These are spacecraft designs using fusion propulsion.

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.

Asteroid Mining Crew Transport

Asteroid Mining
Crew Transport
ΔV51,000 m/s
Exhaust Velocity100,000 m/s
Specific Impulse10,000 s
Propellant Mass Flow0.96 kg/s
Thrust96,000 N
PropulsionD-D Fusion
Propulsion Bus2,000,000 kg
Propellant mass4,000,000 kg
Payload Section
1250 person habitat
+1250 persons (85,000 kg)
+ consumables
+ priority cargo (150,000 kg)
4,000,000 kg
Wet Mass10,000,000 kg
Dry mass6,000,000 kg
Mass Ratio1.67
Single Fusion powerplant6 GW
Number powerplantsx2
Total power12 GW
Thrust power4.8 GW
Waste Heat2.8 GW
Waste Neutrons4.4 GW
Length480 m
Diameter400 m
Centrifuge radius150 m
Radiator surface area19,200 m2

This little gem is from Aerospace Projects Review Blog. Scott Lowther discovered it in a 1981 Boeing report Controlled Ecological Life Support System: Transportation Analysis.

The spacecraft was a fusion powered rocket designed to transport miners to the asteroid belt. 1,250 miners per trip. And a cargo of 150 metric tons.

In case it is not clear, the nose of the ship is in the upper left corner, with the tanks on girders. The tail of the ship is in the lower right corner, with the engine.

In the first picture, note the fly-like object in the upper left corner. That looks as if it was supposed to be a 37 meter long Space Shuttle. The pods on the ends of the centrifuge arms may or may not be shuttle external tanks, 47 meters long and 9 meters in diameter.

This thing is freaking ginourmous.

I made a quick and dirty 3D model in Blender and scaled it to with the assumption that the fly in the corner is indeed a 37 meter Space Shuttle. Hard to do since the drawing of the "shuttle" is just a smudge. Reading the dimensions off my model it says the ship is roughly 480 meters long, has a diameter of 400 meters at the tips of the heat radiators, and the centrifuge has a 150 meter radius. Treet these numbers with grave suspicion, they could be off by an order of magnitude plus or minus.

The report Scott found only mentioned the ship in passing, he has filed a FOIA request for another report that might go into more details. I hope so, it would be nice to have some solid figures to work with instead of all this conjecture and assuming.

For the Terra-asteroid run, the vehicle would boost for 11 days, coast for 226 days and brake for 13 days to rendezvous. Adam Crowl calculates if the jet-power is 4.8 GW and the mass-ratio is 5/3 for a return to Earth mission, then an exhaust velocity of ~100 km/s and a total delta-vee of 51 km/s. That means a mass-flow rate of 0.96 kg/s.

There are absolutely huge heat radiators because the engine has to get rid of 2.8 freaking Gigawatts of waste heat.

The heat radiators are triangular, so that they can stay inside the shadow cast by the anti-radiation shadow shield. This is for three reasons:

  1. parts of the radiator extending out of the shadow could scatter deadly radiation onto the passengers
  2. parts of the radiator extending out of the shadow will suffer neutron activation
  3. parts of the radiator extending out of the shadow will suffer neutron embrittlement.

Dr. Luke Campbell points out that the engine is going to need one heck of a shadow shield, because distance attenuation ain't gonna do diddly-squat. Not against 4.4 gigawatts of neutron radiation it isn't. At a measly distance of 480 meters the 4.4 GW of neutron radiation will still be strong enough to give everybody on board the ship a lethal dose of radiation in about 1/5 of a second.

(Doing my own calculation, assuming a person with a body mass of 68 kilograms, a cross section of 0.445 m2, who is 480 meters away from 4.4 GW of neutrons, I figure that with no shadow shield they will be exposed to about 10 grays per second, or a LD35 lethal dose of 2 grays in 1/5 second)

He goes on to say that even with a perfect shadow shield enough neutron radiation would scatter around it if a heat radiator, another spacecraft, space station, or asteroid was outside of the shadow and close to the ship. This can be dangerous over the 12 or so days of continuous thrust (the 226 days between thrust events would probably allow any acute radiation injury to heal — but the chronic stuff will accumulate from different burns).

Dr. Campbell goes on to say that a good shadow shield would probably have an interaction length on the order of a few centimeters at the 2.2 MeV energy of D-D fusion, so a few meters of neutron shielding would reduce the dose by something like 40 orders of magnitude. By way of comparison, the shadow shield on an old NERVA nuclear rocket was only about 0.25 meters thick.

Assuming that the centrifuge arm is indeed 150 meters long, it can spin at a safe no-nausea 2.5 RPM and produce a full gravity of acceleration.

If those pods are indeed Space Shuttle external tanks, they will have an internal volume of a bit more than 2050 cubic meters (the sum of the LH2 and LOX tank). If all of that is passenger volume (meaning the consumables, cargo, life support systems, and everything else is at the ship's spine), then each of the 1,250 passengers will have about 14.8 m3 to call their own for the 8.3 month journey. This is less than the 17 m3 NASA figures the crew needs for missions longer than six months or so. On the other hand, NASA is talking about crew members, not passengers. They are not actually running the ship, so as long as they don't actually start foaming at the mouth and go berserk, 14.8 m3 is probably good enough. Spartan but managable.

Let's fly further into unsubstantiated fantasy, piling shaky assumptions upon shaky assumptions. Do not take these next figures seriously.

Assuming my hasty 3D model based on a crude sketch is anywhere near accurate, my modeling package says that one of the triangular radiators has a surface area of about 6,400 square meters (counting both sides). This means the entire radiator array has a total radiating surface area of about 19,200 square meters.

This has to cope with the 2.8 gigawatts of heat.

Say that the heat radiators are titanium-potassium heat pipes. These have a specific area heat of 150.22 kWth/m2, so to handle 2.8 GW it will take about 18,640 m2. This is less than the model's radiator area of 19,200 m2 so we are in good shape.

Ti/K heat pipes have a specific area mass of 100.14 kg/m2 so 18,640 m2 worth would have a mass of 1,866,600 kg. This is less than the propulsion bus mass of 2,000,000 kg so we are in good shape.

This means the propulsion bus has 133,400 kg left for the engine and structure. Sounds reasonable to me. But again all of this is fantasy, done for amusement value.

In the data block, the figures in black are from the report, the yellow figures are deduced. Treat the yellow figures with some skepticism.

4.5 Asteroid Base

The mission assumes an asteroid mining operation with a 5000 person habitat. The complex transportation scenario for this advanced mission involves four different vehicles and three separate space bases (refs. 86 and 91).

c. The GEO base serves as the final assembly area for the large fusion rocket system used to propel payloads out to the asteroids. Cargo and propellant are unloaded from electric-powered transfer vehicles sent up from the LEO base. The enlarged OTV used to transfer personnel and priority cargo is designed to transport 441,000 lb (200,000 kg) from LEO to GEO. The complex fusion propulsion system is assembled at the base with the fusion power core, propellant tanks, large thermal radiators, and the personnnel and priority cargo modules. The resulting vehicle, shown in figure 4-11, can transport 1250 passengers and 150 metric tons of priority cargo to the asteroids.

The gross start mass for the resupply mission would be 10,000 metric tons, of which power plant comprises 2000 tons; hydrogen propellant, 4000 tons; and payload, 4000 tons (1250-person habitat plus consumables and priority cargo). The power plant consists of two 6 GW fusion reactors utilizing the deuterium-deuterium fusion reaction. The total power plan provides 4.8 GW of thrust power while radiating almost 2.8 GW of waste heat and 4.4 GW of high energy neutrons.

d. There are two methods the fusion rocket will use to propel vehicles to the asteroid base: fast transfer for personnel and priority cargo, and slow transfer for nonpriority cargo. The manned resupply mission is a fast hyperbolic transfer orbit consisting of an 11-day thrust period to achieve hyperbolic velocity, followed by a 226-day coasting, and a 13-day deceleration to match velocity with the asteroid base. The return mission leaves the asteroid approximately 113 days later for a reverse of the ascent mission.

The second method is used to accelerate unmanned cargo pods on a slow elliptical (Hohmann) transfer orbit out to the asteroid base. Figure 4-12 illustrates the different trajectories. The slower trip takes 130 days longer but costs less than half of what the fast, hyperbolic trip costs. All nonpriority cargo is brought to the asteroid facility in this manner. Empty cargo pods are not returned to Earth, they may be discarded or used in a variety of ways as storage modules or Closed Ecological Life Support System (CELSS) modules.

e. A fleet of two fusion rockets is envisioned. They each make one round trip per asteroid orbit (synodic cycle) to the asteroid mining facility and leave a few days apart. Because of the synodic cycle, the fusion rocket vehicles are delayed at the asteroid base for approximately 113 days, at the GEO location they are delayed approximately 288 days. During these delays the fusion rockets are used to decelerate unmanned cargo pods at the asteroid base and to accelerate the pods at GEO. Cargo pod launches are timed to arrive at the asteroid base shortly after the manned resupply vehicles so that the fusion rockets can decelerate the cargo pods. The rendezvous opportunity (synodic cycle) repeats itself every 928 days. This transportation system allows half of the total crew to be rotated each cycle.

