of pressure and chamber radius from the values previously published. This is due almost entirely to the second term in equation (l) which tends to favor small radii in order to keep the chamber wall strength high. The inclusion of a temperature dependent material strength also allows us to optimize with respect to temperature as seen in Figure 3.

     The vehicle also has a minimum in cost and total mass with respect to the discharge time constant. As we decrease τ the nozzle throat, and thus the nozzle mass becomes large; however, as we increase τ the pressure vessel heating and thus mass becomes large. This minimum is graphically shown In Figure 4a.

     For the rest of the plots of Figure 4 we have held temperature, pressure, χ and ϵ fixed at their near-optimum values of 6750°K, 1000 Atm, 0.7, and 300, respectively. δMc is the mass of the nuclear charge, assumed to be pure carbon for calculations purposes. β is the hydrogen tankage fraction and η is the "tankage" fraction for the nuclear charges. Figure 4e depicts the sensitivity of the cost and total mass on the yield strength of the pressure vessel. We have assumed that the temperature dependence is unaltered and have merely multiplied the strength values given in Figure 1 by the factors shown as the abscissae in Figure 4e. Figure 4f clearly indicates the sensitivity of the overall vehicle performance to the amount of energy absorbed per pulse in the pressure vessel vail. The abcissa point labeled "Base + 1%” indicates that the constant multiplying Y in the first term of equation (l) has been altered to 0.0271, etc.

     Figure 5 shows the dependence of total vehicle mass on the mission demanded of the vehicle. Each point represents the conditions obtaining when the pressure and expansion ratio is held fixed but temperature and χ are chosen such that $ is minimum with respect to both of these. "Mission" is defined here as the specification of both the payload (fixed mass) and the ideal velocity Increment, ΔV, through which this payload must be accelerated.

     In Table 1 we present the various vehicle parameters obtained uhen the Helios vehicle is optimised with respect to pressure, temperature, χ, and expansion ratio for a mission of payload = 100 tons and ΔV = 60,000 ft/sec. Of interest is the indication that the diameter of the pressure vessel is about 10 ft, the nozzle throat radius is 1 ft and the nuclear yield required is 3.6 tons of H. E. equivalent energy.

Table 1
Properties of an
Optimized Hellos Vehicle
Extrinsic (assumed)
MF100 tons
ΔV60,000 ft/sec
τ0.2 sec*
δMc70 lb.
Optimised Variables
po1,000 Atm
Intrinsic (calculated)
Variables at Optimum
Mo1,005 tons
N6,052 pulses
$0.7045 x ($109)
MN25.7 tons
Ms118.2 tons
Îsp1,538 sec
Mp494.2 tons
r1.933 meter
rt52.4 Cm
Y3.6 tons

* We see from Figure 4a that this should also be an "optimised variable" having an optimum value of about 0.27 sec.

From FINAL HELIOS PERFORMANCE CALCULATION by Lawrence Livermore National Laboratory (1965)

Hermes from The Martian

The Martian movie is based on the novel of the same name. Both have the Atomic Rockets Seal of Approval. Enough said.

Warning: this section contains spoilers for the novel and the movie.

Hermes (Novel)
PropulsionIon drive
Acceleration0.002 m/s2
Gravity0.4 g
Gravity TypeBola Spin

Hermes in the Novel

Author Andy Weir based the original mission on Robert Zubrin's Mars Direct proposal. Weir updated Zubrin's chemical rocket to a nuclear-reactor-powered ion drive using argon propellant. You see, a puny chemical rocket has to use Hohmann transfer orbits which have launch windows tied to the synodic period of Mars. That mission would have had a required stay time on Mars of a little over a year. For dramatic reasons, Weir needed the mission capable of being aborted at any time with a return to Terra. The ion drive allowed this.

In Andy Weir's original conception, the Hermes is cone-shaped so it can aerocapture at Mars and at Terra, saving precious propellant mass.

The Hermes has an acceleration of 0.002 m/s2 (2 millimeters per second, per second). Andy Weir said that the delta V budget for the return trip was about 5,000 m/s.

The spacecraft is split down the middle parallel to the long axis. This allows the two halves to separate, attached with cables, so they can spin like a bola to provide artificial gravity. The halves are called "Semicone-A" and "Semicone-B".

The main airlock/docking port is located in the customary place, on the nose.

Andy Weir mentioned that the movie version of the Hermes has quite a different design. But he also noted it was "way cooler-looking than the version I imagined."

flight plan
Terra to Mars124 days
Surface Ops31 days
Mars to Terra241 days
Sol-6 Abort
flight plan
Terra to Mars124 days
Surface Ops6 days
Mars to Terra236 days
flight plan
Terra to Mars124 days
Surface Ops6 days
Mars to Terra236 days
Terra Slingshot0 days
Terra to Mars Flyby322 days
Mars to Terra211 days

An Ares mission begins with 14 uncrewed launches (probably with an Atlas or similar sized booster) dropping airbag-cushioned payloads on Mars. These would each weigh about 1000kg on launch, with up to 600kg of payload to the surface. This includes parts of the Hab and supplies.

The crewed part of the mission is mediated by the Hermes, a large vehicle for deep space with a nuclear powered ion drive designed to fly between Earth and Mars and back. The Hermes is used by every mission and was assembled in Earth orbit at (no doubt) astronomical expense.

The six astronauts of Ares 3, together with their supplies are launched from Earth to Hermes. The Mars Descent Vehicle is launched separately towards Mars at about this time.

Hermes and the MDV travel to Mars, parking in Mars orbit after 124 days in deep space. Hermes remains in orbit, uncrewed, while the MDV flies the astronauts to the surface.

On the surface, the astronauts build their Hab from airbag cushioned cargo drops and perform their mission. After 30 days on the surface, they climb into the Mars Ascent Vehicle and fly back up to orbit, where they meet the Hermes and fly back to Earth, taking 208 days to return.

Hermes parks in Earth orbit and the crew return to Earth in some re-entry vehicle like Orion or Dragon.

The MAV was launched years before, made its return fuel on Mars using electricity and ambient atmosphere, then was used for about 6 hours to get back to the Hermes.

This mission architecture is very credible, given a nuclear powered ion drive, which is technically possible but politically problematic. IMO, the architecture is inefficient given most of the hardware is used only once, Hermes is not self sufficient, and the astronauts spend only 30 days on the surface.

Mars Ascent Vehicle (MAV) is quite large. It had to be soft landed, but even empty weighs much more than all 14 presupply missions combined. Given that NASA has (in the story) developed soft-landing capability for tens of tonnes, it's not clear why stuff as mission critical as the Hab is landed relatively inaccurately in lots of parts. It could be that this simply reduces mission cost and complexity, or that there was no practical way to land something as bulky as the hab (even disassembled) in one piece.

The MAV employs In-Situ Resource Utilization (or ISRU) to make fuel and oxidizer for the return flight. Two (Earth) years of power from a 100W Radioisotope Thermoelectric Generator (RTG) is enough to make 13kg of fuel (methane) and oxidizer (oxygen) from every 1kg of hydrogen (H2) precursor brought from Earth, for a total of nearly 20T of fuel.

At various points of the novel, Weir describes the MAV as weighing 32 metric tons when fully fueled, and standing 27m tall. This implies that it is very long and skinny, which is unnecessary in the thin Martian atmosphere. Not only that, this means a lot of rocket mass relative to the amount of fuel it can carry (spherical rockets are vastly more efficient, absent significant atmosphere). Needless to say that's a bad thing. By comparison, the Falcon 9, a long and skinny rocket by usual standards, is about 70m tall and weighs about 600T on launch. The MAV could easily be a conical shape perhaps 5m wide and 10m tall.

Weir states that it has two stages, though one stage is perfectly adequate for the relatively low delta-V required to reach Low Mars Orbit (4.1km/s). Nevertheless, with a 325s Isp methane-oxygen engine, a two stage system would have a 16T first stage, a 8T second stage, and a 8T orbital module, with an implied mass fraction of 81% fuel vs 19% metal in each stage.

Towards the end of the novel, engineers at JPL describe the MAV as having an unrealistically low launch weight of 12,600kg (12.6T) — similar to a fully-loaded Dragon capsule. So we'll assume this is the dry mass. Let's assume, then, that the orbital module is 8T, the first stage is 3T, and the second stage is 1.6T, empty. The 19.397T of fuel is distributed accordingly, implying an engine Isp of 405s in order to reach 4.1km/s of Low Mars Orbit. This is low for H2/O2 engines, but extremely high for a methane-oxygen engine. Even SpaceX's planned monster Raptor engine has a notional vacuum Isp of 380s.

In order to get to 5.8km/s and intercept the Hermes, the mass of the orbital module needs to be reduced from 8000kg to 4280kg, a reduction of 3720kg. This takes into account adding 780kg of fuel, removing 500kg from the first stage (pulling off an engine), and so on. The accuracy of the numbers indicate that Andy Weir did the math, but it's not clear on what metrics he designed the MAV and its launch system.

More generally, given that the total delta-V needed to get from Mars to Earth is *only* 7.8km/s, a MAV that flies all the way back to Earth is completely possible, though it would probably need to be bigger than the MAV presented here to have adequate life support. But given that the fuel/delta-V is most easily obtained on the surface of Mars, rather than brought from Earth, a direct ascent architecture actually makes a lot of sense.

On Sol 68, Watney points out that NASA never used large RTGs on crewed missions before Ares, but during the Apollo program RTGs were deployed by astronauts to power lunar seismometers. On Sol 69, Watney states that Lewis had buried the RTG for safety reasons. A RTG stashed somewhere on the surface, however, is much less likely to overheat.

Mars Descent Vehicle (MDV). In the Sol 7 log entry, Watney mentions that the Mars Descent Vehicle (MDV) is useless to him for escaping, since its thrusters cannot even lift its own weight. This, of course, refers to its weight when fully fueled. Before landing, much of its fuel has burned off and it can achieve neutral thrust for a hovered landing. Nevertheless, it lacks (by far) the fuel capacity, thrust, and delta-V necessary to fly anything back to orbit!

In Chapter 8, Bruce and Teddy discuss potential MDV modifications. It is strongly implied, though not stated, that the design would not admit the addition of more engine clusters, and they don't have the time to invent a bigger engine. It is likely that this is a narrative device.

Orbital Mechanics

On the first page, Watney states that he was six days into the best two months of his life. Evidently he was confused, as later it's made clear that the surface operations of the mission were only 30 days long.

30 days on the surface is the extent of surface stay permitted under an "opposition" class mission, wherein the astronauts fly by Venus on either the outbound or inbound leg. While shortening the overall mission, opposition class missions significantly lengthen the time spent in space, and also bring the spacecraft much closer to the Sun, increasing the crew's exposure to radiation.

The alternative mission design is the "conjunction" class mission, wherein the crew takes a relatively short 4-6 month Hohmann transfer flight either way, with a ~560 day stay on the Mars surface in between launch windows. Obviously, if Watney had been stranded on a conjunction mission, he would have had no shortage of snacks!

One additional detail is that Weir's spaceship, the Hermes, employs ion thrusters throughout the mission, enabling a wider class of missions and trajectories than the traditional point-and-shoot orbital mechanics described in the previous paragraphs.

In Chapter 16, the Purnell maneuver is discussed, by which the crew can return to Mars fewer days than the 404 it would take Iris to get there. It's probably worth noting that there is a very similar delta-V requirement for Iris to get to Mars vs a resupply probe to reach the Hermes. The advantage of the Hermes approach is that Iris had to be able to do entry, descent, and landing. If this is the case, Iris could also get there faster by borrowing a basic ion thruster package from, say, the Asteroid Redirect Mission (ARM) spacecraft. It's also not clear why all the crew need to return to Mars (aside from narrative reasons) - most or all could return to Earth in the entry vehicle while Hermes takes the Purnell maneuver to Mars to pick up Watney before he starves. Although the remaining crew would then depend on a new entry vehicle being sent up to meet them on their eventual return to Earth.

In Chapter 20, Annie somewhat incongruously asks why Hermes can't wait at Mars for Watney to get there, when it seems he'll be slowed by the dust storm. Venkat points out that Hermes is on a fly by and can't slow down enough to be captured into orbit, but this is not entirely true. On Sol 505, Bruce says to Venkat that Hermes is flying by Mars at 5.8km/s. Mars escape velocity is only 5.5km/s (the Earth's is 11.2km/s by comparison) meaning that a delta-V of only 300m/s is needed to capture into orbit. Given that Hermes can accelerate at 2mm/s/s, a two day burn would be sufficient to capture into a big elliptical orbit, drastically increasing their margin of error. Perhaps, if Hermes slows enough for an orbital capture, its launch window to return to Earth will close too quickly to be useful.

Of course, the MAV was designed to reach Low Mars Orbit, with a delta-V of 4.1km/s. Getting to 5.8km/s is highly non-trivial, as discussed in the previous section describing the MAV. Of course the unmodified MAV has life support, so Watney could wait while Hermes spirals down to 4.1km/s to pick him up (~30 days, because Hermes can't exploit the Oberth effect), while in the modified MAV he gets close to 5.8km/s, making it much easier for Hermes to rendezvous. A MAV that got to, say, 5.2km/s would split the difference nicely. Either way, the most likely explanation is that maneuvering Hermes to do this would make them miss the Earth launch window.

On Sol 543, Beck mentions that the modified MAV will hit 12gs during launch. While they have lightened the primary payload by about 16%, Watney also removed a spare engine, suggesting that the unmodified MAV would hit at least 10gs during launch, which is unlikely for a rocket designed to fly humans! Later, Johanssen reads out a velocity of 741m/s at an altitude of 1350m, which is staggeringly fast, implying an acceleration of 20.7gs. Perhaps she dropped a zero?

When Johanssen and Vogel talk about getting Watney to orbit, what they mean is solar orbit, since Watney will have to escape Mars entirely in order to be intercepted by Hermes.

During the intercept procedure, Ares 3 crew have to think fast to find additional sources of delta-V to move the Hermes close enough to catch Watney as he flies by. The distances and velocities mentioned during this passage in Chapter 26 are correctly calculated and almost entirely realistic.

Watney suggests making a small breach in his suit and using the stored gas as a rocket to close the velocity mismatch of 42m/s. Assuming he has 5kg of gas on board (including reserve tanks) and an exhaust velocity of 400m/s (unlike rocket exhaust, it's not hot) this confers about 17m/s of delta-V, which is just not enough. This idea is transferred to the Hermes, which will spit out its atmosphere to slow down. Assuming Hermes weighs 100T, it would have to lose about 5T of air to make up the required 29m/s of delta-V. At sea-level atmospheric conditions, this implies that Hermes has a volume of 4000 cubic meters, or a floor area of 1300 square meters, or 13,000 square feet, which is the same as a very large house. Perhaps Hermes has large pressurized volumes that aren't used much for habitation? Martinez estimates that the air will take 4 seconds to leave, which implies a relatively small hole, since the shockwave would take about 0.1s to cross a Hermes-sized volume of air. A realistic concept is that Hermes is composed of two large Bigelow inflatable modules each with a diameter of about 12m, such as the BA 2100 habitats. Also worth mentioning that the process of blowing the "Vehicular Airlock" (VAL) will send lots of airlock fragments into space, hopefully missing Watney.

From The Martian: A Technical Commentary by Handmer, Jermyn, Paragano, Lommen, Nosanov (2015)

Hermes (francisdrakex)
Inert Mass60,000 kg
Crew+Payload Mass7,000 kg
RCS Propellant6,000 kg
Argon Propellant29,000 kg
Dry Mass67,000 kg
Wet Mass102,000 kg
Mass Ratio1.52
ΔV24,000 m/s
EngineIon drive
Exhaust Velocity50,000 m/s
Single Engine Thrust5 N
Engines in Arrayx40
Total Thrust200 N
Acceleration0.002 m/s
Length85 m
Span22 m
Reactor Power10 MWth
Reactor Fuel600 kg 239Pu
Gravity0.4 g
Gravity TypeTumbling Pigeon
Gravity Spin3 rpm
Gravity Radius40 m

francisdrakex's version of Hermes

francisdrakex is a talented space artist who took a stab at designing the Hermes. He did an outstanding job if I say so myself, and not just because he was assisted by some data from this website.

The entire spacecraft was designed to fit inside a 5 m payload fairing for easy boosting into LEO.

His design used the "tumbling pigeon" method of artificial gravity, which is always a good choice to reduce the spin rate below nausea levels.

The ion engine array is mounted at the spin center, a classic technique from Stuhlinger's Ion Rocket.

As per standard best practices, the dangerously radioactive nuclear reactor is mounted as far as possible from the habitat module and the crew. The reactor has a set of heat radiators to reject waste heat. The radiators are in a triangular pattern, to keep them inside the shadow cast by the anti-radiation shadow shield.

The habitat module is a standard TransHab inflatable module.

Hermes (Movie)
Length83 m ?
120 m ?
Gravity0.4 g
Gravity TypeCentrifuge
Gravity Spin4.97 to 6.30 rpm
Gravity Radius9.0 to 14.5 m

Movie version of Hermes

The movie Hermes has the engines mounted aft the reactor, and a centrifuge to provide artificial gravity.

Ship's power is apparently from a set of 12 solar cell arrays, straight off of the International Space Station (the brown elongated rectangles). Which seems a bit redundant if you already have a nuclear reactor.

Rhett Allain did some calculations about the gravity centrifuge on the movie version of Hermes, and not surprisingly discovered that there was a bit of artistic license involved.

The novel states that the artificial gravity is 0.4 g. Mr. Allain did some measurements from the movie and figured the centrifuge is spinning at about 1.08 rotations per minute (0.109 radians per second). Unfortunately to produce 0.4 g the radius of the centrifuge would have to be an outrageous 329 meters! According to one of the graphics in the flight center, the Hermes is only 80 meters long.

Mr. Allain made some further measurements from the movie and concluded the centrifuge was about 9.0 to 14.5 meters in radius. To produce 0.4 g it would have to spin at an angular speed of 4.97 to 6.30 rotations per minute (0.52 to 0.66 rad/second). Which is right at the nausea limit.

Anyway 1.08 rpm is six times slower than 6.30 rpm, which is where the artistic license comes in.

Using the movie figures of 14.5 meters in radius and a spin rate of 1.08 rpm, the artifical gravity would be a pathetic 0.02 g, not 0.4.

Another difficulty is that the spacecraft is supposed to slow down at Mars and Terra by using aerobraking. This will require something like the ballute from 2010 The Year We Make Contact. Two of them, one for each braking.


Human Outer Planet Exploration (HOPE) is from the NASA report TM-2003-212349 by Melissa McGuire, Stanley Borowski, Lee Mason, and James Gilland (2003). . Revolutionary Concepts for Human Outer Planet Exploration (HOPE) { slide show }.

This was a given as a design problem for rocket scientists.

The problem was to design a manned mission to the Jovian moon Callisto, transporting a given payload, and returning the crew and scientific samples back to Terra. The payload included an In-situ resource utilization (ISRU) plant capable of cracking Callistonian water ice into hydrogen and oxygen rocket fuel. They assumed that space probe precursor missions had mapped Callisto's surface so that landing sites could be selected in advance, with due respect toward safety, operations, and scientific gain.

Calllisto was chosen as a destination because it is outside of Jupiter's radiation belts, and it has water ice on the surface for propellant production. The purpose of the mission was to establish an outpost and propellant production facility near the Asgard impact site on Callisto.

Several design teams entered the challenge, each basing their spacecraft around a different propulsion system for comparison purposes. The idea was to promote apples-to-apples comparison, as opposed to the sad proliferation of apple-to-oranges comparisons.

Transportation of the specified payload is left up to the mission designers.

Some designs use several unmanned spacecraft to deliver all the payload except the crew and TransHab module. Those arrive on a separate manned spacecraft, which is only dispatched upon successful arrival of the unmanned spacecraft.

Other designs using more potent propulsion systems have a single spacecraft carrying all the payload.


Mass Breakdown
Crew Quarters
40,000 kg
3,933 kg
3-Person Crew Pod
40,000 kg
Surface Habitat40,000 kg
ISRU Plant40,000 kg

The standard HOPE payload is a TransHab crew quarter for six (including consumables and the crew), a Lander to ferry three crew and supplies to and from the surface of Callisto, a surface habitat module to house the three surface explorer crew members, and an In-situ resource utilization (ISRU) plant. The ISRU plant package includes an ISRU factory to crack Callistian ice into fuel for the lander, a reactor to power the ISRU plant and surface hab, and two rovers.

Some of the designs that use weaker propulsion systems and thus have longer mission lengths use two TransHab modules to reduce risk and increase available storage for the increased consumables required.

Payload: TransHab Module

TransHab Mass
Rad shield
6 crew-year
& Spares
TOTAL w/Cont46,268

This is pretty much a bog-standard TransHab habitat module, right off the shelf.

Crew quarters for six crew. Pressurized volume is about 333 cubic meters. Typically includes 15 metric tons of consumables, but varies according to mission length of the particular design.

Payload: Lander

Lander Mass
Tanks &
Life Support2,025
TOTAL w/Cont25,009

The common base section carrying a three person crew pod. Can transport three crew to Callisto surface and back. It can carry down 40 tons to the surface.

Payload: Surface Habitat Module

Surface Module Mass
Tanks &
Life Support10,779
TOTAL w/Cont36,616

The common base section carrying the inflatable surface habitat module. Can house three crew members on the surface of Callisto. It provides shelter and serves as a laboratory.

Payload: In-Situ Resource Utilization Plant

ISRU Plant Mass
ISRU Plant1,782
TOTAL w/Cont37,909

The common base section carrying the ISRU kit. This lands on Callisto the nuclear reactor, two rovers, and the ISRU plant to crack Callistian ice into LH2/O2 fuel for the lander.

This is a 1 MW-thermal reactor using a Brayton power converter to produce 250 kilowatts of electricity. It supplies power to ISRU plant and surface habitat module. Reactor is sited one kilometer away from habitat due to radiation. Alternatively tractors can be used to create hills out of local material to act as radiation shielding and reduce the mass required for reactor shielding and long cables.

Material on Callisto's surface is about 55% water ice and 45% rock. The ISRU plant will consume 215 kW of electrical power while processing 21 kilograms of water per hour into liquid hydrogen and liquid oxygen fuel for the lander. This will produce enough lander fuel for one lander sortie mission between the base and the orbiting ship every 30 days. The created fuel is stored in the fuel tanks of the common base sections of the ISRU plant and surface habitat. The engines of those common base sections are used as spares in case the lander's engines need repairs.

The rovers are equipped with bulldozer shovels in order to scoop and transport ice to the ISRU plant.


ΔV138,000 m/s
Specific Power38 kW/kg
(37,700 W/kg)
Thrust Power111 megawatts
PropulsionFission Fragment
Payload60,000 kg
Wet Mass303,000 kg
Dry Mass295,000 kg
Propellant Mass4,000 kg
Length120 m
Span62 m
Radiator area6,076 m2
Total Power1 GW
Thrust43 N
Isp527,000 s
Exhaust Velocity5,170,000 m/s

Final Report: Concept Assessment of a Fission Fragment Rocket Engine (FFRE) Propelled Spacecraft . This HOPE spacecraft was designed using a Fission Fragment Rocket Engine.


This HOPE mission concept was based around Magnetoplasmadynamic (MPD) Nuclear Electric Propulsion (NEP).

There are three spacecraft: a one-way tanker, a one-way cargo ship, and a round-trip manned ship (the Piloted Callisto Transfer Vehicle or PCTV).

The tanker is unmanned. It transports to Callisto orbit propellant tanks full of propellant that the PCTV will need for the return trip back to Terra.

The cargo vehicle is unmanned. It transports part of the payload to Callisto: the lander, the surface habitat, and the ISRU plant. Both spacecraft will be dispatched on a slow low-energy trajectory to Calliso.

Only after the unmanned vessels successfully arrive at Callisto (especially the tanker) will the PCTV be dispatched. It transports the rest of the payload to Callisto: the 6 crew, life-support consumables, and the TransHab crew quarters. It will use a fast high-energy trajectory to Callisto (in order to minimize consumables and crew radiation exposure) thus arriving with most of its propellant expended. It will replenish its propellant from the tanker for the return trip.

The habitat module is surrounded by tanks for radiation shielding. The tail radiators are cut in a triangular shape, and the outer heat radiators are arc shaped to keep them inside the shadow shield's radiation free zone, to prevent them from scattering radiation into the ship.

The crew will explore Callisto for 120 days, then depart back home to Terra.

HOPE Cargo vehicle

HOPE Cargo vehicle
ΔV20,600 m/s
Specific Power2 W/kg
Thrust Power430 kW
PropulsionMPD thrusters
Specific Impulse8,000 s
Exhaust Velocity78,500 m/s
Wet Mass242,000 kg
Dry Mass182,000 kg
Mass Ratio1.3
Mass Flow1.4 x 10-4 kg/s
Thrust11 n
Initial Acceleration4.6 x 10-6 g
Payload120,000 kg
Length130 m
Diameter55 m

The purpose of this unmanned vehicle is to transport the HOPE payload elements: lander, surface habitat, and ISRU plant. And a couple of propellant tanks for the benefit of the manned spacecraft.

HOPE Tanker

HOPE Tanker
ΔV20,600 m/s
Specific Power2 W/kg
Thrust Power430 kW
PropulsionMPD thrusters
Specific Impulse8,000 s
Exhaust Velocity78,500 m/s
Wet Mass244,000 kg
Dry Mass184,000 kg
Mass Ratio1.3
Mass Flow1.4 x 10-4 kg/s
Thrust11 n
Initial Acceleration4.6 x 10-6 g
Payload103,000 kg
Length135 m

The purpose of this unmanned vehicle is to transport propellant tanks so that the crew vehicle can refuel at Callisto for the trip home.

HOPE Crew vehicle

Piloted Callisto Transfer Vehicle
ΔV26,400 m/s
Specific Power6 W/kg
Thrust Power1.5 MW
PropulsionMPD thrusters
Specific Impulse8,000 s
Exhaust Velocity78,500 m/s
Wet Mass262,000 kg
Dry Mass188,000 kg
Mass Ratio1.4
Mass Flow3.6 x 10-4 kg/s
Thrust28 n
Initial Acceleration1.1 x 10-5 g
Payload79,000 kg
Length117 m
Diameter52 m

Stuhlinger Ion Rocket

Stuhlinger Ion Rocket
Length150 m
Wet mass360 metric tons
Dry massLander ship: 240 metric tons
Cargo ship:170 metric tons
Lander ship: 120 metric tons
Cargo ship: 190 metric tons
Mass of
Mars Lander
70 metric ton
Storm cellar
50 metric tons
Storm cellar
1.9 m
Storm cellar
Rotation rate1.3 rpm
0.14 g
115 MWt
40 MWe
4,300 m2
75 MWt
thrust98 N

Note the similarity of the HOPE MPD Crew Vehicle to this 1962 Ernst Stuhlinger design for a Mars ion-drive rocket. In both cases the engine are at the ship's middle, with triangular heat radiators.

In the mission plan, the expedition would have three spacecraft carrying a Mars lander, and two without. The astronauts would live in the storm cellars for the 20 days it would take to pass through the Van Allen radiation belts. Earth-to-Mars transfer would span mission days 57 through 204. On day 130 the thrust would be changed 180°, brachistochrone style.


This HOPE mission concept was based around Magnetized Target Fusion engines.

This section has been moved here


Isp3,000 to 30,000 sec
Exhaust Vel29,400 to 294,000 m/s
VASIMR engine alpha0.24 kg/kWe
VASMIR PPU alpha0.52 kg/kWe
VASMIR total alpha0.76 kg/kWe
Common 30 MWe VASIMR
Reactor output10MWe
Num Reactorsx3
Total output30MWe
Reactor radiators13,320 m2
Engine radiators900 m2
Crewed 30 MWe VASIMR
Payload mass105,000 kg
Inert mass207,000 kg
Dry mass312,000 kg
Propellant mass118,000 kg
Wet Mass430,000 kg
Mass Ratio1.38
ΔV9,470 to 94,700 m/s
Spin Gravity1/8g
PayloadTransHab module
6 crew
Crewed 30 MWe VASIMR
Payload mass238,000 kg
Inert mass185,000 kg
Dry mass423,000 kg
Propellant mass83,000 kg
Wet mass506,000 kg
Mass Ratio1.196
ΔV5,260 to 52,600 m/s
PayloadCrew Lander
Surface Habitat
ISRU gear
Crew mission return propellant
Vapor core reactor output200MWe
Reactor alpha0.29 kg/kWe
Reactor radiators3,279 m2
Engine radiators1,580 m2
Payload mass105,000 kg
Inert mass368,000 kg
Dry mass473,000 kg
Propellant mass150,000 kg
Wet mass623,000 kg
Mass Ratio1.32
ΔV8,160 to 81,600 m/s
Spin Gravity1/8g

This HOPE mission concept was based around Variable Specific Impulse Magnetoplasma Rocket (VASIMR) propulsion.

This is from Revolutionary Concepts for Human Outer Planet Exploration (HOPE) (2003), Revolutionary Aerospace System Concepts Human Outer Planet Exploration FY02 – Concept Catalog (2003) and VASIMR_Baseline_options.ppt (2003)

The uncrewed cargo mission goes first, starting at Terra-Moon Lagrange point 1 and delivering to Callisto the following cargo:

  • Crew Lander
  • Surface Habitat
  • In situ resource utilization gear
  • Propellant for the crewed ship to return home to Terra

The crewed mission only happens if the cargo mission is a total success, obviously. The crewed mission only carries the TransHab module and the crew. And a second TransHab as a spare.


Begin Spiral from Earth at L1: 12/6/2043
Period of Earth / L1 Spiral: 15 days
Heliocentric Transfer to SOI of Jupiter: 950 days
Period of Jupiter Injection: 232 days
Period of Callisto Injection: 17 days
Arrival at 500km Parking Orbit: 3/29/2047
Total Mission Time: 1214 days (3.3 years)


Begin Spiral from Earth at L1: 2/19/2045
Period of Earth / L1 Spiral: 15 days
Heliocentric Transfer to SOI of Jupiter: 680 days
Period of Jupiter Injection: 170 days
Period of Callisto Injection: 7 days
Arrival at 500km Parking Orbit: 7/5/2047

Stay Time: 30 days

Begin Spiral Out from Parking Orbit: 8/6/2047
Period of Callisto Escape Spiral: 9 days
Period of Jupiter Escape: 140 days
Heliocentric Transfer to SOI of Earth: 670 days
Period of Earth Injection: 20 days
Arrival at Earth / L1: 11/27/2049
Total Mission Time: 1741 days (4.8 years)

The third option utilizes Variable Specific Impulse Magnetoplasma Rocket (VASIMR) propulsion for all vehicles. VASIMR systems heat hydrogen plasma by RF energy to exhaust velocities up to 300 km/s producing low thrust with a specific impulse ranging from 3,000 to 30,000 seconds.

There is significant debate in the advanced propulsion community with respect to the complexity of the engineering challenges associated with the VASIMR system and hence for the purposes of the HOPE study, VASIMR was viewed at a lower state of TRL than MPD thrusters.

VASIMR performance potential was utilized in this option to improve upon the previous option. A single VASIMR propelled vehicle is used to transport the surface systems and return propellant to Callisto as opposed to two. As in the previous scenarios, the tanked/cargo vehicle remains in orbit around Callisto to be used a future propellant depot.

The piloted VASIMR vehicle was fitted with a second TransHab and configured with its main tanks clustered around the rotation axis. The two TransHabs balance each other and are connected by a pressurized tunnel so that the crew can move between them. Like the previous option, there are hydrogen tanks protecting the crew but they do not begin to empty till the last few months of the return mission. The resulting configuration reduces risk by having two crew habitats, the ability to generate artificial gravity throughout the entire mission plus significantly improved radiation protection.

The down side is that the payload masses have gone up due to combining the cargo and tanker vehicles and the piloted vehicle enhancements. The 10 MW that was used for the MPD option is not enough power for the VASIMR option to meet mission requirements. The VASIMR option does close assuming 30 MW on each vehicle resulting in a piloted mission round trip time of around 4.9 years with 32 days at Callisto. The total mission mass is between the previous two options with the benefits of increased safety and robustness.


Source [1]
(Booster + Sustainer)
(to orbit)
145,000 kg
(to Terra escape)
82,000 kg
Stage 1 enginechemical
Stage 1 thrust10,700,000 N
Stage 1 Isp
(sea level)
365 sec
Stage 1
exhaust vel sea
3,580 m/s
Stage 1 Isp
457 sec
Stage 1
exhaust vel vac
4,480 m/s
Wet Mass850,000 kg
Height85.4 m
Diameter8.54 m
Source [2]
(Booster only)
Stage 1 enginechemical
Stage 1
num engine
Stage 1 thrust13,700,000 N
Stage 1 Isp
(sea level)
365 sec
Stage 1
exhaust vel sea
3,580 m/s
Stage 1 Isp
457 sec
Stage 1
exhaust vel vac
4,480 m/s
Wet Mass
394,625 kg
Dry Mass
18,144 kg
Stage 1
Burn Time
70 sec
Height12 m
Diameter8.45 m
Span13 m
Source [3]
(Sustainer only)
Stage 2 engineNTR LH2
Stage 2 thrust5,782,680 N
Stage 2 Isp800 sec
Stage 2
exhaust vel
7,900 m/s
Wet Mass
453,592 kg
Dry Mass
110,000 kg
Height51 m
Diameter8.54 m
Span8.54 m
Source [4]
(Sustainer only)
Num Crew4
Stage 2
39 m
Stage 3
101 m
Total Height140 m
Stage 2 engineNTR LH2
Stage 2 thrust2,700,000 N
Stage 2 power10,000 MW
Stage 3 engineNTR LH2
Stage 3 thrust44,500 N
Stage 3 power170 MW
Leave Terra
propellant burnt
293,300 kg
Arrive Mars
propellant burnt
232,000 kg
Leave Mars
propellant burnt
100,000 kg
Arrive Terra
propellant burnt
21,000 kg
Source [5]
(Sustainer only)
Num Crew4
Trip Time347 days
Height100 m
Leave Terra
Orbital Altitude560 km
Payload26,400 kg
Propellant burnt293,000 kg
Initial Mass721,000 kg
Final Mass428,000 kg
Thrust2,900,000 N
Arrive Mars
Orbital Altitude1,900 km
Payload24,400 kg
Propellant burnt232,000 kg
Initial Mass414,000 kg
Final Mass181,700 kg
Thrust45,000 N
Leave Mars
Orbital Altitude1,900 km
Payload23,200 kg
Propellant burnt100,000 kg
Initial Mass161,000 kg
Final Mass61,010 kg
Thrust45,000 N
Arrive Terra
Orbital Altitude560 km
Payload9,100 kg
Propellant burnt21,000 kg
Initial Mass39,000 kg
Final Mass17,000 kg
Thrust45,000 N

This is from a 1959 study by Krafft Ehricke for Convair. Alas, details are sketchy, and some sources disagree with each other. Indeed some source disagree with themselves. In the table I separate the data as per the sources, so you can be as confused as I am.

The concept is a solid-core nuclear thermal rocket (the "Sustainer") that would do fast reconnaissance to Mars and Venus. A chemical booster lofts it into orbit because even back then NASA was skittish about a nuclear-powered surface-to-orbit stage. The nuclear section is two-staged, with the first stage discarded after trans-Martian insertion. The second stage is used for the three remaining mission segments.

The mission envisioned a fleet of three to four spacecraft, for mutual support.

Independent Lunar Surface Sortie

Independent Lunar Surface Sortie
EngineChemical (LOX/LH2)
Isp~450 sec
Exhaust Vel~4,415 m/s
Orbiter Wet Mass118,200 kg
Orbiter ΔV8,365 m/s
Lander Wet Mass34,600 kg
Lander ΔV4,172 m/s
Total Wet Mass152,800 kg

This is from Future space transportation systems systems analysis study, phase 1 technical report, section 3.3, page 139 (160 in PDF) Boeing, 1975.

When they say "independent", they mean the missions are totally self-contained. They do no rely upon logistics flights, support missions, orbital propellant depots, or anything like that. A few of the mission modes require a space shuttle flight to return the crew to Terra from orbit, but that's it.

There are several configurations, each with advantages and disadvantages. The baseline mission for all configurations is:

  1. Departure from Terra orbit
  2. Transfer to Luna
  3. Entry into Lunar orbit
  4. Lands on Luna four crew and 4,500 kg of payload
  5. Supports a 14 day exploration stay on Luna
  6. Lift-off into Lunar orbit
  7. Transfer to Terra
  8. Entry into Terra orbit
  9. Return crew to Terra surface
Manned lunar exploration (without supporting lunar orbit station)
In-depth exploration of selected lunar areas
  • Four crew capacity mission module
  • Mission module associated with propulsion stage for descent and ascent (i.e., it is a lander)
  • 14 day lunar stay time
  • Includes 2 crew rover with 10 days life support (20 person-days). 100 km out and return range capacity
  • Experiment station with ALSEP type capabilities
  • Lunar far side capability with communication through relay satellite
  • Sample return capacity of 500 kg
  • 30 meter drill capacity
  • Total lunar surface accessible except for To Be Determined

The spacecraft is basically an Orbital Transfer Vehicle (OTV) transporting a Lunar Lander. Pretty much exactly like the Apollo Command/Service Module transporting the Lunar Module. Except much bigger.

