I stand corrected on the matter of who makes Multibodies. Two stages, two contractors, how else would Uncle Sam do it but spread the pork around?
I had the impression that in the early '70s Dark Ages in the wake of the cancellation of the Lunar Apollo program, that the contractors were backing away from NASA biz and so Douglas sold off their rights and blueprints etc cheap to Boeing so Boeing owned the whole thing.
Now I suppose the solids too are made by a third contractor.
----
Another D'Oh moment for me--looking into the whole matter of electric propulsion (trying to get a general sense of which sort of electric propulsion would be flying in the early 2000s and what its general parameters would be) led me to read up on the example of the European
SMART-1 mission, which used SNECMA-built Hall thrusters with xeon as reactant, getting ISP in the ballpark of 1600. The craft massed about 1/3 of a ton, used around 60 or so kg of xeon (or more, I might be picking up just on what was needed to bring it to L-1 and switch over into the Lunar effective gravity well, then it needed more to bring it down to low Lunar orbit)--but took over a year to get to L-1, and then months more to stabilize in LLO. Pretty much as I guessed.
But what I didn't guess was that this epic slow trip started with a boost to a geosynchronous transfer orbit, perigee of 7000 km (implying that its initial launched parking orbit was already unusually high) apogee over 42,000 km. That kick must have been in the ballpark of 2000-2500 m/sec, delivered by chemical rocket, so the propellant for that must nearly have doubled the launch mass, and counting an extra rocket stage to do it, probably more than doubled.
In my naivete all my efforts to estimate the parameters of low-thrust high-ISP electric thrusters did assume starting from LEO parking orbit; I observed that if that were so, a very long portion of the flight would be spent spiraling out in nearly circular widening orbits that would have a radial component that is a slow crawl compared to the orbital speed. The way I looked at it, the tiny acceleration in an essentially tangential direction is multiplied by the radius to get increment of the angular momentum, and the actual osculating orbit that results at any instant would be very close to a circular orbit with the slowly increasing angular momentum. Another way to say this--if the craft has an outward radial component, conservation of angular momentum will slow the tangential component, so the speed outward a given tangential thrust could maintain would be the one where increasing radius slowing the craft down tangentially is compensated by the thrust. Or not quite compensated, because the successive outward circular approximate orbits have decreasing orbital speeds. Thus in the near-Earth region, the radial velocity is negligible compared to the tangential, and even if we seek to maximize rate of increasing orbital energy (by always thrusting in line with the craft's current motion and not at some angle to it) we very closely approximate tangential thrust that applies a steady, and increasing, increment of angular momentum as the craft slowly recedes from Earth.
Now gradually of course the orbital speed is slowing whereas the equilibrium radial speed a given thrust can maintain is rising (since (circular) orbital speed falls with the inverse square root of distance, and angular momentum therefore rises as the 1/2 power of distance, while a given acceleration increments angular momentum in proportion to radius. If we were trying to maximize the rate of increasing orbital energy, we'd therefore be shifting the thruster to angle outward more and more radially--even if we don't do that, the radial component rises and the tangential one falls so that we eventually reach escape velocity with the two speeds being equal and net orbital energy raised to zero--that is the osculating orbit at that point is an escape parabola. So long before then the neat little approximation above would break down. But it gave me a tool to estimate just how long the thrusters have to push to reach a given distance, until the radial speeds start becoming a big fractio of the circular orbital.
But when we start from an elongated transfer to GTO orbit like that, those approximations go out the window. And to be sure, for a solar powered craft starting from LEO they are not much good anyway, because almost half the time a LEO satellite is in shadow and the solar cells are not putting out power. If we were to launch from LEO like that, such a craft would have thrust only on the daylight side, which would include perigee, so basically the impulse could be roughly approximated by a single instantaneous burn equivalent to the time it spends in sunlight times the actual thrust rate. Even summed over 45-50 minutes like that the impulse is tiny so we'd wind up with nearly-circular orbits with perigee staying roughly the same and apogee rising slowly, roughly on the far side of Earth from the Sun. Gradually the craft would be spending more time in sunlight and so its net impulse per orbit would rise--but mostly because the osculating orbital period was rising. Also the slowest part of the orbit would be near the eclipse zone.
