Nuclear Pulse Propulsion and Nuclear Salt-Water Rockets
Of the other propulsion methods, these two seem to be the least complex. The other methods require control systems that dive into the operation of the reactor itself, in some cases to an almost intimate and impossible level (for the constraints).
For a fission-fragment rocket, one would want to monitor the quantity of neutrons being released by the fissionable materials consistently and without much error, else they risk an overreaction of neutron bombardment—a meltdown. These kinds of torches also require some electromagnetic channeling, which itself warrants a subsystem of monitors and controls. Should this fail, you may risk damaging the exhaust nozzle of the rocket and perhaps other things. There are also other constraints of the device's operations one would want to consider monitoring and controlling, such as the temperature of the ionized particles (you're working in the range of potentially tens of thousands of degrees Kelvin, a few thousand, minimum), and the various integrities of the components of the device; if its a rotating reactor, then you have moving parts and specific alignments to worry about.
Gas core rockets have the advantage that they can gain much higher temperatures, thus higher specific impulse. With great power comes great responsibility. A control system for this torch would need to monitor the temperatures of the core's nozzle and containment, which are the structural crutches of the design (for their temperature limitations). One would want to monitor the structural integrity of the containment and control a neutron moderator to increase or decrease the frequency of reactions. This design also requires an effective cooling mechanism, either through external radiators or gaseous/liquid coolant passing through the nozzle. For the sake of maintaining structural integrity or maximizing/minimizing specific-impulse, one would want to control the rate at which this mechanism dissipates heat. In addition to that, it is thought that there needs to be an internal absorption of thermal radiation as well, to control the rate of reactions. Tungsten particles would seem to be the popular choice, and the rate at which these particles are pushed through the reactor would need to be controlled for similar reasons as the cooling mechanism, to delay the degeneration of the device's components.
A fusion rocket has the benefit that it cannot meltdown. Much of the concept of a fusion torch is theoretical. There seems to be two predominant types: inertial and magnetic containment.
- For the first, one would need to inject pellets of fuel into the reactor and ignite them with high-energy beams of photons or electrons, take your pick. You would need a reliable method for injecting the pellet fuel and a reliable method for activating the beams at precisely the right time and at the right intensity. Typically, you'd do these things inside of a vacuum chamber, so the state of vacuum would need to be controlled. For a successful fusion reaction, one would need to control many beam-to-beam energy ratios on very short timescales (perhaps microseconds). This would seem impossible for an analog system to handle.
- For the second, you would require a myriad of control mechanisms and monitors for maintaining a magnetic containment field within small error margins. What you are essentially doing is balancing two powerful forces: a magnetic pressure and a plasma pressure. A fusion reactor is constantly working to overcome turbulence created by high-velocity, flowing ions as they are carried and conducted by magnetic field lines. You have to concern yourself about quantities of the fuel escaping the magnetic containment and interacting with the reaction chamber's walls. All in real-time, and without electronic computers. As for the control surfaces required, in the operation of a fusion reactor, you would need to be a helicopter parent to it, controlling, monitoring, and real-time analyzing just about every aspect of its operation.
Nuclear salt-water rockets look promising. The amount of fuel you feed one of these and the degree of temperature it gets to are almost of no concern. One manner of control would be the fuel-feeding system. The particles constituting the nuclear fuel are said not to diffuse all at the same, or even at a regular, velocity. They occupy a broad range of many orders of magnitude, and so it is possible that free-moving particles of the fuel may backtrack into the fuel-feeding system, potentially damaging or destroying the system. Of course, you'd want to have control of the fuel injection rate anyway, to adjust acceleration and whatnot, but this is another reason to have that control. If a detector detects a large enough quantity of the fuel being accelerated back into the system by its own diffusion rate, then you may want to close the system for a short time and let the disturbance pass. This seems to be the only control surface for this device.
Nuclear pulse propulsion has been experimented with in the past (see Project Orion) and is fairly well-understood. What you are essentially doing is dropping small (or large, if you're looking for a good time) nuclear explosives behind your ship and detonating them at a distance where it is both "safe" and the explosion will impart a good fraction of its momentum onto your craft. You'd need a control to drop the bombs and a control to detonate the bombs. The intermittent stuff is up to your analog computers to figure out. Dropping should be simple. While your craft is experiencing imparted momentum from the previous blast, release another bomb so that your vessel accelerates away from it. My personal recommendation is that you don't put the actual timer (detonation control) on the bomb itself. Let the bomb be activated by some external force, like a focused laser beam vaporizing some part of its surface, a "fuse," or a hypervelocity projectile smacking into it, fired from the ship, a gun perhaps. Why? Safety reasons, of course. Not because it'd be cool to shoot at nuclear bombs all day.