Failure analysis: I think I know what went wrong

I tried to download the numbers from the altimeter today but found no accelerometer data.  The unit obviously had power throughout the flight, however, since the pressure sensor reported a peak altitude of ~4,900 at 17.6 seconds.  Hmmm.  Now I begin re-reading the ARTS2 altimeter manual and I figured out the problem: the unit must be armed when the rocket is in a vertical position!  This may seem obvious but we literally had no other safe way to arm the electronics so this change of plan made sense at the time.  Here's my best guess at a sequence of events:

• I armed the electronics in a horizontal position so the altimeter's accelerometer never calibrated or started measuring.  In retrospect I thought I'd heard it beeping correctly but I must have been wrong.
• Since the altimeter requires acceleration data to fire the ejection charge nothing happened at peak altitude.
• I had set the timer for 22 seconds based on a simulated delay of 21.5 seconds but the actual barometric peak occurred at 17.6 seconds.  This means the rocket continued to arc over for 4.4 seconds longer than was required.
• This extra speed combined with the extremely heavy rocket parts must have caused the zippers.

When I attempt to certify again I plan to:

• Move the switches to the exterior of the payload section so the rocket can be easily armed in a vertical position.  I'd originally designed around this because I wanted to avoid drilling holes and weakening the critical sustainer/payload area.  I also didn't want to risk lengthening the switch wires adding complication. I feel much better about all this now and the switches will mount in the payload section not the top of the electronics bay as originally feared.
• I'll replace the 1/2 " tubular nylon with 1/2" Kevlar with stitched end loops to improve strength after ejection.
• I'll add two Giant Leap "fireball" anti-zipper pads to both shock cords.

I look forward to another "attempt at success" in fall of this year.