There are four flight modes, distinguished by the lunar orbit used by the OTV while the lander is on the surface. The timelines for each are shown in the graph above.

This usually has a short orbit wait of 21 days, including the 14-day surface stay period. But some surface site longitudes require a long orbit wait of 34 days. Add the 6 days total required for transfers from Terra to Lunar orbit and back, and the total mission time is 28 days for the short orbit wait and 41 days for the long orbit wait.
This is where the OTV orbits Earth-Moon Lagrange 2 instead of Luna proper. The advantage is such an orbit allows access to any surface site at any time. It also allows the OTV to act as a communication relay for 90% of far-side surface missions. The draw-backs are the higher mission delta-Vs and slower transfers. It takes 16 days total for transfers from Terra to Lunar orbit and back, and about 3 days each way between halo orbit and the lunar surface. Bottom line is a total mission time of 38 days, regardless of site location.
This permits free selection of lunar orbits since the lunar departure orbit is unconstrained by the lunar arrival orbit. So the average wait is only one day. With a 14 day surface mission and 3 day Terra-Luna transfers, the total mission time is only 22 days. The drawback is there is no habitat module, the crew spends the entire mission inside the cramped Earth reentry vehicle.

The main spacecraft components are:
  • CREW and EQUIPMENT MODULE (CEM): habitat module and the crew
  • CARGO MODULE(S) (CM): the surface payload
  • LUNAR TRANSFER VEHICLE (LTV): transports the CEM and CM to the lunar surface, returns the CEM to lunar orbit at end of surface mission
  • ORBITAL TRANSFER VEHICLE (OTV): transports the above between Terra orbit and Lunar orbit. And back.


This is basically the habitat module. The crew lives in this for the entire duration of the mission, except at the end of an Apollo-Mode mission (see below) where they briefly live in a reentry vehicle during aerobraking and splash-down.

The upper part is the crew section with a capacity of four crew members. The lower part is the equipment section containing electronics, life support, electrical power generators, and consumables. For the return-to-LEO option the typical life support endurance is 30 days (120 person-days), with a maximum capacity of 41 days (164 person-days). The direct-return option has a nominal life support endurance is 34 days (136 person-days).

CEM Mass Schedule
Item41 Day34 Day
CEM inerts4,860 kg4,535 kg
Crew, gear, reserves1,160 kg1,160 kg
Consumables1,900 kg1,575 kg
TOTAL7,920 kg6,695 kg
Consumables Estimation
Itemkg per Person-Day
Breathing O21.2
Cabin Leakage1.5
Power (fuel cells)5.5
LiOH CO2 scrubbers and misc.2.5
(assumed derived from fuel cells)

One of the many options is the type of crew return method. The CEM can just Return to Earth Orbit (REO Mode) and have the crew retrieved and landed on Terra by a Space Shuttle mission. The alternative direct-return option (Apollo Mode) is to carry a Earth Entry Module (EEM) much like the Apollo Command Module and land the crew with an aerobraked reentry.

The return-to-Earth-Orbit option has a larger payload capacity, more life-support endurance, and more options for the Orbital Transfer Vehicle. The drawback is it relies upon an expensive Shuttle mission to retrieve the crew.

The Direct-Return-To-Earth option does not require an expensive Shuttle mission. The drawback is reduced payload capacity, reduced life-support endurance, and only one option for the Orbital Transfer Vehicle.

The Earth Entry Module has a mass of 6,870 kg with no crew, and a mass of 8,030 with crew.


The two cargo modules contain the Surface Exploration Payload. For the 14 day surface mission, 4,535 kilograms of exploration equipment is provided. Any combination of equipment can be be accomodated. Each of the two payload pallets is approimately 4.4 × 4.4 × 6.0 meters (about 116 cubic meters).

Representative Surface Payload
Lunar Rover Vehicle1,900 kg
Transport and deployment pallet500 kg
Surface experiments1,600 kg
Experiments cannisters535 kg
TOTAL4,535 kg

Konstantin Asteroid Miner

Konstantin Asteroid Miner
Num Enginesx5
FuelLiquid Oxygen / Liquid Methane
Thrust1,993,000 N
Total Thrust9,965,000 N
Specific Impulse380 sec
Exhaust Velocity3,730 m/s
Dry Mass346,000 kg
Wet Mass563,000 kg
Mass Ratio1.63
ΔV1,820 m/s
Mission ΔV3,600 m/s
Crew sizex8

This is from the novel Delta-V by Daniel Suarez. Yes, it is fiction but the author did his homework. The list of NASA scientists and rocketry experts in the acknowledgements is quite extensive.

The spacecraft is named after Konstantin Tsiolkovsky. Who you better have heard of or RocketCat is going to give you such an atomic wedgie.

The spacecraft is designed to travel to NEO asteroid 162173 Ryugu and spend four years mining the heck out of it. It will cannibalize three of its five rocket engines in order to construct three unmanned tugs. The tugs will carry large masses of ore back to cis-Lunar space, where they will accelerate the industrialization of space. The ores are of common elements and metals, but the point is they will already be out of Terra's gravity well. This increases their value a thousand-fold.

In the science-fiction novel, the five methalox rocket engines were purchased from Burkett's "Starion Aerospace" (i.e., five BE-4 methalox engines purchased from Jeff Bezos' Blue Origin). Each engine has 400,000 pounds of thrust (1,993,000 N, total of 9,965,000 N). They are fed from a cluster of oxygen and methane bladder tanks. The tanks are shielded from solar heat by a mylar mirror shield, to avoid fuel boil-off losses. The tanks are clustered around the ship's refinery, used to separate the asteroid material into various elements.

The inflatable habitats were built by Ray Halser's company (i.e., TransHabs built by Robert Bigelow of Bigelow Aerospace)

The spacecraft is just over 250 meters from prow to stern.

The prow of the spacecraft has a large solar photovoltaic array and a communication antenna.

Next is a hub containing four docking port, each with a Mule class utility vehicle. The hub also has four suit ports with suits, and four teleoperated drones.

The mid-section has an artificial gravity centrifuge, with three arms each with an inflatable habitat on the end. Two habs are for the crew to live in, the Fab hab is a factory with 3D printers and other manufacturing equipment. The spin radius is 106 meters, the artificial gravity at the bottom of each habitat is 1 g. I calculate the rotation rate to be 2.9 rotations per minute, which is within the spin-nausea limit. The centrifuge has three arms because two arms are not spin-stable. The three arms are at asymmetric angles because the Fab hab is much more massive than the two crew habs. Each arm has a rail-mounted ballast weight used to balance the spin as crew and material moves from hab to hab. The arms must spun down, folded "upward", and locked to the ship's spine before each burn of the main engines. Otherwise the arms might shear off. This is not a problem since there are only two engine burns during the Terra-Ryugu trip.

They need spin gravity because the mission plan calls for them to stay at Rhyugu for four freaking years. Zero gee for that long might not be survivable.

Connecting each habitat to the hub are 2 meter wide tunnels, with electrical/data/plumbing conduits along three sides and composite ladder rungs on the fourth. To save on mass and maintenance an elevator is not used. Instead there is a steel cable with carabiners on a planetary gear. The crew hook their suit harness to the carabiners. The cable can transport a crewperson the length of the tunnel in five minutes.

Next comes a bay carrying four Honey Bee mining robots.

After that comes an ore refinery, surrounded by bladder tanks of methalox fuel, protected from solar heat by a mylar sun shield.

At the bottom is the engine cluster of five methalox engines.

The spacecraft has a dry mass of 346 metric tons and a wet mass of 563 metric tons (mass ratio of 1.63). It was secretly assembled in Lunar DRO for plot related reasons, at a cost of nine billion dollars. The fully amortized cost including construction and R and D was more like twenty-four billion

Every century or so there is an orbital window for an ideal Terra-Ryugu trajectory. According to the novel the next is December 13, 2033. Starting at Lunar DRO during the window a Trans Ryugu Injection will cost only 1.7 km/sec delta-V, have a trip time of 48 days. and a Ryugu Orbital Insertion of 1.9 km/sec. Total of about 3,600 m/s delta-V. The ship was designed to carry the bare minimum propellant for the return trip, but that would not be needed if they can demonstrate an ability to extract oxygen and methane from Ryugu in-situ resource utilization within ten days of arrival. If successful, they can stay and fill their tanks to the brim.

While mining Rhyugu, the Konstantin will hide behind the asteroid, to help shield the crew from deadly solar flares. Only the solar photovoltaic array and the communication antenna will peek over the asteroid. Rhyugu does not have enough gravity to hold the Konstantin in a proper orbit so the ship is forced to use reaction control thrusters in "bang-bang" mode.


Habitat module are pretty close to standard TransHab modules. On the Konstantin, there are two hab modules of 38 metric tons each, both of which house four crew. Modules are 11 meters in diameter, 8.5 meter tall, and are divided into two levels by static-free decking (plus an upper "attic" and a lower "bilge"). The hab walls were fashioned from a 50-centimeter-thick laminate of Kevlar, Nextel, Nomex, Viton, polyethylene foam, and insulation, designed to withstand the impact of micrometeorites. The upper level contained the ship’s galley, medical bay, and living area, while the lower level held four crew workstations, a shower, a bathroom, and life support systems. The aluminum-walled water-lined core contained the crew quarters and doubles as the anti-radiation storm cellar. The storm cellar walls contain 4,000 liters of water as radiation shielding.

The workstations are for among other things the ship's controls. They are basically chairs, the controls are created inside virtual-reality goggles worn by the crew. This saves mass, avoiding the necessity to boost tons of display screens, dials, buttons, wiring, and switches into orbit—not to mention replacement parts.


The mining robots are APIS™ and Honey Bee™ designs created by the real-world TransAstra Corporation, and used in the novel by permission. APIS stands for Asteroid Provided In-situ Supplies. It is a pun on "apis" which is the genus that includes the species honeybee (because like bees Apis efficiently gathers and returns useful resources and then utilizes those resources to perform useful work). A PDF report with more detail about the Honey Bees is available here.

The problem is that asteroids such as Ryugu are more like a pile of gravel in free fall than it is a solid lump of rock. So if you used jack-hammers or explosives you'd just scatter the gravel all over the solar system. Then comes the problem of refining the blasted stuff.

Honey Bees use optical mining, bagging rocks in large sacks and using beams of concentrated sunlight to spall the rock into tiny pieces. Ryugu's orbit has a perihelion of 0.9633 AU and an aphelion of 1.4159 AU. So the solar flux varies from 1.08 to 0.5 that of Terra, or from 1.48 to 0.683 kilowatts per square meter. Each Honey Bee has a pair of 7.5 m radius parabolic mirrors. Therefore each can gather up to 382 to 242 kilowatts of sunlight to spall the ore.

The ship carries drone robots that look like three-legged spiders, with the legs being 18 meters long. These are based on technology created for NASA’s canceled Asteroid Redirect Mission. They walk around on Ryugu's surface looking for likely bolders up to 10 meters in diameter possessing surface spectra of useful elements. When one is found, the drone stradles the bolder with its legs, drills into it with small mandible-like arms, then pushes off Ryugu. Once it has the bolder off Ryugu's surface, it leaves it there and returns to Ryugu to prospect some more.

A Honey Bee then rendezvous with the bolder. It opens up a large bag and wraps it around the bolder. It then focuses its twin parabolic mirrors on the Sun, and directs the beam on the bagged bolder. Frozen volatile ices are vaporized. The vapors are prevented from escaping by the bag. A cold trap condenses the volatiles in a side pouch. The process of vaporizing embedded ice spalls the surface of the bolder into tiny gravel. Eventually the entire bolder has been reduced to a form suitable for the ship's refinery to work on. All the volatile ices have been extracted, but the volatiles that are chemicaly bound in minerals (the "hydrates") are still there.

The ore is taken out of the Honey Bee's bag and put into the spacecraft refinery's reaction chamber. First the refinery heats the gravel to 500°C, causing the hydrates to release their volatiles. This is mostly water vapor, but also includes ammonia, carbon monoxide, nitrogen, methane, hydrogen cyanide, and hydrogen sulfide. These are all valuable chemicals, especially the water. The reaction chamber is spun like a washing machine to condense the volatiles into a liquid and to get them out of the chamber. This liquid is piped into storage for later purification. There hydrogen peroxide and postassium permanganate is used to oxidize out the non-water chemicals, leaving pure water. The water can be used as is, or cracked into hydrogen and oxygen. Hydrogen peroxide and postassium permanganate reagents can be synthesised from asteroid elements.

By this stage all the volatiles in ices and volatiles in hydrates have been extracted. There remains some more stubborn volatiles.

The remaining rock in the reaction chamber now undergoes "benefication", which means concentrating the valuable stuff by throwing away the worthless stuff. This process also extracts the more stubborn valueable volatiles. The first step is to pressurize the reaction chamber with pure hydrogen and heated to 800°C. Stubborn volatiles such as hydrogen, carbon dioxide, sulfur, nitrogen, hydrocarbons, chlorine, sulfuric acid, hydrochloric acid, and assorted amino acids are extracted. These are also spun out from the chamber and stored for later filtration.

At this point the ore in the reaction chamber has had its mass reduced by about half because all the volatiles have been extracted. What remains is the involatile residue. About a third of the residue is valuable iron-nickel-cobalt alloy. This is now purified by an acid leach, which removes everything except the alloy and the silicates. Acid is injected into the chamber, causing a furious chemical reaction. The chamber is spun again to remove the acid and the impurities: phosphorus, sodium, potassium calcium, and magnesium. They are called "impurities" but all are useful elements, they are captured and stored. The acids used in the leach can be synthesised from elements obtained earlier in the process so this is sustainable.

The remainder is iron-nickel-cobalt alloy plus silicates. On Terra this would be purified by smelting in a furnace, but that's a very bad idea in a spacecraft. So an alternate method is used. The idea is to separate the iron-nickel-cobalt alloy into its component elements by successive gasification. The chamber is heated to 100°C, carbon monoxide is injected, and the pressure raised to two atmospheres. Nickel reacts first combining with four carbon monoxide molecules to form nickel tetracarbonyl gas (very toxic). This is removed by spinning the chamber and stored; leaving the iron, cobalt, and silicates. Next the pressure is increased. Now the iron reacts with five carbon monoxide molecules to become iron pentacarbonyl gas (also toxic). It too is spun out of the chamber and stored; leaving the cobalt and the silicates. Finally the chamber temperature is increased to 200°C and the pressure increased to 10 atmospheres. Now the cobalt combines with eight carbon monoxide molecules to form dicobalt octacarbonyl gas (yes, this is toxic too). It is also spun out and stored. At room temperature and pressure the stored nickel and iron carbonyls condense into liquids, which makes them easier to handle (easier than handling red-hot gas at any rate). The stored cobalt carbonyl becomes a powder, but can be turned back into a gas by heating it to a modest 52°C.

The bolder has now been rendered down into valuable water, iron, nickel, and colbalt; plus other assorted chemicals.

What's left is silicate residue, basically sand. The worthless stuff. It currently isn't useful for anything other than bulk radiation shielding. But in the future it may be used to synthesize silicon or glass. And later still there may be a method invented that can economically extract the tiny amount of platinum group metals.

A much more high-tech method of rendering down an asteroidal bolder is vaporizing the thing in a fusion torch and running the resulting plasma through a mass spectrometer (a disassembler). That will sort out all the component atoms nice and neat into separate bins. A pity it requires use of nuclear fusion, ultra-high electromagnetic fields, and other technologies that will take centuries to master. And even longer to make the equipment cheap enough so that this will be an economically sensible method to extract those elements.


Above the central hub is a cluster of four docking ports for the four Mule utility craft. The port cluster is located far enough above the central hub so that the 40° thrust cone from a Mule will not spray the centrufuge arms. Otherwise the exhaust could possibly damage the arms. Immediately above the docking ports are the four suit-ports for the space suits. In between the suits are stored the Valkyrie telepresence drones.

The space suits are entered through suit ports. This means the spacesuits never enter the habitable part of the ship, they are always outside of the spacecraft. The crew gets into the suit by opening the back pack like a fat airlock door.

Why? Because the asteroid regolith is an extreme biohazard. Asteroid regolith is five times finer than talcum powder, and under the microscope looks like tiny razor blades. In the lungs they can enter the blood stream, which will kill you. The regolith which stays in the lungs can cause silicosis, which will kill you. The regolith can abrade gaskets, jam valves, and short-out circuit boards, which can also kill you. Don't let that deadly crap get inside the ship. Since you cannot work in a spacesuit near an asteroid without being freaking covered with death-dust, you can't let your spacesuit inside either. That's why you use suit ports.

Kuck Mosquito

RocketCat sez

This thing looks really stupid, but it could be the key to opening up the entire freaking solar system. Orbital propellant depots will make space travel affordable, and these water Mosquitos are just the thing to keep the depots topped off.

Kuck Mosquito
ΔV5,600 m/s
Specific Power4.8 kW/kg
(4,840 W/kg)
Thrust Power484 megawatts
PropulsionH2-O2 Chemical
Specific Impulse450 s
Exhaust Velocity4,400 m/s
Wet Mass350,000 kg
Dry Mass100,000 kg
Mass Ratio3.5
Mass Flow49 kg/s
Thrust220,000 newtons
Initial Acceleration0.06 g
Payload100,000 kg
Length12.4 m
Diameter12.4 m

Kuck Mosquitoes were invented by David Kuck. They are robot mining/tanker vehicles designed to mine valuable water from icy dormant comets or D-type asteroids and deliver it to an orbital propellant depot.

They arrive at the target body and use thermal lances to anchor themselves. They drill through the rocky outer layer, inject steam to melt the ice, and suck out the water. The drill can cope with rocky layers of 20 meters or less of thickness.

When the 1,000 cubic meter collection bag is full, some of the water is electrolyzed into hydrogen and oxygen fuel for the rocket engine (in an ideal world the bag would only have to be 350 cubic meters, but the water is going to have lots of mud, cuttings, and other non-water debris).

The 5,600 m/s delta-V is enough to travel between the surface of Deimos and LEO in 270 days, either way. 250 metric tons of H2-O2 fuel, 100 metric tons of water payload, about 0.3 metric tons of drills and pumping equipment, and an unknown amount of mass for the chemical motor and power source (probably solar cells or an RTG).

100 metric tons of water in LEO is like money in the bank. Water is one of the most useful substance in space. And even though it is coming 227,000,000 kilometers from Deimo instead of 160 kilometers from Terra, it is a heck of a lot cheaper.

Naturally pressuring the interior of an asteroid with live steam runs the risk of catastrophic fracture or explosion, but that's why this is being done by a robot instead of by human beings.

In the first image, ignore the "40 tonne water bag" label. That image is from a wargame where 40 metric tons was the arbitrary modular tank size.

There are more details here.


# Engines1
2,700 kg
607s @MR 4
545s @MR 6
5,960 m/s @MR 4
5,350 m/s @MR 6
T/W  9.8 @MR 4
12.1 @MR 6
ΔVExpend: 8,030 m/s
Reuse: 7,170 m/s
Expend: 3.9
Reuse: 3.8
Expend: 19,000 kg
Reuse 9,600 kg
Expend: no ship
Resuse: 8,200 kg
Length18 to 20 m
Width4.6 m

The name of the spaceship is LOX-augmented Nuclear Thermal Rocket Lunar Transfer Vehicle, mercifully abbreviated to LANTR LTV.

This is from Human Lunar Mission Capabilities Using SSTO, ISRU and LOX-Augmented NTR Technologies: A Preliminary Assessment (1995). Stanley Borowski of NASA Lewis Research Center was looking for a way to economically send manned expeditions to Luna, that is, with something cheaper than a non-reusable Saturn V rocket.

First off, he outlined the parameters for a true reusuable single-stage-to-orbit booster, instead of that Rube Goldberg Space Shuttle contraption.

Secondly, a more powerful engine that puny chemical rockets was indicated for the spacecraft. A NERVA-like solid core nuclear thermal rocket would be nice. Unfortunately while their specific impulse was a vast improvement over chemical engines, scaling up the blasted things from the putt-putt NERVA prototype so they had halfway decent thrust levels was a problem. It was a pity, since it only needed high thrust at certain parts of the mission. For the rest of the mission it could get by with already achieved levels of thrust. It's too bad there wasn't any way to make the engine shift gears... waitaminute!

There is a way to shift gears, the old LANTR trick! Just inject some supersonic oxygen into the exhaust nozzle like an afterburner and you could increase the thrust by up to 440%. That is good enough, and sure is easier than designing a monstrously huge reactor. Of course this degrades the specific impulse by a drastic 45% but you can't get something for nothing. You only need the afterburner for small parts of the mission, the rest of the time you can have normal specific impulse.

As it turns out, while Borowski didn't actually invent the LANTR concept, he helped develop it and has promoted it for lunar applications.

Thirdly there was that perennial problem of The Tyranny of the Rocket Equation. You can't have a reasonably sized payload as long as you are lugging along all your propellant. This looks like a job for In-situ Resource Utilization. That always gives the Tyranny a swift kick in the gonads.

Luna has a scarcity of hydrogen, but it has oxygen coming out of its ears. Which is just what a LANTR needs. As it turns out, there was a 1993 study looking into this, called LUNOX.

There had been earlier grandiose plans for huge lunar bases with titanic mining and refining installations to exploit Lunar resources. But LUNOX was trying to do this on the cheap, on a more modest scale. Read: on a scale that would NOT give NASA's funders in Congress an acute case of sticker-shock.

An initial lander delivers an oxygen production plant, storage tanks, and a nuclear reactor. A second lander delivers six remote-controlled tractors. Two of them are "loaders", which operators on Terra use to scoop up ilmenite-rich lunar soil and deliver it to the oxygen production plant. They figure that one plant could produce about 24 metric tons of LOX per year, which is enough for three manned missions.

Before a LUNOX site had been established, the project would have make do with ordinary NTR using no LANTR. Due to the state of the art, these would be use-once-and-throw-away spacecraft. But once LUNOX was up and running they could switch to reusable LANTRs and enjoy a much more economical trip to Luna. Using LUNOX would cut the mass to be boosted into LEO in half! And a resuable LANTR LTV can perform up to 20 missions before its nuclear fuel rods become spent.

Expendable LANTR LTV would transport as payload a Lunar landing / Earth return vehicle (LERV) with a crew module. After the LANTR LTV was disposed of, the LERV would land the astronauts on Luna, and after the mission was over transport the crew back into Lunar orbit and send them on their way back home to Terra.

Resuable LANTR LTV would assume there were a supply of LERVs at the lunar site. These would have their name changed to Lunar Landing Vehicles (LLV) since they would not be used as Earth Return Vehicles. Instead, the LANTR LTV would just transport the naked crew modules and surface payload modules. The LLVs would ferry the crew and surface payload from orbit down to the surface, and later ferry crew modules and a tank of refueling LUNOX back up to the orbiting LANTR LTV. One of the surface payload module types would be liquid hydrogen from Terra. Remember that Luna has vast supplies of oxygen, but precious little hydrogen. The LLVs need the stuff.

The expendable LANTR LTV would operate their engines at an oxygen-to-fuel (O/H) mixture ratio (MR) of 4.0, while the resuables would use an MR of 6.0

Design Comparison
Thrust66,700 N66,700 NBRAVO66,700 NCHARLIE
Exhaust Vel9,230 m/s9,230 m/sBRAVO9,230 m/sCHARLIE
MR Thrustn/a221,000 NBRAVO273,000 NCHARLIE
MR Exhaust Veln/a5,960 m/sBRAVO5,350 m/sCHARLIE
# Missions11BRAVO20CHARLIE
INERT MASS7,000 kg8,300 kgBRAVO8,300 kgCHARLIE
Payload20,000 kg
18,500 kg
15,500 kg
8,800 kg
(crew mod
surface pl)
12,000 kg
(surface pl
Terran LH2)
DRY MASS27,000 kg26,800 kg23,800 kg17,700 kg20,300 kg
LH2 Propellant13,000 kg6,500 kgBRAVO6,500 kgCHARLIE
LOX Propellantn/a24,900 kgBRAVO24,900 kgCHARLIE
Refuel LUNOXn/an/aBRAVO+17,100 kg+9,700 kg
RCS Propellant300 kgBRAVO300 kgCHARLIE
Total Propellant13,000 kg31,700 kgBRAVO31,700 kgCHARLIE
WET MASS40,000 kg58,500 kg55,500 kg49,400 kg52,000 kg
Mass Ratio1.482.182.332.792.56
ΔV LUNOX3,6000 m/s7,190 m/s7,810 m/s9,470 m/s8,680 m/s
MR ΔV LUNOXn/a4,640 m/s5,040 m/s5,490 m/s5,030 m/s

Design Bravo and Design Delta have almost identical designs (Delta entries that say "BRAVO" are identical to Bravo entries). Design Charlie is one design with two columns, for two different payload mixes (2nd column Charlie entries that say "CHARLIE" are identical to 1st column Charlie entries).

ΔV LUNOX means delta-V without using LANTR afterburner, and not taking into account refueling at Luna with LUNOX lunar oxygen. MR ΔV LUNOX means delta-V with LANTR, but still not taking into account LUNOX.

DESIGN ALFA (expendable)

This is a pathetic Lunar Transfer Vehicle using only a putt-putt NERVA engine with no LANTR afterburner. It is presented for comparison purposes, so the LANTR designs can point at it and laugh. It was designed to have its components boosted into orbit by a conventional Space Shuttle or a Titan IV rocket.

Alfa has an inadequate delta V of 3,6000 m/s. This means it does not have enough ΔV to do a Lunar Orbit Insertion burn. Instead, unlike the other designs, the poor LERV has to separate and do the burn itself. The extra propellant required really cuts into the LERV payload mass. Alfa has a higher listed payload mass than the other designs, but more of it is LERV fuel and less of it is LERV hardware and payload.

DESIGN BRAVO (expendable)

This was designed to be boosted into orbit by a hypothetical new single-stage-to-orbit (SSTO) rocket with a 9.2 m (30 foot) cargo bay. So the design is split into three 9.2 m long parts (9 m propulsion module, 9 m propellant module, and the LERV).

Unfortunately the only way to make everything fit was to put the propellant module liquid oxygen tank inside the liquid hydrogen tank. This is a bad idea. In September of of 2016 a SpaceX Falcon 9 rocket blew up on the launch pad because a liquid helium tank immersed in a liquid oxygen tank froze the oxygen into solid oxygen. Design Bravo has extra insulation around the oxygen tank (incidentally cutting into the payload mass) but it is still a matter of concern.

Also in a desperate attempt to make everything fit into the booster, they had to use a methane (CH4) fueled LERV instead of the more efficient hydrogen (LH2) fueled LERV (methane tanks are smaller because methane is more dense).

DESIGN DELTA (expendable)

This is basically Design Charlie with the assumption that NASA can get Congress to approve funding for a more spacious SSTO booster with a 13.7 m (45 foot) cargo bay. This allows relocating the liquid oxygen tank outside of the liquid hydrogen tank, so it is no longer a ticking time bomb. It also allows using the more efficient LH2 LERV.


There are two columns for Design Charlie, but they are the same spacecraft with two different payloads. It assumes that there is a supply of reusable LLVs on site at the LUNOX base to ferry crew and cargo back and forth (delivered by prior expendable missions). So Charlie just carries the naked payloads, it does not lug along the LLVs as well.

The "crew delivery" payload has 6.8 metric tons of crew and crew module, plus 2 metric tons of surface supplies for the LUNOX base. The "cargo delivery" payload is just 12 metric tons of surface supplies. Among the surface supplies are liquid hydrogen fuel for the LLVs and other equipment at LUNOX base. Remember the Lunar soil is jam-packed with oxygen but hydrogen is very hard to come by.

The design shown is based on Design Bravo, with the oxygen tank ticking time bomb.

Since the LLVs at LUNOX base can also refuel Charlie with oxygen, Charlie is reusable. It can perform 20 missions before the nuclear fuel runs out. The other designs are disposed of, criminally wasting 95% of their costly nuclear fuel rods. Actually even Charlie is wasting 85% of its fuel rods, but NASA figures attempting to remove the rods for reprocessing is just begging for a nuclear disaster in space.

Charlie just runs its engines at 6.0 MR for the entire mission, to stay within the LH2 and LOX propellant limits.


Normal Growth LCOTV
PropulsionIon Drive
Specific Impulse8,000 s
Exhaust Velocity78,480 m/s
Input power
per Engine
46 kW @
beam voltage
Engine Efficiency82%
Engine Life6,000 hours @
beam current
16 amps
per Engine
20 kg
Thrust per Engine0.7 N
Number of Engines206
Total Thrust145 N
Thrust to Weight5×10-5
Trip Time180 days
Payload Bay227 metric tons
at 100 kg/m3
Power PlantSolar Cell
Solar Cell Area54,416 m2
Structural Mass4,057 kg
Power Plant Mass26,831 kg
System Mass
11,671 kg
System Mass
2,217 kg
Thermal Control377 kg
Avionics520 kg
Growth Margin9,339 kg
Inert Mass55,012 kg
Payload Mass227,000 kg
Dry Mass282,012 kg
Propellant Mass29,744 kg
Propellant Reserves892 kg
Wet Mass312,648 kg
Mass Ratio1.11
ΔV8,190 m/s
Accelerated Technology LCOTV
PropulsionIon Drive
Specific Impulse8,000 s
Exhaust Velocity78,480 m/s
Thrust per Engine5.2 N
Number of Engines26
Total Thrust135 N
Thrust to Weight4.76×10-5 g's
Payload Bay227 metric tons
at 100 kg/m3
Power PlantSolar Cell
Solar Cell Area41,495 m2
Structural Mass2,880 kg
Power Plant Mass22,212 kg
System Mass
1,979 kg
System Mass
2,005 kg
Thermal Control68 kg
Avionics520 kg
Growth Margin5,075 kg
Inert Mass34,739 kg
Payload Mass227,000 kg
Dry Mass261,739 kg
Propellant Mass26,901 kg
Propellant Reserves784 kg
Wet Mass289,424 kg
Mass Ratio1.11
ΔV8,190 m/s

This is from Technology Requirements For Future Earth-To-Geosynchronous Orbit Transport Systems (1979). The unmanned Large Cargo Orbital Transfer Vehicle (LCOTV) transports cargo from Low Earth Orbit to Geosynchronous Orbit. The report also described a chemically powered LEO to GEO transport for priority cargo.

The report was trying to put a price tag on an transport system capable of handling the construction of a large solar power station.

According to a later report, for various reasons, the report concluded the project would require a fleet of 13 LCOTVs. Not all of the fleet would be on line, some would be undergoing maintenance. The ships in the fleet would be allocated such that the SPS project would receive a total of 56 flights, transport a total payload of 29,860 metric tons, make from 1 to 13 flights a year, and transport from 33 to 2010 metric tons per year.

Two designs were considered. The "Normal Growth" design used conservative extrapolations of the state-of-the-art, the "Accelerated Technology" assumed additional money was invested to increase the state of the art.

Langley Hybrid Mars

Langley Hybrid Mars
Inert Mass21,500 kg
Chemical Fuel23,100 kg
Ion Propellant18,500 kg
Total Propellant41,600 kg
Chemical Engine
TypeAerojet R-42
Array sizex10
Isp303 sec
Exhaust Velocity2,970 m/s
Total Thrust890 N
Electric Engine
TypeHERMeS Hall Thruster
Array sizex24
Power Requirement318 kWe
IspHi 3,000 sec
Lo 2,000 sec
Exhaust VelocityHi 29,000 m/s
Lo 20,000 m/s
Thrust/engineHi 0.525 N
Lo 0.675 N
Total ThrustHi 12.6 N
Lo 16.2 N

This is from An Integrated Hybrid Transportation Architecture for Human Mars Expeditions (2015)

Everybody in rocketry is waiting for somebody to invent a torchship.

For a typical Mars mission, designers want an engine with high thrust so the Trans-Mars insertion burn takes only a few minutes. Instead of taking a year, with most of that year spent inside the Van Allen radiation belt irradiating the hapless crew with atomic death-rays.

But at the same time designers want an engine with a high specific impulse, otherwise the design will get eaten alive by the The Tyranny of the Rocket Equation. The result will be a spacecraft looking like a skyscraper composed of fuel with a wretchedly tiny Mars expedition perched on the top.

Chemical engines have fantastic thrust and woefully minuscule specific impulse ("muscle" class). Electric propulsion like ion drives are the exact opposite ("fuel-economy" class). A torch drive would be high in both, but currently there ain't no such animal.

Most Mars missions using chemical propulsion are forced to rely upon staging to cope with the huge fuel requirements. This turns the spacecraft into a disintegrating totem pole, throwing away fuel tanks and engines. Oh, and did I mention that designers wanted the spacecraft to be totally reusable? Staging is almost criminal if you are using nuclear engines, but there are plenty of Mars missions that do just that. Nothing like jettisoning strongly radioactive fission reactors still full of nuclear fuel into eccentric solar orbits where they will be a radiation hazard for generations to come. Tends to earn harsh words from the future generations, it does.

So lacking a torch drive, how does a designer get out of this mess? Presently the best idea is to make a hybrid spacecraft, with two types of engine. The unstoppable Stanley Borowski has a design using solid core nuclear thermal rockets along with an ion drive, using the clever Bi-Modal NTR trick. The Langley design manages to make a chemical-ion hybrid.

The Langley design dramatically lowers the fuel requirements (which is the same thing as raising the payload size), allows the spacecraft to be reusable, and manages to get away with using a Solar Electric Propulsion (SEP) engine with a remarkably low power requirement of only 300 kilowatts. The latter is important because 300 kW can be easily suppled by a low-mass solar array, a megawatt class electric drive would need a nuclear reactor and those things weigh tons.

The chemical engines take care of just the high-thrust burns, and the fuel-economy ion drive performs the burns where thrust is not quite so urgent. If you arrange the trajectory with sufficient cleverness, the fuel mass savings will more than take care of the mass penalty of having two separate rocket engines. And also have no significant increase in the flight times, even if you are using a low-thrust ion drive.

The spacecraft does require a cis-lunar Basecamp in order to reduce the orbital energy, but you can't have everything.

Bottom line: larger mission payload and a totally reusable spacecraft.

To size the spacecraft the designers used software called Glenn Research Center Collaborative Modeling for Parametric Assessment of Space Systems (GRC COMPASS).

One of the design parameters was the ability to do Mars missions with [1] 300 day (10 month) or longer Mars orbit dwell time and [2] less than 1,100 day (3 year) round trip heliocentric mission duration.

The cargo and crew are launched to the basecamp. The crew uses a fast "taxi" transfer which takes eight to ten days, 300 to 400 m/s ΔV, and a lunar gravity assist (LGA). The cargo uses a slow ballistic lunar transfer which takes 120 to 200 days and 0 to 40 m/s ΔV (to maximize cargo delivery capability).

The heart of the system is the Hybrid Propulsion Stage (HPS).

The ion drive uses components designs that were created for the Asteroid Redirect Mission (ARM). The ARM engine buss uses 150 kW of electrical power for 12 Hall Effect ion engines utilizing xenon propellant. The engines can operate in high specific impulse (3000 sec) or low specific impulse (2000 sec) modes, shifting gears.

For the Hybrid, two busses are to be used, with a power requirement of about 318 kWe, 24 engines, and double the thrust.

The ion engines are 12.5 kW HERMeS Hall Thruster, as described in "Development Approach and Status of the 12.5 kW HERMeS Hall Thruster for the Solar Electric Propulsion Technology Demonstration Mission" IEPC-2015-186. HERMeS is an acronym for Hall Effect Rocket with Magnetic Shielding.

HERMeS operating conditions
Discharge Voltage (V)Discharge Current (A)Discharge Power (kW)

Option 1 uses the ARM engines and the ARM structure. Option 2 uses the ARM engines but a new structure is designed so that the HPS can fit into the SLS narrower 8 meter diameter shroud instead of the larger 10 m shroud. Out of an abundance of caution Option 1 was chosen.

Both options use the same solar arrays which provides 435 kWe at Terra's distance from Sol, but only 184 kWe at Mars. The ion engines are not used near Mars so the fact that the solar array does not provide enough power for the ion engines is not a problem. Remember the engines use 318 kWe

Both options use chemical engines: array of x10 Aerojet R-42 engines using storable MMH/NTO fuel with oxidizer-to-fuel ratio of 1.65, nominal specific impulse of 303 seconds and total thrust of 890 Newtons.