86. Advanced Propulsion Systems Concept for Orbital Transfer. 1981. NAS8-33935.

91. Technology Requirements for Future Earth-to-Geosynchronous Orbit Transportation Systems. June 1980. NAS1-15301.

Discovery II

RocketCat sez

This should happen more often. A team of rocket scientist at the Glenn Research Center were inspired by the Discovery from the movie 2001. So they designed one with modern technology that would actually work!

Discovery II
ΔV223,000 m/s
Specific Power3.5 kW/kg (3,540 W/kg)
Thrust Power3.1 gigaWatts
Specific Impulse35,435 s
Exhaust Velocity347,000 m/s
Wet Mass1,690,000 kg
Dry Mass883,000 kg
Mass Ratio1.9
Mass Flow0.080 kg/s
Thrust18,000 newtons
Initial Acceleration1.68 milli-g
Payload172,000 kg
Length240 m
Diameter60 m wide

This design for a fusion propulsion spacecraft is from the NASA report TM-2005-213559 by Craig H. Williams, Leonard A. Dudzinski, Stanley K. Borowski, and Albert J. Juhasz of the Glenn Research Center (2005). The goal was to produce a modern design for the spacecraft Discovery from the movie 2001 A Space Odyssey. The report has all sorts of interesting details about where the movie spacecraft design was correct, and the spots where things were altered in the name of cinematography. The movie ship had no heat radiators, and the diameter of the centrifuge was too small. Arthur C. Clark was well aware of this, but was overruled by the movie people.

Ehricke Fusion Ship

Ehricke Fusion
Fusion beta0.8 to 0.9
Helium 3
4,590 sec to
45,900 sec
45,000 m/s
to 450,000 m/s
Thrust4,450 N to
445 N
Jet Power100 MW
Engine Mass15,000 kg
> 100,000 kg
255,000 kg
Total Mass455,000 kg
Mass Ratio2.3
Length92 m

This is from "Solar Transportation" by Krafft Ehricke, collected in Space Age in Fiscal Year 2001, AAS Science and Technology series, Vol. 10 (1966).

The spacecraft was designed for missions in the 24,000 m/s to 37,000 m/s delta V range. It uses what the report calls a "controlled thermonuclear reactor" (CTR) using Deuterium/Helium 3 (D-3He) fuel (D 40%, 3He 60%). From the description it is a linear magnetic confinement engine. Note how cryogenic radiators are triangular to keep them inside the radiation free shadow cast by the shadow shield.

Note the Thrust Augmentor at the end of the engine. This allows the engine to shift gears, trading off trade thrust for exhaust velocity (specific impulse) and vice versa. It is used for similar reasons that one would downshift with an automobile. In the case of the rocket if it is currently deep in a gravity well, one can increase thrust so you can get out of the well quick and stop paying its per-second gravity tax. The cost is a lower exhaust velocity (which means bad gas mileage) but in this case it is worth it. The Thrust Augmentor shifts gears the standard way, injecting cold hydrogen propellant into the hot exhaust.

Shifting Gears
Jet Power

Remember that the jet power is equal to the exhaust velocity times thrust, divided by two. When the drive changes gears the thrust and velocity change, but the jet power stays the same (100,000,000 watts or 100 megawatts). You can use this equation to calculate other values for thrust and exhaust velocity which are not on the table.

The values for acceleration assume a fully fueled spacecraft with a mass of 455,000 kg.

mDot is propellant mass flow in kg/sec, it is equal to thrust divided by exhaust velocity. Multiply mDot by the duration of the engine burn (in seconds) to calculate how much propellant was expended.

The report gave an example of shifting gears.

Say the spacecraft starts deep in Terra's gravity well in NEO (in a 1.5 hour orbit). If it used high gear (exhaust velocity 450,000 m/s) it would have a pathetic thrust of 445 Newtons, and would take forever to move out of NEO with a miniscule acceleration of 0.0000997 g. Instead it down-shifts to low gear (exhaust velocity 45,000 m/s) for a brawny thrust of 4,450 N. The relatively huge acceleration of 0.001 g will have the spacecraft out of the gravity well in 26.5 hours flat ("out of the well" defined as local parabolic velocity at 28 Earth radii distance). The drawback is it will burn 95 metric tons of propellant but you can't have everything (otherwise you'd have a mythical torchship in your hot little hands). At this point your ship is 95 tons lighter (new mass of 360,000 kg) so the instantaneous acceleration is up to 0.00127 g.

Since the ship is now basically in Sol's heliocentric gravitational field (0.0006 g) it can safely upshift into high gear fuel economy mode. Thrust of 890 N, exhaust velocity of 225,000 m/s. This will accelerate the spacecraft to solar parabolic velocity (42 km/s) in about 60 days, burning about 8 metric tons of propellant and moving the spacecraft a distance of 1.1 AU.

If you now set the engine to "idle" and just coast, you will cross the orbit of Mars in 60 days flat, and the orbit of Jupiter in 300 days. This is just for illustration, you will zip by both planets at 42 km/s. If you actually wanted to be captured into orbit you'd have to do a braking maneuver.

Given a requirement of 37,000 m/s delta V and heavy use of low gear (exhaust velocity 45,000 m/s) I calculate a mass ratio of about 2.3, if my slide rule isn't lying to me. The report didn't go into such details.

The spacecraft also has four of those adorable space taxis.

Exacting Class Starfighter

Exacting Class Starfighter
ΔV7,000,000 m/s
Specific Power450 MW/kg
(450,000,000 W/kg)
Thrust power9 terawatts
PropulsionICF Fusion
Thrust3,000,000 newtons
Exhaust velocity6,000,000 m/s
Dry Mass20 metric tons
Wet Mass40 to 65 metric tons
depending upon fuel
Length60 meters
(Whipple shield)
5 meters
(Internal hull)
4 meters
Heat radiator
30 meters
Heat radiator
5 meters
Power plant50 MW Brayton-cycle
w/argon working fluid
ArmamentUV laser (3 turrets)
Spinal coilgun (2)
Exhaust plume

This is a design by Artist Zach Hajj (a.k.a. Zerraspace), which I found astonishingly good. Personally I cannot find anything scientifically inaccurate with it. The artist mentioned that he used this website as a resource, and I'd say he did his homework.

The structural components of the spacecraft are composed of high-emissivity graffold (folded graphene) scaffolding. The skin is armored with low absorptivity + high emissivity alloy for anti-laser armor, and a Whipple shield to defend against kinetic attacks. For sensors it has frontal and rear IR batteries and several antennae incorporated into the skin.

This is a departure from my regular work towards something more speculative; a true “spacefighter”, a small vessel capable of operating both in space and in an atmosphere. Each one of these is an enormously demanding task on its own, hence the hybrid craft must make a number of compromises to fully operate.

A military vessel first and foremost, the starfighter is not a comfortable ride. Variable thrust and gravity will send the pilot rocking, which the gyroscopic cockpit can only do so much to accommodate, and it has no life support, so he must remain in his space suit at all times. The craft is not meant to hold him for more than a few hours — ideally it is only operated from carriers or planetary bases during skirmishes. That being said, the frontal module can be ejected in case of an emergency, at which solar panels unfurl to provide enough power to operate antenna and coordinate a rescue mission. The starfighter’s minimal design lends to easy conversion to a drone or smart-ship, as this only requires putting a decent computer in place of the cockpit.

In battle, the starfighter serves chiefly as an interceptor or assault craft. As an interceptor, it shoots down incoming missiles and directs fire away from more vital ships. When on the attack, the onboard lasers might not be powerful enough to do significant damage to larger ships, but equipped nukes allow it take down a limited number of opponents of any size. Some militaries even prefer them to capital ships, as many can be built for the same cost as a larger vessel and each loss is less of a hit to the fleet, yet in their numbers they are harder to take down.

Firstly, I set up the ship so as to handle a continuous 24 hours of 5 G acceleration with a mass ratio of 2 when exhaust velocity is 2% c (maximum delta-v 1.5% c), which is probably well beyond how it'll generally operate. I figured it could moderate exhaust velocity by only partially igniting the fuel, letting it get that incredible thrust when needed. If helium-3 fuel is used at maximum exhaust velocity, delta-v is tripled, but then power limits acceleration to 1.5 G at most.

Second, the radiators are deliberately shaped to form a delta-wing when fully unfolded, as NASA pages I found on the matter suggested this was best suited for supersonic flight, and the flaps and slats on the side are used to increase lift at lower speeds where this configuration doesn't work so well. The radiators are also used as an aerobrake while landing, and directly dispose of much of the heat built up.

Lastly, the length required for the coilguns made me think they would have to be stationary mounts, making them somewhat less useful than the ship's missile and laser retinues, so they'd likely more often be used for accelerating missiles than as weapons in their own right. To help compensate, they're open on both ends. When the coilguns are to fire backwards, the projectile is moved to the front end, then accelerated back along the full length, letting it be used both ways.

