Refer to Reaction Mass Calculations for detailed derivations of the formulae used in these calculations.
Note that all of this document assumes the drives in question can customise the reaction mass exhaust velocity for optimal trip times and reaction mass consumption. This is likely to be the case whenever a starship engine is computer controlled.
If you are happy taking 2 weeks, the power requirement drops to 190 GW, and for a 4 week trip, the required power is only 23.8 GW. If you are willing to take twice as much reaction mass, the power levels drop to about 53% of these values. A 900 tonne ship carrying 200 tonnes of reaction mass gives the 18% reaction mass treated as standard in GURPS Space. Note that these are enormous powers compared to the megawatt range of starship power plants detailed in GURPS Space.
All is not lost, however...
What this means is that if you want to use ion drives to go anywhere at sublight speeds in a reasonable amount of time and not devoting 90% or more of your ship's mass to reaction mass, you need to have gigawatt range power plants available for your ships. This implies reducing the cost, mass, and volume of the power plants described in GURPS Space. To do this in a systematic manner:
Note that at these low per-megawatt costs, it becomes ridiculously easy to have ample power for everything else the ship might need. There is no excuse for not having enough power to supply life support, weapons systems, etc, unless you also increase their power requirements to the gigawatt range - which is absurdly unrealistic.
Generating gigawatts of power uses a lot of fuel. Running a 1 GW total conversion plant continuously for a week will consume 6.73 grams of matter. Similarly, an antimatter/matter annihilation plant will consume a total of 6.73 grams made up of half matter and half antimatter. Since hydrogen to helium fusion only releases 0.753% of the available mass energy, a 1 GW fusion plant will consume 863 grams of hydrogen per week. This assumes 100% efficiency, which can never be achieved in practice.
If we assume about 90% efficiency, a 1 GW fusion plant consumes 1 kg of hydrogen per week. This is a significant rate of fuel consumption, and the GURPS Space statement that fusion plants never need refuelling becomes false. However, this amount of hydrogen is ridiculously cheap. Simple electrolysis, using a comparatively trivial amount of energy, can extract 1 kg of hydrogen (and incidentally 8 kg of oxygen) from 9 kg of water.
At roughly 90% efficiency, an antimatter plant will consume 3.5 grams of antimatter per week, per gigawatt. Generating this antimatter of course requires at least the same amount of energy to be invested in its production. Antimatter production is likely to be restricted to large planetside or orbital industrial facilities.
Generating gigawatts from fission plants will burn through fuel rods like there is no tomorrow, and more primitive energy sources will have no hope. If you need gigawatt power, you are probably restricted to fusion, antimatter, and total conversion plants for shipboard use. Fusion is about as hard-science as you can get for these options. Antimatter power plants are feasible using known physics, but will require enormous amounts of energy to generate the antimatter fuel, making it comparitively very expensive. Total conversion is science-fantasy.
The antimatter thermal rocket engines as listed in Vehicles are high-thrust engines suitable for lifting off from planets. As such, they consume prodigious amounts of reaction mass over fairly short operation times. However, for reaction mass efficiency, it is reasonable to assume that the engine can be tuned (or built) to operate as a high exhaust velocity, low-thrust engine similar to an ion engine. In this case, the fuel consumption levels listed in Vehicles should not be used.
The above calculations of antimatter consumption can be applied, using the required power levels of ion drives for desired trip times and reaction masses. The only difference is the antimatter fuel is mixed with the reaction mass, rather than used to generate huge quantities of intermediate electric power. In many respects this a simpler engine mechanism than an ion drive.
As a low-thrust space drive, an antimatter thermal engine must produce high velocity ionised exhaust, rather than hot steam, so the engine itself must be designed to withstand such operating conditions. Using the antimatter thermal engine size and cost formulae from Vehicles will result in an unrealisticaly small and cheap engine. It would be better to use formulae similar to those for ion drives.
Note that recalculating reaction mass and antimatter consumption for the low-thrust regime results in a vastly smaller rate of reaction mass consumption, but can require large amounts of antimatter to generate the required exhaust energies. This is the trade-off in this design.
Pions are likely to be generated in vast quantities in antimatter reactions, but containing them and channeling them to form an effective exhaust would be difficult. Magnetic fields only affect charged pions, the neutral variety would simply travel in a straight line. Furthermore, positively and negatively charged pions are affected in opposite senses by magnetic fields, so directing both sorts in the one direction is quite a challenge! Un-channeled neutral pions would represent a radiation threat to ship passengers, as they would cause highly energetic secondary particle and gamma ray showers as they passed through the ship.
Assuming these problems can be overcome by TL9 or TL10 technology, the
fuel requirements of an antimatter pion engine can be calculated using the
same antimatter requirements as an antimatter thermal engine. Antimatter
pion engines also require an equal amount of normal matter for the
annihilation reaction, but no reaction mass, since the pions generated
in the annihilation provide reaction material.