Lewis Research Center GCNR

This is from Mission performance potential of regeneratively cooled gas core nuclear rockets (1971)

The report notes that while solid-core nuclear thermal rockets have twice the specific impulse of chemical rockets, this isn't enough of an increase for high-energy trajectories with very high payloads. On the other end of the spectrum, ion drives have superb specfic impulse, but their pathetic thrust lead to undesirably long mission times.

The report decided to stop playing around and look at a rocket engine that is more in the middle between solid NTR and ion drives in both specific impulse and thrust: open-cycle gas-core nuclear thermal rocket. Gas-core has about four to twelve times as much specific impulse as solid-core, about sixty times as much thrust, and a hideously deadly radioactive exhaust plume. About as radioactive as if a solid-core rocket had a total nuclear meltdown at a rate of one solid-core rocket per second.

The report ignores the radiation, figuring it is not their department to worry about it. They wanted to analyze the performance of the gas-core NTR to find the circumstances where the performance was so superior that it warranted a more detailed study (including how to deal with the lethal exhaust).

They didn't bother analyzing the safer closed-cycle NTR because it was a typical compromise that wound up with the disadvantages of both and the advantages of neither. For one thing the specific impulse was half that of an open-cycle engine.

Spoiler Alert: the open-cycle gas-core nuclear rocket is so superior to the solid-core that it isn't even a contest. The real eye-popping gains show up at 3,000 seconds of specific impulse, but even 1,500 seconds is impressive. It is worth it to develop the GCNR further.

Their baseline was a pretty standard regeneratively-cooled open-cycle gas-core engine. They assume a specific impulse in the range of 1,000 to 3,000 seconds (exhaust velocity of about 10,000 to 30,000 m/s) which is about seven times that of chemical rockets and about four times that of solid-core NTRs.

But unlike prior reports, they looked into optimizing the thrust levels for a given mission.

Engine Mass

The mass of the gas-core NTR engine is obviously the sum of its parts:

Me = Mmod + Mps + Mtp + Mn

Me: mass of engine (kg)
Mmod: mass of nuclear moderator (kg)
Mps: mass of pressure shell (kg)
Mtp: mass of turbopump (kg)
Mn: mass of nozzle (kg)

They assumed the following relationships:

MH2: hydrogen propellant flow rate (kg/sec) Note the "M" should have a dot over it (because m-dot means mass flow). Unfortunately many computers do not have unicode fonts capable of rendering the character "Ṁ".
F: thrust (N)
Isp: specific impulse (sec)
g: standard value of gravity acceleration (m/sec2) = 9.80665 m/sec2
Isp × g: exhaust velocity (m/s)

MU = MH2 × ( 1 / H2/U)

MU: uranium fuel flow rate (kg/sec)
MH2: hydrogen propellant flow rate (kg/sec)
H2/U: hydrogen-to-uranium flow-rate ratio = 100
1 / H2/U: reciprocal hydrogen-to-uranium flow-rate ratio = 0.01

VU: volume of uranium in core (m3)
Vc: volume of core (m3)
MH2: hydrogen propellant flow rate (kg/sec)
MU: uranium fuel flow rate (kg/sec)

Moderator Mass

Moderator is assumed to be 0.762 meters thick.

Mmod: mass of nuclear moderator (kg)
D: outer diameter of the core/moderator (m) = for this study 3.66 m
ρmod: density of moderator material (kg/m3) = 1,150 kg/m3

Pressure Shell Mass

P: pressure (atm)
Mcr: critical mass in reactor (kg) = 48 kg
F: thrust (N)
Isp: specific impulse (sec)
D: outer diameter of the core/moderator (m) = for this study 3.66 m
MU: uranium fuel flow rate (kg/sec)
MH2: hydrogen propellant flow rate (kg/sec)

t: thickness of pressure shell (m)
P: pressure (atm)
D: outer diameter of the core/moderator (m) = for this study 3.66 m
σ: allowable stress in pressure shell (atm) = 13,600 atm

Mps: mass of pressure shell (kg)
D: outer diameter of the core/moderator (m) = for this study 3.66 m
t: thickness of pressure shell (m)
ρps: density of pressure-shell material (kg/m3) = 8,000 kg/m3

Turbopump Mass

Mtp: mass of turbopump (kg)
MH2: hydrogen propellant flow rate (kg/sec)
P: pressure (atm)
ρH2: density of hydrogen (kg/m3) = 72 kg/m3

Nozzle Mass

Mn: mass of nozzle (kg)
ε: area ratio of nozzle = 300
F: thrust (N)
P: pressure (atm)

Hydrogen Propellant Temperature

This does not help calculate the mass of the engine, but it is needed with other parts of the design.

T: propellant temperature (°C? K?)
P: pressure (atm)
F: thrust (N)
Isp: specific impulse (sec)
D: outer diameter of the core/moderator (m) = for this study 3.66 m
MU: uranium fuel flow rate (kg/sec)
MH2: hydrogen propellant flow rate (kg/sec)


Note that the thrust level has been optimized for each mission.

D: outer diameter of the core/moderator = for this study 3.66 m
Mcr: critical mass in reactor = 48 kg
H2/U: hydrogen-to-uranium flow-rate ratio = 100
σ: allowable stress in pressure shell (atm) = 13,600 atm
ρps: density of pressure-shell material (kg/m3) = 8,000 kg/m3
ρH2: density of hydrogen (kg/m3) = 72 kg/m3
ε: area ratio of nozzle = 300
ρmod: density of moderator material (kg/m3) = 1,150 kg/m3

Mission delta-Vs

The report took the simplistic ideal mission delta-Vs and made a function to account for gravity losses, since in the real world rocket impulse burns are not instantaneous (indeed, with ion drive a burn can take weeks). The function figured in acceleration levels, parking orbit eccentricity, and final hyperbolic excess velocity.

This has to be done iteratively, since once you calculate the real delta-V, you optimize the thrust level to the real delta-V, which means you have to recalculate the real delta-V, which means you have to reoptimize the thrust level… You keep iterating until the function converges on a value.

(Mp)i: propellant mass of ith maneuver (kg)
(M0)i: mass at beginning of ith maneuver (kg)
ex: antilog base e or inverse of natural logarithm of x
ΔV: delta-V (km/sec)
Isp: specific impulse (sec)
g: standard value of gravity acceleration (m/sec2) = 9.80665 m/sec2
Ispg: exhaust velocity (m/s)

(M0)i+1: mass at beginning of next maneuver after ith maneuver (kg)
(M0)i: mass at beginning of ith maneuver (kg)
(Mp)i: propellant mass of ith maneuver (kg)
(Mpstr)i: propellant-structure mass of ith maneuver (kg)
Mjettison: mass jettisoned, such as Mars Lander (kg)
(Mis)i: interstage structure mass of ith maneuver (kg)

(M0)imax: mass at beginning of last maneuver (kg)
(Mp)imax: propellant mass of last maneuver (kg)
Mpstr: propellant-structure mass (kg)
Mts: thrust-structure mass (kg)
Mpay: payload mass (kg)
Me: engine mass (kg)


Mts: thrust-structure mass (kg)
F: thrust (N)


(Mpstr)i: propellant-structure mass of ith maneuver (kg)
(Mp)i: propellant mass of ith maneuver (kg)


(Mis)i: interstage structure mass of ith maneuver (kg)
(M0)i: mass at beginning of ith maneuver (kg)


Me: engine mass OF A SOLID-CORE NUCLEAR ROCKET as an approximation. For GCNR use this equation. (kg)
F: thrust (N)

For Mars and Jupiter missions Mjettison = 136,100 kg

Uncrewed Lunar Ferry Mission

In the uncrewed lunar ferry mission the GCNR starts parked in LEO, does a Hohmann transfer to deliver various amounts of payload (Mjettison) to be placed into a lunar orbit. It still has 50,000 kg of Terra return payload. It then does a Hohmann back into LEO parking orbit. After refueling it is ready to deliver another payload.

The graph in figure 5 compares the performance of the gas core rocket with a 825 sec Isp solid-core Nerva-II, and a theoretical maximum 825 Isp Nerva-II (meaning it has zero engine mass). A specific impulse of 2,000 seconds was chosen for the GCNR. The GCNR leaves both NERVAs in the dust. It needs much less initial mass in LEO for a given payload. Actualy for the 50,000 kg payload to Luna and 50K kg back (blue line in graph) the GCNR can carry close to twice the payload (red line in graph).

The graph in figure 6 shows fuel/propellant consumption, where the GCNR shames the NERVAs even more. With a 50,000 kg payload the GCNR uses about half the the propellant needed by the NERVAs.

Figure 7's graph shows, among other things, that raising the specific impulse from 2,000 seconds to 3,000 seconds only reduces the initial mass in LEO (IMLEO) requirement by 11%. Simply because the propellant requirement goes down with higher specific impulse.

Figure 8a shows how the moderator mass varies with the core diameter. You probably should use the "best guess" curve for required critical mass.

Figure 8b shows that changing the hydrogen-propellant to uranium-fuel ratio (H2/U) doesn't change the IMLEO very much. The report uses H2/U = 100. The IMLEO goes up as H2/U rises because the pressure shell weight increases (see equation), of course as H2/U rises the amount of expensive uranium fuel needed goes down.

Uncrewed Slingshot Mission

In the uncrewed slingshot mission the GCNR starts parked in LEO, the spacecraft accelerates into some amount of hyperbolic excess velocity (V), releases a payload of mass Mjettison, waits until it comes back to Terra, and then burns like heck to circularize into LEO. It still has 50,000 kg of Terra return payload. It can be reused after being refueled.

Figure 9 shows how various factors affect the IMLEO. Mass increases with increase in amount of V, mass decreases with increase of specific impulse. At V=0 if you double the GCNR Isp from 1,500 to 3,000 seconds you will save 20% on IMLEO. At V=5 km/s doubling Isp will save 30% on IMLEO.

The GCNR can point its finger and laugh at the optimistic solid-core NTR. SCRN cannot reach V=5.5 km/s at all, not without staging at any rate. Which coincidentally is the delta-V requirement for a 300 day Mars round trip. A single-stage 1,500 Isp GCNR can manage it. A 3,000 Isp GCNR can do that without even working up a sweat.

Crewed Interplanetary Missions

For crewed interplanetary missions, it assumed that 136,100 kg (Mjettison) of payload is delivered into a 0.9-eccentricity planetary parking ellipse with a periapsis at 1.1 planet radii. An additional 90,700 kg is the command module / habitat module plus the Terra reentry vehicle. The GCNR does not brake into orbit to be reused, it goes streaking by Terra into a far Solar orbit. The crew bails out in the reentry vehicle that aerobrakes to the surface. The analysis did not set a limit on entry velocity, though in reality if it is much over 11 km/sec it is difficult to make a reentry vehicle that won't melt into a fiery blob of molten impure aluminum.

Figure 10 below shows the results for the Mars round-trip mission.

The analysis simplified things by assuming circular coplanar orbits, so this is more of the average performance regardless of synodic period.

In figure 10a, if you fix the IMLEO at 106 kg for all cases, 3,000 sec Isp GCNR can do the mission in 225 days, 1,500 sec GCNR needs 360 days, and the weakling 825 sec SCNR needs a whopping 430 days.

If you fix the trip time at 200 days, the 3,000 sec GCNR needs only one-fifth the IMLEO of the 1,500 sec GCNR. The IMLEOs are all higher than 106 kg, but it might be worth it if you are trying to spare the crew from excessive radiation doses.

Figure 10b shows among other things that the optimum thrust is in the range of 400,00 to 5 million Newtons (4×105 to 50×105 N).

Figure 11 below shows the results for the Jupiter mission.

The possible missions are the circles, ignore the lines connecting them (those are just to label circles with their specific impulse). A mission can be performed in 580 days, or 1,020 days or 1,420 days but not at any other intermediate time values. The circle values are when Jupiter and Terra are in opposition or conjuction at mid-say. At other times the delta-V becomes impossibly high. If you really want the scary mathematical details they can be found in Approximate trajectory data for missions to the major planets.

Because of the circle limits, for the Jupiter mission raising the specific impulse can only reduce the IMLEO. For the Martian mission raising the specific impulse can both reduce the IMLEO and the trip time. But the GCNR can perform the mission faster than the SCNR because the latter is too weak to have access to the columns of circles with lower trip times. It can only do the 1,420 day trip.

In the 1,420 day column, the 3,000 sec Isp GCNR requires only 15% the IMLEO of the SCNR, and 60% of the 1,500 sec GCNR.

Figure 11b shows among other things that the optimum thrust is in the range of 106 to 2×107 Newtons, about twice that of the Mars mission.

Lewis Research Center Ion Rocket

Lewis Ion Rocket
Propulsionion drive
Exhaust velocity20,000 to
100,000 m/s
Return payload
(cabin, etc.)
23,000 kg
Crew supplies
(4.5 kg/day/man)
18,000 kg
Exploration Rocket
(Mars lander)
18,000 kg
Powerplant41,000 kg
Propellant104,000 kg
Wet Mass204,000 kg
Dry Mass100,000 kg
Mass Ratio2.04
ΔV14,000 to
71,000 m/s
Mars mission
500 days
Length120 m

This is from a 1965 study by the Lewis Research Center entitled Space Flight Beyond The Moon.

This is a standard ion rocket powered by a nuclear reactor. The reactor is at the nose, behind a shadow shield and separated from the crew compartment by a 120 meter long boom in order to use distance as extra radiation shielding. The heat radiators are trimmed in order to not extend outside of the radiation shadow, their narrow aspect indicates the shadow shield is minimal in order to save on mass.

Higher ion drive exhaust velocities come at the expense of higher power requirements, which means higher power plant mass, which means lower acceleration. The report mentions that one can calculate the optimal exhaust velocity/power plant mass, but does not go into details. As a general rule a rocket's acceleration should not be below 0.05 m/s2 (5 milligee) or it will take years to change orbits.

Some of the designs depict the crew cabin as two modules, which probably means they spin around the ship's spine for artificial gravity.

Lewis Research Center Mars Landing

This is from Manned Mars landing mission by means of high-thrust rockets (1966).

Yet another nuclear engine design that tries to squeeze out enough delta V so it can actually land astronauts on Mars, instead of attempting to do months of science in five minutes as they go whizzing by. Since they were designing with underpowered NERVA engines with only 850 seconds of specific impulse, they were forced into the irresponsibly cuckoo "nuclear staging" solution just like the Boeing IMIS. Except the Lewis design does not jettison five dangerously radioactive nuclear reactors like used beer cans. It jettisons seven.

And like the IMIS, it cannot afford a NERVA for braking into Terra orbit. They come in hot at full velocity, trusting to the miracle of aerobraking. The crew of seven enters a winged reentry craft and gets to dice with death, gambling that they will run out of delta V before they run out of ablative heat shield. If they lose, perhaps some kid will make a wish on the shooting star which is all that is left of their incinerated bodies.

The mission had a crew of seven. Depending upon the types of optimization, either four or all seven made excursions to the Martian surface. A minimum of three crew was needed for emergency operation of the spacecraft. Normal operation required six crew, operating in pairs. 1/3rd of the day was for spacecraft operational duty, 1/3 was for sleep, and 1/3 was off-duty (recreation, personal chores, scientific duties, and study).

The propulsion is by a series of NERVA style solid-core nuclear thermal rockets using liquid hydrogen as propellant. They have a specific impulse of 850 seconds and a thrust-to-weight ratio of 10. Well, actually T/W of 8 because the liquid hydrogen tanks need thermal protection from the engines or they will boil-off dry in no time. And T/W of 3 if you must have biological anti-radiation shields. The crew is protected from engine radiation by a combination of propellant tanks, separation distance, and command module storm cellar but no biological shields (but offhand it looks like that will not be enough shielding). Somewhat arbitrarily the study authors assumed each engine would have a minimum mass of 3,630 kg (7,260 kg with thermal and biological shields.) The initial acceleration of each engine with a full propellant tank is 0.2 gs.

There was a variant with chemical engines, but it was pretty pathetic.

The report has zillions of trajectories, each optimized for one factor or another. For mission with durations of 600 days and below, the total delta V requirement vary from 26,070 to 41,360 m/s.


As with most designs this is a combination of the spacecraft control room and the anti-radiation storm cellar. Because nobody wants to die horribly by manning an unshielded control room during a solar proton storm, and leaving the control room unmanned is a Really Bad Idea. During normal operations a crew of two or three occupies the command module. All seven can be contained during radiation events for short periods (1 day), four of the crew have to stand.

The command module has a mass of 4080 kg (9,000 pounds), not counting the radiation shielding. It has a volume of 12.7 cubic meters, of which 1.4 m3 are radiation sensitive operating equipment.

The radiation shielding is the chemical rocket fuel and oxidizer used by the Earth reentry vehicle. The fuel is pumped into the vehicle at the end of the mission.


The life support system is wrapped around the command module. The food and oxygen cycles are open, but the water cycle is closed (totally recycled). It assumes each crew consumes per day 1 kg of food, 1.8 kg of water, and 0.95 kg of oxygen (which is a little skimpy on the water). Plus each crew has 9 kg of recycled water per day for washing and other utility purposes.

Cabin air leakage is assumed to be about 0.68 kg per day, and complete air changes are made at three month intervals.

The total life support requirements for 7 crewmembers per day is 1608 kg plus 260 watts of power.


The living module has a mass of 4080 kg (9,000 pounds) not including radiation/meteor shielding and a volume of 156 cubic meters. The breathing mix pressure is 48 kPa (7 psi). To improve reliability the module is divided into two pressure independant units.

The walls hold about 29 kilograms per square meter of reentry vehicle chemical fuel as meteor and mild radiation protection. For full-blown radiation storms the crew retires to the storm cellar. The chemical fuel is held in multiple independant loops. Since the sun only heats the sun-side of the module, temperature is equalized by rotating the meteor shield or circulating the fuel. The desired level of temperature is maintained by adjusting the angle of the module with respect to Sol and by surface coatings.

The living module is connected to the Earth deceleration system by a long boom. The boom rotates fast enough to provide the living module with 0.3 g of artificial gravity (close to the Martian surface gravity of 0.376g). A smaller counter-rotating centrifuge balances the angular momentum. It provides up to 10 g's for the crew to exercise because it would be real nice for the crew to be able to walk when they return to Terra.

The scientific crew might spend their entire duty cycle (all day) in the living module. The control crew have to spend part of their duty cycle in the command module.


The spacecraft requires 7.5 kilowatts of power. A solar Rankine system was selected because nuclear power reactors have too much mass and are too radioactive to be repairable. The Rankine was assumed to have an alpha of 91 kilograms per kilowatt for a mass of 680 kg. 1,360 kg since they carry along a spare.

During the mission the generator dish does its best to shield the cryogentic hydrogen propellant tanks from the burning rays of the sun. Otherwise all the hydrogen will boil off.

Solar photovolatic panels were considered, but in 1966 they were not exactly "mature" technologies.


As previously mentioned the design cannot afford the additional NERVA engines and propellant to brake the spacecraft into LEO. So the designers took the cheap way out and used aerobraking.

The system has a mass of 24,490 kilograms, including the 16,780 kg of chemical fuel which is stored during the mission around the command and living modules as radiation shielding. It carries a payload of 1,090 kg (the "real mission payload") consisting of the crew, the Mars surface samples, and the data.

The system is a winged reentry vehicle with a thick ablative heat shield on its belly (much like NASA's Space Shuttle), a retro-rocket with the chemical fuel, and a meteor shield to protect everything during the long mission.

The winged reentry vehicle has a mass of 4,990 to 7,480 kg for atmospheric entry velocities of 7,930 to 19,810 m/s. It has 1.7 meters of unobstructed interior depth to accomodate the seated crew. Internal volume of 40 cubic meters. Leading edges swept 60°, small radius nose, and maximum attack angle of 23°. Maximum G load of 10 g's, entry corridor depth is 48 kilometers. The planned entry maneuver calls for 8 g deceleration at supercircular speeds and 4 g at subcircular speeds.


Each landing vehicle can carry two crew.

Since the landers use aerobraking most of their mass is the propellant needed to lift off from Mars and return to the orbiting spacecraft.

While the spacecraft is in Martian orbit, the landers separate and shed their meteor/thermal shields. The retrorocket fires to put the lander on entry trajectory, then is jettisoned so it does not obstruct the aerobraking heat shield. The shield burns its little heart out because the Martian atmosphere is like making love in an airlock exceedingly close to vacuum (1/100th of Terra's atmospheric density does not help much with braking). The landing/hover rocket extends from the top and burns at 2 gs of deceleration. It has enough fuel to hover and "translate" for only two minutes so the pilot has to pick the landing spot quickly. Hopefully the shock absorbers are up to the task of absorbing the landing impact.

The takeoff tanks are kept cool enough to avoid boil-off by foils in a vacuum jacket. Ordinarily you couldn't do this on a planet with an atmosphere, but as previously mentioned the Martian atmosphere is not that far from being a vacuum already. The fuel has a specfic impulse of 430 seconds, they suggest Diborane fuel with Oxygen difluoride oxidizer (B2H6 + OF2). Well, at least they didn't do a jackass maneuver like try to use FLOX for oxidizer.

The crew does as much science as they can possibly cram into 40 days. Life support and electrical power have 40 days worth of consumable. If you need a longer mission stay time, you'll have to replace the power and life support with a system as massive as the spacecraft's.

Just before takeoff the lander is stripped of every possible gram of excess mass, because it is not carrying much in the way of extra fuel. This includes ripping off the tank insulation, detaching the landing/hover rocket, ditto the remains of the heat shield and landing system. The lander takes off on a bare minimum delta V trajectory into orbit. Two more tiny boosts allow rendezvous with the spacecraft.

The takeoff section has a mass of 1,540 kg, including two crew plus 227 kg of Mars samples and data. If one lander fails, the other can carry all four crew but no Mars samples.


These are two unmanned landers on a one-way trip to land exploration equipment. Each lander has a mass of 1,360 kg and carries 1,810 kg of payload. The total mass for both loaded landers is 6,340 kg. The payload includes scientific equipment, land roving vehicles and their fuel.

Reference Source

     7. Anon: Manned Mars Exploration in the Unfavorable (1975-1985) Time Period. Vol. H. Summary. NASA CR-53911, 1964.
     8. Ehricke, K. A. : A Study of Early Manned Interplanetary Missions (Empire Follow- On). NASA CR-60375, 1964.
     9. Anon. : A Study of Manned Mars Exploration in the Unfavorable Time Period (1957- 1985). Vol. IH. NASA CR-53668, 1964.
     11. Widmer, Thomas F. : Application of Nuclear Rocket Propulsion to Manned Mars Spacecraft. Proc. AIAA and NASA Conf. on Eng. Problems of Manned Interplanetary Exploration, Palo Alto (Calif.), Sept. 30-Oct. 1, 1963, AIAA, pp. 85-101.
     12. Ragsac, R. V., et al. : Manned Interplanetary Missions. Follow-on Study of Final Report. Vol. 1. Summary. NASA CR-56762, 1964.
     14. Shapland, D. J. : Preliminary Design of a Mars-Mission Earth Reentry Module. NASA CR-56209, 1964.
     15. Dixon, Franklin P., and Neuman, Temple W. : Study of a Manned Mars Excursion Module. Vol. I of III - Pt. I. NASA CR-56182, 1963.

Lewis Research Center Mars Ref

This is from Nuclear thermal rocket workshop reference system Rover/NERVA (1991). The author is Dr. Stanley K. Borowski, who helped design several realistic spacecraft in this section of the website.

The point of this paper was to present a "reference design" to use to measure other design proposals presented at the conference. There was a Mars mission reference and a Lunar mission reference, both using solid-core nuclear thermal rocket propulsion based on NERVA technology.

The standard NERVA is basically liquid hydrogen heated in a nuclear reactor then emitted through a converging/diverging exhaust nozzle to create thrust. The hydrogen is compressed to high pressures by a turbopump. It is then "preheated" by cooling the nozzle, reflector, control rods, peripheral shield, and core support structure. Finally it is injected into the reactor.

One of the problems is where to get the energy to run the turbopump. There are two solutions: the "hot-bleed" cycle and the "full flow topping" or "expander" cycle.

In Hot-bleed, about 3% of the hot hydrogen exhaust emitted from the reactor is diverted (left green arrow) to run the turbopump. It is then either used for roll control or reintroduced into the exhaust nozzle. So it is called "hot-bleed" because it is bleeding off some of the hot stuff and using it to make the turbopumps spin.

In Full Flow Topping, all the preheated hydrogen is diverted (right green arrow) to run the turbopump. Then it is injected into the reactor to create thrust. Full Flow Topping has superior specific impulse compared to Hot-bleed, I presume it is much more difficult to engineer.

The Lewis Mars Reference mission was developed by Borowski in 1991, aimed at the Mars launch opportunity in the far-future year of 2016 (heh). It examined both old-school 1972 and more modern 1991 NERVA engines.

In 1969, Werner von Braun described a Mars mission where the spacecraft had triple NERVA engines, using a 640-day opposition class mission with an 80-day stay at Mars and inbound Venus swingby. The Lewis Mars Reference mission uses a spacecraft with a single NERVA, using a of 434-day opposition class mission with a 30-day stay at Mars and an inbound Venus swingby. The Lewis mission is much easier on spacecraft stress and astronaut exposure to galactic cosmic rays.

The Lewis mission came in two options. The "all propulsive" profile uses extra propellant so at the end of the mission the spacecraft can be braked into Terra orbit for future re-use. The other option is to forgoe the extra propellant, carry seven metric tons of Earth Crew Capture Vehicle (ECCV), and at end of mission have the crew use the ECCV to do a aerobraking landing while the abandoned spacecraft sails off into an eccentric heliocentric orbit, never to be used again. Though presumably in future decades the authorities would want to capture and properly dispose of derelict spacecraft with still radioactive engines littering the solar system.

The base assumptions and ground rules for designing the mission are as follows:





  • 34.94t MTV


  • 500 km x 24 hr2 (EARTH ARRIVAL)



  • 340 m/s (dia > 28.5°)
  • 100 m/s (dia < 28.5°)


  • 560 m/s



1 250 km x 33,852 km = 1 SOL ORBIT = 24.66 HOURS
2 500 km x 77,604 km = 24 HOUR ORBIT


Engine Mass3
Ext Shield Mass4
'90 GRAPHITE NERVA850334/758.004.519.4
'90 COMPOSITE NERVA925334/758.824.520.2
'90 CARBIDE NERVA1020334/759.314.520.7
'90 COMPOSITE PHOEBUS9251112/25021.769.037.65



     VARIES WITH TANK SETS: TMI (~ 13%), MOC (~ 15%), COMMON TEI/EOC (~ 16%)

The reference mission's optimization is focused on reducing the Initial Mass in Low Earth Orbit (IMLEO) {which can be thought of as the wet mass}. This is the reason the spacecraft and mission is built around using a single NERVA engine, since those things are heavy. The engine has a thrust of 334 kilo-newtons (75 klbf).

One engine instead of three means the thrust-to-weight (T/W) ratio goes way down. The spacecraft has a lower acceleration, which means it takes longer to escape Terra's gravity, which means the gravity losses become larger, which means more delta V is needed, which means more propellant is needed.

The way to avoid this vicious cycle is to use the magic of the Oberth Effect. By doing the Terra departure burn at perigee (at the point in the orbit when closest to Terra, more generally periapsis) you actually get some delta V for free (actually the extra delta V comes from the potential energy from the mass of the propellant expended).

In this case, you want to do three burns at periapsis to minimize gravity loss. Refer to the graph above. If the spacecraft has a thrust-to-weight ratio of 0.05, escaping Terra with a single burn at periapsis will cost you 1,500 m/s of gravity loss. But if you do the escape with three separate burns at periapsis, the gravity loss is only 350 m/s.

Naturally if you increase the engine thrust in such a way that the T/W ratio goes up, this will also lower the gravity loss penalty. This is tricky since higher thrust engines generally also have a higher mass. But in this case it is almost impossible since the optimization is focused on lowering IMLEO. You'd somehow have to increase the thrust while keeping the mass the same. Maybe by shifting gears.

The report points out if you swap the 334 kilo-newton NERVA engine for a honking monsterous 1,112 kilo-newton Phoebus engine the spacecraft could do a single burn escape with gravity loss of only 400 m/s (the T/W ratio rises to 0.15). The price is the IMLEO rises from 615 metric tons to 750 metric tons.

Boeing Ref
Lewis Ref
EARTH DEPARTURE2/25/20162/25/20163/15/2016
MARS ARRIVAL7/31/20167/31/20168/19/2016
MARS DEPARTURE8/31/20168/31/20169/19/2016
VENUS FLYBY3/10/20173/10/20173/16/2017
EARTH ARRIVAL5/04/20175/04/20175/23/2017
EARTH DEPARTURE C3 (KM2/SEC2)10.3410.3414.07
MARS ARRIVAL VH (KM/SEC)6.826.825.31


The table above compares the IMLEO spacecraft masses of two older Mars reference missions, and the Lewis all-propulsive (resuable) reference mission with a NERVA operating at a specific impulse of 925 seconds. The impressive part is how the optimized trajectory of the Lewis mission saves about 150 metric tons of IMLEO.

On the left is the Lewis all-propulsive optimized reference ship. On the right is the NASA reference ship. The differences are in the sizes of the various propellant tanks, and the IMLEO. The Trans-Mars Injection Drop Tanks are limited by the payload shroud dimesions of anticipated heavy launch vehicles. The report assumes the limit is 10 meters in diameter by 30 meters in length.

For the four NERVA engine types, the initial mass in low Earth orbit (IMLEO) and the total engine burn time was calculated for the mission. The carbide core has the lowest mass, but the 1,112 kN composite core has the shortest burn time.

The 2016 propulsion-optimized 434-day mission was assumed, along with engine thrust of 334 kilo-newtons (75 klbf) or 1,112 kN (250 klbf), 1000 psia chamber pressure, 500-to-1 nozzle expansion ratio, 3 perigee burn Terra departure (1 burn for 1,112 kN composite), and in reuse mode (i.e., spending extra delta V to avoid discarding the ship).

NERVA Engines
EngineTempIspIMLEOBurn time
GRAPHITE CORE2,350 K850 s725 mt202.8 min
(334 kN)
2,700 K925 s613 mt179.4 min
(1,112 kN)
2,700 K925 s750 mt65.3 min
CARBIDE CORE3,100 K1,020 s518 mt158.4 min
Burn Durations
334 kN
(75 klbf)
1,112 kN
(250 klbf)
122.1 min104 min87.8 min38.2
(# perigee burns)
MOC40.0 min36.8 min33.8 min13.4 min
TEI30.0 min28.0 min26.1 min11.0 min
EOC7.1 min6.9 min6.7 min2.7 min
TOTAL199.2 min
(202.8 min)
175.7 min
(179.4 min)
154.4 min
(158.4 min)
65.3 min

TMI = Trans Mars Injection, MOC = Mars Orbital Capture, TEI = Trans Earth Injection, EOC = Earth Orbital Capture

The report looked at other missions across the synodic period. The chart below assumes the spacecraft uses the 334 kN (75 klbf) composite engine with an Isp of 925 seconds. The chart shows how increasing the initial mass in low Earth orbit (IMLEO) shortens the trip time.

"All Prop" is the all-propulsive mission where extra delta-V is spent to capture the spacecraft into Terra orbit for reuse.

"ECCV" is the mission where the extra delta-V is NOT spent, the crew abandons the spacecraft in the Earth Crew Capture Vehicle (ECCV) and aerobrakes to Terra landing, but the spacecraft goes sailing off into the wild black yonder.

"Split-sprint" is where the cargo is sent in an unmanned spacecraft on a Hohmann conjunction-class trajectory, while the crew goes in a manned spacecraft on a faster high-energy opposition-class trajectory.

Not shown is the dangerous "Hohmann tanker/dual vehicle" mission. This is where the unmanned cargo ship also carries the manned spacecraft's return propellant. Which means if the manned ship arrives only to discover that all the return propellant has leaked out, the crew is doomed.

As you can see, the 2018 All Prop mission has an IMLEO of about 700 metric tons at a 434 day mission. If you wanted to decrease the mission time to 365 days (1 year) you'll have to almost double the IMLEO to about 1350 metric tons.

Los Alamos Mars Mission

This is from Manned Mars missions: A working group report (1986). This was mainly to update earlier Mars mission data, examine the impact of new technologies, and identify new issues that needed new research. In other words a lot of this report is a re-hash of old stuff.

The report starts with the old, tired, golly-gee-whiz stuff about how the Space Shuttle was only the beginning. You ain't seen nuttin' yet, folks! This was followed by the obligatory word-salad of miscellaneous excuses why Mars was the logical destination for our valiant astronauts. Things like "going to Mars is an endeavor that can captivate and motivate a generation of young people from throughout the world, just as the American frontier motivated generations past".


The payload is basically:

  • Mission Module: the habitat module the astronauts live in
  • Mars Excursion Module: vehicle that transports the explorers to the Martian surface, then returns them to the spacecraft in orbit at the end of the surface stay.
  • Scientific Equipment: equipment used for scientific observations on the journey to and from Mars, plus on the Martian surface, in an attempt to cram as much science into the expensive Mars mission as they possibly can.


The report figured that anything with an acceleration below 0.05 g's (dotted line in figure 4.22) would increase the trip time over acceptable duration, so should be eliminated from consideration. So the only acceptable propulsion systems are Orion nuclear pulse, Solid-core nuclear thermal rockets (NTR), or Chemical rockets (cryogenic liquid oxygen/liquid hydrogen)

Orion drive with its fuel tanks full of nuclear bombs is a non-starter. Solid-core NTR would be a vast improvement over chemical rockets, but the technology is not quite ready for prime time. Chemical rockets have a lackluster specific impulse, but at the time the report was written they had been around for half a century. It certainly was a mature technology. Most of the report focuses on chemical rocket propulsion (with occasional sunday-newspaper science-supplement comments about pie-in-the-sky stuff like antimatter rockets. Hey kids, ask your grandparents what a newspaper was).


Just like every other Mars mission design in the universe, there are three basic trajectories: Fly-by, Opposition Class, and Conjunction Class.

Fly-by mission are a waste for a manned mission, since there really isn't anything a manned fly-by can do which couldn't be performed vastly less expensively by a robot probe.

Conjunction class missions have a longer mission duration than Oppostion class (three years as opposed to two). They require a longer stay on Mars than Opposition class (one year as opposed to two months). The advantage is that Conjunction class requires less propellant and less initial mass in LEO (IMLEO).

The report suggests that Opposition class is suited for unmanned cargo delivery missions, and for initial short duration manned missions (when the bugs are still being worked out). Later manned missions might be better suited to use Conjunction class.

Low Enriched Uranium NTP

Low Enriched Uranium NTP
Num Enginesx3
Thrust111,200 N
(25 klbf)
Total Thrust333,600 N
Engine Isp875 sec
Exhaust Velocity8,580 m/s
OMS Isp500 sec
OMS Exhaust Velocity4,900 m/s
RCS [*]
FuelNTO / Hydrazine
Deep Space Habitat Module46,783 kg
Inline Stages
Main Usable Propellant
(sans 4% FPR)
27,761 kg LH2
RCS Usable Propellant4,039 kg NTO / Hydrazine
Dry Mass10,696 kg
Inert Mass
(incl. 4% FPR)
13,075 kg
Wet Mass43,875 kg
Stage Length11.1 m
Stage Diameter7.5 m
Tank Diameter7.0 m
Num Stages3
Core Stage
Main Usable Propellant
(sans 4% FPR)
13,449 kg LH2
RCS Usable Propellant3,000 kg NTO / Hydrazine
Dry Mass26,180 kg
Inert Mass
(incl. 4% FPR)
27,426 kg
Wet Mass43,875 kg
Stage Length19.2 m
Stage Diameter7.5 m
Tank Diameter7.0 m
Num NTP Enginesx3
Mission Times
Terra to Mars160 days
Mars Stay620 days
Mars to Terra160 days
NTP Primary Burns
(sans 4% FPR)
TMI ΔV622 m/s
TMI Duration352 sec
MOI ΔV1,668 m/s
MOI Duration823 sec
TEI ΔV1,352 m/s
TEI Duration479 sec
EOI Δ581 m/s
EOI Duration181 sec
Terra Sphere of Influence ΔVs
Launch to NRHO10 m/s RCS
115 m/s OMS
NRHO to LD-HEO95 m/s RCS
100 m/s OMS
LD-HEO to NRHO46 m/s RCS
70 m/s OMS
Mars Sphere of Influence ΔVs
Plane change, Apo-twist250 m/s OMS

This is from NASA's Nuclear Thermal Propulsion (NTP) Project (2018) and STMD: Nuclear Thermal Propulsion Update (2019)

NASA is no stranger to atomic rockets. The Nuclear Engine for Rocket Vehicle Application (NERVA) ran for almost two decades before it got the axe in 1972. Shifting priorities, political winds and space budget cutbacks lead to NERVA's downfall.