From Exacting Class Starfighter by Zerraspace

Firefly Starship

Firefly Starship
2013 design
ΔV2.698×107 m/s
Wet Mass17,800 metric tons
Dry Mass2,365 metric tons
Mass Ratio7.526
Payload150 metric tons
PropulsionZ-Pinch DD Fusion
Exhaust Velocity1.289×107 m/s
Thrust1.9×106 N
Acceleration0.11 m/s
(0.01 g)
Accel time4 years
Coast time93 years
Decel time1 years
Firefly Starship
2014 design
ΔV2.998×107 m/s
Wet Mass45,000 metric tons
Dry Mass3,000 metric tons
Mass Ratio15.0
Payload150 metric tons
PropulsionZ-Pinch DD Fusion
Specific Impulseone million seconds
Thrust855,000 N
Acceleration0.019 m/s
(0.002 g)
Accel time25 years
Coast time70 years
Decel time5 years
Length~1,0000 m

Icarus Interstellar has a project to design a fusion-rocket based interstellar spacecraft. They call it "Firefly". The technical lead director is Robert Freeland.

Most of the other Icarus fusion designs use inertial confinement fusion. That's because IC fusion is easier to get halfway worthwhile power levels. Magnetic confinement fusion would be nicer but once you get enough nuclear fusion going to to be worthwhile, the magnetic bubble pops like a cheap balloon.

The drawback to IC fusion is that the confinement time is pathetic. The longer you confine the fusion reaction, the more of the fusion fuel actually burns and generates energy. But in IC fusion the first bit of fusion acts to blast the pellet apart, scattering the un-burnt fuel to the four winds.

Back in the olden days of fusion research, the darling was Z-Pinch fusion. You send a bolt of electricity (about 5 mega-amps) down the center of a long tube full of ionized plasma, creating magnetic field which compresses the plasma enough to ignite nuclear fusion. One of the big advantages with Z-Pinch was that the confinement time (and net energy output from the burn) can be increased by simply making the reaction chamber longer.

Unfortunatley, the disadvantage is that Z-Pinch fusion suffers from several hydrodynamic instabilities which disrupt the plasma. So researchers stopped working on it in.

But in 1998 Dr. Uri Shumlak discovered you could eliminate the instabilities if you made the plasma move at high velocities. Based on this work, Z-Pinch was selected for the Icarus design.

The Firefly's long thin tail is the Z-Pinch tube, frantically fusing and radiating x-rays like a supernova. So the starship was given its name for similar reasons as the one on the TV show: it is a flying thing whose tail lights up.

The spacecraft profile is long and skinny, for two reasons:

  • Its cruise velocity is a substantial fraction of the speed of light (4.5c for the 2013 version). This make interstellar dust grains impact with about 9.1×10-4 joules worth of damage, the equivalent of 46,000 cosmic ray photons. You want to reduce the ship's cross section as much as possible to minimize the number of grain impact events.
  • The longer the ship is, the farther the payload can be placed from the deadly radioactive Z-Pinch drive, taking advantage of distance shielding.

Many other starship designs use 3He-D fusion, because all the reaction products are charged particles that can be easily shieldied. The drawback is that 3He is rare, you'd have to harvest the atmosphere of Jupiter for twenty years in order to get enough.

Instead, Firefly uses D-D fusion, since deuterium can be easily found in common seawater. Of course then you have to deal with all the nasty neutrons and x-rays produced by that reaction. Firefly's approach is to forgo the use of massive radiation shields, and instead try to let as much of the radiation escape into space. The Z-Pinch core is almost totally open to space with only a triad of support rails connecting the aft electrode and magnetic nozzle to the rest of the vessel.

Even with that, the waste heat is going to be titanic. That's where the heat radiators come in. Notice how they are the bulk of the ship. Makes the thing look like a garantuan lawn-dart. The radiators use beryllium phase-change technology, and are positioned as close as possible to the heat loads on the tail.

A long conical shield forwards of the reactor core deflects x-rays away from the payload using shallow-angle effects. The electrodes, rails, and other structure near the core are constructed of zirconium carbide (which is capable of surviving the intensely radioactive environment.

The 2014 design had a total length of just under one kilometer, half of which is the fuel tanks. The forward part of the ship uses the old fuel tank in lieu of spine trick in an effort to save on ship mass.

A fission reactor provides secondary power.

Gasdynamic Mirror

There are problems with attempting to confine ionized plasma in a reaction chamber long enough for most of it to undergo nuclear fusion. Either by trying to hold the frantically radiating piece of star-stuff in an impossibly strong magnetic chamber or by bending the chamber into a ring and letting the fusion stuff run around like stock cars on a racetrack. The fusion plasma fights you at every turn.

In the Gasdynamic Mirror propulsion system, they attempt to avoid that by making the reaction chamber a long and skinny tube, so the plasma just travels in a straight line until it is all burnt. The trouble is that the tube has to be really long. High exhaust velocity, remember?

Gasdynamic Fusion Propulsion System for Space Exploration by Terry Kammash and Myoung-Jae Lee (1995) has this useful table:

Sensitivity of Rocket Performance to Total Mass
System 1System 2
ParameterR = 50R = 100R = 50R = 100
Rocket Length (m)31.7515.87213107
Plasma density (cm-3)1×10162×1015
Plasma temperature (keV)1510
Magnetic field (T)11.284.12
Thrust (N)2.52×1043.36×103
Total mass (Mg)1,6111,589400295
Specific impulse (s)1.58×1051.29×105
Specific power (kW/kg)24111114
τRT to Mars (days)106105143124
Mg: mega-grams, another name for tonne or metric ton
R: plasma mirror ratio
τRT: round-trip travel time

Physics Basis For The Gasdynamic Mirror (GDM) Fusion Rocket by T. Kammash and W. Emrich Jr. (1998) had a sample spacecraft:

GDM Propulsion Characteristics
Fuel50-50 detuterium-tritium
Specific impulse (s)1.27×105
Exhaust velocity (m/s)1.25×106
Specific power (kW/kg)13.4
Thrust (N)2.5×103
Thrust power (MW)2.2×103
Payload +
structural mass (Mg)
Engine mass (Mg)101
Dry mass (Mg)380.7
Wet mass (Mg)423
Propellant mass (Mg)42.3
Mass ratio< 1.1
ΔV (m/s)119,000
Initial acceleration (m/s2)0.0059
Mars round trip (days)170
Fuel density (cm-3)1016
Temperature (keV)10
Plasma length
~spacecraft length (m)
Plasma radius (cm)5
Mg: mega-grams, another name for tonne or metric ton
β: ratio of plasma pressure to vacuum magnetic field pressure (how efficient the magnetic field is in confining the plasma so it doesn't leak everywhere)
R: plasma mirror ratio

When Kammash and Emrich calculated round trip mission time to Mars they used a simplistic Brachistochrone trajectory (accel to midpoint - flip - decel to destination), ignoring planet orbital motion and gravity. They chose what they considered to be appropriate values for the efficiencies and masses of the various components to obtain the engine and wet mass, but did not give a detailed breakdown in the report.

Propellant mass was a jaw-dropping less than 10% of wet mass (mass-ratio less than 1.1). Predictably the initial acceleration is a meager 0.0059 m/s2 (about 0.74 snail-accelerations), but what did you expect, a torchship?

Performance Optimization of the Gasdynamic Mirror Propulsion System by T. Kammash and W. Emrich Jr. (1999) had a sample spacecraft (thrust, specific power, and spacecraft length are all about double the above design, wet mass is about triple):

GDM Reference Conditions
Fuel50-50 detuterium-tritium
Specific impulse (s)1.422×105
Exhaust velocity (m/s)1.395×106
Specific power (kW/kg)133
Thrust (N)2.21×104
Thrust power (MW)1.189×104
Dry mass (Mg)1,500
Plasma length
~spacecraft length (m)
Plasma diameter (cm)8
Vacuum β0.95
Plasma density (ion/cm3)2.2×1016
Plasma temperature (keV)10
Gain factor (Q)2.35
Magnetic field
at center (tesla)
Magnetic field
at mirror (tesla)
Fusion power (MW)1.486×104
Injector power (MW)5,064
power (MW)
power (MW)
Neutron power (MW)1.189×104
Neutron wall
load (MW/m2)

GDM Vehicle Weights (Mg)
Thermal Electric Converter55
Direct Electric Converter53
Neutral Beam Injector23
Fuel Cell/Capacitor System35
Tritium Breeding System10
Lithium Shield37
Magnet Cooling System19
Mars Lander60
Total Dry Mass1,500

The fusion plasma is fueled and heated by neutral beam injectors. The injectors are powered from the nuclear fusion via a combination of direct (harvesting charged particles) and thermal (harvesting thermal gradients) energy converters. The direct electric conversion system has an assumed efficiency of 90%, the Brayton cycle thermal electric conversion system has an assumed efficiency of 30%, and the neutral beam injectors have an efficiency of 100%.