Launching into GTO, all of a sudden the craft is in an orbit where it spends most of its time 5 to 7 times as far away from Earth as at perigee, buying longer periods, and less eclipse time as a fraction of the orbital period. SMART-1 was on its way to the Moon and so I suppose its thrust was for most of the time it was on optimized to increase energy, and then later on to increase angular momentum so that its orbit circularized--but looking at illustrations of snapshots of various osculating orbits I can see that it started circularizing early on, that is lowering its eccentricity. With the craft spending most of its time far out on the orbit, where a given thrust in the tangential direction would deliver a bigger increment of angular momentum, whereas at lower orbital speeds (far below the circular orbital speed at those radii) the orbital energy increment was relatively low I can see how this must be so, even if one is trying to maximize the latter rate.
The slow part of a spiral out from LEO is the low part, and while boosting with electric thrust from there would indeed still save a lot of reaction mass, it also takes the ship on a leisurely barge ride through the worst intensities of the Van Allen Belts, particularly the low one that is most intense and that an Apollo-like Lunar trajectory could avoid almost completely since that belt covers a limited range of latitudes. A spiraling path from LEO has it cruising at a snails pace there in the most concentrated part of the belt for months. This prospect alarms me enough when I consider
a considerably higher thrust (but still absolutely low-thrust) fusion pulse drive for manned flight from LEO to Luna. SMART-1's trajectory reminds me that other discussion of electric propulsion I've seen worries about the Van Allen Belt passage issue too--not for any hazard to human astronauts on these very slow passages, but due to damage the radiation there can do to the craft itself, notably to the solar power panels.
One possible solution is to harden the craft to endure the passage--and if we had Dr. Slough's electromagnetically driven fusion pulse rocket ready to hand today, where a 15-20 ton propulsion module could pulse a .4 kg mass of lithium at ISP of 5000 for a 20,000 Newton-second impulse once a minute (maybe once every ten seconds) so that the lunar transfer would be not a year but a few weeks for a craft massing many tens of tons all up, we could similarly harden the craft as we would in any event for solar flares, providing a "storm shelter" shielded by supplies composed mostly of low-atomic-weight elements, thus they too could withstand the Belts despite spending a couple weeks in them. But the alternative, despite the considerable cost, of simply boosting the ship with a high-thrust chemical impulse so it is well on its way, appeals--even though the subsequent spiraling orbits will still hat o many crossings of the belts, at least these will happen at high radial speeds, cutting down exposure time, and the initial boost shaves a considerable amount of incremental boosting time and places the craft where its low-thrust rocket is most efficient. It would seem SMART-1 had a similar option chosen to deal in part with the belts and also to save time, trading off much of the mass savings advantage for these benefits.
So--I was wrongly assuming that Lockheed ITTL was going to use electric thrusters to raise GEO satellites in steady spiraling near-circular orbits from LEO directly. Instead it looks like some kind of hybrid akin to the trajectory SMART-1 followed--despite the mass cost, a chemical boost of some 2500 m/sec initially to put the payload and thruster bus into transfer orbit, then a slow steady circularization of that orbit.
Now looking at SMART-1 and figures for the masses of solar panels versus their power output, it had a single SNECMA PPS-1350-G that massed 5.3 kg, delivered (as used on the mission) up to 68 mN at 1200 watts of power (SNECMA claimed up to 88 at 1500 W) and the Wikipedia page on SMART-1 cites 29 kg for the electrical propulsion system mass (excluding obviously the 80+ kg of Xeon propellant) so that 29 kg may include the solar panels. (It might not since those panels obviously had other functions too). Another Wikipedia page says 300 W/kg "are available" from space-optimized panels, and while this might not reflect the state of the art when SMART-1 was launched it suggests just 4 kg can account for SMART-1's power needs for this thruster. Say it's 36 kg all up for a system that can use 1200 W to put out 72 mN and we have 1/2 kg all up per m/N, which suggests to me a dedicated heavy bus massing 2 tons could put out 4 Newtons, using 60 of these thrusters drawing 72 kW from a quarter-ton solar panel array. In proportion then a 20 ton all up craft, comprising the two ton bus, 5 tons of xeon propellant, and 13 tons of payload, could be moved from LEO to low Lunar orbit after being boosted some 2500 m/sec using an 18 ton hydrogen-oxygen booster in 15 months on the same path as SMART-1.