The problem is trying to make a Mars mission spacecraft using conventional chemical rockets is like attempting to make a rubber-band powered passenger airplane.

A Mars expedition spacecraft with only chemical rockets can only abort back to Terra within five days of the start of the mission. After that it is committed to the mission. By contrast, a spacecraft using NTP can abort any time from the start up to three months into the mission.

As a ray of hope, in 2018 NASA started a new atomic rocket program, the Nuclear Thermal Propulsion Project, lead by Doyce “Sonny” Mitchell. NASA is partnering with BWXT Nuclear Energy, Inc. on engine development.

One of the obstacles is that the original NERVA project used highly-enriched uranium. Otherwise know as "weapons-grade." The powers-that-be are hysterically afraid of Highly Enriched Uranium (aka "Weapons-Grade") falling into the Wrong Hands. HEU drastically increases the development costs and security regulations.

BWXT Nuclear is well aware of this, and has figured out how to make a solid core nuclear thermal rocket engine using Low Enriched Uranium (LEU). As it turns out there is so much energy in uranium that using HEU is a bit of over-kill, if you pardon the expression. LEU has more than enough power to send a spacecraft to Mars while simultaneously soothing the fears of the powers-that-be.

Chemical engines are hard pressed to make the Terra-Mars section of the mission is less than 258 days. A trio of 111 kN BWXT engines can do it in 160 days flat. That's a bit more than three months less, which is three less months of freefall muscle atrophy and cosmic radiation exposure. Not to mention three months less of life support consumables mass that can be reassigned to mission scientific payload. Actually six months less if you count the return trip.


  • NTP: Nuclear Thermal Propulsion. Another name for Nuclear Thermal Rocket

  • 4% FPR: Flight Performance Reserve, the extra propellant carried in case things go wrong. In this case it is an additional four percent of the propellant mass.

  • OMS: Orbital Maneuvering System, engines used for orbit-to-orbit manuevers

  • RCS: Reaction Control System, the spacecraft's attitude jets

  • TMI: Trans Martian Insertion, engine burn to depart Terra for Mars

  • MOI: Mars Orbit Insertion, engine burn to brake into Mars Orbit

  • TEI: Trans Earth Insertion, engine burn to depart Mars for Terra

  • EOI: Earth Orbit Insertion, engine burn to brake into Terra orbit

  • NRHO: Near Rectilinear Halo Orbit, a stable spot to assemble spacecraft components or for a ship to await the arrival of the crew. The object sort of orbits around EML1 or EML2

  • LD-HEO: Lunar Distant High Elliptical Orbit. An orbit with a perigee near LEO and an apogee near Luna. Starting in this orbit helpfully reduces the delta-V needed for TMI.

The report has a vague reference to an Orbital Maneuvering System (OMS) engine that has a specific impulse of 500 seconds. It is used for manuevers within Terra and Martian spheres of influence. I'm not sure if this is a separate engine or a different mode of the nuclear engine.

Luna from Destination Moon

Retro Rocket Article
PropulsionLiquid Core NTR
Specific Impulse1,050 s
Exhaust Velocity10,300 m/s
Wet Mass226,000 kg
Dry Mass45,000 kg
Mass Ratio5.0
ΔV16,600 m/s
Thrust11,000,000 N
Thrust Power56 gigawatts
Initial Acceleration5g
49 m/s2
Engine Mass9,000 kg
Structure Mass27,000 kg
Payload Mass9,000 kg
Propellant Mass181,000 kg
Dry Volume70 m3
Total Volume100 m3
Length46 m
Ladder Length25 m
Body Diameter5.6 m
Wing Span21 m
Short story
Dry Mass30,246 kg
(32.34 tons)
Propellant Mass196,550 kg
(217.66 tons)
wet mass)
Wet Mass226,796 kg
(250 tons)
Mass Ratio7.69
Initial Acceleration5g
49 m/s2
Thrust11,110,000 N
Exhaust Velocity9,144 m/s
(30,000 fps)
Isp932 sec
ΔV18,640 m/s

Inspired by a post by Retro Rockets I took a look at the classic spaceship Luna from the movie Destination Moon (1950). With Robert Heinlien as technical consultant, this movie was the most scientifically acurate one since Frau im Mond (1929). It held the throne for 18 years, until it was supplanted by the movie 2001: A Space Odyssey (1968).

For the specifications I used data from Spaceship Handbook and the Retro Rockets article. I then massaged the figures until they were internally consistent.

Spaceship Handbook calculated that a round trip mission to the surface of Luna would take about 16,480 m/s of delta V. So that's our performance limit for the mission. In addition, it will have to have a thrust-to-weight ratio greater than 1.0, since it has to lift off from Terra's surface. The movie specifies 5 gs, which translates to 11,000,000 newtons.

The movie specified that the reaction mass was water, not liquid hydrogen. While this does simplify the tankage, it does cut the exhaust velocity/specific impulse in half. Heinlein apparently wanted storable propellant, not liquid hydrogen with all its cryogenic coolers and other complications.

A solid-core nuclear thermal rocket engine is not going to be able to crank out enough delta V, not with water at the specifed mass-ratio it ain't. But the liquid-core Liquid Annular Reactor System (LARS) will do nicely. It can jet out liquid-hydrogen propellant at 20,000 m/s or better, so it can probably manage to hurl water at 10,300 m/s. That will give the Luna a delta-V of 16,600 m/s, just a tad larger than the required 16,480 m/s for the Lunar mission. More than enough, assuming you don't waste a lot of delta V during the landing.

The movie says the structural mass is 27 metric tons, which makes it 60% of dry mass. Nowadays NASA vessels typically have a structural mass of 21.7% of structual mass. 60% is a bit extravagant but believable with 1950's technology. If you made the structure NASA-light, you could add about 17 metric tons to the payload. The payload is the crew, equipment, life support, acceleration couches, and controls.


      “Loading is coming all right,” answered Bowles, “provided the trucks with the oxygen aren’t late.”
     “You should have flown it in.”
     “Quit jittering. The trucks are probably in Cajon Pass this minute.”
     “Okay, okay. Power plant, Doc?”
     “I haven’t broken Ned Holmes’ (Atomic Energy Commission's Tamper-Evident) seal on the atomic pile yet. The water tanks are filling, but they’ve just started.” (spacecraft uses water as propellant)

     The great ship was ringed with floodlights spaced inside the bull pen. It was still framed by the skeleton arch of the Gantry crane, but the temporary anti-radiation shield which had surrounded its lower part down to the jets was gone; instead there were posted the trefoil signs used to warn of radioactivity—although the level of radiation had not yet become dangerously high.

     But the power pile was unsealed and the ship was ready to go. Thirteen-fifteenths of its mass was water, ready to be flashed into incandescent steam by the atomic pile, to be thrown away at thirty thousand feet per second.

(ed note: 13/15th of wet mass is propellant mass. Means mass ratio = 7.69

exhaust velocity 30,000 fps = exhaust velocity 9,144 m/s = specific impulse of 932 sec

ΔV (delta-V) = 18,640 m/s. Spaceship handbook calculated Luna would need 16,480 m/s for the moon round trip so we have enough. )

     High up in the ship was the control room and adjacent air lock. Below the air lock the permanent anti-radiation shield ran across the ship, separating the pressurized crew space from the tanks, the pumps, the pile itself, and auxiliary machinery. Above the control room, the nose of the craft was unpressurized cargo space.
     At its base triangular airfoils spread out like oversize fins—fins they would be as the ship blasted away; glider wings they would become when the ship returned to Earth with her tanks empty.

     Time checks had been completed with Muroc, with White Sands and with their blockhouse. The control room was quiet save for the sighing of air-replenishing equipment, the low hum of radio circuits, and stray sounds of auxiliary machinery. The clocks at each station read 3:29—twenty-four minutes to H-hour.

     The four were at their stations; two upper bunks were occupied by pilot and co-pilot; the lowers by power engineer and electronics engineer. Across the lap of each man arched a control console; his arms were supported so that his fingers were free to handle his switches without lifting any part of his body against the terrible weight to come. His head was supported so that he might see his instruments.
     Barnes could hear sirens, rising and falling, out on the field. Above him in the TV screen, the Luna stood straight and proud, her head in darkness.
     “And ten!
     “And nine!
     “And eight!”—Barnes licked his lips and swallowed.
     The word was lost in sound, a roar that made the test blast seem as nothing. The Luna shrugged—and climbed for the sky.
     Barnes felt himself shoved back into the cushions. He gagged and fought to keep from swallowing his tongue. He felt paralyzed by body weight of more than half a ton; he strained to lift his chest. Worse than weight was noise, a mind-killing “white” sound from unbearable ultrasonics down to bass too low to be heard.
     The sound Dopplered down the scale, rumbled off and left them. At five effective gravities they outraced their own din in six seconds, leaving an aching quiet broken only by the noise of water coursing through pumps.

(ed note: initial acceleration = 5g = 49 m/s2 )

     Corley joined Bowles. Barnes hesitated; he wanted very badly to see, but he was ashamed to leave Traub working. “Wait,” he called out. “I’ll turn ship and we can all see.”
     Mounted at the centerline of the ship was a flywheel. Barnes studied his orientation readings, then clutched the ship to the flywheel. Slowly the ship turned, without affecting its motion along its course. “How’s that?”
     “Wrong way!”
     “Sorry.” Barnes tried again; the stars marched past in the opposite direction; Earth swung into view. He caught sight of it and almost forgot to check the swing.
     Power had cut off a trifle more than eight hundred miles up. The Luna had gone free at seven miles per second (11,265 m/s which is Terra escape velocity); in the last few minutes they had been steadily coasting upwards and were now three thousand miles above Southern California. Below—opposite them, from their viewpoint—was darkness. The seaboard cities stretched across the port like Christmas lights. East of them, sunrise cut across the Grand Canyon and shone on Lake Mead.

     Traub was saying in a steadier voice, “Spaceship Luna, calling White Sands. Come in White Sands.”
     At last an answer came back, “White Sands to Spaceship—go ahead.”
     “Give us a series of radar checks, time, distance, and bearing.”
     A new voice cut in, “White Sands to Spaceship—we have been tracking you, but the figures are not reasonable. What is your destination?”
     Traub glanced at Barnes, then answered, “Luna, to White Sands—destination: Moon.”
     “Repeat? Repeat?”
     “Our destination is the Moon!”
     There was a silence. The same voice replied, “ ‘Destination: Moon’—Good luck, Spaceship Luna!”

     Later the radio claimed his attention. “White Sands, calling Spaceship—ready with radar report.”
     The first reports, plus a further series continued as long as White Sands and Muroc were able to track them, confirmed Barnes’ suspicion. They were tracking “high,” ahead of their predicted positions and at speeds greater than those called for by Hastings’ finicky calculations. The difference was small; on the autopilot displays it was hardly the thickness of a line between the calculated path and the true path.
     But the difference would increase.
     “Escape speeds” for rockets are very critical. Hastings had calculated the classical hundred-hour orbit and the Luna had been aimed to reach the place where the Moon would be four days later. But initial speed is critical. A difference of less than one percent in ship speed at cutoff can halve—or double—the transit time from Earth to Moon. The Luna was running very slightly ahead of schedule—but when it reached the orbit of the Moon, the Moon would not be there.
     Doctor Corley tugged at his thinning hair. “Sure, the cutoff was mushy (when you turn off the engine the thrust does not instantly stop, it sort of gradually tapers off. This was not according to the design), but I was expecting it and I noted the mass readings. It’s not enough to account for the boost. Here—take a look.”
     Corley was hunched at the log desk, a little shelf built into the space between the acceleration bunks. He was strapped to a stool fixed to the deck in front of it. Barnes floated at his shoulder; he took the calculation and scanned it. “I don’t follow you,” Barnes said presently; “your expended mass is considerably higher than Hastings calculated.”
     “You’re looking at the wrong figure,” Corley pointed out. “You forgot the mass of water you used up in that test. Subtract that from the total mass expended to get the effective figure for blast off—this figure here. Then you apply that—” Corley hesitated, his expression changed from annoyance to dismay. “Oh, my God!”
     “Huh? What is it, Doc? Found the mistake?”
     “Oh, how could I be so stupid!” Corley started frenzied figuring.
     “What have you found?” Corley did not answer; Barnes grabbed his arm. “What’s up?”
     “Huh? Don’t bother me.”
     “I’ll bother you with a baseball bat. What have you found?
     “Eh? Look, Jim, what’s the final speed of a rocket (the delta-V), ideal case?”
     “What is this? A quiz show? Jet speed times the logarithm of the mass ratio. Pay me.”
     “And you changed the mass ratio! No wonder we’re running high.”
     “We both did—my fault as much as yours. Listen; you spilled a mass of water in scaring off that truckload of thugs—but Hastings’ figures were based on us lifting that particular mass all the way to the Moon. The ship should have grossed almost exactly two hundred fifty tons at takeoff; she was shy what you had used—so we’re going too fast.”

(ed note: wet mass = 250 tons = 226,800 kg. Shortly before lift-off they did a half-g burn of superheated steam to discourage a truckload of Soviet agents from sabotaging the launch. The expended propellant altered the mass ratio)

     “Huh? I wasted reaction mass, so we’re going too fast? That doesn’t make sense.” Barnes hooked a foot into the legs of the stool to anchor himself, and did a rough run-through of the problem with slide rule and logarithm table. “Well, boil me in a bucket!” He added humbly, “Doc, I shouldn’t have asked to be skipper. I don’t know enough.”
     Corley’s worried features softened. “Don’t feel that way, Jim. Nobody knows enough—yet. God knows I’ve put in enough time on theory, but I went ahead and urged you to make the blunder.”
     “Doc, how important is this? The error is less than one percent. I’d guess that we would reach the Moon about an hour early.”
     “And roughly you’d be wrong. Initial speed is critical, Jim; you know that!”
     “How critical? When do we reach the Moon?” Corley looked glumly at the pitiful tools he had with him—a twenty-inch log-log slide rule, seven place tables, a Nautical Almanac, and an office-type calculator which bore the relation to a “giant brain” that a firecracker does to an A-bomb. “I don’t know. I’ll have to put it up to Hastings.” He threw his pencil at the desk top; it bounced off and floated away. “The question is: do we get there at all?”
     “Oh, it can’t be that bad!”
     “It is that bad.”

     From across the compartment Bowles called out, “Come and get it—or I throw it to the pigs!”
     But food had to wait while Corley composed a message to Hastings (professor at the university who calculated the original course using room-sized computers). It was starkly simple: OFF TRAJECTORY. USE DATA WHITE SANDS MUROC AND COMPUTE CORRECTION VECTOR. PLEASE USE UTMOST HASTE—CORLEY.
     The Luna plunged on; Earth dropped away; radio signals grew weaker—and still no word from Hastings. Corley spent the time trying endlessly and tediously to anticipate the answer he expected from Hastings, using the tools he had.
     “Doctor Hastings calling.”
     “Oh, fine,” Corley acknowledged. “Slide out of there and let me take it.”
     Hastings’ report was short but not sweet. They would reach the Moon’s orbit where planned, but more than fifty hours too soon—and would miss the Moon by more than 90,000 miles!
     Barnes whistled. “Hot pilot Barnes, they call me.”
     Corley said, “It’s no joke.”
     “I wasn’t laughing, Doc,” Barnes answered, “but there is no use crying. It will be tragic soon enough.”
     Traub broke in. “Hey—what do you mean?”
     “He means,” Bowles said bluntly, “that we are headed out and aren’t coming back.”
     “On out? And out—out into outer space? Where the stars are?”
     “That’s about it.”
     “Not that,” Corley interrupted, “I’d estimate that we would reach our farthest point somewhere around the orbit of Mars.”
     Traub sighed. “So it’s Mars, now? That’s not so bad, is it? I mean, they say people live on Mars, don’t they? All those canals and things? We can get another load of water and come back.”
     Corley raised his voice. “Please! Everybody! I didn’t say we weren’t going to get back. I said—”
     “But you—”
     “Shut up, Red! I said this orbit is no good. We’ve got to vector west, toward the Moon. And we’ve got to do it at—” He glanced at a clock. “Good grief! Seven minutes from now.”
     Barnes jerked his head around. “Acceleration stations, everybody! Stand by to maneuver!”

     For forty hours they fell toward the Moon. The maneuver had worked; one could see, even with naked eye, that they were closing with the Moon. Some two hundred thousand miles out the Luna slid past the null point between Earth and Moon, and began to shape her final orbit. It became evident that the correction vector had somewhat overcompensated and that they were swinging toward the Moon’s western limb—“western” as seen from Earth: the Luna’s orbit would intersect her namesake somewhere on the never-yet-seen far side—or it was possible that the ship would skim the far side at high speed, come around sharply and head back toward Earth.
     Two principal styles of landing were possible—Type A, in which a ship heads in vertically, braking on her jets to a landing in one maneuver, and Type B, in which a ship is first slowed to a circular orbit, then stopped dead, then backed to a landing when she drops from the point of rest.
     “Type A, Jim—it’s simplest.”
     Barnes shook his head. “No, Doc. Simple on paper only. Too risky.” If they corrected course to head straight in (Type A), their speed at instant of braking would be a mile and a half a second and an error of one second would land them 8000 feet above—or below!—the surface.
     Barnes went on, “How about a modified ‘A’?”
     Modified Type A called for intentionally blasting too soon, then cutting the jets when the radar track showed that the ship hovered, allowing it to fall from rest, then blasting again as necessary, perhaps two or three times.
     “Confound it, Jim, a modified ‘A’ is so damned wasteful.”
     “I’d like to get us down without wrecking us.”
     “And I would like us to get home, too. This ship was figured for a total change of twelve and a half miles per second. Our margin is paper thin.”
     “Just the same, I’d like to set the autopilot to kick her a couple of seconds early.”
     “We can’t afford it and that’s that.”
     “Land her yourself, then. I’m not Superman.”
     “Now, Jim—”
     “Sorry.” Barnes looked at the calculations. “But why Type A? Why not Type B?”
     “But Jim, Type B is probably ruled out. It calls for decelerating at point of closest approach and, as things stand now, ‘closest approach’ may be contact.”
     “Crash, you mean. But don’t be so damned conventional; you can vector into a circular orbit from any position.”
     “But that wastes reaction mass, too.”
     “Crashing from a sloppy Type A wastes more than reaction mass,” Barnes retorted. “Get to work on a ‘B’; I won’t risk an ‘A.’ ”
     Corley looked stubborn. Barnes went on, “There’s a bonus with Type B, Doc—two bonuses.”
     “Don’t be silly. Done perfectly, it takes as much reaction mass as Type A; done sloppily, it takes more.”
     “I won’t be sloppy. Here’s your bonus: Type A lands us on this face, but Type B lets us swing around the Moon and photograph the back side before we land. How does that appeal to your scientific soul?”
     Corley looked tempted. “I thought about that, but we’ve got too little margin. It takes a mile and a half of motion to get down to the Moon, the same to get up—three miles. For the trip back I have to save enough mass to slow from seven miles a second to five before we dip into the atmosphere. We used up seven to blast off—it all adds up to twelve. Look at the figures; what’s left?”
     Barnes did so and shrugged. “Looks like a slightly fat zero.”
     “A few seconds of margin at most. You could waste it on the transitions in a Type B landing.”
     “Now the second bonus, Doc,” Barnes said slowly. “The Type B gives you a chance to change your mind after you get into a circular orbit; the straight-in job commits you beyond any help.”
     Corley looked shocked. “Jim, you mean go back to Earth without landing?”
     Barnes lowered his voice. “Wait, Doc. I’d land on the Moon if I had enough in tanks to get down—and not worry about getting up again. I’m a bachelor. But there’s Mannie Traub. No getting around it; we stampeded him. Now it turns out he has a slew of kids, waiting for poppa to come home. It makes a difference.”

     Traub’s features worked in agonized indecision. “But it’s not my business to decide!” Bowles spoke up suddenly. “You’re right; it’s not!” He went on, “Gentlemen, I didn’t intend to speak, because it never crossed my mind that we might not land. But now the situation demands it. As you know, I received a coded message.
     “The gist was this: our trip has caused grave international repercussions. The Security Council has been in constant session, with the U.S.S.R. demanding that the Moon be declared joint property of the United Nations—"
     “As it should be," Corley interrupted.
     “You don’t see the point, Doctor. Their only purpose is to forestall us claiming the Moon—we, who actually are making the trip. To forestall us, you understand, so that the United States will not be able to found a base on the Moon without permission—permission that is certain to be vetoed.”
     “But,” pointed out Corley, “it works both ways. We would veto Russia establishing a base on the Moon. Admiral, I've worked with you because it was a way to get on with my life's ambition, but, to be frank, using the Moon as a rocket launching base—by anybody—sticks in my craw.”
     Bowles turned red. “Doctor, this is not an attempt to insure the neutrality of the Moon; this is the same double-talk they used to stop world control of atomics. The commissars simply want to tie us up in legalisms until they have time to get to the Moon. We'll wake up one morning to find Russia with a base on the Moon and us with none—and World War Three will be over before it starts.”
     “But—Admiral, you can’t know that."
     Bowles turned to Barnes. “Tell him, Jim."
     Barnes gestured impatiently. “Come out of your ivory tower, Doc. Space travel is here now—we did it. There is bound to be a rocket base on the Moon. Sure, it ought to be a United Nations base, keeping the peace of the world. But the United Nations has been helpless from scratch. The first base is going to belong to us—or to Russia. Which one do you trust not to misuse the power? Us—or the Politburo?"
     Corley covered his eyes, then looked at Bowles. “All right,” he said dully. “It has to be—but I don’t like it."
     Traub broke the ensuing silence with “Uh, I don’t see how this ties in with whether we land or not?"
     Bowles turned to him. “Because of this: the rest of that message restored me to active duty and directed me to claim the Moon in the name of the United States—as quickly as possible. We would have what the diplomats call a fait accompli. But to claim the Moon I have to land!”

     Barnes locked eyes with him. “You can take your authority and—do whatever you think proper with it. I'm skipper and will stay so as long as I'm alive." He looked around. “All hands—get ready for approach. Doc, go ahead with trial calculations, Type B. Mannie, warm up the pilot radar. Bowles!"
     Finally Bowles answered, “Yes, sir."
     “Rig the autocamera in the starboard port. We’ll take a continuous strip as we pass around the far side."
     “Aye aye, sir.”
     Traub leaned from his couch and peered out the starboard port. “lt’s just like the other side.”
     Barnes answered, “What did you expect? Skyscrapers? Co-pilot, how do you track?”
     “Speed over ground—one point three seven. Altitude, fifty-one point two, closing slowly."
     “Check. I project closest approach at not less than twenty-one—no contact. What do you get?”
     “Closer to twenty, but no contact."
     “Check. Take over orientation. I'll blast when altitude changes from steady to opening."
     “Aye aye, sir!"
     The Luna was swinging around the unknown far face of the Moon, but her crew was too busy to see much of the craggy, devil-torn landscape. She was nearing her closest approach, traveling almost horizontally. She was pointed tail first, ready to blast back from a top speed of a mile and a half a second to a circular orbit speed of a mile a second. At Barnes‘ order Bowles gave his attention to placing her axis precisely horizontal.
     The television screen read “View Aft”; in its center was a cross mark lying over a picture of the mountainous horizon they were approaching. He jockeyed the ship against the re- action of the flywheel, then steadied her by gyros when one cross line held steady on the horizon.
     Barnes set his controls on semiautomatic, ready both to fire and cut off with one punch of the firing button. Into his autopilot he fed the speed change he wished to achieve. Altitude dropped to forty miles, to thirty, to less than twenty-five. “Power plant,” Barnes called out, “stand by for blasting!"
     “Ready, Jim,“ Corley reported quietly.
     “Everything sweet, Skipper."
     Barnes watched ground speed with one eye, the radar altimeter with the other … twenty-three, it said … twenty-two … twenty-one and a half…
     Twenty-one point five … twenty-one point four—point four again—and again. Point five! and crawling up. His finger stabbed at the firing button.
     The blast was fourteen seconds only, then it cut off, but in the same mushy fashion which it had before. Barnes shook his head to clear it and looked at his board. Altitude twenty-one point five; ground speed, one plus a frog‘s whisker—they were in orbit as planned. He sighed happily. “That’s all for now, troops. Leave everything hot but you can get out of your hammocks."
     Bowles said, “Hadn’t I better stay and watch the board?"
     “Suit yourself—but they won’t repeal the law of gravitation. Doc, let’s see how much juice we have left." He glanced at a clock. “We've got an hour to make a decision. It will be almost half an hour before Earth is in sight again."
     “I don’t like the way she cuts off," Corley complained.
     “Quit fretting. I used to have a car that sounded its horn every time I made a left turn." Bowles got a container of coffee, then joined Traub at the starboard port. They peered around the automatic camera and watched the moonscape slide past. “Rugged terrain," Bowles remarked.
     Traub agreed. “There’s better stuff going to waste in California." They continued to stare out. Presently Bowles turned in the air and slithered back to his acceleration couch.
     Mannie came to the desk. “Mannie,” Barnes said, pointing at a lunar map, “we figure to land spang in the middle of the Earthside face—that dark spot, Sinus Medii. It's a plain.”
     He was interrupted by Bowles’ voice. “Captain! We are closing, slowly."
     “Are you sure?”
     “Quite sure. Altitude nineteen point three—correction: point two … closing.”
     “Acceleration stations!”
     Barnes was diving toward his couch as he shouted. Traub and Corley followed him. As he strapped down Barnes called out, “Co-pilot—get a contact prediction. All hands, stand by for maneuvers." He studied his own board. He could not doubt it; they were in something less than a perfect circle.
     He was trying to make a prediction from his display when Bowles reported, “I make it contact in nine minutes, Captain, plus or minus a minute.”
     Barnes concentrated. The radar track was jiggling as much as five or ten percent, because of mountains below them; the prediction line was a broad band. As near as he could tell, Bowles was right.
     “What now, Captain?” Bowles went on. “Shall I swing her to blast forward?" A slight nudge would speed up the ship, in effect, lift her, permit her to fall around the Moon rather than curve down.
     It would also waste reaction mass.
     Nine minutes . . . nine hundred miles, about. He tried to figure how many minutes it would be until they raised Earth over the horizon, ahead.
     Seven minutes, possibly—and Earth would be in sight. A landing at Sinus Medii was impossible but they still might land in sight of Earth without using more precious water to correct their orbit. Co-pilot, steady ship for deceleration. Sing out when you see Earth."
     “Aye aye, sir!"
     “There’s Earth!“
     Barnes glanced up, saw Terra pictured in the TV screen, rising behind a wall of mountains. Bowles went on, “Better land, Jim. You'll never get over those mountains."
     Barnes did not argue; their altitude was barely three miles now. He shouted, “Stand by. Red, start swinging as soon as I cut off."
     “Fire!” He stabbed the button. This maneuver was manual, intended only to stop their forward motion. He watched his ground-speed radar while the ship shivered—nine-tenths. . . seven . . . five. . . four. . . three . . . two . . . one. . . six-hundredths.
     He jerked his finger off just before it dropped to zero and prayed that a mushy cutoff would equal his anticipation.
     He started to shout to Bowles, but the ship was already swinging.
     Earth and the horizon swung up in the TV screen and out of sight.
     For a crawling ten seconds, while they fell straight down, the Luna crept into position for a tail-first landing. They were less than three miles up now. Barnes shifted scale from miles to feet and started his prediction.
     Bowles beat him to an answer. “Contact in seventy-two seconds, Skipper."
     Barnes relaxed. “See the advantage of a Type ‘B’ landing, Doc," he remarked cheerfully. “No hurry—just like an elevator."
     “Quit gabbing and get us down,” Corley answered tautly.
     “Right,” Barnes agreed. “Co-pilot, predict the blast altitude."
     His own hands were busy to the same end.
     Bowles answered, “Jim, you going manual or automatic?”
     “Don’t know yet." Automatic firing was quicker, possibly more certain—but that mushy cutoff could bounce them like a pingpong ball. He steadied cross-hairs on his autopilot display and read the answer: Blast at five two oh feet. What do you get, Red?"
     “Check." Bowles added, “That’s less than three seconds blast, Jim. Better make it automatic."
     “Tend to your knitting."
     “My mistake."
     Nearly forty seconds passed and they had fallen to eleven thousand feet before he decided. “Power plant, set for manual landing. Co-pilot, cover me at five hundred feet.”
     “Jim, that’s too late,” Bowles protested.
     “You will be covering me all of a tenth of a second after I should fire."
     Bowles subsided. Barnes grabbed a glance at the TV screen; the ground under them seemed level and there was no perceptible drift. He looked back at his board. “Correction—cover at five ten.”
     “Five ten—right."
     The seconds clicked past; he had his finger poised over the button when Bowles shouted, “Jim—look at the screen!"
     He looked up—the Luna, still carrying a trifle of drift, was now over a long crack, or rill —and they were about to land in it. Barnes jabbed the button.
     He let up at once; the Luna coughed to silence. The rill, canyon, or crevasse was still in sight but no longer centered. “Co-pilot—new prediction!"
     “What happened?" Corley demanded.
     “Prediction,” Bowles chanted, “blast at…at three nine oh."
     Barnes was adjusting verniers for his own prediction as Bowles reported. “Check,” he answered. “Cover at three seven oh." He threw one glance at the TV screen. The crevasse was toward the edge of the screen; the ground below looked fairly smooth. Unquestionably the ship had a slight drift. All he could do was hope that the gyros would keep them from toppling. “Brace for crash!”
     480—450—400— He jabbed the button.
     The terrible pressure shoved his head back; he lost sight of the altimeter. He caught it again—190—150—125—
     At “fifty" he snatched his finger away and prayed.
     The jet cut off sloppily as always. A grinding jar slammed him more deeply into the cushions. The ship lurched like an unsteady top—and stayed upright.
     Barnes found that he had been holding his breath a long time.

From DESTINATION MOON by Robert Heinlein (1950)

Lunar Transportation System

Lunar Transfer Vehicle
Module Core
Inert Mass8.1 t
Propellant Mass7.0 t
Crew Module
(incl. crew)
8.4 t
Height14.4 m
(incl. drop tanks
less cargo)
15.2 m
Lunar Transfer Vehicle
Drop Tanks
Isp481 s
Num Engines4
Thrust89,000 N
Total Thrust356,000 N
Inert Mass5.8 t
Propellant Mass129.8 t
Lunar Excursion Vehicle
Isp465 s
Num Engines4
Thrust89,000 N
Total Thrust356,000 N
Inert Mass5.8 t
Propellant Mass22.4 t
Crew Module
(incl. crew)
4.4 t
Cargo Mass
15 t
Cargo Mass
33 t
Height8.5 m
Body Width
(less cargo)
7.5 m
Landing Gear
11.3 m

This is from Report of the 90-Day Study on Human Exploration of the Moon and Mars (1989) and from

This transport system has two components: the Lunar Transfer Vehicle (LTV) and the Lunar Excursion Vehicle (LEV). The LTV transports crew and cargo between Low Earth Orbit (LEO) and Low Lunar Orbit (LLO), part of the cargo could be a LEV. The LEV transports crew and cargo between LLO and the lunar surface.

Lunar Transfer Vehicle

The LTV is a "one and one-half" stage design, with a reusable core surrounded by expendable propellant tanks. This reduces the propellant load by about 10% compared to a single-stage reusuable vehicle. The core contains the propulsion/avionics module, the main propellant tanks, the aerobraking shield, the crew module (if any), and other assorted subsystems.

The LTV and LEV are boosted into orbit in a single heavy-lift launch vehicle. The LTV will be boosted with the core fully fueled, but the LEV will only be partially fueled due to the payload limit of the launch vehicle. The four fully loaded drop tanks will be boosted into orbit by two subsequent heavy-lift vehicles. The crew and any cargo modules would be boosted by the space shuttle.

Some in orbit assembly will be required: adding drop tanks to LTV core, the eight peripheral aerobrake segments attached to the LTV aerobrake shield core, and the cargo modules added to the LEV.

The LTV does a trans-lunar injection burn, and jettisons two empty drop tanks. It brakes into LLO and drops the two remaining empty drop tanks. It then acts as a staging base in LLO for the LEV.

If there is already an empty LEV in LLO parking orbit waiting to be reused, the LTV loads it with propellant, consumables, and attaches new cargo modules.

When the LEV has performed its mission, the LTV does a trans-Earth burn using the core propellant tanks.

It circularizes itself into LEO using aerobraking instead of propellant, at a considerable savings in initial mass required in LEO at mission start. After each mission the aerobrake shield is refurbished and verified at the International Space Station. The aerobrake shield can be reused for five missions.

The optional LTV crew module provides habitable support for the crew for the 4 day translunar trip and up to 7 days for the return to the space station. Naturally the crew can override the automatic rendezvous and docking system. Crew module obtains electricity from the LTV, has a two-gas open-loop environmental control and life support system, has a galley, zero-gravity toilet, and a personnel hygiene station.

The crew module has docking ports fore and aft, passing through one is an intravehicular activity (no space suit required). There is no airlock, so extravehicular activity requires all the crew to don space suits and depressure the entire module. There is enough repressurization gas carried for 2 EVAs.

The crew module carries a storm cellar with walls filled with water radiation shielding. The water is vented before aerobraking to save wear and tear on the aerobrake shield.

Lunar Excursion Vehicle

In "reusable" mode, the LEV can transport 15 metric tons of payload to the lunar surface (along with a crew and crew module), and return to LLO. It can be reused up to five missions.

In "expendable" mode the LEV can transport 33 metric tons of payload to the lunar surface (with neither crew nor crew module) and stays on the surface forever after.

If the cargo load is small enough, an unmanned LEV with no crew module has enough of an automatic pilot to be able to land, discharge cargo, return to orbit, and rendezvous with the orbiting LTV.

The LEV and LTV shares a lot of systems designs to reduce development and testing time (such as engines, cryogenic RCS, avionics, software, communiciation equipment, fuel cells, etc.).

When the LEV is parked in lunar orbit and abandoned, it is powered by solar arrays. On the lunar surface, the propellant system is designed for 30 days. For longer stays it will require surface support (from in-situ resource utilization).

The LEV's crew module is related to the LTV crew module, but with some differences. It has no storm cellar. It transports four crew members between the LTV and the lunar surface. During landing operations two crew members have landing control panels and windows, the other two are in shock webing and just have to be patient and stare at the walls.

The LEV's crew module's systems are in a quiescent state, except for 4 days during descent/ascent missions (2 days during descent and initial surface operations, 2 days for preparation and ascent to orbit). While quiescent the crew module has no interal power, thermal control, or propellant conditioning. Bottom line is either the descent/ascent mission only lasts 4 days, or there has to be support systems available on the lunar surface (a lunar base in other words).

Just like the LTV crew module, the LEV crew module has no airlock and only enough repressurization gas for 2 EVAs.


Mass Schedule
Total Length16.16 m
Max Diameter7.62 m
PropulsionChemical LOX/LH2
Glider Length9.30 m
Glider Mass9,163 kg
Booster Rocket Thrust26,700,000 N
Mass Schedule
a. Body3,402
(1) Structure1,588
(2) Heat Shield1,814
b. Wing Group907
(1) Structure363
(2) Heat Shield544
c. Control System352
(1) Aerodynamic272
(2) Attitude79
d. Environmental Control694
(1) Equipment Cooling63
(2) Structure Cooling426
(3 ) Cryogenic Storage205
e. Landing Gear318
f. Instruments & Displays91
g. Electric Power System272
h. Guidance & Navigation181
i. Communications113
j. Furnishings & Equipment386
(1) Seats & Restraints102
(2) Decompression Chamber79
(3) Equipment Compartment136
(4 ) Miscellaneous68
k. Life Support181
1. Crew (3 men)272
m. Radiation544
n. Abort System1,361

This is from Lunar Expedition Plan from the headquarters Space Systems Devision of US Air Force Systems Command (1961).

Please do not confuse this with LUNOX which has a similar name, that is about using lunar mined oxygen with LANTR nuclear thermal rockets.

This was a US Air Force project started in 1958 designed to seize the "high ground" of space, about three years before the start of NASA's Apollo program. The idea was to ensure that the first lunar landing was made by the US (showing those pesky Soviets who is top dog in space), establish a 21-crew lunar USAF base, and use the base like a spy satellite to keep an eye on US enemies. And maybe even site a few nuclear missiles at the lunar base.

The project was cancelled in 1961, mostly due to the USAF's increasing preoccupation with how the Viet Nam War was devouring their budget. Contributing factors were that it would cost x1.5 the Apollo program, as the study progressed it became obvious the scheduled time-line was incredibly overoptimistic, President Kennedy wanted the moon race to be a civilian effort not a military one, and only an idiot would put a spy satellite 400,000 kilometers away from Terra's surface on Luna when the satellite would have a much better view from a 200 kilometer Low Earth Orbit. The only thing more stupid is siteing nuclear missiles at a lunar base, since it provides zero advantage over siteing them on Terra but does inflict lots of severe handicaps.