The charged particles from the fusion plasma will be split evenly between thrust and direct electric conversion.

The energy required to jump-start the engine will come from a capacitor bank charged by a set of fuel cells. The bank produces a 1,000 MW-sec pulse of electricity to start the fusion reaction. The capacitors have a charge storage density of 36 kJ/kg.

The fuel cells also provide station keeping power whien the fusion engine is shut down. When the fusion engine is started, some power is tapped to recharge the fuel cells.

The fusion engine has a radiation shield composed of liquid lithium. This is also part of the tritium breeder system.

The habitat module is a set of eight modified International Space Station modules, sized to proved long term accommodations for a crew of six. Storm cellar is a room with walls lined with foodstuffs, water, and waste products. The life support system is not fully closed, oxygen/carbon dioxide cycles are closed but foodstuffs are considered to be consumables.

Power budget for habitat module is 100 kW, which is a pitance compared to the megawatts demanded by the fusion engine. When the engine is shut down the 100 kW comes from the fuel cells.

The lander is a resuable launch vehicle having a vertical takeoff, vertical landing design. It includes a small in-situ fuel processing system powered by a nuclear reactor. Perhaps something like Zubrin's NIMF.

Hedrick Fusion Spacecraft

Hedrick Fusion Spacecraft
EngineTandem mirror
Helium 3
Thrust3,678 N to
37,500 N
105 sec to
200 sec
Exhaust Velocity981,000 m/s
to 1,962 m/s
Specific Power
(inc. radiators)
1.2 kWthrust/kg
(833 kg/MW)
Fusion Power1,959 MW
Input Power115 MW
Thrust Power1,500 MW
Thermal Power
(not useable for
plasma thrust)
574 MW
Engine Mass1,250,000 kg
Engine Length113 m
Midplane Outer
1.0 m
Neutron Wall
0.17 MW/m2
Central Cell
magnetic field
6.4 T
Electron Density1.0×1021 m-3
3He to D
density ratio
87 keV
Ion temperature105 keV
Fuel Ion
6 sec
Ion confining
270 kV
ΔV? m/s
Living Modules
4 cm
Diameter4 m
Length7.3 m
3,300 m3
4.2 rpm
80 m
Habitat Ring
×36 Living
320,200 kg
Shadow Shield3,890,000 kg
×4 Mass
1,924,000 kg
×4 Radial Arms178,000 kg
Payload6,692,000 kg
Total Hab
Ring Mass
13,000,000 kg
Mass Schedule
Ring Mass
13,000,000 kg
Engine Mass1,250,000 kg
750,000 kg
Fuel Mass1,730,000 kg
Wet Mass16,730,000 kg
Mass Ratio1.12

This is from the report Mars manned fusion spaceship (1991). It uses a Tandem mirror engine

There are 36 living modules composing the centrifuge ring. Each module is 4 meters in diameter and 7.3 meters long. The hull is an aluminum-lithium alloy 4 centimeters thick to shield from galactic cosmic radiation. So at a rough guess there is about 3,300 cubic meters of pressurized habitable volume.

Module types include airlocks, bathrooms, bedrooms, cafeteria, controls, library, life support, recreation, recycling, research, saferoom (storm cellar in case of solar proton storms I guess), and storage.

There is a pressure-tight spacedoor between each module. It is a damage control device to allow isolating a module in case of hull rupture/depressurization, toxic gases, or fire. Doors will handle 14.6 psi of pressure, low temperature, and will close automatically. To reduce mass each door is a sandwich of an aluminum honeycomb 13.5 cm thick between two sheets of titanium each 0.25 cm thick. The door is 187 cm high by 93.1 cm wide with a total mass of 10.7 kg. Corners are rounded to prevent curling and to press equal force around edge of seal. Doors are on tracks and can be opened/closed by spring, electric motor, or manually. If there is a pressure difference the door cannot be opened. Assuming a minimum 55 kPa atmospheric pressure to prevent suffocation, and a hull puncture the size of an entire door (1.67 m2), the doors have to shut within 0.03 second to keep the two living modules adjacent to the breached modules above 55 kPa, and within 1.14 second to keep the entire habitat ring above 55 kPa.

Not shown is any sort of a landing craft, which presumably would be parked on the ring hub, nose-to-nose. Without a lander, the entire trip is kind of pointless.

The modules are on the rim of a centrifuge 80 meters in diameter rotating at 4.2 rpm to provide an artificial gravity of 0.8g. This provide enough gravity to reduce bone decalcification, and is below the 6 rpm spin nausea limit. This puts the modules under a shear stress of 22 MPa, which the aluminum-lithium allow can easily handle.

The centrifuge ring is supported by four radial supports. Each is 38 meters long, with an out side diameter of 4 meters with a 13.2 centimeter thickness.

As with all centrifuges, astronauts and other objects moving around will unbalance the centrifuge and make it unstable. The four centrifuge radial support arms have movable masses ("mass elevators") which dynamically ensure the centrifuge center of mass stays positioned on the centrifuge center of rotation. Assuming a maximum imbalance of 52.5% to 47.5%, and a radial arm length of 40 meters, each movable mass will need to be 481,000 kilograms. They will be made of cast iron, cylindrical with a radius of 1.8 meters and a length of 6.15 meters. To avoid problems with coriolis acceleration, the movable masses should have a velocity of no higher than 0.1 m/s when they are moving to correct an imbalance.

The tandem mirror fusion reactor is composed of 25 magnetic mirror cells. Each cell has 4 belt radiators for removing waste heat. There are 100 belt radiators total.

The specific power is 833 kg/MWthrust, which is about an order of magnitude worse than the later 3He-D Mirror Cell design (64 kg/MWthrust)

Maximum radiation dosage from the fusion reactor that the astronauts can be safely exposed to was set at 2.5 millirem per hour (0.025 millisievert/hr).

A shadow shield is set adjacent to the living modules. The shield is a ring-shaped steel tank full of boric acid. The shadow shield is 1.75 meters thick along the line of radiation flux, and has a total mass of 3,890,000 kilograms. The tank walls are 5 centimeters thick, so about 5% of the total shield mass is steel tank.

Alternatively the shadow shield can be placed so it encases the long reactor cylinder (a "reactor cover shield"). This would be lighter, but now the shadow shield has to cope with the intense waste heat from the reactor. The shield would be 1.37 meters thick and have a lower mass of 1,970,000 kilograms, plus the mass of the cooling system.

One calculation predicted a Terra-Mars trip would take 178 days at an acceleration of 1.6×10-4 g and a payload fraction of 0.40. But when I look at the report's reference for that statement, I discover that they are quoting a 1964 book by Ernst Stuhlinger (the designer of the Mars Umbrella Ship) called Ion Propulsion for Space Flight. In other words the report writers did not actually calculate the performance parameters of the Tandem Mirror fusion reactor.

Hyde Fusion Rocket

Hyde Fusion Rocket
PropulsionInertial Confinement
Thrust40,000 N
Exhaust Velocity2,650,000 m/s
Thrust Power54 gigawatts
Engine Specific
110 kW/kg
Ignition Rate
100 Hz
Magnetic Nozzle
Engine Mass
Magnetic Nozzle
coil and matrix8.7 metric tons
anti-burst structure8.5 metric tons
coolant coil8.1 metric tons
neutron shield44.4 metric tons
gamma-ray shield56.3 metric tons
lithium coolant27 metric tons
heat radiators13 metric tons
sub total166 metric tons
Driver module
0.520 metric tons
Driver module
0.435 metric tons
Driver module
0.955 metric tons
x200 Driver modules191 metric tons
Driver system
12 metric tons
Driver system
optical system
6 metric tons
Driver system
11 metric tons
Power transmission
5 metric tons
Compulsator12 metric tons
Capacitor banks25 metric tons
Driver system
sub total
262 metric tons
Total428 metric tons
Misc. Mass
Cargo Payload
(VIP Payload)
1,458 metric tons
(50 metric tons)
Cargo Fuel
(VIP Fuel)
650 metric tons
(2,058 metric tons)
Fuel Tank16 metric tons
Thrust Truss20 metric tons
Shadow Shield
17 metric tons
Auxiliary reactor5 metric tons
Total Mass
Wet Mass2,594 metric tons
Cargo Mass Ratio
(VIP Mass Ratio)

This is from A Laser-Fusion Rocket for Interplanetary Propulsion by Roderick A. Hyde (1983). I apologize for any mistakes but the document appears to be scanned from a poor photocopy of a pre-print that was almost unreadable. As you can see from the diagrams below.

Pellets of fusion fuel (with a coating of propellant) are injected into the reaction point at a rate of 100 pellets per second. There they are imploded by the Driver using a 2 megajoule? pulse of laser radiation from a krypton fluoride laser (which is only 6% efficient). The laser pulse is divided into 8 laser beams which are reflected by mirrors to converge at the reaction point from all directions. The laser pulse compresses the pellet, igniting the fusion reaction. Two krypton fluoride lasers will be used at 50 Hz, alternating pulses to make an effective pulse rate of 100 Hz. The Magnetic nozzle directs as much as it can of the exploding pellet's plasma energy into producing rocket thrust, and prevents as much as it can of the plasma energy from frying engine components.

