Derek Lyons wrote:

Yes, I'd like to see some more. I find it hard to understand how PWM
can fly on a strategic missile, but be 'risky state of the art' a
decade later.

I'm not an electronics guy but I *was* a systems design engineer. I can
foresee lots of problems with the use of certain technologies in certain
applications, even technologies considered "mature" for that other
application. Case in point: PWM on Poseidon may have had different specs -
data access rates, interface requirements, vibration tolerance,
acceleration tolerance, acoustic environmental tolerance, etc. Oh, yeah -
lifetime and reusability, as well. An SLBM controller may get by with a
MTBF of 60 minutes total use, perhaps, including test cycles. The SSMEs
may have had specs calling for a hundred times that. Furthermore, the SSME
controllers probably have a much higher real-time computational load on
them than do similar data devices on a solid-fueled SLBM. Of course, these
are just suppositions on my part, albeit educated ones based on knowing
first hand that the design specifications make all the difference in
something like this. The prior discussion of the problems with Viking's
systems should have made that clear.

Another more interesting topic, and one I *do* know about, was the SSF MDMs.
At the time, the state-of-the-art PC was a 386DX running at about 33 MHz.
Yet MDMs were spec'd at 286's running at 12 or 16 MHz (can't remember the
details). The software people desperately wanted to change the MDM spec to
the known, "reliable" and much more powerful 386 to cope with code- and
feature-bloat, but at the time, there were no widely-available rad-hardened
386's available, period. So using "proven" 386's in the design simply
wasn't an option they could do without incurring lots of cost and
development effort.

In point of fact, I believe the MDM specs were eventually bumped up but by
then I was out of the loop and into law school. I have no idea what the
flight hardware is now but I *do* know that if FEL had been in the '95 time
frame that it was when I joined the program, every one of dozens of MDMs
would havbe been 286 boxes running at 16 MHz or slower, in spite of
"proven" technological alternatives.

--
Herb Schaltegger, B.S., J.D.
Reformed Aerospace Engineer
Remove invalid nonsense for email.