So that the project wasn't a total waste, the USAF released a partially declassified report in order to disconcert the Soviets. Look what we were planning, and you didn't even know. Makes you wonder what else we are up to, eh Khrushchev? You better watch your step.

If the project had actually created a real rocket, the advantage was that the US would have the technology for a great launch vehicle very similar to the Space Shuttle. A booster with a LOX-LH2 core, huge strap-on solid rocket boosters, and a flying re-entry vehicle. This would be a great set of technologies for future space programs. Instead we went with the Apollo program, which gave us the Saturn V rocket which will never be used again, the Lunar Module which isn't good for much except allowing three crew and a minuscule payload to visit the lunar surface, and the slightly less useless Apollo command/service module.

Having said that, the LUNEX design had some severe problems.

Mainly because it was using a direct ascent design. This is when you boost the lunar spaceship into orbit as one piece, land the entire clanking mess on Luna, then blast-off the entire clanking mess (less the landing rockets) directly into a return to Terra.

The point is that every gram counts. Particularly during landing and lift-off. Each extra gram is going to require tons more fuel. This is why the Apollo mission didn't send down to the surface the command/service module along with the lunar module, because it wasn't needed on the moon. Only an idiot would design a mission like that, or a designer terrified of having the astronauts being forced to perform an orbital rendezvous with the return spacecraft.

The fuel required for LUNEX's direct ascent was so outrageous that the spacecraft could not carry enough. The program would have to use an unmanned cargo rocket to land 20 tonnes of fuel first. Once they were in place, LUNEX could land next to the cargo rocket and refuel. Or accidentally land too far from the cargo rocket and inflict a public relations nightmare on the USAF as the stranded astronauts waited for their oxygen to run out. The cargo rocket also carries a landing beacon to provide terminal guidance for the crewed spacecraft. The cargo rocket is basically just the LUNEX's landing stage with the fuel tank cargo perched on top, it is missing the launch stage and the reentry glider.

Assuming everything went according to plan, the spacecraft would return to Terra. Shortly before arriving all parts of the spacecraft would be jettisoned except for the hypersonic lifting body. This would have to enter Terra's atmosphere at a blistering 11.3 kilometers per second, aerobrake while hoping they run out of excess velocity before they run out of heat shield, and land on a aircraft runway at Edwards Air Force Base. This would be a huge engineering task since basic data on such hypersonic reentry had not been explored yet. Did I mention already how incredibly overoptimistic the development time-line was?

At the start of the mission, the booster does not send the spacecraft in Terra orbit. No, it is boosted directly into trans-Lunar trajectory (the same way as the Apollo missions). The trip to Luna will take about and one-half days. The spacecraft will arrive pretty much still with a mass of 61,000 kilograms. All of which will have to be delta-Veed about 2,737 meters per second to land the entire contraption next to the uncrewed cargo ship, since the landing legs can only handle about 6 m/s of jolt.

The crew will spend five days exploring, then lift-off directly into their Terra return trajectory.

When aerobraking, strict accuracy is required. The re-entry angle must be within ± 2° of optimal. Too steep and the crew will be flattened by the gee forces while being incinerated by the heat. Too shallow and they will ricochet off the atmosphere into an eccentric orbit. By the time they can make another attempt, the crew will run out of air and/or suffer a lethal radiation dose while cruising the Van Allen belt.

The 2.5 day trip to Luna and 2.5 day trip back home was selected as optimum. Longer flights would have more problems with life-support and guidance. Shorter flights would need excessive amounts of fuel.

Lighter and Tanker

Specific Impulse450 s
Exhaust Velocity4,410 m/s
Wet Mass56,300 kg
Dry Mass25,898 kg
Inert Mass898 kg
Payload25,000 kg
Mass Ratio2.17
ΔV3,410 m/s
Mass Flow31.8 kg/s
Thrust140 kiloNewtons
Initial Acceleration0.25 g
Length18.3 m+engine
Diameter≈4.57 m
gas core NTR
Specific Impulse3,600 s
Exhaust Velocity35,000 m/s
Wet Mass433,000 kg
Dry Mass268,000 kg
Mass Ratio1.61
ΔV16,730 m/s
Inert Mass
(dry mass - payload)
108,000 kg
Payload Mass total160,000 kg
Payload Mass Hydrogen
(less tankage)
139,000 kg
Mass Flow100 kg/s
Thrust3,500 kiloNewtons
Initial Acceleration0.6 g
Length37 m+engine
Diameter≈18 m

These two designs are from The Resources of the Solar System by Dr. R. C. Parkinson (Spaceflight, 17, p.124 (1975)). The Lighter ferries tanks of liquid hydrogen from an electrolyzing station on Callisto into orbit where waits the Tanker. Once the Tanker has a full load of tanks it transports them to LEO. All the ships are drones or robot controlled, there are no humans aboard. The paper makes a good case that shipping hydrogen from Callisto to LEO would eventually be more economically effective than shipping from the surface of Terra to LEO, with the break-even point occurring at 7.8 years. Please note that this study was done in 1975, before the Lunar polar ice was discovered, and probably before the ice of Deimos was suspected.

Warning: most of the figures in the table are my extrapolations from the scanty data in the report. Figures in yellow are sort of in the report. Use at your own risk.

The tanker uses a freaking open-cycle gas-core nuclear thermal rocket. This is an incredibly powerful true atomic rocket, but it is only fractionally more environmentally safe that an Orion nuclear bomb rocket. The report says it should be possible to design it so the amount of deadly fissioning uranium escaping out the exhaust is kept down to as low as one part per 350 of the propellant flow (about 300 grams per second), but I'll believe it when I see it. Since it is used only in deep space we can allow it, this time. The report gives it an exhaust velocity of 35,000 m/s, which is about midway to the theoretical maximum.

The lighter can get by with a more conventional hydrogen-oxygen chemical rocket. It will need an acceleration greater than Callisto's surface gravity of 1.235 m/s2, for safety make it 1.5x the surface gravity, or about 1.9 m/s (0.6g).

The four major Galilean moons are within Jupiter's lethal radiation belt, except for Callisto. The black monolith from 2010 The Year We Make Contact only told us puny humans to stay away from Europa, so Callisto is allowed. If you want ice that isn't radioactive, you've come to the right place. It is almost 50% ice, and remember this is a moon the size of planet Mercury. That's enough ice to supply propellant to the rest of the solar system for the next million years or so. Europa has more, but it is so deep in the radiation belt it glows blue. Callisto is also conveniently positioned for a gravitational sling shot maneuver around Jupiter to reduce the delta-V required for the return trip to Terra.

The report says that the requirements for an economically exploitable resource are:

  1. It is not available in the Terra-Luna system
  2. It must provide more of it than the mass originally required to be assembled in Terra orbit at the outset of the expedition
  3. It must be done within a reasonably short time (the break-even time)

Hydrogen fits [1], or at least it did until the Lunar ice was discovered. [2] and [3] depend upon the performance of the vehicle.

There are three parts. First is the Tanker, which is an orbit-to-orbit spacecraft to transport the hydrogen back to LEO and brings the expedition to Callisto in the first place. Next is an electrolysis plant capable of mining ice, melting it into water, cracking it into oxygen and hydrogen, and liquefying the hydrogen. Last is a Lighter which is an airless lander that ferries liquid hydrogen from the plant on Callisto to the orbiting Tanker.

The report decided to use modular cryogenic hydrogen tanks that would fit in the Space Shuttle's cargo bay. They would have to be about 18.3 meters x 4.57 meters, about 300 cubic meters capacity. The report has a filled tank massing at 26,000 kg, with 22,000 kg being liquid hydrogen and 4,000 kg being tank structural mass. Examining the drawing of the tanker, the front cluster is composed of four tanks while the rear has nine, for a total of thirteen. The tanker will have a length of two tanks plus the length of the rocket engine, 37 meters plus rocket. The rear has tanks arranged in a triangular array about four tanks high. So a diameter roughly 18 meters or so.

The lighter carries a single tank, so it is roughly one tank in diameter, and one tank long plus the fuel tanks+engine length. It will need a large enough liquid hydrogen/liquid oxygen chemical fuel capacity to lift off from Callisto to the tanker and land back on Callisto.

The report figures that the electrolysis plant can produce hydrogen for about 39 kW-h/kg, that is, each kilogram of hydrogen in the plant requires 39 kilowatt-hours. Figure it needs more electricity to liquefy the hydrogen, and more to produce the liquid oxygen needed by the lighter, for a total cost of 50 kW-hr/kg for liquid hydrogen delivered to the orbiting tanker. So a 2 megawatt nuclear reactor could produce 350 metric tons of hydrogen per year. Launch windows back to Terra occur every 398.9 days.

Once the lighter has made enough trip to fully load the tanker, the tanker departs for LEO. It will use some of the hydrogen for propellant, some will be the payload off-loaded at LEO, and enough will be left to return the tanker to Callisto. The amount of payload is specified to have a mass equal to 37% of the fully loaded mass of the tanker. It also specifies that the inert mass fraction of the tanker is 25% of the tankers fully loaded mass.

The report had an esoteric equation that calculated the mass of the lighter and electrolysis plant as a percentage of the tanker mass in order to be economically viable. It turns out to be 13% of the fully loaded mass of the tanker. When the expedition is launched the tanker will carry the lighter, the electrolysis plant, and enough propellant so that the total mass is 52.9% of the fully loaded mass (i.e., it departs half empty). The lighter will have its tanks full.

Five years later, upon arrival at Callisto, the lighter lands the electrolysis plant on a prime patch of ice. It then starts the cycle: patiently waiting for the plant to fill the payload tank and the fuel tanks, boost the payload to the tanker, then land back at the plant to start again.

In context: this was one of a series or articles I wrote for Spaceflight at the encouragement of the then editor, Ken Gatland, triggered off in the dark days following abandonment of the Apollo programme by a discussion at the BIS as to what would be needed to make spaceflight self-supporting. The first article was published in Spaceflight 1974 p.322 under the title "Take-Off Point for a Lunar Colony." There was then a second on "The Colonization of Space" (S/F 1975 p.88) and a couple of subsequent ones on Lunar Colonies (S/F 1977 p.42/103). Later, when they invented the first spreadsheets, I did some speculation on how the economics of everything might fit together economically in a big input-output model which got published as "The Space Economy of 2050 AD" in JBIS v.44, p.111 (1991) which also appeared in my book Citizens of the Sky (1989) later. It is unlikely that I was consistent through all of this — my opinions develop with time — and by the 1991 period I was heavily in to the economics or reusable launchers and what would happen if the models were pushed to very high flight rates.

Going back to "The Resources of the Solar System", I'm not sure how much detail I managed at the time. I remember that there were a couple of things influencing me at the time. One was the concept of a gas core nuclear engine (GCR) which might have a specific impulse of about 3500 sec (35 km/s). To really move around the Solar System you need a high thrust-to-mass engine with this sort of specific impulse, and GCR had the interesting property of using hydrogen as propellant. (Ion motors can meet the specific impulse, but to do a similar job would require a power-mass ratio several orders better than anything we could consider then or even today — VASIMIR suffers the same problem). Nowadays I might put my money more on a pulsed-fusion system (see “Using Daedalus for Local Transport,” JBIS, 62, p. 422-426 (2009)) — note NOT using helium-3, which would change the model significantly.

The second thing at the time that influenced me — at a time when the Space Shuttle was still a paper vehicle — was that the Space Shuttle payload bay was just about the right size (15 ft × 60 ft) to carry a full liquid hydrogen tank (there are reasons now why it wouldn't which led to the abandonment of design work using Centaur as an upper stage) — so my modular design was based around using that as a standard tank. For use in long duration space missions the tank would have to have some sort of active cooling system to keep the hydrogen from boiling away, but given that you could then ship LH2 around the Solar System on slow, economical trajectories like modern oil tankers on Earth. Once you have rerfuelling stations at either end interplanetary flight becomes a lot easier and you can think of using higher speed trajectories for special cargo like human beings.

From personal email from Dr. Parkinson (2014)

All the other figures in the table are ones I've extrapolated from the few figures given in the report.

A plausible figure for nuclear power generation is 0.12 Megawatts per ton of generator. This would make the electrolysis 2 MW power reactor have a mass of 16,000. This is close to the 25,000 kg mass of a payload tank. So to simplify, assume the electrolysis rig with liquefaction gear and all masses a total of 25,000. This also ensures that the lighter is capable of landing it.

The tanker's inert mass fraction is 25%, and hydrogen payload is 37%. This means the dry mass is 62%, which means the mass ratio is 1.61. With an exhaust velocity of 35,000 m/s, this yields a total delta-V of 16,730 m/s. I am unsure if this is enough for a Callisto orbit-LEO mission followed by a LEO-Callisto orbit mission. Not without a heck of a gravitational sling-shot it isn't. Or I could have made a mistake in math.

Note both the payload and the propellant is hydrogen, stored in the same array of tanks. If the inert mass fraction is 25%, then the payload+propellant mass fraction is 75%. If there are 13 tanks each of 25,000 kg, then the total is 325,000 kg. If this is 75% of the wet mass, the actual wet mass is 433,000 kg. If the payload is 37% of the wet mass, it is 160,000 kg. If a hydrogen tank is 87% hydrogen and 13% tankage, the amount of hydrogen payload is 139,000 kg.

On the initial trip, the tanker carries the electrolysis plant and the lighter (with no payload, but with full fuel tanks). This is 13% of the wet mass or 56,300 kg. If the electrolysis plant is 25,000 kg, the lighter (with no payload) must be 31,290 kg. The lighter payload is one payload tank at 25,000 kg. So the lighter wet mass is 56,290 kg.

The lighter needs a delta-V of 3,414 m/s (Callisto-surface-to-orbit + orbit-to-Callisto-surface). Chemical fuel has exhaust velocity of 4,410 m/s. This means the mass ratio has to be 2.17. This implies the dry mass is 25,898 kg. Subtract the 25,000 kg payload, and there is 898 kg for the structure and the engine. Seems a little flimsy to me, perhaps 25,000 kg is a bit to generous for the payload tank.

Tank is scaled to fit in Space Shuttle cargo bay. At least the the proposed size of the bay in 1975 when the report was written, it was later reduced in size.

  • 4.55 m wide × 18.2 m long
  • Mass 26 metric tons
  • LH2 Mass 22 metric tons

Notes, from left to right:

  1. Docking Ring
  2. Limited amount of pressurization equipment round head end, also radar transponder
  3. Strong ring with attachment points
  4. Recirculating and pressurization pipes
  5. Strong ring with attachment points
  6. Docking Ring
  7. Fill valve connect. Associated propellant management equipment
  8. Radiator

Manned Mars Explorer

Mass Schedule
98,000 kg
70,000 kg
Tether System13,000 kg
Power System20,700 kg
32,000 kg
Comm Satellite3,500 kg
Crew Command
50,000 kg
(Scenario 1)
3,412,800 kg
(Scenario 2)
3,812,800 kg
(Scenario 1)
3,700,000 kg
(Scenario 2)
4,100,000 kg

This is from a student project for Master of Architecture candidates at the University of Houston: Manned Mars Explorer project: Guidelines for a manned mission to the vicinity of Mars using Phobos as a staging outpost.

After weighting all the options, chemical propulsion was chosen. Nuclear electric had too many drawbacks.

The tyranny of the rocket equation led them to go with reliability over redundancy. Equipping the spacecraft with back-up units for all critical systems cuts too much into payload mass. Instead they went with single units that were super-duper fault tolerant.

Medical issues dictated supplying the crew with a full one-Terran-gravity. An elaborate bola system was designed. The system resists twisting via a unique spreader system and four tether configuration. Spin grav is used for the trans-Mars coast and the trans-Terra coast. The tether is reeled in before each propulsive manuever to prevent the spacecraft from destroying itself by the crack-the-whip effect.


An opposition class Venus inbound swingby was used for the trajectory. About 300 days are spent travelling to Mars. It spends 60 days in Mars orbit. It uses the Venus inbound swingby leg to travel to Terra LEO which takes 210 days. This trajectory was chosen due to relatively short overall mission and Mars stay time. It does however require more delta V than conjunction class trajectories.

The sixty day exploration period is mostly focused on Phobos and Deimos, but there is a segment where a crew of three is sent to the surface of Mars for seven days. Staging bases will be set up on the moons, and they will be assesed for deposits of water ice and other valuable in-situ resource utilization goodies.

MPV is the Manned Planetary Vehicle, the spacecraft. CCV is the Crew Command Vehicle, a small auxiliary spacecraft carried as payload.

Mission Phases:

  1. Low Earth Orbit construction
    1. Vehicle assembly
    2. Crew training
  2. Trans-Mars injection
    1. Propulsive maneuver
    2. Communication satellite deployment
    3. Spin-up
    4. Power system deployment
    5. Tether system deployment
    6. Trans-Mars coast
    7. De-spin
    8. Power system retrieval
    9. Tether system retrieval
    10. Communication satellite retrieval
  3. Mars circularization
    1. Propulsive maneuver
    2. CCV surface operations
    3. CCV return to MPV
  4. Trans-Earth injection
    1. Propulsive maneuver
    2. Communication satellite deployment
    3. Spin-up
    4. Power system deployment
    5. Tether system deployment
    6. Trans-Earth coast
    7. De-spin
    8. Power system retrieval
    9. Tether system retrieval
    10. Communication satellite retrieval
  5. Earth orbit capture
    1. Propulsion stage, CCV, and MPV separation
    2. Propulsion stage remains in hyperbolic orbit
    3. CCV propulsively circularizes at LEO with crew
    4. MPV aerobrakes into Space Operations Center orbit


The mission spacecraft is composed of the following components:

  • Pressurized Environment System: the habitat module
  • Power System: provides electricity
  • Structural System: the ship's spine
  • Folding Aerobrake System
  • Four-Tether System: provides 1 gee of spin gravity
  • Staged Propulsion System

It carries the following payload:

  • Crew Command Vehicle (CCV): transports explorers to Phobos, Deimos, and Martian Surface
  • Communication Satellite

Pressurized Environment System

This is the MPV's habitat module. It is composed of a habitation module, a laboratory module, a safe haven (storm cellar) and connecting tunnels. The large modules will have three airlock section, each with two means of egress: one to another pressurized airlock section, and the other to either the exterior or another airlock section. Space suits will be stored next to each exterior egress for use in a planned EVA or emergency escape to the CCV. The storm cellar will accomodate the 6 member crew for 12 days. The anti-radiation walls are 10 cm thick aluminum.

The mass budget for the system is 98,000 kg.

Power System

The power system mass is budgeted at 20,700 kg. It is specified to provide 150 kW constant power. It uses a set of solar dynamic power systems rated at 128 kg/kW. This system cannot be used during propulsive maneuvers or planetary eclipses. During these periods power is supplied by fuel cells.

Structural System

The structural system is rated to withstand forces of 3.5 g under propulsion and 1 g under spin gravity. The layout with pressurized environment allows optimal thrusting through the center of gravity. The system budget is 32,000 kg.

Folding Aerobrake System

The aerobrake system is used at the end of mission, to brake the MPV into the orbit of the space station. This saves a whopping 50% of propellant. The system assumes that the crew has already departed in the CCV, to remove the mass of the CCV and to spare the crew a fiery death in case aerobraking fails.

The system consists of the aerobrake (in two folding sections), a transferrable strutural pallet, folding mechanism, and fuel for Terra orbit circularization and rendezvous with the space station.

It is also used as a movable counterweight mass while in spin-grav mode.

If the aerobrake is non-functional, the MPV would enter a hyperbolic trajectory and enter a wild solar orbit.

The mass budget is 70,000 kg.

Four-Tether System

This is a bola type of spin gravity, providing 1 gee of gravity to the habitat module. Bolas are much more lightweight than centrifuges using girders or something. Much easier to boost into orbit as well. However it is prone to dangerous oscillations and perturbations. If the cables snap, the habitat module will be separated from the propulsion system, and fly into the big dark heading for a lonely doom for the unfortunate crew.

The students designed a four-tether system to deal with oscillations.

The rotation speed was limited to 2 rpm. For 1 gee his means the spin radius at the habitat module will have to be about 224 meters. The spin center will of course shift as propellant is burnt and the center of gravity changes.


While the students were designing they realized that aerobraking the MPV at Earth left little counterweight mass for the return rotation cycle. So they created two scenarios with different arrangements, and different mass budgets.

Scenario 1

Total delta V: 10,475 m/s

Trans-Terra injection propulsion stage containment mass is retained for counterweight mass on return leg rotation cycle.

Aerobrake is transferred from MPV to propulsion stage for counterweight mass on return leg rotation cycle.

Trans-Mars injection separated into two propulsive maneuvers.

Scenario 2

Total delta V: 12,599 m/s

Trans-Earth injection propulsion stage containment mass is retained for counterweight mass on return leg rotation cycle.

Aerobrake remains attached to MPV and is not transferred as in S1

Aerobraking maneuver has a propulsive assist = 610 m/sec which requires more fuel than S2

Trans-Mars injection is one propulsive maneuver.

Falure analysis showed that up to two tether could snap simultaneously without catastrophic failure. In this case the habitat module would be reattached to the propulsion system, and the crew would just have to endure zero gee for the rest of the mission.

The mass budget for the tether system is 13,000 kg.

Propulsion System

The student's analysis showed that a nuclear-electric propulsion system was unworkable, so they used a plain vanilla liquid hydrogen / liquid oxygen chemical rocket. Unsurprisingly the failure analysis revealed that the propulsion system fails anytime after trans-Martian injection the crew faces a death sentence. Even with a free return flyby, the ship will run out of consumables years before it return to Terra.


The CCV is a little auxiliary spacecraft carried as payload. The mission re-uses the heck out of it, to get their money's worth out of the mass it eats up. The crew occupies it during all propulsion burns made by the MPV. It ferries explorers to Phobos and Deimos. It lands a team of three explorers on the Martian surface. And at mission's end it transports the crew to the space station while the unmanned MPV aerobrakes into parking orbit.

The CCV is used by the crew when the MPV does burns because the habitat module does not have any acceleration couches (every gram counts!). Having said that, the couches can be rotated 180° between two configurations, since the direction of "down" is different between MPV burns and CCV landing on Mars. Without rotation, MPV burns would feel like the couches were attacked to the ceiling with the astronauts in a most uncomfortable "eyeballs-out" position.

CCV mass is budgeted at 50,000 kg.

MARCO Fast Crew Transfer Vehicle

This is from Design of a Fast Crew Transfer Vehicle to Mars. It is a 1988 design project done by the students of Dr. Wallace Fowler of the University of Texas at Austin department of Aerospace Engineering and Engineering Mechanics.

This is a fast crew transfer vehicle (FCTV) capable of a 150 day transfer from Terra to Mars, a 40 day stay time at Mars, and a 150 day transfer from Mars to Terra. It is used in the second half of a split mission to Mars, where an unmanned cargo vehicle (the Barge) is sent ahead with the bulk of the payload. Only upon successful arrival will the manned mission be launched.

The FCTV is designed for a crew of six with an endurance of one year (2,190 person-days of life support). Four crew modules (similar to the proposed Space Station Common Modules) are for crew habitats and support equipment. There is a command module that doubles as a storm cellar.

The propulsion is chemical, five Space Shuttle main engines (SSME) with a total delta-V of 27,600 m/s. Actually the spacecraft will expend all its fuel in the first leg of the mission, and will refuel from the Barge in Mars orbit for the second leg.

The ship's power supply is an SP-100 nuclear reactor producing 300 kWe of power using an Alkali Metal Thermoelectric Converter (AMTEC). It will generate approximately 2 MWt of waste heat that will be removed by two rotating bubble membrane radiators (RBMR). Each RBMR can radiate 10.1 MWt, has a radius of 2 meters, and a mass of about 100 kg.

Dates and ΔV's for Departure Opportunities
DateTotal ΔV
Mars Orbit

Mars Base Camp

This is from MARS BASE CAMP UPDATES AND NEW CONCEPTS by Lockheed-Martin (2017).

This Mars base mission concept was released about one hour before SpaceX released their Mars mission concept. The Lockheed mission relies heavily upon NASA's Space Launch System (SLS). A person of suspicious mind would find the timing questionable. It would seem to be for the purposes of stealing SpaceX's thunder.

The background is that the SLS has suffered development delays, cost overruns, and criticism that it is not really needed so should be cancelled. On the other hand it provides lots of jobs in states controlled by powerful senators. Meanwhile the United Launch Alliance (ULA) {which just so happens to include Lockheed-Martin} had a monopoly on boosting USAF payloads even though their boost price kept rising. SpaceX filed a lawsuit which they won, and then proceeded to boost the USAF payloads at a mere 20% of ULA's price tag ($90 million as opposed to $460 million). ULA's VP of engineering made public comments that ULA was quote "resentful of SpaceX" unquote. He latter resigned.

The point being that Lockheed-Martin does not like SpaceX very much.

In a related development NASA made an announcement it was looking into a Deep Space Gateway in cis-Lunar orbit (EML-3). The interesting facts are that the Gateway's components are carefully sized so they cannot be boosted by SpaceX's rockets {only by the as-yet nonexistent NASA SLS}, and that the project has been criticized as having no purpose. Well, no purpose other than giving the SLS a reason to exist, that is. And trying to sabotage SpaceX.

But I digress.

The Mars Base Camp is a crewed vehicle established in Mars orbit. From it a crew of 6 astronauts can perform excursions to Deimos and Phobos, perform telerobotic exploration of the Martian surface (including sample returns), produce LH2/LOX fuel via solar-powered electrolysis from water (either delivered from Terra by unmanned Water Delivery Vehicles {WDV} or from ISRU ice from Martian Moons), and allow astronaut sorties to the surface via reusable Mars Ascent/Descent Vehicles (MADV).

Elements of the camp are pre-placed before the arrival of astronauts, such as the lab, the center node, two excursion modules, two MADV and one or more Mars-orbiting cryogenic fuel depots. These are transported by unmanned dual-mode stages. These have solar electric propulsion for long-period delta V, and chemical thrusters for RCS.

The crew transfer vehicle stack consists of a habitat module, two cryo-stage propulsion systems, and two Orion spacecraft (NASA's proposed Multi-Purpose Crew Vehicle, not the nuclear bomb powered kind. There are too many freaking spacecraft named "Orion").

The crew transfer vehicle carries 6 crew. After the mission is over the transfer vehicle is docked to the Deep Space Gateway while the crew returns to Terra by Orion reentry. The transfer vehicle can be serviced and reused.

The idea is NOT to perform a stupid "flags and footprints" stunt like Apollo, spending billions of dollars to let some clown walk on the moon and then nothing. The Mars Base Camp is intended to establish reusable infrastructure to sustain long-termed crewed operations at Mars. Each mission should lay the groundwork for the next, otherwise the intermittent funding of NASA can lead to large post-mission gaps. As of this writing the post-mission gap after the last Apollo lunar mission has grown to 45 years with no end in sight. In addition every piece of equipment should be reusable because US lawmakers frown on multi-million dollar pieces of gear that are used once then thrown away like high-tech toilet paper.

As much as possible existing technologies should be used, because the long development times for non-existing technologies is just begging to get the entire project axed by penny-pinching congress-critters.

It goes without saying that crew safety is paramount. Astronauts dying a low agonizing death in space will be a public-relations nightmare that NASA is unlikely to survive. No single point of failure is to be allowed, which means everything should be redundant. Two Orions, two crew quarters, two MADV landers, etc. The movie The Martian was nice science fiction, but unlikely to have a happy ending in the real world.

The report also specifically states the Mars Base Camp must be used to prove that the Deep Space Gateway is not a gigantic boondoggle, even if it is. About two pages of the entire report is devoted to listing various ways the DSG can be used to assist the development of the Mars Base Camp.

The MADV landers can each make multiple sorties to the Martian surface, provided there is enough orbital propellant depots to refuel them. The fuel is generated as needed from water by electrolysis. This is because liquid hydrogen and liquid oxygen has an annoying habit of boiling, with the need to waste fuel by venting the vapor to space or the freaking fuel tanks explode. The longer the liquid fuel sits in the tanks, the more you lose to boil-off. The water is supplied by unmanned water delivery vehicles, from NASA if need be but the report has the pious hope that commercial suppliers of water will spring into being. They can produce water by mining various in-situ sources (asteroids, Martian surface, but ideally from the moons of Mars) and deliver it to Mars Base Camp propellant depots. And give NASA the bill for their services.

In between manned mission the unmanned central part of the Mars Base Camp remains in Mars orbit.

  • Lab Module
  • Center Node
  • Deimos/Phobos Excursion Module
  • x2 solar electric propulsion stages and related solar panels

Since the crew transport uses Hohmann transfers, the crew will have to stay at Mars for one synodic period (11 months) before they can head for home. The report notes that due to the availability of the MADV landers, the crew spends much of that time in orbit, inside the Mars Base Camp instead of spending the entire time on the surface as in most all other proposed Mars missions. This lowers the mission cost since you do not have to land the entire crew habitat module and all the support equipment. It also eliminates the failure mode where the habitat module is pre-positioned but the crew lander accidentally planets at a distance from the hab mod which is too far to walk.

The Crew transfer vehicle carries enough fuel for the round trip to Mars and back, and enough surplus fuel for either [a] Two sortie missions to the Martian moons or [b] One sortie mission to the Martian surface. If any more sorties are desired, water will have to be delivered by water delivery vehicles and electrolyzed into fuel. Missions to the moons are performed with an Orion-Excursion System-Cryogenic Propulsion Stage stack split off from the crew transfer vehicle. Missions to the surface are performed by a Mars Ascent/Descent Vehicle. Surface sorties will transport 4 crew to the surface while 2 crew remain in orbit in the base camp.

The Crew transfer vehicle is designed to have zero boil-off, but the MADV and WDV are not.


Water delivery vehicles can theoretically be of any size, but the report assumes they carry 52 metric tons of water. The report calls this a 50 MT Class WDV. Two of these carry enough water to fuel one MADV sortie. They have a solar-electric propulsion stage that doubles as a 375 kW solar-powered electrolysis plant. 375 kW at Terra orbit, it drops to 160 kW at Mars orbit due to the inverse-square law. WDV carries:

  • Tankage for 52,000 kg of water (full when launched)
  • Tankage for 40,000 kg of LOX/LH2 (6:1 ratio, empty when launched)
  • Solar-electric propulsion stage (375 kW at Terra orbit, 160 kW at Mars orbit)
  • Water electrolysis system (powered by SEP stage)
  • Navigation and communication systems

Upon arrival the WDV is captured by one of the Mars Base Camp's robotic arms, the WDV then starts electrolyzing the water. Two WDV working in parallel can create enough fuel for one MADV sortie in about 2.5 months. The base camp can hang on to two WDV at a time. If multiple sorties are planned, it will be better to have the WDV to cluster slightly ahead or behind of the base camp orbit to create an orbital propellant depot. This will mean the MADV will have to detach from the base camp and make a short trip to the propellant depot to fuel up.


The Mars Ascent/Descent Vehicles are sized for a 4.7 km/s entry velocity and have a mid-L/D profile. As previously mentioned they carry 4 crew, while 2 crew remain in orbit. They are totally reusable, which means no inflatable aerodynamic decelerators or other device that cannot be repackages and refurbished. Parachutes could be used were it not for the regrettable fact that the thin Martian atmosphere would require a prohibitively large chute to land a 100+ metric ton spacecraft. Retro thrust must be used.

The MADV has six RL-10 engines for propulsion. Wikipedia says they have 110,000 Newtons of thrust each for presumably a total of 660,000 Newtons.

Once landed the MADV will be home for the crew of 4 for the next 10 days (actually sols), though it has a 50% contingency margin for a total of 15 sols in case of emergency. It has a payload of 2,500 kg of scientific experiments. A 2 person airlock is adjacent to the 2 person mechanical lift on the spacecraft side which transport the crew to the planet's surface. It has enough consumables for two 2-person EVAs per sol.

The fuel tanks contain more fuel than is needed for a descent and ascent. The remaining fuel is used in fuel cells for power generation. It requires 780 m/s delta-V to land and 5,200 m/s delta-V to ascend, for a total of 6,000 m/s delta-V. The MADV has a propellant mass fraction of 74% (which I calculate to imply a mass ratio of 3.85).

Two water delivery vehicles (with 40,000 kg of LOX/LH2 each) can give the MADV enough fuel for one sortie. The report is not specific as to exactly how much is required, so the MADV requires something between 40,001 kg and 80,000 kg fuel for a sortie (probably the full 80,000). If so with the propellant mass fraction I calculate the wet mass of the MADV is about 108,000 kg. If true this is approximately the same mass as the Shuttle Orbiter (without the external tank and solid rocket boosters).

I did some pixel measuring on the cut-away view. Assuming that the human figures were the standard 1.77 meters tall (and the diagram is accurate), the MADV is approximately 25 meters tall (about 31 meters shorter than the Space Shuttle Stack).

Mars DRA 5.0 NEP

Mars DRA 5.0 NEP
Mission TypeConjunction Class
Reactor Power2.5 MWe
Specific Impulse5,000 sec
Exhaust Vel49,000 m/s
Engine Power
280 kW
Num Enginesx10
Xenon Propellant>60,000 kg
Delta V>24,000 m/s
Num Crewx6
Total Mission980 days
Mars Stay400 days

This is from Human Exploration of Mars Design Reference Architecture 5.0, Addendum #2.

This was from a NASA study to make some basic Mars exploration missions using a variety of propulsion systems. The idea was these could be used to measure the performance of other design proposals and/or be used as a springboard to create new designs.

This particular design reference was the one using nuclear electric propulsion: ion drive in other words. Since it is nuclear powered you will notice the spacecraft is coated with heat radiators. The design uses a 2.5 megawatt reactor and ion drives with a specific impulse of 5000 seconds.

The Mars Lander is sent ahead in an uncrewed cargo mission and waits in High Mars Orbit for the crewed mission to arrive.


The spacecraft can be boosted into orbit with only two launches. The first carries the crew habitat and ion drives, the second carries the nuclear reactor. Both carry a portion of the xenon propellant. They have chemical propulsion systems based on Space Shuttle OMS engines used by the sections to rendezvous and dock. The two sections have small solar photovoltaic arrays to provide power until the two components connect. Then the heat radiators and boom deploy and the reactor powers up.


The as yet uncrewed spacecraft starts its slow spiral to Terra escape velocity. This allows the spacecraft to spend lots of time passing through the deadly Van Allen radiation belt since there is no crew aboard to be harmed. This is a common problem with low-thrust/high-specific-impulse propulsion systems such as ion drives. How long will this take? About 354 days.

At the 354 day mark the spacecraft is almost at Terra escape velocity. The six person crew is delivered to the spacecraft in a small Orion spacecraft. The crew frantically checks out the ship, they can abort back to Terra within the next 11 days but after that they are stuck on the ship for the duration of the mission.

The trip to Mars takes 293 days. Upon arrival the ion drive will take 21 days to spiral in to a High Mars Orbit (20,000 km circular orbit, similar to Deimos). The crew then descends in the Mars Lander (send in an earlier uncrewed mission and waiting in orbit) then spends the next 400 days doing all the science they possibly can.

At the end of their stay, the crew uses the ascent module to return to the orbiting spacecraft. It will take another 21 days to spiral out into Terra Transfer Insertion. The trip home will take 215 days.

Upon Terra arrival the crew bails out in the Orion and aerobrakes to a landing at about 13 kilometers per second. The uncrewed spacecraft starts the long process of braking into a storage orbit, there to wait to be used for a follow-on mission.

Mars Expedition Spacecraft

This is from a NASA Manned Spaceflight Center (renamed the Johnson Space Center in 1973). The study was done in 1963. I have not been able to find lots of hard details, but there is some information in David Portree's monograph Humans to Mars on pages 15 to 18, available here.

It travels in a Hohmann transfer to Mars, separated into two parts spinning like a bola for artificial gravity. In Mars orbit, the heat shield, laboratory, and rendezvous ship separate and land. After a forty day stay, the astronauts use the rendezvous ship to climb back into orbit and travel to the mother ship. After the journey back to Terra, the astronauts land via the re-entry module.

Mars NEP with Artificial Gravity

This is from a document entitled Human Exploration Of Mars: Artificial-Gravity Nuclear Electric Propulsion Option, 15 July, 2003 which apparently has vanished from the face of the web so throughly that I cannot find it any more (but thoughtfully labeled "Internal NASA Use Only"). But you can find much of the details in this earlier report. Actually both are not so much reports as they are series of slides. Earlier details can be found in Preliminary Assessment of Artificial Gravity Impacts to Deep-Space Vehicle Design.