That does not sound so great but I'd think payloads to GEO would typically be a lot smaller than 13 tons; say only 4 tons, then the dry mass falls to 6 from 15, the orbital delta-V the solar thruster must achieve is--well, that's a bit tricky, but say 2/3 what SMART-1 had to accomplish, so the propellant would be reduced to under 2 tons. Now we have 8 tons all up to boost to GTO with an 8 ton chemical rocket, 16 tons all up in LEO versus 38, and we still have some reserve propellant that might get the solar thruster back to LEO. How long would the trip out take? If the delta-V is 2/3 and the thrust to mass ratio initially 2.5 times greater we have maybe 4 months out. That would imply, thrusting about 60 percent of the time, propellant consumption of 1850 kg, so no we wouldn't have a lot of reserve left after all. To spiral the 2 ton bus back down to LEO--well, we wouldn't want to try to recover it from a GTO orbit so we can't just reverse the path out.
Trying to figure it as a direct boost from LEO using the thrusters all the way is tricky, especially considering the gradually reducing effect the shadow of the Earth has. I'd guess the thrust time and hence propellent consumption would more than double, to 4 tons say, and if we want to get the bus back, we'd need 1.3333 tons of extra propellant, which would raise the outbound propellant need to 5. Say 6.5 tons of propellant, 12.5 all up, it takes 8 months to boost the 6 ton payload-bus combo up to GEO--but then, the bus massing about 1/3 that it will only take a third that time to bring it back, so a bit over a year round trip, and the customer is waiting 8 months from launch for their payload to go on line. But, assuming the bus has already been launched before, just 10.5 tons need to be launched, and Lockheed gets their bus back.
Now notice, the SNECMA early-2000s Hall thruster model only gets 1600 ISP. I gather that the theoretical upper limit of Hall thrusters might be pushed to double that. For the same power input, that cuts the already weak thrust in half, but it also would cut the propellant masses needed down. Say now we only need .75 tons to bring the bus back down, cutting 7.5 outbound tons to 6.75; with doubled mass efficiency we need only 2.25 for the outbound leg, or 9 all up, down to 7 tons launched to LEO, just 3 of which is propellant, 4 being GEO payload. The whole stack is now 6/7 previous mass and .9 at GEO meaning we should shave some time off the outward boost. But wait! At constant power we've cut the thrust in half, so now it takes the better part of two years!
However, not to worry, we can probably up the power--a factor of 4 will restore the original mass flow rate, and now the doubled ISP appears as double instead of half thrust. The price we pay for that is needing to quadruple 72 KW to 288, raising the array from a quarter ton to a full ton, adding 750 kg to the 2 ton bus. Well, that will undo a lot of the mass saving above but anyway we only launch the bus once.
Now we need a full ton of propellant to bring the 2.75 bus back down; we need to arrive at GEO with 7.75 tons altogether; we need 2.6 tons of xeon for the upward leg, or 10.35 tons in LEO, 7.6 were launched for this mission, 4 of payload, 3.6 of propellant.
Allowing for the different masses changing the time profile versus doubled thrust, I figure the whole cycle now takes 5 months, with the payload arriving at GEO well before 4 months have passed.
Turning instead to a chemical alternative, specifying that it too is a 2 ton dry bus that can boost at ISP 440 (typical of Centaur) and must be returned to LEO from a GEO delivery--I'm going to leave out inclination change which implies that LEO is equatorial, hence our launch was from Kourou. I didn't try to account for inclination change above, so I'm skipping it now! But these American launchers are going to want to launch to 30 degrees or more inclination and I think that adds 500 m/sec delta-V requirement at GEO circularization (and sending the bus back down to a 30 degree inclination too). Skipping that 1000 m/sec in the process merrily, if we want to return a 2 ton bus from GEO to LEO, it needs to go through 1500 m/sec deorbiting burn up high and then lose 2500 m/sec at perigee, approximately--4000 m/sec all up, so the two ton bus rocket needs a bit over 3 tons fuel just to get itself back home. Thus the load going "up" is 4 tons payload plus 5 tons bus and return fuel, or 9, this too needs to go through delta-V of 4000 at least, so again we need to add a bit over 1.5 or 13.5 tons, 22.5 all up in LEO to deliver 4 tons to GEO and return the bus--which comes back to LEO in just 8 hours plus however long it takes to fuss around in GEO to set things just right there, plus possible coasting time up to 24 hours to position the returning bus to the right part of its low equatorial orbit. 20.5 of the total had to be launched this time around; the two ton bus rocket needs to hold 14.5 tons of hydro-lox propellant. I suppose that is roughly in the right ballpark for a plausible rocket. Over 4 times the payload to GEO is fuel.
