Thanks Brett, I often wondered the sequence of events that cause the Columbia disaster, things like that weigh heavy on the heart, that and "Challenger".......I wondered long and hard why the shuttle design didn't include a cockpit escape pod. Thanks again for the explanation!
Cockpit escape pods don't work very well, and are very heavy. For it to be of any use, it has to more-or-less be a separate complete spacecraft all by itself. On the way up you have to handle everything from ground level to nearly orbital velocity/altitude. If you abort just short of orbital velocity, you might be up there for 45 minutes, and then have to do essentially a full reentry sequence, meaning heat shield, parachutes, some sort of attitude control, oxygen, scrubbers, etc. With solid motors, early on in the boost you have to also consider how you get out of the way of the rest of the stack, since abort thrust termination takes a while and leaves a lot of debris to escape. Doable but far from trivial and every ounce takes away from payload. You can see what sort of a fireball you have to escape from looking at Challenger. In fact, on that one, the crew compartment *did* stay together to sea impact, although it is expected that most or all were unconscious from the original shock of the breakup. That was on one very narrow set of conditions, and you would still have to have had some sort of parachute for this HUGE chunk and some way of stabilizing it before chute deployment. Higher and you need a heat shield just for abort.
Boost stage escape systems on Mercury and Apollo took essentially the entire crew compartment off at 20Gs for the entire 1st stage firing, and the escape motor was ejected shortly after 1st stage cutoff. After that you just flew the same capsule back down, including on Apollo, possibly canning the second stage and going to orbit with just the third stage and SM propulsion, just SM propulsion, or canning the whole thing, using the SM motor to do a retro burn and then coming back - all the way to aborts they did on Apollo 13 where they used the LM propulsion. If the failure on Apollo 13 had happened on Apollo 8 (no LM) they all would have died very soon, same thing if it happened AFTER a landing on the way back.
Gemini was different in that it had no escape system. Just ejection seats for a limited portion of the boost. Same as the Shuttle test flights where they had ejector seats for the flight deck. Both were very limited capability and the one time the flight rules said they needed to eject (Gemini 6 pad abort), they didn't.
The Russians have a similar abort system to Mercury/Apollo and have used it at least once.
On the way back in, there has never been much of a failure tolerance. If the heat shield was damaged you were dead. Soyuz and Gemini have contingency "loss of attitude control" mode that that comes into play if the attitude control fails, and instead of flying a lifting entry, they do a ballistic rolling entry. That raised the re-entry loads from maybe 6 Gs to upwards of 13-14Gs. Normally they fly it at an angle to the airstream due to a CG offset, and then roll it until it is "nose-up" to extend the landing/reduce the G loads, and "nose-down" to shorten the landing/increase the G loads. Apollo flew "head down" to start until they slowed enough to not skip off, then rolled to a constant 6gs, rolled
"head up" head-down at reduced G, then back down to 6Gs as I recall, then adjusted from there to hit the carrier target.
For Orion it goes back to very much like the Apollo system with a Soyuz-like escape tower. I don't work Orion (now, anyway) but I would bet they are going to do the same flown re-entry as Apollo, with a back-up rolling entry for orbital missions. I would bet that you can't do a rolling entry from escape/lunar speeds, too much Gs/too narrow a window. Last I saw, the first manned Orion test flight was going to be circumlunar.
Brett
p.s. the real problem with Columbia (and the other shuttles) was the side-mounted orbiter, externally-insulated tank, and the TPS designed with NO tolerance for debris hits. The specification for the tolerable debris hits was ZERO, i.e. there was no requirement to tolerate damage, and a hard requirement to prevent debris hits for everything else. They saw hits from the first flight and in reality it was decently robust with not-infrequent pitting and occasional airframe damage around the tiles. The RCC is pretty tough to bird strikes (at the altitude you get birds, the speed is still low) and other damage, too, but obviously not bulletproof. The tank insulation was on the outside where it both was prone to damage, could flake off, and could get wet then frozen into blocks of ice. And of course the denser it got from water absorption, the more like it was to separate under vibration and load. You can put the tank insulation on the inside or outside - the Saturn V 2nd stage had it on the outside and the third stage on the inside - but the shuttle tank had it on the outside, mostly because it was easier. Of course on the Saturn V, even if the foam shed (which it almost certainly did), it *didn't matter* very much. If it sheds off the Shuttle tank, most of the places it doesn't have much effect, but on the "upper" surface near the nose, the foam can hit the orbiter TPS. Even then, if it's dry, it still probably doesn't hurt anything since its so light. If it's saturated with water, different story, as we saw.
They convinced themselves that since they had had a bunch of hits in the past, it didn't matter, and they didn't pay excess attention to the problem. Even after they saw the hit, they figured it was just likely to be a maintainence problem. And in any case, there wasn't any fall-back position, they were going to run out of supplies if they stayed on-orbit and waited for an emergency (unplanned) rescue mission. And the chances of throwing together a rescue entailed substantial risk all by itself.