The Mars Crew Transfer Vehicle is basically a "tumbling pigeon" spacecraft with an ion drive powered by a nuclear reactor (nuclear-electric propulsion or "NEP"). The report mentions Ion thrusters, MPD thrusters, and VASIMR. At this point in the study they are assuming a specific impulse of 4,000 seconds (exhaust velocity of 39,200 m/s)


There are two basic mission classes: Short-Stay (opposition class) and Long-Stay (conjunction class). The important difference is once the astronauts have landed, how long is it until the launch window arrives? "Launch window" translates as "the date when the spacecraft had better depart for home or the astronauts will all die a lingering death from running out of oxygen". The report suggests that the first few mission will be short-stay because they are less risky. The longer you stay the bigger the chance that something will go wrong (vital equipment wears out, astronaut doing a surface excursion breaks their leg, somebody develops appendicitis, things like that).

The report decided the mission needed the following characteristics:

  • Initial missions limited to 18-24 month round trip (allows lesser performance engine to be used. Also minimizes steering requirements, which is a problem with huge fragile spinning artificial gravity ship)
  • Three months stay in Mars system
  • “Split mission” –no “Mars-specific” cargo sent out with crew (meaning the Lander is sent on an unmanned mission to High Mars Orbit first, it is not carried along with the manned mission. Means spacecraft can be generalized, not forced to be optimized to different types of Landers)
  • Assembly Orbit: Low Earth orbit 700 km (easier to assemble)
  • Departure/return point: High Earth orbit 90,000 km (requires less delta V and less propellant, assumes the presence of local orbital shuttles)
  • Destination: High Mars orbit 90,000 km (requires less delta V and less propellant)
  • Piloted vehicle stack less than 200 tons initial mass (less stress on spacecraft spine, less delta V, less propellant)

Prior to the Mars Crew Transfer Vehicle (MCTV) making the journey to Mars, the Mars Lander travels to High Mars Orbit unmanned and under remote control. Naturally if the lander fails to make it or arrives damaged the manned mission is scrubbed.

The components of the MCTV are boosted into 700 km LEO by a series of launches, and the components are assembled in orbit. It is then boosted into the 90,000 km HEO.

Crew is delivered to the MCTV by Earth Neighborhood transport infrastructure (XTV), which is some system of orbital transports. It then departs on the 10-odd month transit to High Mars Orbit.

At Mars Orbit, the MCTV rendezvouses with the Mars Lander and is loaded with the surface exploration crew. The poor Mikeys stay in orbit while being forgotten from the pages of history. The Lander lands, and the surface crew starts their 30 day surface stay. The Mikeys send them a constant stream of Mark Watney jokes.

At the end of 30 days the surface crew rides the Lander up to rendezvous with the MCTV. If something catastrophic happens during the lift off, the Mikeys will never forgive themselves for the jokes. Assuming all goes well the Lander is jettisoned and the MCTV departs on its ten-odd month trip home. In HEO it is met by an XTV craft and the crew is returned to Terra.

The mass returned to Terra is 89 metric tons.

Trajectory Sensitivity Analysis

The purpose of the trajectory sensitivity analysis is to determine tradeoffs and sensitivities of key trajectory parameters including:

  • Earth departure altitude
  • Mars parking orbit altitude (and ultimately Mars lander size)
  • Stay time in Mars vicinity
  • Useful time on Mars surface
  • Total trip time
  • Earth return altitude

Trajectory Assumptions:

  • Earth Departure Orbit: 700 km altitude
  • Earth Return Orbit: vary from 30,000 to 90,000 km altitude
  • Mars Parking Orbit: vary from 500 to 17,200 km altitude
  • Stay Time in Mars Orbit: calculated to sum time in Mars vicinityto approximately 90 days (Resulted in stay times at Mars in orbit from 37 to 77 days)
  • Total Trip time includes spiral time from LEO to high Earth orbit

Nuclear Electric Propulsion Vehicle System Assumptions:

  • Power: 6 MW
  • Specific Impulse (Isp): 4000 sec
  • Thruster efficiency: 60%
  • Tankage Fraction: 5% (metal tank mass as percentage of propellant mass)

Mission Assumptions:

  • Mass returned to Earth: 89 metric tons
  • Launch Date: 2026
  • Stay time in Mars space: approx 90 days (Resulted in stay times at Mars in orbit from 37 to 77 days)
  • Mission Total Trip time goal: 700 days

Limiting Orbit Assumptions (for sensitivity trade)

  • Earth departure orbit altitude : LEO of 700 km
  • Earth return orbit altitude: vary between 30,000 - 90,000 km
  • Mars parking orbit altitude: vary between LMO of 500 km and aerosynchronous


  • Missions of 700-days round trip are possible with limits on Earth and Mars orbit altitude choices
  • Total trip time does not equal total crew time (Note: The astronauts will ascend to the NEP vehicle once it’s in the high earth altitude via a the XTV)
  • Trade studies needed to evaluate choice of Mars parking orbit with respect to Ascent/Descent vehicle versus NEP vehicle performance
  • Further analysis needed to evaluate proximity to Sun on return leg

Mars Crew Transfer Vehicle Engine

The ion drive may be either electromagnetic or electrostatic.

Mars Crew Transfer Vehicle Body

Lander and Aerobrake

Mars Crew Transfer Vehicle Artificial Gravity

The entire point behind this study was to discover the optimum way to give artificial gravity to an ion-drive spacecraft. Prolonged microgravity missions do horrible things to the health of the crew. Mars missions tend to have over-long wait times to start with. Limiting Mars missions to the ones with the shortest duration drastically reduces the available mission trajectories in a given decade. Ion and other electric powered drives only exacerbate the problem with their absurdly low accelerations. This particular design is going to take three extra months just to accelerate up to Terra's escape velocity (chemical and nuclear thermal propulsion reaches escape velocity in a matter of minutes). That is long enough for the crew to lose 4.5% of their bone mass.

The problem is that the standard artificial gravity architectures have problems on a spacecraft that uses rockets for propulsion. And these problems are also exacerbated by low-acceleration drives.

It all boils down to Thrust Vector Control (TVC).

Each of the mission's maneuvers contains a specifed Axis of Acceleration. To perform the maneuver the spacecraft's thrust axis has to be exactly on the axis of acceleration. Before the maneuver the spacecraft has to be rotated so the thrust axis is in the proper orientation, and during the burn the thrust axis must be monitored and corrected if it drifts off the specified acceleration axis.

The problem is that the spacecraft's spin-gravity section acts like a gargantuan momentum wheel. This gyrostabilizes the ship and will fight your attempts at TVC tooth and nail. This is referred to as the Rotational Angular Momentum problem.

Spin Gravity Concepts
  • In spin section (entire ship) the habitat is counterweighted by the reactor and power conversion system

  • The entire ship rotates as a unit, there are no segments without rotation.

  • The majority of TVC is by pointing the entire vehicle
  • No rotating problematic joints, megawatt power connections or fluid piping

  • Power conversion can take advantage of operating in a gravity field
  • The vehicle angular momentum must be continuously vectored during TVC in order to deal with the rotational angular momentum problem.

  • Heat radiators have to be designed to operate in a gravity field.

  • It is challenging to design methods for crew ingress, crew egress, and ship docking to a spinning object.
Mars NEP with Artificial Gravity
(this section)
  • In spin section (everything but engine modules) the habitat is counterweighted by the reactor and power conversion system

  • Thrusters are de-spun and gimbaled for TVC
  • TVC is decoupled from rotational angular momentum, thus avoiding the rotational angular momentum problem.

  • Power conversion can take advantage of operating in a gravity field
  • Design is faced with the daunting problem of transferring megawatts of electricity and kilograms of propellant across a rotating joint. For months.

  • Potential cyclical loading of rotating joints can shatter them.

  • Heat radiators have to be designed to operate in a gravity field.

  • It is challenging to design methods for crew ingress, crew egress, and ship docking to a spinning object.
Boeing STCAEM Mars NEP
  • Habitat modules spin for gravity, the rest of the spacecraft is stationary.

  • The two habitat modules act as counterweights.

  • The thrusters are gimbaled for TVC.
  • TVC is decoupled from rotational angular momentum, thus avoiding the rotational angular momentum problem.

  • Heat radiators can take advantage of operating in a zero-gravity field

  • Crew ingress, crew egress, and ship docking can be easily done to a stationary docking port.
  • There are inefficiencies in duplicating habitat modules.

  • Allowing crew to transfer between two spinning modules is a problem.

  • Potential cyclical loading of rotating joints can shatter them.

  • Power conversion have to be designed to operate in a zero-gravity field.

  • Design is faced with the problem of transferring kilowatts of electricity across a rotating joint.
Hedrick Fusion Spacecraft

The problem can be avoided by de-spinning the spin-grav section of the ship for the duration of the thrust. Sadly, since the thrust is more or less on for the entire trip, this kind of defeats the point of giving the ship spin-grav in the first place.

Ox Cart and Beanie Cap avoid the rotational angular momentum problem by de-spinning the engines from the spin gravity sections of the ship. The spin plane is aligned with the interplanetary trajectory plane. The main draw-back is the engineering and maintenance nightmare of the rotation joints.

Fire Baton is trying a new approach. The entire ship spins in order to avoid those nasty rotation joints. Instead it tries to precess the entire ship in order to aim the thrusters for TVC.

To lock a spacecraft or other object solid with gyrostabilization you actually need three spinning gyros at 90° angles to each other. A spin-grav ship only has one spinning object. So instead of being locked in place, if you push on it the spinning thing will undergo precession. Which is a fancy word meaning the object rotates unexpectedly at a right angle to the direction you push it. Try playing with a spinning gyroscope and you'll quickly discover this.

The report did an analysis and discovered that thrust vector adjustments came in two classes: very slow rates and moderate rates. The very slow rates were changing the vector less than two degrees per day (during the heliocentric trajectory). The moderate rates were changing the vector fifteen degress per day (during Terra departure and during midcourse thrust reversals). This means that two different steering strategies can be used. For flipping the main engines to point the opposite direction (for braking) a third strategy can be used.

In both strategies, the mechanism is to thrust in a direction at right angles to the desired steering direction, to precess the thrust axis in the desired direction (see "resulting precessional yaw rate" in diagram above). The thrust has to be done intermittently, when the thruster is pointed in the correct direction by the ship's spin (see "thrusting arc" in diagram above). If the thrust is appled every 180° of a spin-grav rotation, very slow rates require 3 Newtons and moderate rates require 15 Newtons.

The three strategies are:

  1. firing the control thrusters (RCS)
  2. differentially throttling the main ion thrusters (the two banks are throttled in an unbalanced manner)
  3. firing tangential RCS to spin the ship 180° on its long axis

The report tried all the combinations, and concluded that the lowest propellant consumption was if:

  1. Very slow rates precession (∼2°/day, 3 n) was performed with differentially throttling the ions engines ±5%
  2. Moderate rates presession (15°/day, 15 n) was done with the RCS
  3. Spinning 180° on the long axis was performed with tangential RCS

Over the entire mission the report calculates these strategies will require approximately one extra metric ton of propellant (1,074 kg). And no nasty rotation joint needed.

Mars One

EngineSpace Shuttle
Uprated Engine
Thrust2,043,000 N
Isp452.3 sec
4,437 m/s
Engines in
TMI Stage
Engines in

This design comes from the fictional book THE MARS ONE CREW MANUAL by Kerry Mark Joëls, former NASA employee and co-author of the SPACE SHUTTLE OPERATOR'S MANUAL. He notes that the design is assembled out of parts from twenty years of NASA Mars mission studies, which means some of the components were designed with different assumptions and may not fit together well. But the figures are at least in the ballpark. He acknowledges the help of quite a few scientists at the California Space Institute, the Jet Propulsion Laboratory, The Center for Earth and Planetary Studies, the Ames Research Center, the Johnson Spaceflight Center, and Lockheed Missiles And Space.

The Trans-Mars-Insertion (TMI, Earth Departure) stage is to the left. It comprises about 60% of the entire spacecraft's Initial Mass in LEO (IMLEO). After the burn the stage is discarded and the spacecraft coasts to Mars.

Actually, according to my slide rule, it will have to comprise a bit more than 65% of the IMLEO, assuming it has to produce 4,650 m/s delta V using LOX/LH2 fuel with an Isp of 452.3 sec. 65% is just for the TMI stage's fuel, more will be needed for the mass of the engines, tanks, and structure. 65% is a mass ratio of 2.85

The Propulsion Stage is immediately to the right. It is used for Mars-Orbit-Insertion (MOI, Mars Arrival), Trans-Earth-Insertion (TEI, Mars Departure), and Earth-Orbit-Insertion (EOI, Earth Arrival).

There are four spacecraft modules: two habitat modules (HAB 1, HAB 2), one laboratory module (LAB), and one storage module. Each are 15.2 meter long by 4.3 meters in diameter. They each have a mass of 32,200 kilograms, with some slight variation from module to module. The modules are attached to the central tunnel/docking assembly. The central tunnel has six hatches: four for the modules, one for the Mars Excursion Module (MEM), and one for an EVA airlock in case the crew has to enter space to repair part of the spacecraft.

Located at the tips of the habitat modules are the Deployment Platforms. Each platform holds one Mars drone airplane, one comsat, an assortment of Venus atmospheric probes for the Venus fly-by, and various Mars hardlander and surface penetrator probes. Buried in the center is the inter-module transfer tunnel.


The top portion of each HAB module is the Control Center. Each HAB module control center is identical, and each can perform all functions necessary for the operation of the main ship. The main computers are located in the control center, allowing the center to be a communications control room and a classroom for training and simulation. An airlock hatch connects the center to tunnels leading to other modules.

The control center is divided into three work-station areas. The command station is for the mission com- mander and pilot (or for the lander commander and pilot). The science station is for the chief science officer. The experiment stations are for the science specialists. Each station has computer input and output capability, a Reaction Control System (RCS) control panel, an environmental control system panel, and communications and navigation instrumentation.

The control center doubles as the anti-radiation storm cellar.

  • A Science Station
  • B Suit Storage
  • C Command Station
  • D Experimental Station

The Health and Hygiene (H & H) deck consists of the Health Maintenance Facility (HMF), the shower, the lavatory, and the waste management (toilet) compartments.

The health maintenance facility is a combination intirmary, exercise area, and health diagnostic center. It contains an array of exercise and medical equipment. It can also be used for in-flight emergencies.

The HMF takes up about half the deck area, with the other half divided roughly in thirds for the other functions. The deck is 2.1 meters high and has three communications stations and extra ducting to remove unwanted odor and humidity (i.e., so it doesn't stink like a gym locker room).

  • A Exercise
  • B Diagnostic
  • C lnfirmary
  • D Shower
  • E Lavatory
  • F Waste Management

The wardroom in each HAB module is a communal room for meals, meetings, entertainment, and some class activities. The galley area is for food preparation and storage. Video monitors and a portable computer port allow the area to be used for training, group telecommunications, and showing videodisc movies and other programs. Should an emergency arise, the crew from the other HAB module can use the wardroom as sleeping quarters.

  • A Group Activities
  • B Recreation
  • C Galley

Each of the two HAB modules has color-coded sleep compartments. Each compartment is wedge-shaped, measuring 2.1 meters along the main wall, 1 .5 and 1.6 meters along the side walls, and 0.6 meters along the sliding door. There is a 0.09-cubic-meter storage cabinet on one side and a small 0.045-cubic-meter cabinet near the door. The sleep restraint is on the same wall as the small cabinet. A heavy Velcro-secured curtain provides privacy The 2.4 meter ceiling makes the room a bit more spacious, and the back wall has a tack board for pictures, charts, or other personal items. There is a communications substation in each compartment with a headset, an alarm, and a small digital readout with emergency codes.

Below the sleep center is a combination storage/life-support area. The life-support area contains your air-handling equipment, your electrical distribution center, your water-treatment equipment, and your solid waste pretreatment equipment. The chamber has soundproofing and the machinery is specially mounted to reduce noise and vibration.

Four banks of storage cabinets are also included. They contain emergency rations as well as normal food stores, hygiene supplies, and personal efiects. The storage areas are spaced similarly to the storage module. One bank of cabinets opens to the life-support area. This group of lockers contains spare parts, maintenance information, and tool kits.

An observation area which permits visual inspection of the deployment palette is located near the hatch area at the end of the compartment.

  • A Storage
  • B Life Support




Trans-Mars-Insertion burn accelerates the ship's initial velocity in LEO (28,160 kph) to 44,900 kph. My slide rule says this represents a delta V of 16,740 kph or 4,650 meters per second.

The stay of the astronauts on the Martian surface is 30 days (29.2 sols)


The Mars Rover is sufficiently massive that it is transported to Mars in a separate uncrewed vehicle. It only has enough delta V for the Trans Martian Insertion burn: 4,694 m/s. For entering Mars orbit and landing the rover it relies totally on aerocapture and aerobraking.

Mars Umbrella Ship

RocketCat sez

There is just something about this surreal design that gets to you. People who briefly saw the deep space umbrella in 1957 still remember it. Totally unlike any other spacecraft you've ever seen. That is, except for science fiction ships from artist who also were haunted by the blasted thing.

Not a bad ship either. Except that pathetic one-lung ion drive is so weak that it takes a third of a year to reach orbit halfway between Terra and Luna. I'm sure we can do better than that today. Swap it out for a VASIMR or something and you'll have a ship that can go places and do things!

Stuhlinger Umbrella Ship
Total Travel
6.30×107 sec2 years
Payload1.50×108 g150 tons
Ionic Mass
2.20×10-22 g
Specific Power5.00×104 erg/sec g0.95 kW/kg
6.70×10-2 cm/sec26.7×10-5 G
Initial Mass
7.30×108 g730 tons
3.65×108 g367 tons
Power Planet
2.15×108 g215 tons
Total Power
1.15×1015 erg/sec114.5 MW
Total Electrical
2.29×1014 erg/sec22.9 MW
in Jet
2.06×1014 erg/sec20.9 MW
Driving Voltage4.88×103 volts
Total Ion
4.22×103 amperes
1.15×104 cm115 metres
Total Length8.50×103 cm85 metres
8.40×106 cm84 km/s
Thrust4.85×107 dyne48.5 kgf
*dm/dt5.77×100 g/sec0.00577 kg/s
*delta-vee5.82×106 cm/sec58.2 km/s
Values from paper. * values derived by Adam Crowl

Unusual spacecraft designed by Ernst Stuhlinger in 1957, based on a US Army Ballistic Missile Agency study. It made an appearance in a Walt Disney presentation "Mars and Beyond". 4 December 1957. David S. F. Portree, noted space history researcher and author of Wired's Beyond Apollo blog, managed to uncover the identity of Dr. Stuhlinger's report for me, it is NASA TMX-57089 Electrical Propulsion System for Space Ships with Nuclear Power Source by Ernst Stuhlinger, 1 July 1955. Thanks, David!

Detailed blueprints of this spacecraft can be found in the indispensable Spaceship Handbook by Jack Hagerty and Jon C. Rogers, or are available separately.

The spacecraft resembled a huge umbrella, with the parasol part being an enormous heat radiator.

At the very bottom is a 100 megawatt (thermal power) fast neutron nuclear reactor, mounted on a 100 meter boom to reduce the radiation impact on the crew habitat. A fast neutron reactor design was chosen because they can be built will a smaller mass and smaller size (reducing the size of the shadow shield). The reactor is capped with a shadow shield broad enough to cast a shadow over the entire heat radiator array. The part of the shadow shield closest to the reactor is 1.8 meters of beryllium. This stops most of the gamma rays, and slows down the neutrons enough that they can be stopped by an outer layer of boron. The shadow shield has a mass of 30 metric tons, and coupled with the boom distance it reduces the radiation flux at the habitat ring to 10 fast neutrons per second per cm2 and 100 gamma rays per second per cm2.

The liquid sodium will be carried in pipes constructed of molybdenum. The reactor will have a specific power around 0.1 kW per kg. It contains 0.6 cubic meters of uranium enriched 1.7%, and has a mass of 12 metric tons. No moderator or reflector is required. "Cool" liquid sodium (500° C) enters the reactor and leaves the reactor hot (800° C) at the rate of 300 kg/sec. The reactor contains 600 molybdenum pipes with an inner diameter of 1.8 cm and a length of 1 meter. Electromagnetic pumps move the liquid sodium, since it is metallic. Such pumps are used since the only way to make pumps that will operate continuously for over a year with high reliability is to have no moving parts. The pumps will consume about 100 kW.

The hot sodium enters the heat exchanger, where it heats up the cool silicon oil working fluid. The now cool liquid sodium goes back to the reactor to complete the cycle. The heat exchanger is used because silicon oil is more convenient as a working fluid, and because the liquid sodium becomes more radioactive with each pass through the reactor. The heat exchanger contains 3000 tubes for liquid sodium, with a total length of 1,800 meters and an inner diameter of 1.3 cm. The silicon oil is boiled into a vapor at 500° C under 20 atmospheres of pressure.

The hot oil vapor travels up the boom to a point just below the umbrella. There it runs a turbine which runs a generator creating electricity. The turbine is a low-pressure, multi-stage turbine with a high expansion ratio. Silicon oil was selected since it can carry heat and simultaneously lubricate the turbine, since this has to run continuously for over a year. Silicon oil is also liquid at 10° C, allowing the power plant to be started in space with no preheating equipment. The oil has a specific heat of about 0.4 cal per g per degree C, a heat of vaporization of 100 cal per g, a density of 1 g per cm3. If the umbrella heat radiator is at a temperature of 280° C, this implies that about 100 kg/sec must flow through the turbine. The feed pumps will consume about 200 kW. The total mass of the working fluid in the entire system will be about 8 metric tons.

Newton's third law in the turbine causes the section of the spacecraft from turbine upwards to rotate, including the ring habitat module and the umbrella heat radiator. The spin rate is about 1.5 rotations per minute. The generator is cooled by small square heat radiators mounted on the habitat ring.

The boom below the turbine is counter-rotated so it remains stationary. This is because the boom has the ion engine. If the boom was not counter-rotated, the ion engine would also rotate. The result would be a stationary ship behaving like a merry-go-round, spinning in place while spraying ions everywhere like an electric Catherine wheel.

The hot silicon oil vapor is injected into the central part of the rotating umbrella heat radiator (the radiator feed), and centrifugal force draws it through the radiator. The cooled oil is collected at the rim of the radiator, and pumped back to the reactor to complete the cycle. The rotation of the ring habitat module provides artificial gravity for the crew. The habitat ring is in the central part of the umbrella.

The umbrella heat radiator will have a temperature of 280° C. The silicon oil vapor will be reduced to the low pressure of 0.1 atmosphere, to reduce the required mass of heat radiator. The ship will be oriented so that the umbrella is always edge on to the Sun, for efficiency. The diameter of the umbrella will be about 100 meters, constructed of titanium. The wall thickness is 0.5 mm, the thickness of the disk is 6 cm near the center and 1 cm near the rim. The umbrella is composed of sectors, each with an inlet valve near the center and an outlet valve at the rim. If any sector is punctured by a meteorite, the valve will automatically shut until repairs can be made. The other sectors will have to take up the slack.

The electricity runs an ion drive, mounted on the lower boom at the ship's center of gravity. The ion drive uses cesium as propellant since that element is very easy to ionize. Cesium jets have a purplish-blue color. The umbrella section and the reactor have about the same mass, since the reactor is composed of uranium. The habitat ring has a bit more mass, this is why the ion drive is a bit above the midpoint of the boom.

Cesium has a density of 1930 kilograms per cubic meter. The spacecraft carries 332,000 kilograms of cesium reaction mass. This works out to 172 cubic meters of reaction mass, which would fit in a cube 5.6 meter on a side. Which is about the size of the block in between the ion drive and the landing boat, the one with the boom stuck through it. (ah, as it turns out my deduction was correct, now that I have the original report to read)

However, cesium propellant is now considered obsolete, nowadays ion thrusters instead use inert gases like xenon. Cesium and related propellants are admittedly easy to ionize, but they have a nasty habit of eroding away the ion drive accelerating electrodes. Xenon is inert and far less erosive, it is now the propellant of choice for ion drives.

Mounted opposite the ion drive is the Mars landing boat. It is attached so its center of gravity is along the thrust axis. This ensures that the umbrella ship's center of gravity does not change when the landing boat detaches. The landing boat uses a combination of rockets and parachutes to reach the surface of Mars. The upper half lift off to return to the orbiting umbrella ship.

The habitat ring has an outer radius of 19.5 meters, an inner radius of 15 meters, and a height of 6 meters (according to the blueprints). If I am doing my math properly, this implies an internal volume of 2,900 cubic meters, less the thickness of the walls. At a spin rate of 1.5 rotation per minute, that would give an artificial gravity of about 0.05g.

Above the umbrella and habitat ring is an airlock module containing two "bottle suit" space pods. Above that is a rack of four sounding rockets with instruments to probe the Martian atmosphere. At the top is the large rectangular antenna array.

The spacecraft is much lighter than an equivalent ship using chemical propulsion, and has a jaw-droppingly good mass ratio of 2.0, instead of 5.0 or more. However, the spacecraft's minuscule acceleration is close to making the ship unusable. It takes almost 100 days to reach an orbit only halfway between Terra and Luna. At day 124 it finally breaks free of Terra's gravity and enters Mars transfer orbit. It does not reach Mars capture orbit until day 367, but it takes an additional 45 to lower its orbit enough so that the landing boat can reach Mars. All in all, the umbrella ship takes about 142 days longer than a chemical ship for a Mars mission, due to the low acceleration. Which is bad news if you are trying to minimize the crew's exposure to cosmic radiation and solar proton storms.

The design might be improved by replacing the ion drive with an ion drive with more thrust, or with a magnetoplasmadynamic, VASIMR or other similar drive invented since 1957.

In his paper, Dr. Stuhlinger proposed that the Mars expedition be composed of a fleet of several ships. The Mars exploration equipment would be shared among all the ships. In addition, there would be some "cargo" ships. These would only carry enough propellant for a one-way trip, so they could transport a payload of 300 metric tons instead of 150. They would be manned by a skeleton crew, who would ride back to Terra on other ships.


      The space ship, traveling from the earth satellite toward Mars, will not land on Mars but will end its voyage in a circular orbit about 600 miles (970 km) above the Martian surface.

     For the third phase of the trip, a winged landing craft will be detached from the orbiting ship. It will reduce its orbiting speed by rocket power and enter a downward trajectory. After a long glide through the Martian atmosphere, it will land either on skids like a glider, or by parachute and counter rockets.

     At the end of the exploration period on Mars, the landing boat will take off from the planet by rocket power and will join the space ship, which is still orbiting about the planet at an altitude of 600 miles. The crew will transfer back to the ship and will make the return trip to the earth satellite. The last portion of the expedition, the hop from the satellite to earth, will be done by one of the winged fourth stages of the commuter rockets.

     The longest part of the voyage will be the section between the satellite orbit around the earth and the orbit around Mars. The space ship will be tailor-made for the conditions prevailing during this voyage. Quarters for the crew will be sealed and provided with an artificial atmosphere. The ship will not be streamlined since it travels only through perfect vacuum. The thrust of the rocket motors need not lift the vehicle against the gravity forces, since these forces are exactly balanced by centrifugal forces in any object that moves around in a satellite orbit. Even a relatively low thrust will enable the ship to leave its original satellite orbit and to enter into a trajectory which finally approaches the Martian ellipse.

     The space ship will be assembled in the satellite orbit close to the space station. All components of the ship, its equipment, and the fuel needed for the round trip, must be carried into the satellite orbit by the three-stage commuter rockets. These carrier rockets must overcome the earth's gravity and atmospheric drag, and they must impart to their payloads the orbital velocity of about five miles per second. This earth-orbit operation proves to be the most costly part of the entire Mars expedition. For every pound of payload, about 160 pounds of take-off weight must be invested in the commuter rockets. The space ship designer will, therefore, make the greatest effort to build his vehicle as light as possible. Furthermore, he will plan the expedition in such a way that any components which become unnecessary during the voyage, such as empty tanks, containers, supports, and even instrumentation, can be disposed of immediately. Ship and crew should finally arrive back in the earth satellite orbit with a bare minimum of equipment and reserves.

     By far the largest part of the take-off weight of such a space ship will be made up of the propellants. The attempt to reduce the mass of a space ship, therefore, leads immediately to an investigation of its propulsion system. The basic rocket equations show that the performance of a rocket engine is mainly determined by the exhaust velocity of the gases from the combustion chamber. In rocket engines based on a chemical reaction between the propellants, the exhaust velocity is intimately related to the temperature inside the combustion chamber. The temperatures at which modern combustion chambers operate are close to the maximum temperatures which can be expected from chemical reactions. There is not much hope that the performance of chemical rocket motors can be improved much beyond the point at which we have arrived today.

     If seems, however, that another type of reaction motor holds some promise for use in an interplanetary vehicle. If the velocity of the exhaust particles is not produced by the heat energy of a chemical reaction, but by electrical fields, much higher exhaust velocities can be obtained. The amount of electrical energy which can be imparted to a given mass of exhaust material is much greater than the energy which can be given to the same mass by a chemical reaction. Also, the electrical field would direct the exhaust particles in such a way that they would not strike the thrust chamber walls. Hence, the wall heating problem in an electrical engine would be considerably less than in a chemical engine.

     An electrical propulsion system would require the ionization of a suitable propellant material. It would also require a primary power source, the conversion of the primary power into electric power, and a thrust chamber in which the electric power is applied to accelerate the ions.

     A detailed study of the feasibility of an electrical propulsion system has already been made. This study has proved that an electrical system is feasible and that an electrically propelled space ship would be much lighter than a ship with a chemical propulsion system. The electrical system would, however, be definitely restricted to space vehicles traveling between satellite orbits because the thruster of an electrical propulsion system would always be so small that it could never lift the vehicle from the surface of a planet against the gravity forces. The propulsion system would operate continuously, first accelerating the ship, and later decelerating it by reversal of the thrust direction. In the Mars trip, tor instance, the ship's velocity would increase steadily up to the point of thrust reversal and then decrease to such an amount that, upon the approach to Mars, the ship would be captured by the planet's gravitational field. The primary power source must generate power throughout the time the ship is traveling. The total length of travel for a round trip to Mars will be of the order of two years.

     The basic assumptions underlying the design of the electrical space ship are a payload of 150 tons and an acceleration of about 10-4g. The payload includes the crew, with equipment and sufficient supplies of oxygen, water, food, living quarters, observation instruments, and the landing craft with equipment for the crew to subsist on Mars. The minimum acceleration must be great enough to complete the round trip in a reasonable time and to allow the expected corrective maneuvers during flight. A nuclear reactor is chosen for a primary power source. It is a "fast" reactor for weight-saving reasons, containing twelve tons of uranium. Its U-235 content is enriched to about l.7%. The reactor heat is absorbed by a cooling system employing sodium-potassium as a coolant.

     The reactor is located at a point about 250 feet away from the living quarters. It is shielded by a thick layer of beryllium and a sheet of boron so that the strong neutron and gamma radiations are kept away from the living quarters. The heat-energy contained in the sodium-potassium is transferred to a working fluid (silicon oil) in the heat exchanger. Steam produced in the heat exchanger drives a turbine which is coupled to an electric generator. The steam leaving the turbine enters a large radiation cooler where it condenses again. From there, the fluid is pumped back to the heat exchanger.

     The material best suited for the propellant is one that can be ionized easily and has a high yield. An alkaline metal like rubidium or cesium will be chosen. The atoms of these metals are ionized with almost l00% efficiency when they strike a hot surface of platinum foil. A temperature of about 200°C is enough to produce a sufficient vapor pressure of the alkaline element. The vapor enters an ionization chamber containing hot platinum grids, and the ions are extracted from the chamber by an electric field. This field accelerates the ions in the thrust chamber to a velocity of about 50 miles per second. They leave the propulsion system in a steady flow, representing an electric current. The electric power, as determined by this current and the potential difference through which the ions pass in the thrust chamber, must be provided by the electric generator.

     The maximum current density which can be obtained in the thrust chamber is limited by space charge effects. These effect also influence the formation of the jet of ions which extends from the thrust chamber into empty space. An unlimited beam of ions—even a beam of a noticeable length in fact—would be impossible. The space charge would act back on the thrust chamber and would neutralize the accelerating field.

     In order to produce and maintain a continuous flow of particles out of the propulsion system, the ions must be electrically neutralized soon after they leave the thrust chamber. Fortunately, this neutralization can be achieved rather easily. When the alkaline atoms come in contact with the heated platinum grid one negative electron jumps off every atom, leaving a positive ion behind. The electrons enter into the platinum foil. These electrons must be expelled from the ship — otherwise, the ship would quickly assume a strong negative charge which would prevent any further expulsion of positive ions through the thrust chamber. The natural way to neutralize the ions is to expel the electrons in the immediate vicinity of the ion thrust chambers. The ion beams mix shortly behind the end of the thrust chambers, where the electrons and ions form a neutral plasma. In this way, the strong space charge of the exhaust jet is avoided.

     An expulsion chamber for electrons consists of a hot filament which emits electrons, and a field of about 200 volts potential difference. One ion thrust chamber has a diameter of about one inch, and a total of many thousand thrust chambers will be needed to produce the thrust required for a space ship. The ion chambers and the electron chambers are tightly packed so that neutralization of the ions occurs at a distance of about one inch behind the thrust chambers. It is assumed that the power plant and thrust chambers are in operation during the entire trip, either accelerating or decelerating the vehicle.

     A schematic of the nuclear reactor, the heat exchanger, the turbogenerator, and the ion and electron chambers is shown in Figure 4. The largest component of the propulsion system will be the radiation cooler. The optimum size of the cooler is one for which the total mass of the power generating system is a minimum, based on a given electrical power output and a given temperature of the hot steam. The figure characterizing the specific power of the power plant, measured as power output divided by the total mass, proves to be one of the decisive figures from which the design of an electrical space ship must start. This figure was found to be of the order of 0.1 kw per kg. With that figure, an assumed acceleration of at least l0-4g, and a total payload of 150 tons, the design data for a space ship capable of going to Mars and back can be derived.

     The detailed study shows that for any given set of the four parameters -— payload, minimum acceleration, specific power, and destination planet — optimum values are found for propellant mass, total power, and accelerating voltage. With these optimum values, the total initial mass of the space ship is a minimum. Values different from these optimum figures would result in a heavier ship.

     The following design data were determined for the ship:

Total initial mass730 tons
Propellant mass365 tons
Total electric power23 megawatts
Accelerating voltage4380 volts
Exhaust velocity50 miles per sec.
Total thrust110 pounds

     The travel time of this ship to Mars would be a little over a year; the time for the trip back, a little less than one year. The ratio of total initial weight to payload is less than 5 to 1, which is a very favorable figure for a rocket propelled vehicle.

     The structural design of the ship will take into account the absence of atmospheric drag and appreciable acceleration forces. Structural elements will be very light The proposed design is shown in Figure 5. It is symmetrical around the longitudinal axis with the reactor at one end and the living quarters at the other. As soon as turbine and generator start to turn, the entire ship revolves slowly in the opposite direction. The rotation of the ship, which continues as long as the turbogenerator turns, is very desirable, because it makes the condensed fluid in the cooler flow to the outer rim, from where it can be pumped back conveniently to the heat exchanger. Also, the crew in the toroidal living quarters will sense at least a little gravity, simulated by the centrifugal force. The thrust chamber with propellant tanks will be mounted in such a way that the thruster force always goes through the center of gravity of the entire ship. The landing craft for Mars will be attached to the thrust chamber itself, with the thrust vector pointing through the center of gravity. The thrust vector will normally be parallel to the tangent of the trajectory.

     The flight path of a space ship with an electrical propulsion system differs from that of one powered by chemical rocket motors. The acceleration of an electrical ship is only a small fraction of one g. Its propellant consumption and mass ratio are smaller than in a chemically powered ship. The time of propulsion is much longer. As mentioned above, the electrical propulsion system operates during the entire trip except for a few powerless periods of short duration which are needed for corrective maneuvers. The electrical ship's trajectory will not follow an elliptical path, but segments of spirals.

     At first the ship spirals around the earth (Fig. 6), and its distance from the satellite station increases very slowly (after two hours, it will not be more than 20 miles away). After one hundred days of steady spiraling, its distance from the earth will be l00,000 miles — about halfway to the moon — and it will have completed 376 revolutions around the earth. A few days later, its speed and distance from the earth will have become so large that the ship is no longer restrained by the earth's attractive force. Its trajectory will flatten out, making a transition to a large spiral around the sun (Fig. 7).

     On the 195th day, the thrust unit will be rotated through 180 degrees and the ship starts to decelerate. If it did not, it could never be captured by the Martian gravitational field. The deceleration leads the ship gradually into an inward spiral about the sun.

