In article ,
Charles Talleyrand wrote:
...The disadvantage is that with 4-ish engines on the lower stage
you probably cannot tolerate an engine failure...

You can come close. The Saturn V, with five engines on the first stage,
had only a couple of brief time windows in which it was unable to survive
a single engine failure. Mind you, in many cases it would not be able to
complete its mission, but at least it would remain under control and give
you time to think about whether there was any way to salvage the mission.

You really need 6-8 engines to be able to fly the mission with an engine
out at any time. (The Saturn I twice carried on successfully after losing
one of eight first-stage engines -- once a deliberate test, once a real
failure.)

...For example, is motor design cost a large part of the
overall vehicle cost?

It's a significant part of the vehicle cost, although how much depends on
details. It also tends to take longer than vehicle development, so if you
start them at the same time, the vehicle will be waiting on the engines.
(Both the F-1 and the J-2 were started before the Saturn V. The SSME was
a major cause of shuttle development delays.)

Are most failures due to motor failures?

"Most" is an overstatement, but at least some assessments have found that
it's the single largest cause. (The number of data points is not large
and how you classify some of them is rather subjective.)

How many engines you can afford to *lose* is an important number. If the
answer is "none", you want as few engines as possible, ideally 1. But
with most reasonable sets of assumptions, overall vehicle reliability
improves substantially once you start being able to lose one and carry on,
because the fault tolerance outweighs the larger number of engines.

(One of the assumptions you have to make, of course, is what fraction of
engine failures are catastrophic, since one of *those* may cause damage
you can't survive even if the loss of thrust would be okay. However,
except in cases of gross abuse -- e.g., failing to shut the engine down
when the tanks run dry, allowing the high-speed turbopumps to suck air --
failures of fully-developed liquid-fuel engines are almost always benign.)

Is a
single large motor likely to weigh less and/or have a higher ISP than
a few smaller (but still large) motors?

Other things being equal, it's not likely to make much difference.
Neither thrust/weight nor Isp depends much on size, except at the extreme
low end, where performance tends to deteriorate some (at least for
orthodox design approaches).

Mind you, larger engines are notoriously more prone to combustion
instability. And their inflexible geometry can be hard to fit into
a vehicle. But those are secondary issues, "mere engineering". :-)

Basically, anyone have any good arguments for either choice?

Depends on your priorities. A demand for low cost pushes toward
simplicity. A demand for high reliability pushes toward fault tolerance.

P.S. Is it reasonably easy to tailor an engine to
atmosphere or vacuum operation with changes to the engine bell;
things like turbopump and cooling systems can remain the same?

There are some compromises involved. (In particular, if you want the
cooling system to stay the same, you probably can't regeneratively cool
the changeable part of the nozzle.) So it won't be quite as good as
engines built specifically for particular conditions. But it has been
done; the penalties are not huge.
--
MOST launched 30 June; first light, 29 July; 5arcsec | Henry Spencer
pointing, 10 Sept; first science, early Oct; all well. |