Teensy Complete Lockup

Status
Not open for further replies.

brianmichalk

Active member
I've been posting here over the last few months about issues I've been having with my Teensy brass case feeder controller.

I came to the conclusion that RF noise is stopping the Teensy3.2. One of my customers switched to different brass primers and began having problems, and I have not been very successful duplicating the issue in my lab. I ran my machine for 5.5 hours before it locked up, but I am able to see noise on an oscilloscope. I had problems on a different project with spurious interrupts ... again involving brass. At that time I measured 20v peak to peak pulses in the about 20nS range. This was in all pins on the Teensy.

A couple of months ago, I made a board revision with a copper pour ground, and switched from tantalum capacitors to ceramics on the voltage regulator. This is when I began getting reports of lockups that I could not duplicate. I was sure it was my software. I am using the watchdog timer, and even the watchdog stops working when the CPU latches up. I fly wired bypass capacitors on my digital lines, and for now that seems to be working, but its not proof that the problem is solved. All of my digital IO is Hi-Z, externally biased. I do have one analog signal.

I suppose that when voltages go out of spec, nothing is guaranteed anymore. The only way to reset the Teensy was pressing the program button and loading firmware, or power cycling it. The RST pad on the bottom of the board, might be an option. Either get another watchdog module and have it reset the Teensy. How well would an RC network do for this RST pin? When the board locks up, all of the digital pins go to a LOW state. If this behavior is consistent, then I could assert logic HIGH. When the RC discharges, the board is reset, but I think that would require the pad to transition from low to high.
 
Status
Not open for further replies.
Back
Top