     On the 276th day, the thrust will be switched again to acceleration, and this last maneuver carries the ship gently into the Martian ellipse. It arrives there on the 347th day. If the entire trip has been timed correctly, the ship will approach a point on the Martian ellipse where Mars is located at that time. If the ship should arrive too late or too soon, it will merely turn the thrust vector slowly toward the sun or away from the sun. By doing this, it manages to stay in the Martian ellipse with overspeed or underspeed. In the first case, it will approach Mars from the rear; in the second, it will be approached by Mars. The approach of the ship to the Martian ellipse and the capture are shown in Figure 8. A few thrust maneuvers, as indicated in the figure, will be necessary to direct the ship into a spiral around the planet. Otherwise, it would crash on Mars or pass the planet in a hyperbolic trajectory.

     On the 402nd day, the ship will have descended on the spiral to an altitude of 600 miles above the surface of Mars. The crew shuts off the motor and prepares for the exploration of the planet. The correct time to start the return trip will be 472 days away. This long waiting period gives the crew ample time to observe Mars closely by telescope and rocket probes, to descend to the barren surface with a winged landing craft, to explore the landscape and study the mysteries, and finally to return to the orbiting space ship by means of the rocket-powered central part of the landing craft.

     The trip back to earth will be similar to the earth-Mars trip. It will begin with 42 days of spiraling around the planet. Then, a decelerating period follows which puts the ship into an inward spiral around the sun. Subsequent acceleration adapts this spiral to the earth's ellipse. A few capture maneuvers follow, and a narrow spiral around the earth ends the long trip. After a total time of three and a quarter years, the crew arrives again in the orbit of the satellite station. A short shuttle trip takes them down to earth.

     The continuous operation of the propulsion system makes the guidance of the space ship easy. At no time will there be a need for unusual accuracy of presettings or aiming. Corrections can be introduced any time as soon as the trend toward a deviation becomes noticeable. During the spiraling around the earth or Mars, for example, a period of powerless orbiting can be introduced in case a time delay should be needed. If the ship should be late, it can gain time during the spiraling phase by opening the throttle a little more.

     Navigation likewise will not impose insurmountable problems. The ship will keep a constant watch of the earth, Mars, Venus, Jupiter and the sun. The directions to these celestial bodies with respect to the direction of one of the fixed stars will be continually measured and recorded by automatic star trackers. The actual positions of the sun and the planets in a coordinate system fixed to the stars are accurately known from the astronomical almanac. With these two sets of data, the instantaneous position of the ship can always be found. In fact, the ship's coordinates are continuously computed from the star tracker readings and compared with the expected coordinates. If any deviation should occur, corrective measures will be taken immediately.

     In spite of the fact that relatively simple techniques are available by which a space ship can be propelled, guided, and navigated through interplanetary space, a number of questions remain which might appear much more difficult to solve. Meteors and cosmic rays present a danger unknown to earth-bound beings who are well protected by the atmospheric shield.

     Maintaining an artificial atmosphere so human beings can live inside the living quarters and the space suits and work comfortably sounds like a tremendous problem. A total travel time of two full years duration spent in the monotonous seclusion of the ship's living quarters may seem a psychological impossibility. But things are not as bad as they might seem. Small meteors, which are infrequent, can be shielded off by an absorbing bumper. It consists of a thin sheet of metal around the ship. Larger meteors are very rare. If one of them should punch through the wall of the ship, the doors of the damaged compartment close automatically, and in most cases the damage can be repaired before a real disaster develops. If a vital part of the ship's machinery should be destroyed, the crew abandons that ship and boards one of the other ships (there will be a total of about 10 ships traveling together in one expedition). If a man should be hit — well, the probability for such an accident is about comparable to the probability that a man loses his life on this earth in some kind of accident.

     We still know little about the dangers of cosmic radiation in outer space. But we do know that these dangers are much smaller than previously assumed, and we may be confident that ways and means of efficient protection will be available before the first trip to Mars begins. After all, the manned satellite station will represent an excellent research laboratory to study all the effects of outer space, including weightlessness, artificial atmosphere, and life in confined quarters. By the time the first space ship leaves the satellite for Mars, its voyage will be much better prepared, in every respect and detail, than was Columbus' expedition when he started out to find this continent.

     The probability of the safe return of the spacefarers to earth will be greater than for many a daring and courageous team who set out in the past seeking new lands.

     The crew on an interplanetary ship will have more comfort and more space to move around in than the crew on a modern submarine. They will stay in constant contact with earth by radio and television. The men to be selected for the expedition must be of excellent health and stability. They will be persons of the scientific type who combine the love of adventure with the craving for scientific knowledge — men who can forget their personal desires in favor of the idea of a great technical and scientific achievement. Men of this nature will not mind spending two quiet years traveling on board a space ship. In his normal life, such a man always carries with him a backlog of unfinished scientific work which he cannot find the time to complete. If he is given the opportunity of two full years of undisturbed time to study and work on his pet projects, this prospect will be for him one more dream to come true when the first space ship takes off for its long voyage to Mars.

From OUTLOOK TO SPACE TRAVEL Space Journal #1 Summer 1957 by Ernst Stuhlinger (1957

Master artist Nick Stevens has recreated the umbrella ship in a series of images. Click to enlarge.

Blender artist Owen Egan is making his own recreation of the original Disney animation. I am quite impressed, looks just like the original.

I am not quite the artist that Nick Stevens and Owen Egan are, but I had to try my hand at it. Click to enlarge.


Mars Transportation System
University of Minnesota
Mass Schedule
Engines, fuel tanks, and RCS57.0 mT
MTV Habitat Module63.9 mT
Biconic MEV143.4 mT
Ascent/descent vehicle26.9 mT
Aeroshell (Mars)39.7 mT
Aeroshell (Terra)12.5 mT
DRY MASS359.1 mT
Propellant705.2 mT
(with 15% contingency)
1,118.2 mT
Mass Ratio3.11
ΔV12,245 m/s
Engine typeLow Pressure Nuclear Thermal Rocket
Number of enginesx3
Thrust per engine113 kN
Nozzle throat diameter1m
Nozzle length12.9 m
Nozzle area ratio60
Flow rate10.5 kg/s
Chamber pressure1 atm
Chamber temperature3,200°K
Specific impulse1,100 sec
Exhaust Velocity10,791 m/s
Mass per engine1,977 kg
Thrust / Weight ratio6
Thermal power525 MW
Fuel region power density4.6 MW/liter
Reactor geometryspherical shell
Reactor core ID/OD50cm/100cm
Reactor moderatorBe + Zr
Reactor reflectorBe + Graphite
Reflector thickness10cm
Flow directionRadial Flow
Fuel materialUC-ZrC
Fuel form1mm beads
Fuel matrix melting point3,700°K
Fuel density in fuel bed0.5g U235/cm3
Fuel loading70 kg U235
Fuel assembly typeAxial/Radial/Axial
Number of fuel assemblies120

This is from Conceptual design of a Mars Transportation System (1992). This was a year-long senior design course at the University of Minnesota, done in conjunction with NASA Marshall Space Flight Center and several major aerospace corporations

The design parameters was a spacecraft capable of transporting a six person crew to the Martian surface, providing for a sixty day surface stay, then transporting the crew back to Terra.


As with most spacecraft of this type, the Mars Transfer Vehicle (MTV) is all built around a long truss (150 meters) which is the thrust frame or ship's spine. It is a long truss because the radiation from the engines can kill the crew, and distance costs less in terms of ship mass than lead anti-radiation shields. The entire length of the spacecraft is 185.07 meters.

At the bottom of the truss are attached three solid-core nuclear thermal rockets, at the top is the habitat module. Just above the engines are eight propellant tanks full of liquid hydrogen, sized so they can be conveniently jettisoned at various stages of the mission. The Biconic Mars Excursion Vehicle (MEV) is an uncrewed cargo lander that transports to the Martian surface exploration gear and a surface base. Above the habitat is the Ascent/Descent vehicle that transports the crew to the surface, nestled in the Mars Aerobraking Shell. After the surface mission is complete, the Ascent component transports the astronauts back up to the MTV. At the end of the mission when the ship approaches Terra, the crew reenter the ascent component, dock with the Terra Aerobraking Shell, and uses it to land on Terra. The rest of the rocket goes sailing off into an eccentric Solar orbit, to be a radioactive hazard for the next few thousand years.

Assembling this entire clanking mess in orbit will be a challenge, given the limited boost capacity of current heavy-lift vehicles. The report goes into this in great detail, but I won't bore you with it.


This is the spacecraft's habitat module for the crew. It is welded to the top of the ship's truss, it does not ever leave the ship. The module is 16 meters long and has two stories. It contains 160 square meters of floor space and an internal volume of 375 cubic meters. This breaks down to 62.5 cubic meters per crew, which is safely above the 25 m3 minimum recommended by NASA.

In the center of the lower level is the storm cellar where the crew shelters in the even of a solar proton event. This doubles as the airlock and storage area of the ship's primary computer. The cellar's dimensions are 2m x 5m x 5m, with walls of 7cm solid aluminum. The overall mass will be 16.64 metric tons. This seems a bit inadequate to me, my slide rule say the walls will be about 189 kilograms per square meter, while other references say it should be closer to 5,000 kg/m2.

As with most such designs, it provides artificial gravity by the tumbling pigeon method. Outbound the effective artificial gravity with be 1.0 g (2.9 rpm) while inbound it will be 0.5 g (3.3 rpm). The distance of the habitat module from the spin center will change, since the ship's center of gravity moves as propellant is expended.


The uncrewed cargo lander has the bi-cone shape which is peculiarly effective for aerobraking on Mars. Bi-cone means the shape is a fat cone perched on top of truncated narrow cone. After aerobraking it lands using oxygen-methane rockets.

It ferries down to the Martian surface the Mars habitat, and a variety of exploration equipment including several rovers and a nuclear reactor. The crew stays in orbit until they are quite sure the MEV landed with all the cargo intact. The cargo lander will stay on the Martian surface for the rest of eternity, or at least until it is canabalized by future Martian colonists.

The Mars habitat is rated for six crew for 60 days (360 person-days).


The Ascent/Descent vehicle transports the crew from the orbiting MTV to the Martian surface, transports the crew back to the MTV at the end of the surface exploration phase, and finally transports them from the hurtling MTV into a circular orbit around Terra. Then they will rendezvous with the International Space Station or some ferry rocket which will eventually bring them home to Terra. A ballistic landing on Terra was rejected because of the huge heat shield required and the need for decelleration levels that might kill the crew.

Assuming the astronauts do not have the misfortune of living out some science fiction nightmare, with civilization vanishing in a zombie apocalypse or something. Even worse, ending with the astronauts on the beach staring at the ruins of the Statue of Liberty sticking out of the sand.

To save on fuel mass, aeroshells are used by the Ascent/Descent vehicle to shed velocity both for the Mars landing and insertion into low Earth orbit. Two aeroshells are carried, since the first one will be used up during Mars landing.

The fuel for descent is carried in tanks bolted to the Mars aeroshell. The ascent vehicle will leave the landing gear on the Martian surface, and will jettison spent fuel tanks on the way up.


The outbound leg will take about 200 days, the stay at Mars will be 60 days, and the inbound leg will be 250 to 270 days. Total mission duration will be from 510 to 530 days.

Marshall Manned Mars Mission

This is from the Marshall Space Flight Center workshop Manned Mars Missions Working Group Papers. This was to examine the impact of new technologies on existing proposals and to identify new technological issues. So there were actually several closely related designs instead of one single concept.

Martin Mars Mission System Study

Martin MMSS
EngineNTR Solid Core
5,000 MWth
Isp900 sec
Exhaust Vel8,830 m/s
10,000 kg
Rad Shield
10,000 kg
25,000 kg
25,800 kg
185,000 kg60,000 kg
Dry Mass255,800 kg130,800 kg
225,000 kg
Wet Mass480,800 kg355,800 kg
Mass Ratio1.882.72
ΔV5,574 m/s8,836 m/s
Transfer Time
220 to
300 days
100 to
170 days

This is from Manned Mars System Study (MMSS): Mars transportation and facility infrastructure study. Volume 2: Technical report (1990).

The basic design used a conventional liquid oxygen—liquid hydrogen cryogenic propulsion system, and was quickly mired in the boil-off problem. Plus the propellant mass was sizable.

They did a quick analysis of solid-core nuclear thermal, nuclear-electric with MPD thrusters, and solar-electric. The electric versions had a lower Initial Mass In LEO (IMLEO) but much longer transfer times. The nuclear-thermal on the other hand was far superior to the chemical cryogentic, with lower IMLEO and shorter transfer time.

Assumed ΔVs
Terra Departure ΔV3,800 m/s4,500 m/s5,500 m/s
Mars Arrival ΔV1,500 m/s2,500 m/s4,400 m/s
Mars Departure ΔV1,500 m/s2,500 m/s4,400 m/s
Terra Arrival ΔV3,800 m/s4,500 m/s6,200 m/s
TOTAL ΔV10,600 m/s14,000 m/s20,500 m/s
Flight Time
(each way)
200—330 days120—170 days80—120 days

For the cryogentic chemical vs nuclear thermal analysis, used a sample mission of a roundtrip voyage from LEO to a 250 km × 1 sol Martian orbit, with delta-Vs as per the above table. In addition the analysis assumed:

  • 60,000 kg roundtrip payload
  • Aerobraking at Terra and Mars
  • Aerobraking mass fractions are 15% for Conjunction mission, 20% for Medium energy mission, and 30% for High energy mission
  • NTR engine mass of 15,000 kg for Conjunction mission, 20,000 kg for Medium energy mission, and 30,000 kg for High energy mission
  • Staged engine burns
  • Cryogenic rocket stages with mass fraction of 0.9
  • Cryogenic rocket engine specific impulse of 470 seconds
  • NTR engine specific impulse of 900 seconds

The analysis used IMLEO as the metric, the lowest IMLEO wins. The results were:

Initial Mass in LEO
Cryo/no aerobrake958,000 kg2,479,000 kg23,211,000 kg
Cryo/aerobrake317,000 kg555,000 kg1,567000 kg
NTR/no aerobrake289,000 kg480,000 kg1,408,000 kg
NTR/aerobrake195,000 kg282,000 kg547,000 kg

As the missions increase in energy, so do the benefits of the NTR. However for High Energy missions, the NTR/no aerobrake is quite close to the mass of the Cryo/aerobrake. For that mission the NTR should use aerobraking to have a clear advantage over Cryo.

METTLE Mission To Europa

METTLE Mission
Length200 m
Hab Ring Radius45 m
Engine Power2.5 MW
Specific Impulse
(high gear)
29,969 s
Exhaust Velocity
(high gear)
294,000 m/s
(high gear)
17 N
Specific Impulse
(low gear)
2,956 s
Exhaust Velocity
(low gear)
29,000 m/s
(low gear)
172 N
Num Enginesx8
Total Engine
Power Req.
~20 MWe
Total Thrust
(high gear)
136 N
Total Thrust
(low gear)
1,376 N
Reactor Power5 MWe
Waste Heat
15 MWt
Num Reactors6
Total Reactor
30 MWe
Total Reactor
Waste Heat
90 MWt
Habitat Ring300 MT
Descent Vehicle300 MT
Remaining S/C1,400 MT
ΔV29,000 to
32,000 m/s
Mass Ratio1.10 to 1.12

This is from a student study Human Missions to Europa and Titan - Why Not? (2004)

The spacecraft has an overall length of 200 meters with a habitat module in the form of a ring with a 45 meter outer radius.

The thrust frame supports eight VASIMR engines mounted as four pairs (the report was skeptical about how much the VASIMR could actually vary its thrust, so they played it safe and varied it by assuming non-variable thrust and using multiple engines). Engine heat radiators are mounted betwen the engines.

The six nuclear reactors are mounted on three reactor support booms, two reactors per boom. Each reactor can produce 5 megawatts of electricity for a total of 30 megawatts (!!?!). The engines only require 20 MW so the mission can survive the loss of one boom. The reactors produce 90 MW of waste heat total, so the booms are coated with heat radiators. It is hard to tell from the diagram, but it looks to me like they have four radiator panels per boom, which drastically reduces the efficency to about 70% since the panels are shining heat into each other.

The reactors have shadow shields, you can tell because the heat radiators have been trimmed to stay in the shadow. Unfortunately if the radiators indicate the outline of the safe radiation shadow, to my eye it appears that the habitat ring is sticking into the deadly radiation zone.

The reactor support booms also houses the superconductive magnets and plasma injectors which create the artificial magnetosphere to protect the spacecraft from radiation.

From the point where the booms attach to the spacecraft's spine back to the aft end are mounted six propellant tanks containing argon reaction mass for the engines.

On the fore end of the spacecraft is the payload: the habitat ring and the Europa landing craft.


The habitat module is a ring with a 45 meter outer radius. It spins at 3 rpm to provide 0.45 g of artificial gravity. It is composed of 24 modules. Each has a floor space of 43 square meters, for a total of 1,032 square meters. The relatively large number of modules is to allow redundancy in module function, and to allow emergency isolation in case of depressurization or fire.


As always the delta V cost goes up when you lower the trip duration. See the graphs below. However the report warns that the graphs were calculated simplistically for an "impulse burn" which can only be performed by a high-thrust rocket, not a VASIMR which needs a long period of constant thrust. So the delta V cost should be given an additional margin of 20% to 30% more so as to take care of gravitational losses.

The report says for the outbound journey they selected a delta V of 14 km/s and 15 to 18 km/s for the return journey, giving a total flight time of 2.7 to 3.9 years. I tried drawing lines for these on the charts below but they do not seem to fit. Anyway according to my slide rule this implies an economical mass ratio of from 1.10 to 1.12.

Michael Nuclear Pulse Battleship

RocketCat sez

Oooooh, Yeah!!! The Orion-drive Michael Battleship is the biggest meanest son-of-a-spacer in the cosmos! Well, maybe second to the Project Orion Battleship.

Just look at that bad boy! Can't you just see that unstoppable titan blazing into orbit on a pillar of multiple nuclear explosions, ready to kick that alien bussard ramjet's buns up between its shoulder blades? The drawback to Orion-drive is that it don't scale down worth a darn. So they didn't even try. No "every gram counts" worries here, they freaking chopped the main guns off the freaking Battleship New Jersey and welded them on!

Any casaba howitzer weapons? Naw, spears of nuclear flame are too feeble. They are using full-blown freaking Excalibur bomb-pumped x-ray lasers! Not infrared, not visible light, not even ultraviolet. X-rays. Just like Teller intended.

What's that you ask? What about the pumping bomb? Well, this is an Orion-drive, moron. That's whats driving the ship. Spit out a few Excaliburs, they aim their hundreds of laser rods on their targets, then the next pulse unit simultaneously thrusts the ship and energizes the graser beams. Another jumbo-sized order of crispy-fried elephant, coming right up!

Still have megatons of payload allowance left over? Well, how about carrying a small fleet of gunships with nuclear missiles? And all four space shuttles?

The look on the elephant's faces was priceless! Michael is coming. And is he pissed!

Battleship Michael
PropulsionOrion Drive
Height226 meters
Diameter113 meters
Massbetween 35,000
and 50,000
metric tons

Warning: spoilers for the book Footfall by Larry Niven and Jerry Pournelle to follow. On the other hand, the novel came out decades ago in 1985. I mean, in the novel the U.S.S.R. still exists. It takes place in the far flung future year 1995.

Footfall is arguably the best "alien invasion" novel ever written. Just like The Mote in God's Eye is arguably the best "first contact" novel ever written. But I digress.

Aliens (called "Fithp") who look like baby elephants arrive from Alpha Centauri in a Bussard ramjet starship (hybrid Sleeper ship and Generation ship). The starship is named "Message Bearer." They immediately ditch the Bussard drive module into the Sun, destroying it. If the Fithp are defeated, the humans can jolly well build their own Bussard drive from scratch to travel to Alpha C and attack the Fithp homeworld.

The Fithp evolved from herd animals, unlike humans. They have a very alien idea of conflict resolution. When two herds meet, they fight until it was obvious which one was superior. Then everybody immediately stops fighting, and the inferior herd is peacefully incorporated into the superior tribe as second-class citizens. Fithp do not comprehend the concept of "diplomacy".

They make the unwise assumption that human beings operate the same way. Big mistake!

The Fithp have somewhat superior technology compared to humans. They attack and seize the Russian space station (the ISS was not started until 13 years after the novel was written), annihilate military sites and important infrastructure with rods from God, then invade Kansas. The Fithp think "Look, humans. We are obviously superior. Now is the time to stop fighting and be peacefully incorporated into our herd." The Fithp calmly wait for the human surrender.

Humans don't work that way (and they have no idea that the Fithp have such a bizarre way of interacting). They savagely counterattack with the National Guard and three US armored divisions. The Fithp are taken aback, and beat off the counterattack with orbital lasers and more rods from God. The humans respond with a combined Russian and US nuclear strike on Kansas, obliterating the Fithp invasion force and most of the Kansas heartland.

The Fithp start panicking. What is it going to take to make these crazy humans surrender?

Finally the Fithp decide to forgo all half-measures. They drop a small "dinosaur killer" asteroid on Terra. The asteroid is called "The Foot." This causes global environmental damage, and more or less kills everybody living in India. Surely this will make the humans surrender!

The Fithp obviously don't know humans very well.

The humans have their backs to the wall, since surrender is not in their nature. The US president has a tiger team of advisers, who were drawn from the ranks of science fiction authors. After all, they are the only experts on alien invasions (in the novel, the various advisers are thinly disguised versions of actual real-world authors. Nat Reynolds is Larry Niven, Wade Curtis is Jerry Pournelle, and Bob Anson is Robert Anson Heinlein). They have got to find a way to carry the battle to the enemy: the orbiting starship and the fleet of "digit" ships. But how do you get thousands of tons of military hardware into orbit quickly enough not to be shot down while in flight using only technology they can develop in a dozen months?

There is only one answer. Project Orion. Old boom-boom. And to heck with the limited nuclear test-ban treaty that killed the project in 1963.

Orion has already been developed. Orion is mass-insensitive, it doesn't care if you are boosting tens of thousands of metric tons. This also means you can use quick and dirty engineering, since you are not stopping every five minutes trying to shave off a few grams of excess mass. You don't have to spend a decade trying to engineer featherweight kinetic energy weapons, just go tear the gun turrets off the Battleship New Jersey and weld 'em on. You can also carry a fleet of gunboats. And all four space shuttles.

The gunboats are going to be quick and dirty as well. Spaceships built around a main gun off a Navy ship, firing nuclear shells. Yes, a spinal mounted weapon

What about the Orion drive battleship's main weapon? Heh. Another cancelled project rises from the grave.

Back in the days of the Strategic Defense Initiative, Edward Teller et al came up with Project Excalibur. What was that? No less than bomb-pumped x-ray lasers. But wait, what about the bomb you need to pump the laser? Well, Orion is an nuclear-bomb-powered drive, remember? Make the propulsive bombs do double duty.

The weapons are called "spurt bombs." Dispensers on the pusher plate eject a flight of the little darlings. The spurt bombs unfurl their 100 laser rods apiece and aim them at Fithp ships. The next nuclear pulse unit is positioned, then detonated. This simultaneously gives thrust to the spacecraft, and pumps all of the spurts bombs. The Fithp ships are sliced and diced by a hail of x-ray laser beams. Spurt bombs look like fasces, "bundles of tubes around an axis made up of attitude jets and cameras and a computer."

Note that the nuclear pulse units will have to be specially designed. Standard Orion pulse units are nuclear shaped charges, designed to channel 80% of the x-rays upwards into the pusher plate (well, to create a jet of plasma directed at the plate but I digress). The battleship's pulse units need to be designed to also direct x-rays at the spurt bombs.

What is the battleship's name? Michael of course. The Biblical Archangel who cast Lucifer out of heaven.

The Michael launches through a cloud of Fithp digit ships, cutting them to pieces but suffering serious damage. The Fithp defecate in their pants and frantically rip the starship out of orbit and start running away. Their superior acceleration make escape possible, up until the point where the crew of the Space Shuttle Atlantis commits suicide and rams the starship. The main drive is damaged, and their acceleration is no longer higher than the Michael. Who catches up and starts pounding the living snot out of the starship.

There is something breathtaking about the Michael that captures the imagination of science fiction fans. On pretty much every single online forum about spacecraft combat, it isn't long until somebody brings it up. There have been many examples of fans trying to make blueprints, illustrations, or even scratch-build models of the battleship.

The original Michael diagram was made by Aldo Spadoni, president of Aerospace Imagineering. Mr. Spadoni is an MIT educated mechanical/aerospace engineer with over 30 years of experience designing and developing advanced aerospace vehicle and weapon system concepts (with most of the more advanced work being classified). He is also a personal friend of Larry Niven and Jerry Pournelle.

Mr. Spadoni did the Michael diagrams around 1997, working directly with Niven and Pournelle. They went through several iterations to arrive at the resulting diagrams.


However, this does bring up a good point that Scott alluded to. Footfall is a novel of course, not an engineering proposal for a space battleship. You glean details regarding the various Footfall spacecraft from the conversations of characters in the story, many of which are not experts wrt what they are describing. As Scott also pointed out, there are inconsistencies in the descriptions that are either intentional or simply mistakes on the part of the authors. Thus, the design of the Footfall spacecraft are open to interpretation.

As an engineer and concept designer, I particularly like the way Larry and Jerry write their stories. They provide enough big picture detail to determine the general design direction for their concepts, but leave the smaller details and the system integration issues to anyone willing to take a crack at envisioning their concepts. Fun stuff! So, I think my overall design captures what the authors intended, but many of the details are open to different interpretations, as some of you have done here.

As I move into discussing some of Michael’s details, I want to note that my primary design goal was to be true to the novel and the authors’ intentions as I understood them. I have my own vision of what a space battleship might look like, as I’m sure many of you have. But that’s not the subject of this design exercise.

As did Scott (Lowther), I struggled to determine Michael’s overall dimensions, given the novel’s inconsistencies. Whatever they wrote, Larry and Jerry envisioned the most compact possible vehicle that would get the job done. Note that Scott is showing an older version of my drawing that shows Michael with the shock absorber array fully compressed along with incorrect dimensions. The final dimensions I came up with are somewhat larger, on the same order as those Scott mentions in a separate post.

Regarding the comment that this is a slick ILM Hollywood design, I think this is reading quite a lot into a hemisphere, a rectangular prism and a shallow cone! Perhaps the commenter is confusing vehicle configuration design with render quality. These drawings were never intended to portray Michael’s actual exterior finish, surface markings, etc. These drawings were created way back when using an ancient vector-based illustration software application called MacDraw Pro. They look pretty awful and it’s certainly not the way I would render Michael today. In hindsight, I should have left them as line drawings and avoided the use of MacDraw Pro’s primitive shading tools.

Regarding the battleship-derived gun turrets, I agree with Scott’s assertion that the text of the book is vague in this area. But based on my discussions with Larry and Jerry, the authors definitely intended for Michael to include two of the full-up 16-inch Iowa class turrets, as well as some smaller gun turrets, not the guns alone.

Regarding the Shock absorbers array configuration, I disagree with you guys. Thinking that Michael is a straight extrapolation of the conventional Orion design configurations is incorrect. The primary purpose of the shock absorber array is, of course, to smooth out the “ride” for the payload/passenger portion of the vehicle. Most of the Orion designs were configured for non-military applications, whereas Michael is a maneuvering warship with massive nuclear pumped steam attitude thruster arrays. In addition to primary Orion thrusting, Michael will be subjected to multi-axial mechanical loads that are NOT along the longitudinal axis of the ship. Also consider that Michael’s design incorporates a pusher “shell” that is far more massive as a fraction of total vehicle mass than the typical Orion pusher plate design. When Michael is thrusting under primary propulsion while engaging in combat maneuvers, an angled shock absorber array design is a good choice for handling the inevitable side loads and for stabilizing the shell wrt the passenger/payload “brick”. Consider a high performance off road vehicle, which must provide chassis stability while the wheels and suspension are being subjected to loads from many directions. You don’t see any parallel straight up and down shock absorbers in the suspension system, do you?

If you look carefully at me design, you can see that that central shock absorber is longitudinal and more massive than the rest. This one is primarily responsible for handling the Orion propulsive loads. Perhaps it should be a bit beefier than I’ve depicted it in the original drawing. The remaining angled shock absorbers handle some of that propulsive load while also providing multi-axial stability. Admittedly, these 2D drawings don’t convey the Shock absorber array configuration that I have envisioned very well.

Since the time these drawing were created, I’ve discussed Michael with Larry and Jerry on a number of occasions. I’ve reconsidered and refined many of Michael’s technical aspects and I’ve designed a more detailed and representative configuration, including an updated shock absorber array. I’m also involved in creating my own high fidelity 3D model of Michael with a few fellow conspirators. I’m looking forward to sharing that with everyone at some point.

(ed note: one of those "few fellow conspirators" was me. Another was Andrew Presby, who is featured on one or two pages of this website.)

From a comment on the Unwanted Blog by Aldo Spadoni (2012)

Around 2010 Andrew Presby and I were commissioned by Aldo Spadoni to turn his Michael blueprints into 3D renders. Click for larger images.

Scott Lowther, author of Aerospace Projects Review is working on a book about nuclear space propulsion. Of course he wouldn't dream about leaving out the coolest Orion Drive spacecraft of all.

Now, strictly by the novel, the Michael is a mile high, which is ludicrous. The protagonists would have to have built a mile-high dome to cover it, which the aliens might have found a bit suspicious. In the diagrams below, Mr. Lowther shows the "large" Michael (one mile) and the more reasonable "small" Michael (1/8th mile).

Master artist William Black also had to turn his formidable talents on the Michael.


Nuclear pulse propulsion battleship Michael from the novel Footfall by Larry Niven & Jerry Pournelle.

“Michaels nose was a thick shield … armored in layers: steel armor, fiberglass matting, more steel armor, layer after layer of hard and nonresiliant soft.” —from Footfall, pg. 446 and 472

“Two great towers stood on the curve of the hemispherical shell, with cannon showing beneath the lip, aimed inward. Four smaller towers flanked them. A brick-shaped structure rose above them. The Brick was much less massive than the Shell, but its sides were covered with spacecraft: tiny gunships, and four Shuttles with tanks but no boosters. The bricks massive roof ran beyond the flanks to shield the Shuttles and gunships.”  —from Footfall, pg. 432

Michael is one of the Orion based concepts I knew I would have to take a run at sooner or later. I referenced the novel, extensively, and Scott Lowther condensed all the design bits he gleaned from Footfall into an Excel spreadsheet, available here, for a project he set aside. The spreadsheet is an excellent guide to all the passages describing Message Bearer, the digit ships, Michael, the stovepipes and Shuttles, and it proved invaluable in my effort.

Most people are probably familiar with Aldo Spadoni's visualization of the iconic warship from Niven and Pournelle’s novel, but for those who are not, Aldo’s drawings are available here.

What I’ve done is meet the Aldo Spadoni design half-way with my own interpretations. My intent was to complement Aldo’s design-thought without entirely rewriting it, keeping in mind what Aldo had to say about the process. One point Aldo raised in conversation on Scott Lowther’s blog is in regards to who is providing description in various scenes from the novel.

Aldo Spadoni: “Footfall is a novel of course, not an engineering proposal for a space battleship. You glean details regarding the various Footfall spacecraft from the conversations of characters in the story, many of which are not experts [with regard to] what they are describing. As Scott also pointed out, there are inconsistencies in the descriptions that are either intentional or simply mistakes on the part of the authors. Thus, the design of the Footfall spacecraft are open to interpretation.”

Aldo makes a good case for the distinctive angled shock absorbers of his design, and I’ll provide his commentary below, the sticking point for me, however, is the parabolic pusher plate Niven and Pournelle describe—early design work on Orion solidly ruled out a parabolic pusher. With shaped-charge nuclear pulse units the parabolic plate will only heat up while offering almost no thrust advantage. Heating and impact stress on the pusher would be of no small concern, the bombs necessary to loft something the scale and mass of Michael would not be the tame little devices used to propel a dinky NASA/USAF 10-meter Orion. Heating is the cost of even partially containing the ionized plasma resulting from nuclear detonation.

Orion works because the plasma is dynamically shaped (as the explosion happens) by the specially designed shaped charge nuclear explosive, X-rays are channeled by the radiation case in the instant before the weapon is vaporized, these exit a single aperture, striking and heating up a beryllium oxide channel filler and propellant disk (tungsten), resulting in a narrow conical jet of ionized tungsten plasma, traveling at high velocity (in excess of 1.5 × 10⁵ meters per second). This crashes into the pusher plate, accelerating the spacecraft. The jet is not physically contained by the pusher, and contact with the pusher is infinitesimally brief, so the pusher is not subject to extreme heating during thrust maneuvers. So, while offering very little performance difference compared to a flat pusher design, the parabolic plate would need regenerative cooling in the bargain, adding weight and complexity to the system. Engineering such a pusher plate would be fraught with difficulties, and conditions under which Michael is built, in my opinion, rule out any eccentric messing with the baseline system. A legion of Ted Taylors would already be kept busy night and day with the mere task of readying a conventional Orion designed under such circumstances—for delivery under a one year drop-rocks-from-orbit-dead deadline.

As Aldo points out, the text of Footfall leaves room for different interpretations and here is where I took some of Aldo’s design-thought and creatively merged it with my own toward the end of addressing the design as presented in the novel. (No, not the army of Ted Taylor clones inhabiting a maze of cubicles in some deep bunker somewhere—that’s just me.)

It occurred to me that what Aldo had done (following Niven and Pournelle’s description), was move the functions of the Orion standard propulsion module down, mounting them directly on the top of the plate, so really it’s a built up intermediate platform/propulsion module. What I’ve done is run with that thought: I chose to treat the entire pusher plate as an early large Orion: a dome sitting on flat pusher plate, concentric rows of toroidal shock absorbers surrounding a core array of gas-piston shock absorbers. There is no central hole-and-bomb-placement-gun-protection-tube in my design (but there is an anti-ablation oil spray system). Instead, pulse units are shot by bomb placement guns mounted to fire around the edge; exactly as in Aldo’s design (the early large Orion had rocket assisted bombs riding tracks on the exterior of the spacecraft—imagine the show that would make). The body of the “dome” in my design is stowage for tanked pressurization gas (for the shock absorbers), anti-ablation oil, and perhaps a reserve number of pulse units.

I’ve retained the scheme of duel pulse unit magazines. Niven & Pournelle called them “thrust bomb” towers. Four “spurt bomb” towers are also mounted to the base—the “spurt bomb” Niven and Pournelle describe is a type of bomb-pumped laser using gamma-radiation rather than X-rays. All of my towers are a good deal beefier than those on Aldo’s design. Narrative in the novel describes the “thrust bomb” towers as doing double duty, providing an extra layer of armor and shielding for the CIC/control room, the nerve center of the spacecraft, which is located in the lower portion of the Brick, wedged between two large water tanks (and two nuclear reactor containment vessels). The water tanks are frozen at lift-off, providing Michael with an ample heat-sink.  

As I mentioned above, Aldo makes an excellent case for the angled shock absorbers on his design, his description below:

Aldo Spadoni: “Most of the Orion designs were configured for non-military applications, whereas Michael is a maneuvering warship with massive nuclear pumped steam attitude thruster arrays. In addition to primary Orion thrusting, Michael will be subjected to multi-axial mechanical loads that are NOT along the longitudinal axis of the ship. … When Michael is thrusting under primary propulsion while engaging in combat maneuvers, an angled shock absorber array design is a good choice for handling the inevitable side loads and for stabilizing the shell [with regard to] the passenger/payload “brick.” Consider a high performance off road vehicle, which must provide chassis stability while the wheels and suspension are being subjected to loads from many directions. You don’t see any parallel straight up and down shock absorbers in the suspension system, do you?

If you look carefully at my design, you can see that that central shock absorber is longitudinal and more massive than the rest. This one is primarily responsible for handling the Orion propulsive loads. … The remaining angled shock absorbers handle some of that propulsive load while also providing multi-axial stability.”

Scott Lowther (of Aerospace Projects Review) offers this insight in regards to angled shock absorbers:

Scott Lowther: “I remain unconvinced at the off-axis "angled" shock absorbers, but they seem to be the popular approach. However, if you do go that route, you have to deal with the central piston in the same way... ball joints fore and aft. *All* the pistons must be free to swing from side to side. If one, even the central one, is locked, then either the pusher assembly cannot move sideways *thus negating the value of the angled shocks), or it'll simply get ripped off its mounts the first time there's an off-axis blast.

Given that the ship is clearly described as having nuclear steam rockets for attitude control, I don't see the value in off-axis blasts for steering. But... shrug.”

I spent a good deal of time reproducing Aldo’s shock absorber array because frankly I think it is brilliant, going back and forth between Aldo’s drawings and my file … in the end the detail would be invisible, so I created a cutaway render with two of the “spurt bomb” towers removed to reveal the system.

True to the novel Michael’s main guns are the 16"/50 caliber Mark 7 gun and turret taken directly off the New Jersey. There is a good deal of discussion (on Scott’s blog and elsewhere) on the suitability of the guns and turrets—the mounting is rotated ninety degrees to vertical relative to the orientation turret, guns, and loading mechanisms were designed for—however, Aldo is quite clear that mounting the full turrets “as is” reflects the author’s intention, and so I’ve kept to their vision in this regard.

