Henry Spencer wrote:
In article ,
Pat Flannery wrote:
We've read up on your "Brown Bess" booster concept; if you were going to
make an unmanned expendable SSTO, how would you go about it, and what
propellant combo would you use?

First, as Richard observed, I'd drop the "unmanned". If it's reliable
enough for expensive cargo, it's reliable enough for people. Like, for
example, me. :-)

(There are people who suggest building relatively unreliable rockets to be
used for bulk cargo -- water, fuel, etc. -- only. I don't think this
actually works out well. You still need moderately good reliability, say,
80-90%, if only to avoid being fined for littering :-). I don't see a
significant cost or complexity advantage to be had from the difference
between that and the 98-99% of conventional expendables. If you can
dependably get 80-90%, it should cost very little extra to hit 98-99%.)

(Getting to 99.9% is harder, as witness the fact that no existing
expendable has definitely achieved it -- there are a few uncertain cases
where moderate production runs simply had no failures -- with the
*possible* exception of the Soviet-era Soyuz launcher. It should be
feasible, given careful design, a high flight rate, and automated
production. Even 99.99% is probably not out of reach for expendables,
if you sweat hard on things like systematic process improvement. Beyond
that is strictly reusable territory.)

The real major dividing line is reusable vs. expendable. Here, by
definition, we're talking expendable.

After that is the big question of whether whoever's paying for it has
constraints to impose: use existing engines, no Russian subsystems, a
minimum payload size, etc. They also might have opportunities to offer,
e.g. use of shuttle-ET production facilities. Many of these things can
severely constrain the design. Assume none of this.

The major subsystem question is engines: buy or build? Buying means you
don't have to get into the engine-development business, which saves a lot
of trouble and may look less risky to potential investors. There are some
downsides: (a) it's a lost dimension of competitive advantage, (b) the
choice of existing engines is somewhat limited and can severely constrain
design choices (in particular, ruling out many unconventional approaches),
and (c) buying engines tends to be expensive and to involve a lot of
hassles. I'd favor build, if only to relax design constraints.

The major specs issue is, how much payload to what orbit? Orbital
inclination affects delta-V requirement by determining how much help you
get from Earth's spin. The big question for orbital altitude is whether
the orbit is low enough for a direct-ascent trajectory -- continuous burn
all the way up, like Gemini or Apollo -- or requires a Hohmann ascent like
the shuttle, injecting into an elliptical orbit and then doing a final
insertion burn at apogee. Hohmann ascent would always be more efficient
if the atmosphere didn't get in the way. In real life, direct ascent
usually incurs little penalty up to 300-400km, but gets rapidly worse
thereafter. The nice thing about direct ascent is no engine restart.
And even with Hohmann ascent, you pay a price for higher orbits. Absent
outside constraints (e.g. cargo delivery to ISS), I'd favor direct ascent
to 250-300km, high enough to last a little while and give the payload time
to maneuver higher or be picked up by a tug.

As for how much payload... depends on whether there's a specific mission
constraint. If not, I would favor relatively small payloads, giving a
small launcher and frequent flights, and relying on orbital infrastructure
(assembly base, tug, fuel depot) to assemble larger systems. Smallness
actually is not that important -- launcher cost scales much more strongly
with complexity, thinness of margins, and closeness to the leading edge of
technology than with sheer size -- but frequent flights are beneficial in
many ways.

*How* small depends on how much inconvenience you're willing to accept.
There are cutoff points where inconvenience rises sharply because you can
no longer launch particular objects in one piece, plus a general slow rise
in inconvenience as orbital assembly operations multiply.

If you want at least the option of launching people, that obviously sets a
minimum size. For serious orbital operations, I see a high payoff for
being able to launch a two-man ferry spacecraft, sort of a stripped-down
Gemini, in one piece: it lets you have one pilot and one passenger, so
the passenger doesn't need exhaustive training in emergency procedures for
the ferry. Gemini weighed a bit under 4t, with early-1960s technology and
greater capabilities than the ferry really needs. An aggressive modern
design could come in quite a bit lighter.

