Spontaneous motor skipping?

[Damaku250]

We appear to be having an odd problem with our drive motors.
Following an odd voltage drop resulting in the loss of control of our bot in the quarter-finals at WAT, we had discovered that one CIM and its accompanying Jag were noticeably hotter than the others.
After removing the electronics and offending gearbox, getting home and testing them on benchtop, we've noticed two things:
1) after removing our CIMs from the gearbox the one that was overheating is also making an odd clicking and is noticeably louder than the other. [Resolved]
2) that our CIMs "jump" and will randomly stop for a random duration between 1/4 and 1/2 a second. When this happens, all the motors jump simultaneously, and the lights on all the Jaguars flash yellow/amber for a brief period while the motors stall.

We HAVE tried changing the DSC and 9403 cRIO modules, to no avail.

[Al Skierkiewicz]

Craig,
Sounds like you have a shorted winding in the CIM motor. No way to fix that, just replace.

[Damaku250]

Originally Posted by Al Skierkiewicz

Craig,
Sounds like you have a shorted winding in the CIM motor. No way to fix that, just replace.

duly noted, but would this cause the second part of the problem?

[Al Skierkiewicz]

Yes, you are drawing down the 12 volt supply which draws down the DSC power that in turn plays havoc with the PWM outputs.

[Damaku250]

Originally Posted by Al Skierkiewicz

Yes, you are drawing down the 12 volt supply which draws down the DSC power that in turn plays havoc with the PWM outputs.

We have now removed the borked motor and are testing with the good one. We are still experiencing "skipping". What could be causing this?

[Al Skierkiewicz]

Craig,
Is the robot signal light also blinking out at the same time?

[Damaku250]

Originally Posted by Al Skierkiewicz

Craig,
Is the robot signal light also blinking out at the same time?

No, the RSL is staying on and we do not appear to be experiencing power interruptions on any other components on the board.

[Al Skierkiewicz]

You are using the same Jaguar? Suspect that as well.
Are any of the lights on the PD winking out at the same time?

[Damaku250]

Originally Posted by Al Skierkiewicz

You are using the same Jaguar? Suspect that as well.
Are any of the lights on the PD winking out at the same time?

We have tried the motor with all 4 different jaguars mounted on our board. All jaguars experience an interruption at the same time.

The lights on the PDB are staying solid.

We have tried 5 different versions of code, all of which are not affecting the problem.

[Al Skierkiewicz]

Craig,
Is this a new problem? Have you had this all season or just recently? It is possible that you have the dreaded watchdog issue. Someone posted last night that there is an issue with Classmate power settings as well. I believe the Classmate issue occurs at 15-20 second intervals while the watchdog is something longer.

[eugenebrooks]

Watch the little screen on the lower left of the
driver station display. Do you see a momentary
"watchdog not fed" message correlated with
the motor glitches? If you see it, you have the
dreaded watchdog issue Al mentions.

Eugene

[Azores]

It is definitely not a 15-20 second cycle, we've had the CIM running for up to 1:30 before it skips before. It may be a Watchdog issue - I'll go run another test and keep an eye out for the Watchdog issue.

This is, as far as we know, a completely new issue however. We hadn't noticed this motor skipping problem for all of build season or all of the practice and seeding matches at Waterloo.

[Azores]

Originally Posted by Azores

It is definitely not a 15-20 second cycle, we've had the CIM running for up to 1:30 before it skips before. It may be a Watchdog issue - I'll go run another test and keep an eye out for the Watchdog issue.

This is, as far as we know, a completely new issue however. We hadn't noticed this motor skipping problem for all of build season or all of the practice and seeding matches at Waterloo.

Just ran the test. Watchdog is indeed cutting out...

[TubaMorg]

Yes we had exactly the same symptoms early in our build season. At first it was manifested in a compressor that wouldn't go up to 120psi, then we noticed the intermittent Watchdog Not Fed error. It also caused the motors to hitch and jerk when they should just be running. We widened the setpoint for the watch dog to .4 seconds and made sure every bracketed piece of code feeds the dog and haven't had an issue since.

[Azores]

Widened the setpoint? I'm not sure what you mean by this. A delay before first feeding it?