In the novel the guns are described firing a nuclear artillery round, this would be a modern version of the W23 15-20 kiloton nuclear round. The Mark 23 was a further development of the Army's Mk-9 & Mk-19 280mm artillery shell. This was a 15-20 kiloton nuclear warhead adapted to a 16 in naval shell used on the 4 Iowa Class Battleships1. 50 of these weapons were produced starting in 1956 but shortly after their introduction the four Iowa's were mothballed. The weapon stayed in the nuclear inventory until October 1962. Presumably under war conditions a new production run would produce the numbers necessary for Michael’s assault on Message Bearer.

Secondary batteries: a generic turret roughly based on the secondary turrets of the Iowa class.

Missile launchers based on the MK-41 Vertical Launching System (VLS).

The “Battle Management Array” is a set of phased-array radars and tight-beam communications antenna for passing targeting information to Michaels secondary spacecraft, all mounted to a pair of shock-isolated cab, each riding its own set of shock absorbers, one mounted atop each “thrust bomb” tower.  A fall-back set of communications antenna and radar are mounted beneath the overhang of the forward shield atop the Brick.

I’ve gone with the dimensions Scott arrived at, which Aldo confirmed in his comments on Scott’s blog: Length:742’ Diameter: 371’.

Different opinions have been offered in regards to Michael’s mass, between 35,000 and 50,000 tons have been opinioned on Scott Lowther’s blog. Pournelle was quoted as saying 2 million tons on one occasion, and 7 million tons on another.

Michaels launch, in the novel, is shortened for reasons of narrative brevity; one character wonders if there were perhaps 30 or more nuclear detonations. Putting Michael in orbit would require 8 minutes of powered flight and about 480 bombs lit off at one bomb per second.

The novel is clear that Michael carries four Space Shuttles mounted to their external tanks sans their SRBs. The number of Gunships is less clear. Nine Gunships are described as destroyed in combat, an unspecified number survive to confront Message Bearer in the final scene. Designing the most compact spacecraft necessary to fill the role, my Gunship measures 100 feet in length, 25 feet in diameter. At these dimensions, 14 Gunships total can be comfortably mounted to Michaels flanks.

For detail on my Gunship design see my following post, Gunship.

1 W23

From Michael by William Black (2015)

Gunship from Larry Niven and Jerry Pournelle's novel Footfall. See my related post Michael for additional detail.

“They take one of the main guns off a Navy ship. Wrap a spaceship around it. Not a lot of ship, just enough to steer it. Add an automatic loader and nuclear weapons for shells. Steer it with TV.” —from Footfall, pg. 354

In the novel these Gunships are referred to as “Stovepipe’s.” I was far less concerned with designing to match that narrative description than I was with designing the most compact spacecraft possible capable of the mission described. Michaels construction (including all its auxiliary spacecraft and subsystems) takes place in secret under wartime conditions, perhaps the moniker is derived from a code name picked randomly (that’s how the 1958 Project Orion was named), or perhaps dockworkers handling the vehicle sections, packed in featureless cylindrical shipping containers strapped to pallets, named the craft, and it stuck. See Aldo Spadoni’s commentary on character-delivered descriptions on my Michael post.    

I built my Gunship around the 5"/54 caliber Mark 45 gun.

Nuclear Round

The nuclear round fired by the Gunship would be something akin to the UCLR1 Swift, a 622 mm long, 127 mm diameter nuclear shell, weighing in at 43.5 kg.

In 1958 a fusion warhead was developed and tested. At its test it yielded only 190 tons; it failed to achieve fusion and only the initial fission explosion worked correctly. There are unconfirmed reports that work on similar concepts continued into the 1970s and resulted in a one-kiloton warhead design for 5-inch (127 mm) naval gun rounds, these, however, were never deployed as operational weapons. See paragraph 9 (not counting the bulleted list) under United States Nuclear Artillery.

Gunship Crew & Crew Module  

The text of the novel is unclear on the number of crew manning the Gunships, but my opinion is no more than 2 would be required, and dialogue in the novel tends to back this up. The loading mechanism is automated, so only targeting and piloting skills are involved. Considering urgency involved in readying Michael, I doubt an entirely new capsule, man-rated for spaceflight, would be considered. Michaels designers would fall back on tried and tested designs and modify them as required. In this case a stripped down Gemini spacecraft and its Equipment Module fits the bill nicely. The life support system matches the mission requirements. Leave off the heat shield (these are one-way missions), and reaction control system—the capsule never operates separate from the Gunship rig. Mount targeting and firing controls for the gun. Probably a single hatch rather than Gemini’s double hatch, and internal flat-screen displays rather than viewports—looking on this battle with naked eyes would leave the astronaut seared, radiation burned, and blinded.

“The exhausts of the gunboats were bright and yellow: solid fuel rockets.” —from Footfall, pg. 454

Eight SRBs akin to the GEM-40 allow options: they could be fired in pairs, allowing four separate burns, or two burns of 4, or a single burn of all eight – needs depending. The SRBs are strapped around a ten foot diameter 40 foot long core containing ample tank stowage for hypergolic reaction control propellants, pressurization gas, and nitrogen for clearing the breech and gun barrel. The reaction control system is used to aim the gun; propellant expenditure would be prodigious.

1UCRL - University of California Radiation Laboratory

From Gunship by William Black (2015)

Mini-Mag Orion

Mini-Mag Orion
PropulsionMini-Mag Orion
Thrust1,870,000 n
Exhaust Velocity157,000 m/s
Thrust Power147 GW
Pulse Unit Energy340 GJ
Nozzle Efficiency87.1%
Nozzle Mass199.6 metric tons
Pulse Rate1 per second

Data is from Mini-Mag Orion Program Document: Final Report from Ralph Ewig's website.

The nuclear pulse Orion drive propulsion system had both reasonably high exhaust velocity coupled with incredible amounts of thrust, a rare and valuable combination. A pity it was driven by sequential detonation of hundreds of nuclear bombs, and required two stages of huge shock absorbers to prevent the spacecraft from being kicked to pieces.

Andrews Space & Technology tried to design a variant on the nuclear Orion that would reduce the drawbacks but keep the advantages. The result was the Mini-Mag Orion.

First off, they crafted the explosive pulses so each was more 50 to 500 gigajoules each, instead of the 20,000 gigajoules typically found in the nuclear Orion. Secondly they made the explosions triggered by the explosive charge being squeezed into critical mass using an external power source instead of each charge being a self-contained easily-weaponized device. Thirdly they made the blast thrust against the magnetic field of a series of superconducting rings (Magnetic Nozzle) instead of the nuclear Orion's flat metal pusher plate.

In the standard nuclear pulse Orion, the pulse units are totally self-contained, that is, they are bombs. Since this makes it too easy to use the pulse units as impromptu weapons (which alarms the people in charge of funding such a spacecraft) a non-weaponizable pulse unit was designed. The Mini-Mag Orion pulse unit has the fissionable curium-245 nuclear explosive, an inexpensive Z-pinch coil to detonate it, but no power supply for the coil. The Z-pinch power comes from huge capacitor pulse power banks mounted on the spacecraft, i.e., the pulse unit ain't anywhere near being "self contained". The banks have a mass of a little over seven metric tons, far too large to use in a weapon (especially one that explodes with a pathetic 0.03 kilotons of yield). The Z-pinch coil should be inexpensive since it will be destroyed in the blast.

For a 50 gigajoule yield (with a burn fraction of 10%), the nuclear explosive is 42.9 grams of curium-245 in the form of a hollow sphere 1.27 centimeters radius (yes, I know the diagram above says the compression target is 0.47 centimeters radius. I think they mean the compressed size). This is coated with 15.2 grams of beryllium to act as a neutron reflector. According to the table below, a 120.7 gigajoule yield uses 21 grams of curium, which does not make sense to me. Usually you need more nuclear explosive to make a bigger burst. I guess the pulse units in the table have a larger burn fraction. The Z-pinch will squeeze the curium sphere from a radius of 1.27 centimeter down to 0.468 centimeters, leading to a chain reaction and nuclear explosion. Since curium-245 has a low spontaneous fission rate, the pulse unit will need a deuterium/tritium diode to provide the triggering neutrons. The pulse units will be detonated about one per second (1 Hz).

The Z-pinch needs 70 megaAmps of electricity. This is 70 million amps, which is a freaking lot of amps. The trouble is that you cannot lay big thick cables to the Z-pinch coil in the pulse unit. The cable will be vaporized by the nuclear explosion, which is OK. But a vaporized massive cable composed of heavy elements will drastically lower the exhaust velocity. This is very not OK. Remember that one of the selling points of the Mini-Mag Orion is the high exhaust velocity. Reduce the exhaust velocity and Mini-Mag Orion becomes much less attractive.

So instead of heavy cables the pulse unit uses gossamer thin sheets of Mylar (20 μ thick). I know that Mylar is usually considered an insulator, but 70 megaAmps does not care if it is an insulator or not. The report calls these Mylar cables Low Mass Transmission Lines (LMTL). They have a total mass of only 2 kilograms, which is good news for the exhaust velocity.

The 70 megaAmps go from the pulse power banks to permanent electrodes mounted on the magnetic nozzle. These take the form of five meter diameter metal rings. Two rings, positive and negative, just like the two slots in an electrical wall socket. The pulse unit proper is a minimum of 0.0244 meters diameter (double the 1.27 centimeter radius). So the LMTL has to stretch from the permanent electrodes to the pulse unit. This makes a five meter diameter disk of Mylar with with the grape sized pulse unit in the center. Actually two stacked Mylar disks (positive and negative) separated by about 2 centimeters of space (g0 in diagram above) so they won't short circuit. Ordinarily you'd use an insulator to prevent a short, something like, for instance, Mylar. Unfortunately here you are using Mylar as the conductor so instead you need a gap. The edge of each Mylar disk has an aluminum rim, each making contact with one of the magnetic nozzle's two permanent electrodes.

To place the pulse unit in the proper detonation point inside the magnetic nozzle, the pulse unit has to be five meters lower than the permanent electrodes in the nozzle. This forces the Mylar LMTL to be an upside down cone instead of a flat disk.

The pulse unit, Mylar LMTL and the aluminum rims are all vaporized during detonation. The magnetic nozzle with its permanent electrodes remain.

There are two power supplies: the steady-state reactor and the pulse power banks.

The reactor is the "charger." It charges up the superconducting magnetic nozzle, and gives the pulse power banks their initial charge. Finally it supplies power to the payload (including the habitat module). In the reference designs below, it outputs 103 kilowatts, has a mass of 9 metric tons, and is expected to supply 50 kilowatts to the payload. It takes 1 hour to give the pulse power banks (main and backup) their initial charge, and takes 39 hours to charge up the superconducting magnetic nozzle. Since the nozzle uses superconductors, its charge will last a long time before it leaks out.

The reactor has to supply 192 megajoules over one hour to charge up the main and backup pulse power banks. The reactor has to supply 7,446 megajoules over 39 hours to charge up the superconducting nozzle.

The tiny bombs need 70 megaAmps in 1.2 microseconds in order to detonate, but the reactor can only produce that many amps in one hour. The standard solution is to use capacitors, which can be gradually filled up but can dump all their stored energy almost instantly. This is the pulse power banks, a Marx bank of capacitors.

The reactor takes half an hour to charge up one pulse power bank, one hour to charge up the bank and the backup bank. The bank discharges all that energy into the pulse unit to detonate it. A separate system in the magnetic nozzle converts about 1 percent of the explosion into electricity and totally recharges the pulse power bank. For subsequent detonations, the reactor is not needed, the detonating bombs supply the power.

In the reference design, the pulse power banks hold 96 megaJoules per bank, there is a main bank and a backup bank for a total of 192 megaJoules, each bank has a mass of 3.5 metric tons, main and backup bank have a combined mass of 7.1 metric tons. The banks have to sustain a pulse unit detonation rate of 1 per second (1 Hz).

The backup bank is in case of a misfire, resulting in a lack of a recharge for the main bank. The still-full backup bank takes over energizing the pulse detonations while the reactor starts slowly re-charging the main bank.

Since the electrical system will be operating at megawatt levels, it will need a sizable set of heat radiators (Thermal Management System). By "sizable" we mean "up to 30% of the spacecraft's dry mass." In the first reference mission, the radiators have to handle 2,576 kW of waste heat, with the radiators having a mass of 15,456 kg and a surface area of 7,728 square meters.

The heat radiators are tapered in order to keep them inside the shadow cast by the radiation shadow shield. This keeps the radiators relatively free of neutron activation and neutron embrittlement. It also prevents the radiators from backscattering deadly nuclear radiation into the crew compartment.

The engine core and feed mechanism will have to inject the pulse units into the detonation point at rates of up to 1 Hz. It too will need redundancy and a minimum of moving parts.

In the second diagram above:

  1. Cycle begins. A pulse unit is at the detonation point with its LMTL contact rings touching the magnetic nozzle's permanent electrodes. Both blast doors are closed. The nozzle is fully extended.
  2. 70 megaAmps detonates the pulse unit. The explosion transmits force into the magnetic nozzle, producing thrust. 1% of the blast energy is converted into electricity which re-charges the pulse power bank. The nozzle moves upward along the feed system as part of the compression cycle. Meanwhile, the upper blast door opens to allow the next pulse unit to enter the feed system.
  3. The explosion plasma dissipates. The nozzle continues to move upward. As the next pulse unit enters the feed system, the upper blast door closes.
  4. The lower blast door opens. The nozzle reaches its highest position. The fresh pulse unit is injected into nozzle at the detonation point with a velocity matching the nozzle, LMTL contact rings of pulse unit touching nozzle's permanent electrodes. The lower blast door closes as the nozzle starts to travel downward along the feed system. When the nozzle reaches it lowest point, a new cycle begins.

The report had three sample "Design Reference Missions", and created optimal spacecraft using MiniMag Orion propulsion. As it turns out, the spacecraft for mission 1 and mission 2 were practically identical, so they only showed the two ship designs.

Design Reference Missions

DRM-1: Crewed Mars Mission:
50 kWe, 100 km/s Δv, 100 ton payload, 90 to 100 days one way trip time.
DRM-2: Crewed Jupiter Mission:
50 kWe, 100 km/s Δv, 100 ton payload, 2 years one way trip time.
DRM-3: Robot Pluto Sample Return:
50 kWe, 150 km/s Δv, 5 ton payload, 8 years one way trip time.

DRM-1/DRM-2 Spacecraft

DRM-1 Mass Budget
Mission delta-v100 km/s
Specific Power347 kW/kg
(347,400 W/kg)
Thrust Power87 gigawatts
Payload Mass100,000 kg
Specific Impulse9,500 sec
Exhaust Velocity93,000 m/s
Power System Mass (Charge)9,038 kg
Power System Mass (Pulse Banks)7,115 kg
Heat Radiators15,456 kg
Magnetic Nozzle Mass102,893 kg
Propellant Mass481,625 kg
Dry Mass (no remass, no payload)150,300 kg
Burnout Mass (no remass)250,300 kg
Ignition (Wet) Mass731,924 kg
Payload Fraction0.137
Propellant Fraction0.66
Dry Mass Fraction0.21
Power System - Pulse Banks
Peak Compression Current89 MA
Capacitor Voltage170 kV
Energy per Bank96 MJ
Capacitor Energy Density54 kJ/kg
Capactior Mass (one bank)1,779 kg
Pulse Bank Mass (total for 2 banks)7115 kg
Power System - Charge Power
Pulse Bank Charge Time60 minutes
Nozzle Charge Time39 hours
Pulse Banks Energy Content192 MJ
Nozzle Energy Content7,446 MJ
Payload Power Requirement50 kW
Power Output Electric103 kW
System Power Density11.4 W/kg
Thermal to Electric Efficiency0.04 fraction
Total Mass9,038 kg
Heat Radiators
Waste Heat Load2,576 kW
Area per Watt3 m2/kW
Mass per Area2 kg/m2
Radiator Area7,728 m2
Radiator Mass15,456 kg
Engine Performance
Specific Impulse9,500 sec
Exhaust Velocity93,164 m/s
Nozzle Efficiency0.45 fraction
Coupling Efficiency0.55 fraction
Pulse Yield120.7 GJ
Pulse Unit Mass6.9 kg
Standoff Distance5.9 m
Fission Assembly Mass21 g
Firing Rate1 Hz
Mass Flow6.9 kg/s
Thrust642 kN
Power29,894 MW
Gain563,631 ratio
Alpha (specific power)222,251 W/kg
Maximum Acceleration0.26 g's
Minimum Acceleration0.09 g's

DRM-3 Spacecraft

DRM-3 Mass Budget
Mission delta-v150 km/s
Specific Power551 kW/kg
(551,300 W/kg)
Thrust Power87 gigawatts
Payload Mass5,000 kg
Specific Impulse9,500 sec
Exhaust Velocity93,000 m/s
Power System Mass (Charge)9,067 kg
Power System Mass (Pulse Banks)7,115 kg
Heat Radiators15,505 kg
Magnetic Nozzle Mass102,895 kg
Propellant Mass630,963 kg
Dry Mass (no remass, no payload)152,723 kg
Burnout Mass (no remass)157,723 kg
Ignition (Wet) Mass788,686 kg
Payload Fraction0.006
Propellant Fraction0.8
Dry Mass Fraction0.19
Power System - Pulse Banks
Peak Compression Current89 MA
Capacitor Voltage170 kV
Energy per Bank96 MJ
Capacitor Energy Density54 kJ/kg
Capactior Mass (one bank)1,779 kg
Pulse Bank Mass (total for 2 banks)7115 kg
Power System - Charge Power
Pulse Bank Charge Time60 minutes
Nozzle Charge Time39 hours
Pulse Banks Energy Content192 MJ
Nozzle Energy Content7,447 MJ
Payload Power Requirement50 kW
Power Output Electric103 kW
System Power Density11.4 W/kg
Thermal to Electric Efficiency0.04 fraction
Total Mass9,038 kg
Heat Radiators
Waste Heat Load2,576 kW
Area per Watt3 m2/kW
Mass per Area2 kg/m2
Radiator Area7,752 m2
Radiator Mass15,505 kg
Engine Performance
Specific Impulse9,500 sec
Exhaust Velocity93,164 m/s
Nozzle Efficiency0.45 fraction
Coupling Efficiency0.55 fraction
Pulse Yield120.7 GJ
Pulse Unit Mass6.89 kg
Standoff Distance5.9 m
Fission Assembly Mass21 g
Firing Rate1 Hz
Mass Flow6.89 kg/s
Thrust642 kN
Power29,894 MW
Gain560,167 ratio
Alpha (specific power)222,122 W/kg
Maximum Acceleration0.41 g's
Minimum Acceleration0.08 g's

MOVERS Orbital Transfer Vehicle

Specific Impulse880 s
Exhaust Vel8,600 m/s
Thrust134,000 N
Endurance7 days
(21 person-days)
PowerFuel cells
Hab Module1,361 kg
Command Module363 kg
Power Systems
1,814 kg
RCS472 kg
and Rendezvous
471 kg
3,583 kg
NTR Engine1,814 kg
Shadow Shield3,856 kg
Propellant Tanks2,994 kg
w/ rad shielding
9,024 kg
DRY MASS25,753 kg
Payload0 kg
DRY MASS25,753 kg
Propellant42,317 kg
WET MASS68,070 kg
Mass Ratio2.64
ΔV8,380 m/s
Payload6,804 kg
DRY MASS32,557 kg
Propellant54,968 kg
WET MASS87,525 kg
Mass Ratio2.69
ΔV8,540 m/s

This is from Conceptual Design of a Manned Orbital Transfer Vehicle (1988). The function of the spacecraft was to deploy, recover, and repair satellites. Those things are expensive, it would be a vast saving to repair and refurbish satellites in place instead of sending up an entire new satellite. The report was prepared by the Modular Orbital Vehicle Engineering Research Society (MOVERS) of the University of Virginia.

The design criteria specified an ability to deliver and retrieve a payload of 6,800 kg from geosynchroneous orbit. A crew of three, life support for seven days, support for extra-vehicular activites. In addition the basic spacecraft should be adaptable to Terra-Luna missions with payloads up to 36,290 kg. This will be done by attaching more modules and propellant tanks.

The basic spacecraft has a delta V of about 8,400 m/s. Varying amounts of propellant are carried depending upon the payload mass, if any.

The elongated tanks are the main propellant tanks. There are four. Dimensions are 11.9 meters long by 4.5 meters in diameter. They carry a total of 42,317 kg of propellant (10,579 kg each), enough for flying with zero payload and 8,380 m/s of delta V.

The spherical tanks are the secondary propellant tanks. There are four. Dimensions are 4.6 meters long by 4.5 meters in diameter. They carry a total of 12,654 kg of propellant (3,162 kg each). With both the main and secondary tanks filled there is a total of 54,968 kg of propellant, enough for flying with 6,800 kg of payload and 8,540 m/s of delta V.

A sample mission servicing a Telstar satellite in GEO requires

Enter waiting ellipse775 m/s2 hours
11 minutes
Enter Hohmann transfer
Transit to Telestar
1,658 m/s5 hours
16 minutes
Match velocity with Telstar1,834 m/s< 14 minutes
Satellite servicing04.5 days
Enter Hohmann transfer
Transit to Terra LEO
1,834 m/s5 hours
16 minutes
Enter waiting ellipse
Dock with space station
2,169 m/s1 hour
43 minutes
TOTAL8,270 m/s5 days
2 hours
29 minutes


This is the heart of the spacecraft, how it earns its keep. The system can resupply fluid consumables to orbiting spacecraft (RCS fuel and coolant) as well as replace malfunctioned or obsolete components. It utilizes a waldo arm but has a backup of a Manned Manuvering Unit to allow an astronaut to go EVA and fix things manually. The entire servicing system is modular and be be detached from the core orbital transfer vehicle.


This supports the Remote Manipulator System (RMS) or waldo arm, which is basically the same as the one on the Space Shuttle. The RMS is approximately 15 meters long, it can safely manipulate a satellite up to 9 meters away. The notch in the ESM supports the waldo arm during periods of acceleration, so the blasted thing does not snap like a twig.

The ESM also has a cubby for the Manned Maneuvering Unit (MMU) or astronaut rocket backpack. The cubby has tanks of nitrogen propellant to recharge the MMU. There is also a cubby for the as-yet not developed Flight Telerobotic Servicer (FTS, a remote-controlled repair drone), for now the cubby is empty.


This contains up to six 1.1 meter diameter spherical tanks. These will be filled with whatever is needed to fill the empty tanks of the satellite being serivced, be it hydrazine, water, liquid helium, or whatever. Presumably future satellites will be equipped with fluid transfer connections that are established when the OTV docks. Older satellites will need the poor astronaut to go EVA, grab an umbilical from the FRS and manually fill up the satellite's empty tanks.


This is a satellite workstation, i.e., a place to tie down the blasted satellite so you can do repairs without it floating all over the place. The FSS is located along the ship's long axis so the center of gravity stays on the center. The FSS will have aids like astronaut foot restraints, propellant resupply umbilicals, power cables, and jacks for component diagnosis, testing, and checkout. The base of the FSS is a rack to store replacement satellite modules, with power feeds to keep the modules alive.


Module is 2.4 meters long by 4.5 meters in diameter. It houses all the command and control modules as well as spacesuits and other necessary equipment for EVA operations. Including the airlock. It is designed so that there is enough room in the main compartment for two astronauts to don spacesuits and allow both to enter the airlock.

The avionics and reaction control system (RCS) are more precise than most spacecraft, since they have to locate, track, and dock with relatively tiny satellites. They are described in excruciating detail in the report, but I won't bother to repeat it here. The equipment is 1988 vintage, which is laughably obsolete by now. Even if it is space-rated.


Module is 9.1 meters long by 4.5 meters in diameter. The interior consists of 21 service modules each one meter wide. These are installed on the four walls, leaving a 2.1 meter square opening in the center for the crew. The crew quarters are clustered at one end, they are wider than the service modules. Therefore the 2.1 m square center contracts to a 1.5 meter square hallway leading to the command module.

Each crew quarter displaces one and a half service bays and encloses an area of 4.3 cubic meters. The sleep restraint and personal use console are oriented parallel to exploit the free-fall environment. There is also a small window for recreational viewing. There are four rooms: three are crew quarters, one is designated as a "safe haven."

The wardroom provides space for a multi-use table and allows a large viewing window in the sidewall. The entire crew can occupy the wardroom simultaneously for eating or conferencing. It can also be used by off-duty crew for conversation or recreation.

The Environmental Control And Life Support System (ECLSS) supplies 54 kg of water per day, 18 kg per crew. Of that 18 kg per crew 6.8 kg is for drinking and food preparation while 11.1 kg for personal hygiene and wash water.

A treadmill and bicycle ergometer are provided as exercise machines to combat calcium bone loss and muscle tone depletion.

The personal hygiene facility provides privacy, contains accidental spills and controls odor. It has facilities for shaving, oral hygiene, hand/partial body washing, and a backup urinal for use in the event that the waste management compartment is occupied. It does NOT have a shower. That takes up too much room and has problems containing all the free-floating globules of water.

The waste management compartment is the toilet. It is much like the system that was used on the Space Shuttle.

The galley contains equipment for frozen, refrigerated and ambient food storage. Meal preparation subsystems include microwave/convection ovens, hot and cold water dispensers, utensil storage and pull-out counters. Clean-up and housekeeping is supported by inclusion of a trash compactor and stowage, and a convenient hand washer.

The hull of the module has a 5g/cm2 aluminum radiation shield. Dosage from the Van Allen radiation belts is estimated at 0.35 Sieverts. In case of a solar proton storm the spacecraft will aim its radiation shadow shield (atop the nuclear engine) to face the sun in lieu of equipping the ship with a full blown storm cellar. This will not provide as much radiation protection as a storm cellar, but the design cannot afford the savage reduction in payload mass.

On board power is supplied by a pair of H2-O2 fuel cells. They provide 12 kW at 2.78 VDC normally, and 16 kW at 26.5 VDC under emergency conditions. The fuel supply is 354 kg of liquid oxygen and 42 kg of liquid hydrogen. The pair of cells also create 44 kg of water per day. Solar power was too large and required constant panel adjustemnt. Nuclear power was too dangerous. Primary batteries had too low an alpha, batteries with enough watts would weigh too much. So fuel cells were chosen.

Avionics1.9 to 2.361 kW
Navigation0.8 kW
Crew Systems2.7 to 2.75 kW
Docking Equip.2.2 kW
Waldo Arm3.75 kW
TOTAL15.35 to 15.861 kW


Solid-core nuclear thermal rocket with a specific impulse of 880 seconds (exhaust velocity of 8,600 m/s) and 134,000 Newtons of thrust. Engine is 4.3 meters long with a diameter of 1.2 meters. Engine mass is 1,814 kg, the shadow shield mass is 3,856 kg.

MSFC NTR Mars Mission

This is from Nuclear Thermal Propulsion Mars Mission Systems Analysis and Requirements Definition (2007), a study by the Marshall Space Flight Center (MSFC) Advanced Concepts Office. The topic of the study was given [1] A Mars Mission and [2] Solid-core nuclear thermal rockets, what range of options where there and how did they compare?

The options depended upon a few key design decisions:

MARS STAY: Short-stay or Long-stay. Short-stay means a 30 to 70 days stay at Mars, and the total mission takes 600 days. Long-stay means a 550 day stay at Mars and the total mission takes 900 days.

ENGINE: All-propulsive or Aerocapture. This boils down to whether you save on propellant by using aerocapture or not. Both cases use solid-core NTR for Trans-Mars Injection (TMI). But for Mars Orbit Injection (MOI) the All Propulsive uses the NTR while the Aerocapture uses an aeroshell heat-shield and a close pass through the Martian atmosphere. In addition the All Propulsive uses NTR for Trans-Earth Injection (TEI) but the Aerocapture has to use chemical rockets due to packaging restrictions within the aeroshell. I think what the report is trying to say is that the NTR engine is to big to fit in the aeroshell, but an auxiliary chemical thruster will.

SPACECRAFT: All-up-mission or Split-mission. All-up means the entire mission is performed by one huge spacecraft. Split-mission uses a fleet of piloted and cargo spacecraft. Generally the uncrewed cargo ships are sent ahead. Only if they all arrive in Mars orbit is the piloted ship sent to join them. All-up and piloted split-missions are round-trip. Cargo split-missions are one-way.

So the design cases the study investigated were:

CASE 1: Short-stay, All-up, All-propulsive

CASE 2: Short-stay, Split-mission, All-propulsive

CASE 3: Short-stay, All-up, Aerobraking

CASE 4: Short-stay, Split-mission, Aerobraking

CASE 5: Long-stay, Split-mission, Piloted: All-propulsive, Cargo: Aerobraking


Obviously missions with crew are round-trip, while uncrewed cargo missions are one-way. So all-up missions and piloted split-mission are round-trip.

Uncrewed cargo split-missions are one-way, sent in advance of the piloted ships. Naturally the piloted ships are not sent until the cargo ships successfully arrive, otherwise what's the point?

All missions depart Terra from a 407 km circular parking orbit and are inserted into a 250 km by 33,793 km elliptical Mars orbit with a period of one Martian day.

When using aerocapture for Mars orbit insertion, the Martian altitude is assumed to be 125 km and maximum allowed arrival speed is 7.350 km/s (hyperbolic excess speed 5.450 km/s).

All missions (including all-propulsive) use aerocapture for Terra return, with a maximum allowed hyperbolic excess speed of 6.813 km/s.

As previously mentioned, the All-propulsive spacecraft use NTR for both TMI and TEI. But the aerobraking spacecraft use NTR for TMI and auxiliary chemical thrusters for TEI.

The report did an analysis and concluded that a 2033 mission start had the lowest Initial Mass In LEO (IMLEO) within a few decades, so that was chosen for the all-up and piloted split-missions. The cargo split-mission is started in 2030 so any failure will give enough advanced warning to abort the 2033 piloted mission.

2033 Piloted Mission Trajectory Data
Terra Departure
Mars Arrival
Earth Arrival
Aerocapture at Mars

2030 Cargo Mission Supporting Piloted Mission Trajectory Data
Terra Departure
Mars Arrival
Aerocapture at Mars

V is hyperbolic excess velocity (km/s). Ventry is atmospheric entry velocity (km/s). ΔV is delta-V, change in velocity require to perform specified maneuver.


Short-stay, All-up, All-propulsive

IMLEO602,000 kg
# Enginesx1
Main Engine
TypeSolid NTR
Thrust Nominal1,100,000 N
Thrust Range1,100,000 to
1,600,000 N
Isp Nom875 s
Isp Rng875 to 900 s
T/W Nom8.35
Engine Life Nom60 min
Engine Life Rng60 to 120 min
# Burns Nom10
# Burns Rng10 to 15
Engine Dia7 m
Engine Length15 m

Case one is for a short-stay at Mars. There is only a single spacecraft carrying everything. The payload is a habitat module with crew, a Crew Exploration Vehicle (CEV) and a Mars lander with its own atmospheric entry aeroshell. The spacecraft uses tumbling pigeon artificial gravity to create at least 0.3 g's. A nuclear thermal rocket engine is used for all maneuvers, including Trans-Mars injection, Mars orbit insertion, and Trans-Earth injection. When the spacecraft approaches Terra at the end of the mission, the crew abandons ship in the CEV and aerobrakes to a splash-down.

The main design drivers of this case was the propellant tanks and the overall vehicle length required to generate the minimum artificial gravity.

To conservatively avoid spin nausea you'll want to spin at 4 rpm and have the vehicle length be around 33.6 meters. If that is too long, you can force the astronaut to train, spin at 6 rpm, and get the vehicle length down to 15 meters. I'm just spitballing but looking at the image, if it is an isometric image, and the transhab is 10 meters tall, the spacecraft is about 106 meters long. Assuming the center of gravity in the center, the spin radius is a luxurious 53 meters. No spin nausea problems there. Even better, the center of gravity is probably closer to the engine because of the payload and the heavy nuclear engine. This means the spin radius for the hab module is even longer.


Short-stay, Split-mission, All-propulsive

IMLEO Piloted376,000 kg
IMLEO Cargo268,000 kg
# Enginesx1
Main Engine
TypeSolid NTR
Thrust Nominal670,000 N
Thrust Range645,000 to
1,600,000 N
Isp Nom875 s
Isp Rng875 to 900 s
T/W Nom7.52
Engine Life Nom60 min
Engine Life Rng60 to 120 min
# Burns Nom10
# Burns Rng10 to 15
Engine Dia7 m
Engine Length15 m

Case two is for a short-stay at Mars. An uncrewed cargo vehicle transports the lander to Mars parking orbit about 2.5 years before the piloted vehicle transports the astronauts. It uses its NTR for MOI.

The piloted vehicle carries only the habitat module, the CEV, and a transfer node for docking. It too uses its NTR for MOI. The propellant tanks are arranged to optimize the center of gravity for tumbling pigeon operations. Upon arrival in Mars orbit it docks with the cargo vessel using the transfer node.


Short-stay, All-up, Aerobraking

IMLEO439,000 kg
# Enginesx1
Main Engine
TypeSolid NTR
Thrust Nominal890,000 N
Thrust Range823,000 to
1,60,000 N
Isp Nom875 s
Isp Rng875 to 900 s
T/W Nom7.98
Engine Life Nom60 min
Engine Life Rng60 to 120 min
# Burns Nom5
# Burns Rng5 to 10
Engine Dia7 m
Engine Length15 m

Case three is for a short-stay at Mars. There is only a single spacecraft carrying everything. The NTR performs the TMI maneuver, the MOI is done by aerocapture, and the TEI maneuver is performed with a chemical stage.

The payload is carried in two separate aeroshells. Shell 1 holds the habitat module, the CEV, the transfer node, and the TEI chemical stage. Shell 2 is integrated with the Mars lander. When the vehicle approaches Mars, both aeroshells abandon the nuclear stage (letting it sail off into the wild black yonder) and both shells separately aerocapture into MOI. They then temporarily dock in Mars orbit using the transfer node, before the lander departs for the Martian surface. The lander uses its integral aeroshell a second time to get to the surface.

While tumbling pigeon gravity is provided on the Mars-bound leg of the mission, it cannot be used on the Terra-bound leg. The tumbling only works with a long spacecraft length. Unfortunately the ship's length is drastically shortened when it jettisons the nuclear stage. The crew will just have to suffer through free fall for the trip home.


Short-stay, Split-mission, Aerobraking

IMLEO Piloted290,000 kg
IMLEO Cargo198,000 kg
# Enginesx1
Main Engine
TypeSolid NTR
Thrust Nominal450,000 N
Thrust Range330,000 to
1,600,000 N
Isp Nom875 s
Isp Rng875 to 900 s
T/W Nom6.59
Engine Life Nom60 min
Engine Life Rng60 to 120 min
# Burns Nom5
# Burns Rng5 to 10
Engine Dia7 m
Engine Length15 m

Case four is for a short-stay at Mars. An uncrewed cargo vehicle transports the lander to Mars parking orbit about 2.5 years before the piloted vehicle transports the astronauts. It performs TMI with its NTR engine. The lander abandons the nuclear stage when approaching Mars and uses its aeroshell for MOI. It then waits patiently for the astronauts to arrive.

The piloted vehicle carries only the habitat module, the CEV, a transfer node for docking, and TEI chemical stage. All are housed in an aeroshell. Exactly like the cargo vehicle it performs TMI with its NTR engine, jettisons the nuclear stage when approaching Mars, and uses the aeroshell for MOI. It docks with the lander using the transfer node, then the explorers travel to the Martian surface.

Like case three, artificial gravity is only available in the Mars-bound leg of the mission.


Long-stay, Split-mission, Piloted: All-propulsive, Cargo: Aerobraking

IMLEO Piloted293,000 kg
IMLEO Cargo154,000 kg
# Enginesx1
Main Engine
TypeSolid NTR
Thrust Nominal330,000 N
Thrust Range200,000 to
450,000 N
Isp Nom875 s
Isp Rng875 to 900 s
T/W Nom6.59
Engine Life Nom60 min
Engine Life Rng60 to 120 min
# Burns Nom5
# Burns Rng5 to 10
Engine Dia7 m
Engine Length15 m

Case five is for a long-stay at Mars. It uses conjunction class trajectories instead of opposition class. Two uncrewed cargo vehicles are used. One delivers a Mars habitat to the surface, the second delivers a Mars lander into orbit.

Mars DRA 5.0 Lander

This section has been moved here.

Atomic Rockets notices

This week's featured addition is BOEING DS-1 SATELLITE INTERCEPTOR


This week's featured addition is ICAN-II FUSION ROCKET

Atomic Rockets

Support Atomic Rockets

Support Atomic Rockets on Patreon