> The escape rocket is vectored along a path away
> from the direction of the launch vehicle.
That presupposes that the capsule is uncoupled from the
upper stage. If the S1 booster is still accelerating,
separation may be impossible, regardless of the intended
escape vector.
I'm sure the NASA sliderules are working on all of this,
but right now, this looks a lot like the ejection seats
on STS-1 and the bail-out procedure for recent flights.
Successful scenarios were unlikely, and even they had
very narrow windows of execution.
I'd rather see NASA man-rate one of the ELVs, and run it
until the first space elevator is up.
That's probably true. There would be some failure modes that would be difficult. John Glenn's escape rocket would have provided thrust in excess of what would be needed as a minimum to get away from the Atlas under full thrust. Apparently the escape tower had quite a kick. Lots of gee force.