Dr. Hyde estimates that this engine can carry 1,500 metric tons of payload, with an average trip-time of 6 weeks to Mars, 3 months to Jupiter, and 1 year to Pluto.


For the fusion fuel inside the pellet, you want as much of the energy to be in hot plasma as possible. Any neutrons and x-rays produced are wasted energy, since they contribute nothing to the thrust and cause damage to the ship and crew. The report has a long section discussion the relative merits of various fusion fuels, but Dr. Hyde settles on Deuterium-Deuterium fusion. The pellets contain 15 milligrams of deuterium salted with 36 micromoles of tritium.

About 30% to 50% of the deuterium fuel will burn, the rest will be wasted. Deuterium has a specific energy of 345 megajoules per milligram. The engine is designed for 2000 megajoules per pulse, so for deuterium at 40% burnup each pellet will require 15 milligrams of deuterium. The pellet of deuterium will be coated with propellant to increase thrust (increasing the propellant mass flow mdot) so that the total pellet mass is 150 milligrams. Pretty much any element can be used for propellant, Dr. Hyde used spare deuterium. This means that by varying the amount of extra propellant the engine can shift gears, i.e., trade exhaust velocity for thrust and vice versa. The on-board pellet factory can change this on the fly.

1280 megajoules will be in the form of charged particle fusion energy for thrust, about 710 megajoules will be wasted in the form of x-ray and neutron radiation (330 MJ of x-rays, 380 MJ of neutrons).

Magnetic Nozzle

The primary task of the thrust chamber's magnetic nozzle is to convert the exploding plasma into thrust. The secondary task is to generate the power required to energize the lasers in the driver. The tertiary task is to breed tritium for the fusion pellets, since the blasted stuff has an unstable half-life of barely 12 years.

A superconducting coil creates a magnetic field that reflects the exploding plasma. The coil is encased in neutron radiation shields, and has a heat radiator to keep the radiation shields from melting. The radiators are on the part of the coil farthest from the ignition point.

In a latter design described in a document I have as yet failed to lay my hands on, there are two coils instead of one: the S-coil and the B-coil. The S-coil is above the ignition point and the B-coil is below. The S-coil is smaller with a denser magnetic field, to encourage the plasma blast to exit out the weaker B-coil. The coils resemble slices of a cone

Since there is not much that can be done to stop the harmful x-rays and neutrons, the idea is to make the magnetic nozzle as "transparent" as possible. It is an open framework where all the components occupy a small fraction of the solid angle as seen from the fusion pellet explosion (just see the diagram below). You want to make all the components "edge on" to the explosion, which is why the coils look like conic sections. In other words, they are blade shields.

As the exploding plasma expands against the magnetic field of the thrust chamber, the field is moved. Induction coils (next to the nozzle coils) harvest this motion to generate electricity for the driver. 33 megajoules of electricity is generated, and stored in a compulsator flywheel for the laser driver to use for the next laser pulse.

Other engine designs try to turn all the plasma into electricity and use that to run an ion drive or something. This adds penalty mass in the form of the generator, and reduces the power available by the inefficiency of the generator. Dr. Hyde thinks it is better to just use the plasma directly as thrust.

The tritium breeder has to produce a minimum of 36 micromoles of tritium per pellet explosion. Otherwise the tritium supply will be operating at a loss, and will eventually run out. This is done with a tube full of liquid lithium-6 and lithium-7 with a loop near the detonation point. The lithium converts the neutrons from each blast into tritium. The trouble is that with the design of the lithium blanket there was a pathetic breeding ratio of only 54%, which mandates a neutron interception fraction of 0.055 in order to make 36 micromoles of tritium. The bottom line is that the liquid lithium will be heated at a hideous rate of 4.2 gigawatts, needing a huge heat radiator to prevent the rocket from vaporizing.

This means that the tritium breeding places a floor on vehicle heating; pellets should have a little tritium as possible. Dr. Hyde set it at 36 micromoles for reasons too complicated for me to understand.

The lithium (Li) loop tritium breeder is placed on the inner side of the magnetic nozzle coils, so they can soak up the neutrons and partially shield the coils. The loop and the coil are wrapped by a blade-shaped metal skin, which acts as an eddy current shield.

The magnetic nozzle is a high-energy type, unlike the low energy nozzle on a Daedalus starship. This means the magnetic field contains about five times the energy of the exploding pellet (radius of 6.5 meters with a current of 22 MA), where the Daedalus magnetic field is weaker than the pellet. High-energy types are more efficient at converting pellet explosion energy into thrust.

This is an axially-symmetric nozzle which means a jet of hot plasma will escape along the long axis, i.e., right into the ship's backside. A smaller magnet is used to deflect this jet so it misses the ship's derrière

The magnetic nozzle is 65% efficient at converting the exploding plasma into thrust. Coupled with the 1280 megajoules of plasma energy per pellet detonation means that the entire engine converts about 42% of the total pellet energy into thrust.

The coil will have to be a superconductor, if for no other reason because 22 MA of current will vaporize a conventional coil. Dr. Hyde specified a vanadium-gallium (V3Ga) superconductor with a 15.8 Tesla peak coil field at 4.8 K temperature will have a current density of 270 kA cm-2. The coil will be embedded in a matrix of vanadium and aluminum. The coil and matrix will have a radius of 6.5 meters and a mass of 8.7 metric tons.

However, remember that the same charge of magnetic field repel each other. 15.8 freaking Teslas will be doing their darndest to expand the coil (read: make the coil violently explode in all directions). The technical term is "magnetic bursting force." This will be resisted by 8.5 metric tons of structural composite. So the total coil+structure mass is 17.2 metric tons.

Keeping the coil cooled down to 4.8 K when it is being exposed to 2 gigawatts of neutron and x-ray energy is somewhat of a challenge. The lithium loop will remove the heat created by neutron radiation in addition to breeding tritium. Dr. Hyde figures it can handle the 2000 watt heat load.

Even worse, some of the neutrons that enter the coil's lithium hydride (LiH) radiation shield will scatter right back out, thus they can hit the coil from its unprotected rear side. The radiation shield will have to cover one side of the coil that is not in direct line of sight of the detonation point (the side where radiation can be reflected off the payload's radiation shield right back at the coil).

And then there is x-rays and gamma-rays, requiring a lead coating on the radiation shield.

So the lithium loop will require an 8.1 metric ton refrigerator to reject the heat plus 10 tons of liquid lithium, the lithium hydride neutron radiation shield is about 44.4 metric tons, and the gamma-ray radiation shield is about 56.3 metric tons.

Total magnetic nozzle mass: 126 metric tons.

As previously mentioned the magnetic nozzle system has to cope with 4.2 gigawatts of waste heat, from x-rays hitting the lead shield and neutrons hitting the lithium loop. The lithium will be the thermal working fluid to move the heat to the heat radiators (then it will have its tritium harvested). It will be pumped by a 20 megawatt MHD pump utilizing the thrust chamber's magnetic field. At a given time there is 10 metric tons of liquid lithium inside the thrust chamber sopping up neutrons, but the total system has 27 metric tons. This includes the liquid lithium in the long pipes leading to the heat radiators, and inside the radiators themselves.

The heat radiators are an array of 7,800 heat pipes, each 11 meters long and using lithium at 1500 K. The array mass is 40 metric tons.

Driver system

The pellets are imploded by 2 megajoules worth of laser beam applied in 10 nanoseconds. Dr. Hyde considered electron beams, but they are hard to focus on a tiny pellet and also cause nasty bremsstrahlung radiation. Proton beams have no bremsstrahlung, but since the detonation point is going to be ten to twenty meters away the proton beam will bloom due to electrostatic repulsion and be impossible to focus on the pellet.

Dr. Hyde considers free electron lasers (FEL), carbon dioxide (CO2) lasers, neodymium-doped glass (Nd:glass) lasers, and krypton fluoride (KrF) lasers. After a long discussion he figures the krypton fluoride laser is the least bad option.

Due to problems with heat rejection speed, Dr. Hyde decided to go with two laser systems alternatively firing at 50 Hz instead of one laser system firing at 100 Hz.

Each system will require 100 laser amplifiers (see diagram above). Each amplifier has a rotating cylinder with its lasers and a non-rotating heat pipe radiator. There are five lasers in the cylinder, firing at a rate of 10 Hz. Actually the cylinder is more like gas filled tube with five laser "buckets" on the rim. Between pulses the hot laser gas is exchanged with cool gas in the core, and the heat is rejected by liquid sodium heat pipes. The heat pipes radiate at a temperature of 900 K.

