Rick Jones wrote in
:

I'm just a lowly member of the peanut gallery, but everytime I see
someone talking about waiting just a little longer I think about
people using sleep() as a syncronization mechanism in a computer
program. Is there no other way to be certain the stages are far
enough apart before igniting the next stage?

That's not exactly the problem--the separation impulse imparted by
the pneumatic pusher/springs/whatever is supposed to do it. Here the
first stage unexpectedly caught up with the second--it wasn't
supposed to do that. And the fix will be simply to let the whole,
unseparated, rocket coast a second or three longer before separation.
They do know the separation mechanism works from the second flight,
but in both cases misbehavior of the first stage caused unwanted
post-separation contact which led to mission failure.

Stay tuned, this drama may not be over yet...I certainly hope that
the Merlin engine doesn't require modifications to minimize shutdown
misbehavior.

Strictly speaking is that true? Did they test payload separation?
How about parachute recovery of the first stage?

Apparently the payload separation mechanism was tested on the second
flight and worked, despite tumbling.

No joy on parachute recovery in any of the flights; this time the
parachute apparently was toasted by the second stage engine "fire in the
hole". I'm curious as to how recovery is going to be feasible with the
engine having to endure splashdown impact and exposure to salt water...

Initially I think SpaceX just wants to see how the system works in
practice and to get some of the hardware back for post-flight analysis.
Actual reuse may take some practice...

The planned fix is to increase the 'dead' time before separation to
allow the Merlin 1C engine to finish sputtering out. I assume they
wanted a very short coasting time to minimize gravity losses,

Just how long _can_ one let things coast in these sorts of situations?
Or does this:

I'd guess many seconds, but if the first stage engine cutoff roll torque
and/or shutdown 'burp' causes rolling/tumbling, this could be an issue
in safe separation and trajectory correction. I don't think the
second stage RCS has the control authority to manage the whole stack.
Adding a retro pack or RCS to the first stage seems counterproductive.
The idea is to keep things as simple as possible.

http://www.spacex.com/falcon1.php#second_stage

"A single SpaceX Kestrel engine powers the Falcon 1 upper stage. A
highly reliable and proven TEA-TEB pyrophoric system is used to
provide multiple restart capability on the upper stage."

imply that they can already restart after a long coast?

Good question, could be slightly useful if the LOX boiloff and venting
isn't an issue.

--Damon