For serious orbital operations, the other thing that it would be nice to
launch in one piece is a minimal habitation module. Perhaps inflatable...
but with an expendable SSTO, a "wet workshop" approach using the spent
stages is also very attractive. Say: Launch #1 carries a life-support
module with consumables, integrated with the spent stage and with a
docking hatch at the top. Launch #2 carries a multi-hatch docking node
integrated with the spent stage, and a tug; the tug maneuvers it to mate
with #1, and sticks around to supply attitude control and reboost. A
ferry docks with one of the ports on the node, and you're in the
space-station business. (Actually berthing would be better than docking,
but that's a detail.) How much does each load have to weigh? That would
need more study, but it's interesting to note that the Apollo-Soyuz
Docking Module was about 2t.

Could this sort of scenario be done with payloads of 2t or less?
Probably, but it might get pretty tight. 5t should be lots. Let's be
mildly aggressive and set the payload at 3t. I'd want to look into
infrastructure issues -- size of manufacturing machinery, size of
facilities, etc. -- and if it didn't look like a slightly bigger launcher
would cross any boundaries that made things significantly harder or more
expensive, make it bigger just on general principles. Materials etc. cost
very little; the infrastructure issues are the main things that make a
launcher cost more just because it's bigger, and they mostly rise in
sudden jumps, not in a steady slope. And far more people have regretted
making a launcher a bit too small than have ever regretted making it
slightly too big.

Anyway, let's cut to the chase -- this has already taken rather longer
than I meant to spend on it :-) -- and look at the launcher. This is
based on some past thought but without rigorous calculation for this
particular design problem.

Shape: a plain cylinder with a cone on top, or possibly a two-slope cone
like the nosecone for Apollo 5 (which has lower drag and more usable
volume) -- simple to make, simple to analyze. More the proportions of
say, a Jupiter than a Delta -- the shorter, fatter shape has a bit more
drag but is a lot stiffer and less prone to bending problems. Very light
tanks, probably pressure-stiffened like the old Atlas. Likewise for the
nose -- that was done on Atlas for SCORE and some other flights. (Here
the nose stays on until reaching orbit, after which it hinges up and over
to expose the payload, staying on the rocket so it goes back down when the
rocket deorbits itself.) Either aluminum alloy or composite -- that would
need more investigation. Composites are stronger and lighter, but more
hassle to make, and there might be minimum-gauge issues with such light
sheets, and composite LOX tanks are still iffy.

Pressurization in the tanks is just enough for structural purposes, i.e.
not very much. Boost pumps at the bottom of the tanks, or possibly the
bottom of the feed lines, add enough pressure to prevent cavitation in
the main pumps. (This approach is out of fashion but it has been done
successfully in the past; it avoids having to make the tanks stronger and
heavier to permit higher pressures.) One interesting option is to make the
boost pumps jet pumps, recirculating a bit of the output from the main
pumps to the jets in the boost pumps. (That too has been done.)

The oxidizer is LOX -- cheap and dense. The fuel is probably propane --
slightly better performance than kerosene, less tendency to leave oily
residues and otherwise misbehave, and it's still liquid and quite dense at
LOX temperatures.

Finally, for engines, I'm partial to the idea of an aerospike with a ring
of small individual chambers. The small chambers help keep the scale of
most engine-development facilities down. The aerospike provides altitude
compensation and also permits a light, compact nozzle with a very high
expansion ratio in vacuum. Expander or gas-generator cycle, preferably
the former if enough heat can be had. (It's been done with propane.)
Post-separation attitude control with propane cold-gas thrusters, and
deorbit by dumping residual propellants through the engines.

This is where hydrogen shines over the 'lesser fuels'.

With the lesser fuels, you just barely make it to orbit, and any fuel
you do have left over, you waste to deorbit the booster to then burn up
in the atmosphere, which is nearly 90% of your usable payload mass,
already delivered to 100 percent of orbital velocity. That's just nuts.

With hydrogen, you get there, and then some, with plenty to spare.

Lesser fuels make the hydrogen 100/1 rule look good.

--
Get A Free Orbiter Space Flight Simulator :
http://orbit.medphys.ucl.ac.uk/orbit.html