A module has a mass of 955 kilograms, of which 520 kilograms is laser and 435 kilograms is heat radiator. 100 modules per laser system and 2 laser systems means 200 modules are needed. Total mass is 191 metric tons.

In addition, the laser systems will need 12 metric tons of connecting trusses, 6 metric tons of optical system to combine and plus-stack the beams, and 11 metric tons for the oscillator system.

The laser driver will require 33 megajoules of energy per pulse. As described above, energy will be harvested from the engine and stored in a compulsator. 33 MJ will be extracted from the compulsator and placed in a capacitor bank, much like a camera strobe.

The power transmission lines connecting the engine and the compulsator have a mass of 5 metric tons, the compulsator has a mass of 12 metric tons, and the capacitor banks mass 25 metric tons.

Miscellaneous Components

There will be an neutron+gamma ray radiation shadow shield located 20 meters from the detonation point to protect the payload region. It will subtend a 3° half-angle, plus thin fins to shadow the thrust chamber heat radiators. Unfortunately the magnetic nozzle coil will scatter some radiation over the edge of the shield. Right behind the shield will be a magnet to deflect the naughty plasma jet.

The fusion pellets have to travel from the rear of the shadow shield to the detonation point during inter-pulse time. This means they have to have a speed of 2 kilometers per second. They will be propelled by a magnetic accelerator or laser ablation. Extreme precision will be required. In practice a pellet might be deliberately delivered slightly off axis from the detonation point in order to do thrust vectoring.

Behind the payload region is the pellet factory. It will take deuterium, tritium from the tritium breeder, and propellant and manufacture them into pellets. Dr. Hyde did not bother designing this but said he doubted it would be massive.

For cargo missions, Dr. Hyde figures the spacecraft will require 650 metric tons of deuterium fuel. If the acceleration is always below 0.1 g then the mass of the fuel tank would be about 16 metric tons.

The engine will be off in between missions and during coasting, so the engine will generate no power. An auxiliary nuclear fission reactor will be provided for housekeeping power and to restart the propulsion system. A 1 megawatt reactor with a mass of 5 metric tons will do.

There will be a truss to transmit thrust from its origin at the magnetic nozzle coil up to the payload. It has a mass of 20 metric tons. It has 8 primary thrust-bearing members. As with the lithium pipes (heat transfer and tritium breeding) the thrust-bearing members will be shadowed by the magnetic nozzle coil shield until the members reach the inner edge of the thrust chamber radiator (50 meters from the detonation point). At that location the members are laterally tied together by a structural ring, then fan out towards the laser driver radiating array. Upon reaching this site, the truss no longer has circular symmetry but is instead biased towards the radiation plane of the rocket. The thrust bearing members are hollow pipes tied together by a lateral truss to avoid buckling. It is rated for a maximum of 5,000 kilonewtons thrust.

Vehicle Performance

The important performance numbers to look at are the ratio of maximum exhaust power to engine mass (power-to-mass ratio) and the exhaust velocity.

The main limit on the power-to-mass ratio is the heat rejection capacity of the laser driver and thrust heat radiators (if you run the engine at a higher rate than the radiators can cope with the ship will melt or vaporize). The secondary limit is the actual mass of the engine.

Exhaust velocity is limited to a maximum of about 2.6×106 m/s due to the energy limits of deuterium fusion. In practice it will be further limited by the energy wasted producing neutron and gamma-rays, inefficiency of magnetic nozzle converting plasma energy into thrust, and most importantly the fraction of the pellet mass used to implode the fuel.

The engine is pretty lousy for interstellar propulsion at least 20% c due to the the exhaust velocity (20% c is greater than 2.6×106 m/s so mass ratio will be ugly).

However the engine will be marvelous for interplanetary travel. In that role the main limit is the power-to-mass ratio. Given a good ratio, the engine can be optimized to increase the thrust a little bit at the expense of the exhaust velocity (shifting gears). As a rule of thumb you want the thrust and wet mass of the spacecraft to be such that it can crank out a minimum of 5 milligees (0.05 m/s2) of acceleration. Otherwise the spacecraft will take years to change orbit. It is worth it to up the thrust enough to allow this even if you are robbing the exhaust velocity.

w = ƒτ * sqrt( (2 * Ek) / mp )

P = (mp * ν * w2) / 2

P = ν * Ek * ƒτ2

P = (F * w) / 2

αp = Pjet / Me

F = mp * ν * w

mDot = mp * ν


w = exhaust velocity (m/s) {2,650,000 m/s} which elsewhere in this site is symbolize by Ve
P = jet power, thrust power (W) {54,100,000,000 W = 54.1 GW} which elsewhere in this site is symbolize by Fp
αp = power-to-mass ratio or specific power (W/kg) {110,000 W/kg = 110 kW/kg}
F = thrust (N) {40,000 N}
ƒτ = efficiency of magnetic nozzle in converting charged particle energy into jet energy {0.65}
Ek = charged particle fusion energy (J) {1,280,000,000 J}
mp = pellet mass (kg) {0.00015 kg}
ν = pellet repetition rate (Hz) {100 Hz}
Me = mass of engine (kg) {486,000 kg}
mDot = propellant mass flow (kg/s)
sqrt(x) = square root of x
x2 = x squared

In the following tables, a "VIP Mission" is one with the shortest possible trip time, but with a microscopic payload. A "Cargo Mission" is one with a longer trip time in exchange for a reasonable cargo. In the cargo mission, given the total starting mass of the spacecraft (2,592 metric tons), 4/16ths is fuel/propellant mass (648 metric tons), 9/16th is payload mass (1,458 metric tons), and 3/16ths is engine mass (486 metric tons). The VIP mission has the same total starting mass and engine mass. The difference is that the payload mass is reduced and the fuel mass is increased.

Over a given mission the exhaust velocity and thrust is varied by changing the pellet mass.

Note that an acceleration of 1 g is 981 cm/s/s. 5 milligees is about 5 cm/s/s

VIP Missions
Table 4
Distance (AU)
Transit Time (dys)9.439.8153.9
Speed Max (km/s)165339667
Acceleration Max
Exhaust Vel
start (km/s)
Exhaust Vel
end (km/s)
Pellet Mass
start (gm)
Pellet Mass
end (gm)
Cargo Missions
Table 5
Distance (AU)
Transit Time (dys)22.293.6362
Speed Max (km/s)70144284
Acceleration Max
Exhaust Vel
start (km/s)
Exhaust Vel
end (km/s)
Pellet Mass
start (gm)
Pellet Mass
end (gm)

The tables above were calculated by with the following equations, whose implications I have not fully digested.

For interplanetary travel, the capability of an Inertial-confinement Fusion Rocket (IFR) is limited more by its power-to-mass ratio, than bi its exhaust velocity. The limitation on exhaust power translates into a bound on the product of exhaust speed "w" and acceleration "a", i.e., on aw. While a large value of w will eventually enable the rocket to reach a high speed, the low a means that doing so takes up a lot of time and distance. When the goal is to travel a given distance "D" as quickly as possible, the optimum technique involves accepting a lower w value in exchange for higher acceleration. This dialing of w can be accomplished in an IFR by placing excess propellant mass outside of the fusion pellet. The extra material lowers the exhaust velocity of the pelet, while increasing its impulse. Pellet-nozzle expansion calculations have been performed for different overall pellet masses, and shown no change in nozzle efficiency.

We have used three models, of increasing sophistication and opacity, to analyze the performance of this IFR. The first is the classic power-limited model, in which gravity and exhaust velocity limits are neglected. This case is easy to solve, and indicates the the pertinent scaling and operational modes. Next the w constraint is included. The resulting zero-gee motion can also be analytically solved, but in less useful form. This solution is then used as the starting point when numerically solving the 2D problem in which solar graviyt is included along with the w and aw constraints.

By neglecting solar gravity and vehicle exhaust velocity limits, we gain a simple insight into the performance capabilities of an IFR. Suppose one wishes to travel a distance "D" in time "T", starting and stopping at rest. The rocket has an initial mass of "M0", of which the powerplanet accounts for a fraction "β" and is characterized by a power-to-mass ratio of "η". The optimum tradeoff between a and w occurs for the time dependent acceleration:

The payload fraction "λ" can be shown to be given by

Note that D, T, and η appear only in the dimensionless parameter α. The trip time is seen to vary with the 2/3 power of distance, and with the inverse cube root of the power-to-mass ratio. There are two interesting operational modes suggested by Eq. 2. The "VIP" mode yields the shortest possible trip time, but a vanishing payload. For this mode:

For economical operation, one is willing to accept a longer trip time in exchange for a large payload fraction λ. The "Cargo" mode results from maximizing the payload throughput λ / T by optimizing over T and the choice of β. This optimum occurs at α = 1/16 and β = 3/16, resulting in a payload fraction λ = 9/16 and a fuel fraction of 1/4 (4/16). For this mode:

In Table 4 we demonstrate the VIP mode capabilities of this rocket, and show Cargo performance in Table 5. For purposes of comparison, the VIP mode numbers assume the same powerplant fraction, β = 3/16, which is optimal for cargo carrying; so the payload of the Cargo mission is swapped for more fuel in the VIP mission. The rocket exhaust power and exhaust velocity are given by:

(ed note: those two equations were already presented above)

For the current design, the nozzle efficiency ƒτ is 0.65, so the power P is 54.1 GW. Using the powerplant of Table 3, we find a power-to-mass of 110 W gm-1. The examples illustrated in Tables 4 and 5 span the range of solar system mission; Martian close approach to show high acceleration capability, Pluto transit to show the opposite extreme, and an average Jupiter mission. The Tables list the distance, trip time, maximum speed, maximum acceleration, the exhaust speed at the beginning and end, and the overall pellet mass at beginning and end.