Could it have been possible for the bus to skip the 2500 m/sec braking to LEO by aero-skipping off the atmosphere to lose it instead? Say we insist on keeping 500 m/sec as maneuvering reserve, but save 2000. This means we deorbit 2.25 tons through 1500 m/sec; we need the booster to mass 3.2 now instead of 5, so up mass to GEO is 7.2, and 1.5 times that is 10.8 so the tank needs to hold 12 tons of propellant; we have all up mass of 18 tons or 4 and a half less than an all-rocket strategy demands; we only need to launch 16, of which 4 is payload; 3 times the payload to GEO is fuel.
So--pushing the envelope toward two ambitious stretches of state of the art--one where we double the ISP of a Hall thruster but maintain mass flow by quadrupling power throughput and hence double thrust; the other where we must design the two ton (dry) chemical booster to fly through an atmospheric aeroskip maneuver to LEO and therefore it needs both careful aerodynamic design and non-ablative TPS that doesn't get degraded in long stays in orbital space, we can contemplate a tradeoff of a slow transfer of almost 4 months and then a month wait to get the bus back, but using less propellant mass than payload, versus having to launch 4 times the payload mass to deliver fuel but getting the bus back later the same day.
If we deny either of these perhaps dubious advances, we can compare an even slower electric bus that takes the better part of a year to cycle but requires only 1 2/3 the payload mass in propellant, versus a very conventional Centaur type bus that requires we launch over 5 times the payload mass to orbit to deliver it to GEO and return the bus.
Note something else about the electric thrusters when frowning at their very slow transit times--we can add propellant tankage to them relatively cheaply in mass terms, and by paying an even bigger penalty in transit time, push bigger payloads to a given goal, or a given payload farther. The chemical buses, on the other hand, seem just about right-sized for this 4-ton payload I pulled out of the air intuitively. In the conventional case I think about the 14.5 tons of hydro-LOX I guessed is as much as 2 tons of structure, less engine, thrust structure and docking ring can hold; to push a bigger payload we either need more buses or to push it to a closer target. The aero-skipping version needs significantly less propellant, but designing a structure that can hold that lesser fuel load and also safely aeroskip every time probably will cut the tankage available down.
Comparing the extremes, the advanced 3200 ISP Hall thruster that still takes 4 months to deliver the goods to GEO versus the almost completely conventional Centaur with a docking ring, we need to launch just under twice a 4 ton payload to get that slow delivery accomplished, versus the old-fashioned Centaur tug that can offer same-day delivery, but requires five times the payload mass to be launched. Is the cost of launching the extra 12 tons worth saving by waiting 4 months? If not, saving less mass by waiting 8 months is clearly a bad deal, and if not, Hall thrusters as electric transfer vehicles are a non-starter for commercial business. They remain desirable for what they have been used for OTL thus far, which is deep space exploration where their slow accelerations cease to be such a drawback.
And then, electric propulsion will have to await a new power source to be useful for anything other than station-keeping and niche applications. Nuclear fission seems unlikely to be mass-effective, competing with solar panels. By the way, can we do that trick I did with the Hall thruster and just quadruple power again to get double ISP and double thrust? Well for one thing that would if possible mean taking a one ton solar array and making it 4 tons, or tripling the bus mass from the original version, so clearly we would be hitting diminishing returns. And no we can't; I was taking the high-end figure of on-line estimates of the maximum theoretical ISP for Hall thrusters and maybe exceeding it too. We can't expect it to be physically possible to double the ISP again. We might still consider doubling the power to double the bus mass all up to 5 tons and speeding up transfers a bit, but again that way lie diminishing returns.
It looks to me like we are just going to have to rely on good old hydrogen-oxygen engines and keep launching lavish amounts of propellant to service them. In fact, it looks like maybe trying to recover a Centaur-sized rocket is kind of marginal economically speaking, considering the extra propellant we need to keep launching to get it back. It might make more sense to just make the rocket cheaper and dispose of it once used.