The potential of this IFR for solar system propulsion is graphically illustrated by the trip times shown in Tables 4 and 5. A quick trip to Mars can be made in 9 days, while even in Cargo mode, Pluto can be reached in a year. In Cargo mode, the rocket can deliver 1,500 metric tons per mission; while the VIP method still permits delivery of ≈ 50 metric ton payloads.

While enlightening, the above analysis is incomplete. The acceleration profile of Eq. 1 requires a zero acceleration and infinite exhaust velocity at the midpoint of the trajectory. Hence, the exhaust speed constraint will be violated during these trajectories. This will certainly occur in the middle of the trips, and for longer missions can occur throughout the journey. When a limitation on w is imosed on the trajectory optimization problem, its solution is no longer as transparent as Eq. 2; but analytic results can still be derived.

(ed note: for more details see the report)

HOPE (Z-Pinch Fusion)

Z-Pinch HOPE ship
Specific Power6.6 kW/kg
(6,553 W/kg)
Thrust Power3.6 gigawatts
PropulsionZ-Pinch fusion
Specific Impulse19,346 s
Exhaust Velocity189,780 m/s
Mass Flow0.2 kg/s
Thrust38,120 N
Dry Mass552,000 kg
Payload150,000 m
Length126 m
Diameter47 m
Mars 90 day mission
Wet Mass635,227 kg
Mass Ratio1.15
Total ΔV27,500 m/s
Mars 30 day mission
Wet Mass887,300 kg
Mass Ratio1.61
Total ΔV93,200 m/s
Jupiter mission
Wet Mass663,100 kg
Mass Ratio1.20
Total ΔV36,100 m/s
550 AU mission
Wet Mass738,400 kg
Mass Ratio1.34
Total ΔV57,200 m/s

Z-Pinch Pulsed Plasma Propulsion Technology Development Final Report. Like all the others, they started with the HOPE study, replaced the HOPE Magnetized Target Fusion engine with their own engine, and compared the two.


ΔV100,000 m/s
Specific Power34 kW/kg
(34,400 W/kg)
Thrust Power11.9 gigawatts
PropulsionAntimatter Catalyzed
Specific Impulse13,500 s
Exhaust Velocity132,000 m/s
Wet Mass707,000 kg
Dry Mass345,000 kg
Mass Ratio2
Mass Flow1.36 kg/s
Thrust180,000 newtons
Initial Acceleration0.255 g
Payload82,000 kg
Length72 m
Diameter190 m wide

This design for an antiproton-catalyzed microfission/fusion propulsion spacecraft is from the University of Pennsylvania.

Fuel pellets have 3.0 grams of nuclear fuel (molar ratio of 9:1 of Deuterium:Uranium 235) coated with a spherical shell of 200 grams of lead. The lead shell is to convert the high energy radiation into a form more suited to be absorbed by the propellant. Each pellet produces 302 gigajoules of energy (about 72 tons of TNT) and are fired off at a rate of 1 Hz (one per second). The pellet explodes when it is struck by a beam containing about 1×1011 antiprotons.

A sector of a spherical shell of 4 meters radius is centered on the pellet detonation point. The shell is the solid propellant, silicon carbide (SiC), ablative propellant. The missing part of the shell constitutes the exhaust nozzle. Each fuel pellet detonation vaporizes 0.8 kilograms of propellant from the interior of the shell, which shoots out the exhaust port at 132,000 meters per second. This produces a thrust of 106,000 newtons.

The Penn State ICAN-II spacecraft was to have an ACMF engine, a delta-V capacity of 100,000 m/s, and a dry mass of 345 metric tons. The delta-V and exhaust velocity implied a mass ratio of 2.05. The dry mass and the mass ratio implied that the silicon carbide propellant shell has a mass of 362 metric tons. The wet mass and the thrust implied an acceleration of 0.15 m/s2 or about 0.015g. It can boost to a velocity of 25 km/sec in about three days. At 0.8 kilograms propellant ablated per fuel pellet, it would require about 453,000 pellets to ablat the entire propellant shell.

It carries 65 nanograms of antiprotons in the storage ring. At about 7×1014 antiprotons per nanogram, and 1×1011 antiprotons needed to ignite one fuel pellet, that's enough to ignite about 453,000 fuel pellets.

Magneto Inertial Fusion Drive Rocket

Fusion Drive Rocket (FDR)
Engine Specs
PropellantLithium foil
PowerD-D Fusion
Delay Between
Fusion Pulses
13.3 sec
Specific Impulse5,000 sec
Exhaust Velocity49,000 m/s
Mass Ratio1.7
ΔV26,000 m/s
Jet Power36 MW
Thrust1,470 N
Propellant mass flow0.03 kg/s
Specific Power
(Jet power/inert mass)
2.4 kW/kg
Fusion Gain200
Fusion Input Power400 kW
Mass Budget
Structure3.4 metric tons
Propellant Tank0.1 metric ton
FRC Formation0.5 metric ton
Propellant Feed1.2 metric tons
Energy Storage
(capacitors 1 J/g)
2.5 metric tons
Liner Driver Coils0.3 metric tons
Switches and Cables1.2 metric tons
Solar Panels
400kW at Mars
@200 W/kg
2.0 metric tons
Thermal Management
(heat radiators)
1.1 metric tons
Nozzle0.2 metric tons
Margin (20%)2.5 metric tons
15 metric tons
(Crew Habitat)
33 metric tons
(Lander Aeroshell)
16 metric tons
(Lander rockets)
14 metric tons
PAYLOAD MASS63 metric tons
78 metric tons
Propellant56 metric tons
134 metric tons
Initial Acceleration0.011 m/s
(0.001 g)
Payload fraction47%
Tankage fraction10%
Terra-Mars Transit82.9 days
Stay Time30 days
Mars-Terra Transit97.1 days

This is from Electromagnetically Driven Fusion Propulsion, J. Slough, Pancotti, A., Kirtley, D., Votroubek, G., International Electric Propulsion Conference, IEPC-372 (2013).

This fusion ship uses John Slough et al's innovative Magneto Inertial Fusion engine. The engine has most of the advanages of both magnetic confinement and inertial confinement fusion with very little of the disadvantages. Instead of trying to crush the fusion fuel with wobbly magnetic fields or a spherical firing squad of lasers, then attempting to heat the propellant by toasting it over the explosion, it tries a more clever method.

Instead it magnetically crushes a foil ring of lithium metal (called a "liner") such that it also crushes a blob of fusion fuel in a mighty fist of metal. Then as the fusion explosion occurs, the lithium becomes propellant, conveniently totally enclosing the blast and doing a fantastic job of converting fusion energy into propellant energy. This also allows the rocket to use open cycle cooling, so it does not need acres of heat radiators like other fusion rockets.

In addition, carrying inert rolls of compact lithium foil propellant is infintely easier than trying to keep huge cryogenic tanks of liquid hydrogen from boiling dry.

Each lithium foil liner has a mass of something between 0.28 to 0.41 kilograms. The crush speed is about 3 kilometers per second, resulting in a solid cylinder of lithium with a fusion explosion going off in its heart.

Using a fusion rocket means the entire mission will take about 210 days, instead of the 540+ days you'd need with a fission rocket. Which means the crew won't max out their career radiation exposure limit in one lousy mission.

For the specified mission a specific impulse of 5,000 seconds is needed. A complicated equation (in the paper) calculates this mandates a fusion gain factor of 200. This means if you want 36 megawatts of jet power coming out of the engine, you will have to feed in 36,000,000 / 200 = 180 kilowatts of electricity. This can be supplied with a low-mass arrangement of a solar photovoltaic array charging a capacitor energy storage bank (instead of being forced to use a weighty nuclear reactor or something). Remember that sunlight is weaker at Mars orbit, you'll need an array rated for 400 kilowatts at Terra orbit in order to eke out 180 kW at Mars.

The crew habitat module appears to be a bog-standard inflatable TransHab module, which is hardly surprising. It can house a crew of six for about 18 months in about 33 metric tons. The rest of the payload is a 30 metric ton Mars Excursion Vehicle. Basically it is using the payload package specified in NASA's Mars Design Reference Architecture 5.0.

And this rocket is totally resuable, unlike those insane designs using fission rocket staging.


Thrust Power118 GW
ΔV80,000 m/s
Chamber Temp100,000,000°C
Exhaust Velocity952,000 m/s
Thrust248,000 N
Propellant Mass
0.26 kg/s
Length1.3 km
Lander Mass
152 metric tons
Food Mass57 metric tons
Oxygen Mass80 metric tons
Engine Mass??
Structral Mass??
Dry Mass>289 metric tons
Propellant Mass<111 metric tons
Wet Mass400 metric tons
Mass Ratio<1.38

The Pegasus is from a BBC science fiction "documentary" called Space Odyssey: Voyage to the Planets. The science was remarkably hard in the show, though details were scanty.

The ship design inspired the Antares for the much more mediocre show Defying Gravity.

The only hard facts about the Pegasus I could find were:

  • The ship was "as long as 12 football fields." British football pitch ×12, length is 1,300 meters
  • The ship mass is 400 metric tons
  • The engine is powered by nuclear fusion
  • Engine chamber temperature is "100 million degrees". Since this is a British show I'm going to presume Celsius, not Fahrenheit.
  • Propellant is liquid hydrogen, heated by the fusion reaction
  • Engine has 158 million horsepower. I calculate that as 118 gigawatts thrust power
  • Ship has an aeroshield on the nose, presumably for aerobraking. Shield is composed of steel, carbon fibre, and beryllium.
  • Ship has a "top speed" of 288,000 km/s. I'm going to presume that means a deltaV of 80,000 m/s
  • Before the mission huge orbital propellant depots were sited around each planet by unmanned ships to increase ship's delta V
  • Crew size of five
  • Mission duration six years
  • Interior volume could hold 10 jumbo jets. Best estimate of interior volume of a 747 (that I could find with a five-minute Google search) is 1,035 cubic meters, external volume close to that. So ship volume 10,350 cubic meters.
  • Consumables are 57 tonnes of food and 80 tonnes of oxygen.
  • Cargo is five landing vehicles (each optimized for a different planet) with combined mass of 152 tonnes. Plus unmanned probes.

There is a centrifuge for artificial gravity. It provides 0.5 g's.

There is a storm cellar to shelter from solar proton storms. In addition, there is an artificial magnetic field to protect against Jupiter's radiation belts.

Thus endeth the canon knowledge.

Now here come my wild conjectures. Using similarly wild assumptions. Feel free to change my assumptions and recalculate.

Since this is a fusion engine that heats hydrogen, it is an Afterburner Fusion Engine.

It is stated that the wet mass is 400 metric tons. We do not know the mass of all the dry mass elements but the masses we are given total to 239 metric tons. The actual dry mass will be bigger.

Mass Ratio is wet mass divided by dry mass. So we know the maximum mass ratio is 1.38, it will be lower than that.

Exhaust velocity is Ve = ΔV / ln[R] where Ve is exhaust velocity, ΔV is delta V, and R is mass ratio. ΔV is 80,000 m/s, R is 1.38 (or less) so Exhaust velocity = 952,000 meters per second.

Thrust is F = (2 * Fp) / Ve where F is thrust, Fp is thrust power, and Ve is exhaust velocity. Ve is 952,000 m/s, I'll assume that Fp is 118,000,000,000 watts (118 gigawatts), so Thrust = 248,000 Newtons.

Acceleration is A = F / Mc where A is acceleration and Mc is ship's current mass. Ship's initial mass is 400,000 kg (400 metric tons) and thrust is 248,000 N so Initial Acceleration = 0.62 m/s or 0.063 g

Propellant mass flow is mDot = F / Ve, so Propellant Mass Flow = 0.26 kilograms per second.


Santarius Fusion Rocket

Tandem Mirror Engine
Helium 3
Thrust3,678 N to
37,500 N
105 sec to
200 sec
981,000 m/s
to 1,962 m/s
Specific Power
(inc. radiators)
1.2 kW/kg
Fusion Power1,959 MW
(2 TW)
Input Power115 MW
Thrust Power1,500 MW
Thermal Power
(not useable for
plasma thrust)
574 MW
Total Mass1,250,000 kg
Total Length113 m
Central Cell
Outer Radius
1.0 m
Neutron Wall
0.17 MW/m2
Central Cell
magnetic field
6.4 T
1.0×1021 m-3
3He to D
density ratio
87 keV
Ion temperature105 keV
Fuel Ion
6 sec
Ion confining
270 kV

This is from Lunar He-3, fusion propulsion, and space development by John Santarius (1992). It uses a Tandem mirror engine.

Dr. Santarius figures that Deuterium-Helium 3 is the best choice for fusion fuel. Deuterium-Deuterium has a lower power density. Deuterium-Tritium reaction emits lots of deadly neutrons and would require more radiation shield mass. Hydrogen-Boron is too difficult to ignite and produces almost all of its power as thermal bremsstrahlung radiation instead of the more desirable fast charged particles. Helium 3-Helium 3 is also far too difficult to ignite.

The fusion reaction chamber is linear, with a magnetic mirror closing each end. The mirror at the exhaust nozzle is weaker, so the star-core hot fusion reaction products shoot out the nozzle (we hope). Dr. Santarius puts it "Thrust is produced by driving one end cell more vigorously to increase axial confinement on that end, thereby unbalancing the end loss of plasma." Which is more precise than what I said, but harder to understand.

Each central cell has a 6.4 Tesla magnet made of a Niobium-Titanium (NbTi) superconductor.

The magnetic mirror end-cell magnets are much stronger. Each mirror has a 12 Tesla Niobium-tin (Nb3Sn) magnet, and a 24 Tesla composite magnet (16 Tesla from a Nb3Sn magnet plus 8 Tesla from a normal conducting copper electromagnet energized with 8 megawatts of power).

The engine can "shift gears" (trade thrust for specific impulse) over an unusually broad range by using three different operating modes. Specific impulse ranges from 105 seconds to 200 seconds, while the thrust-to-weight ratio varies correspondingly from 3×10-4 to 0.03 (thrust of 375 Newtons to 37,500 Newtons).

  • Fuel Plasma Exhaust: The fusion reaction products are also the propellant. This has the highest specific-impulse/exhaust velocity, but the lowest thrust since the propellant mass-flow is so minuscule.
    See Pure Fusion Engines
  • Mass-Augmented Exhaust: a low-field magnetic valve is added and reaction mass is injected into the fusion reaction. This increases the thrust by upping the propellant mass flow, at the cost of cooling the exhaust which lowers the specific impulse.
    See Afterburner Fusion Engines
  • Thermal Exhaust: This uses the fusion reaction's thermal radiation (bremsstrahlung and synchrotron) to indirecty heat a blanket of reaction mass, which becomes the thrust exhaust. This has the lowest specific-impulse but highest thrust, similar to chemical propulsion.
    See Dual-Mode Fusion Engine

Dr. Stuhlinger notes that high-thrust mode allows fast human transport while high-specific-impulse mode allows cargo vessels with large payload ratios. He compares these to sports cars and trucks, respectively.

Mass Budget Terra-Mars mission
9 month each way
mass in metric tons
ChemicalD-3He Fusion
Payload (each way)11,80011,800
Fusion Reactor1,000
D-3He fuel burned0.08
Nonpayload mass orbited47,2003,000

The point being that D-3He fuel is so compact and energetic that the entire fusion spacecraft is 44,200 metric tons lighter than the chemical spacecraft.


Wet Mass6,000,000 kg
Dry Mass1,835,000 kg
Mass Ratio3.27
ΔV200,000 m/s
Thrust2.4 × 105N
Exhaust Velocity170,000 m/s
Thrust Power20.4 gigawatts
Specific Power11.1 kW/kg
D-T Fusion
Width170 m
Height100 m

VISTA is the Vehicle of Interplanetary Space Transport Application, from a study by the Lawrence Livermore National Laboratory. It looks like a tiny flying saucer in the diagrams but it is actually freaking huge. Blasted spacecraft is taller than Godzilla.

Tiny pellets with a deuterium-tritium compount core surrounded by about 50 grams of propellant drop out of the bottom of the cone. At the pellet target position a battery of laser modules zap the pellet with enough energy to initiate a fusion explosion. The propellant blast bounces off the 12-Tesla superconducting magnetic coil to provide thrust. Thrust is throttled by varying the pellet detonation rate from 0 to 30 detonations per second.

With 100 metric tons of payload, VISTA can travel to Mars and back to Terra in six months flat.

Unfortunately about 75% of the fusion energy is wasted, creating no thrust (escaping as neutrons and x-rays). But the remaining 25% is more than powerful enough to give the ship 200 kilometers per second of delta V. The spacecraft is shaped like a cone in an attempt to minimize how much of the wasted energy hits the ship as deadly radiation (only about 4% of the wasted energy irradates the spacecraft).

Atomic Rockets notices

This week's featured addition is Magnetio Inertial Fusion Drive Rocket

This week's featured addition is Thunderstrike Antimatter Rocket

This week's featured addition is the Antimatter Energy

Atomic Rockets

Support Atomic Rockets

Support Atomic Rockets on Patreon