Drive Motor Drop-out

Team 4926 would like some help with ideas to solve an electrical problem on our 2018 competition chassis. Although we are done for the FIRST season, we have a group of motivated students that want to “get to the bottom of it.”

So on their behalf here’s the conditions and symptoms:

  1. Tank Drive
  2. 2-stage ball shifter in high gear
  3. 2 CIMs per side
  4. Driving around in our shop on “legal” carpet with AM high grip wheels
  5. 1 talon SRX per side
  6. 1 Victor SPX per side (set to follow the talon)
  7. Occasionally the left side motor currents go to zero as in the included and linked image.

Can you all give us some things to check? Cue the music that goes with “why did the combustion guy start an FRC team?” The one thing we can’t do is burn stuff!

What do the signal lights on the motor controllers do when the motors stop? If they stay green/red, it’s a connection problem between the motor and motor controller. If they go orange, it’s a problem with the CAN connection (or a code problem). If they turn off, it’s a problem between the PDP a and motor controllers.

Is this an instrumentation issue, or a drive issue? That is, do the left side motors go dead?

Assuming this is a drive issue, the simplest thing for this to be would be for the left side SRX to stop receiving power and/or a control signal. In addition to checking for loose wires, you may want to see if swapping that breaker for a new one solves the issue; it may be tripping and resetting itself. (If you do this, label the one you pull and toss it if this solves the problem.)

Thanks guys. We’ll check these and report back. Do breakers really wear out? We may swap them to the other side to see if this follows the breakers. The old “A-B-A” test sequence. Tried and true.

We are planning to spend lots of time this week debugging this so keep suggestions coming!

From our experience at GSD, yes. The breakers (which we hadn’t bought more of since our rookie year) on our drive train were popping several times per match, making our drive train die for ~7s, until we swapped in some new ones from 95 (Grasshoppers).

Before Pine Tree we changed our climber’s gear ratio, which resulted in higher current draw. During Finals 1 I noticed that we kept on falling (since we don’t use a rachet and instead stall the motor), and I guessed that it was because we were using an old breaker, and after swapping one in from B.E.R.T. (133), we were able to climb without falling in Finals 2.

The more you break your breakers, the less resistant they will be to further breaking, so I’d recommend checking if that’s the cause of your issue since from your logs your DT seems to go out for roughly the amount of time that it would take for the breaker to reset.

charge the battery? :slight_smile:

Indeed, but how does a low state of charge cause this problem?

Hopefully someone who knows the details will reply…but the system is designed to shed loads as battery voltage drops, to keep the roborio and radio working.

If the battery voltage is really 12.0 volts at rest, then you need to charge it! and don’t expect the robot to function properly, until you do

Look at the current just before the “drop-out”. Is that really showing more than 60 amps? It looks like the circuit breaker is doing its job.

I would check for mechanical issues that are making the motors work harder than they should.

GeeTwo’s initial question is key - I’m presuming the actual problem you’re chasing down is “left side of drivetrain occasionally stops turning”, not “logs look weird but robot drives just fine”.

As others have said too, I’d start with watching motor controller lights.

Other experiments to try:
–swap motor controllers from left to right (and vice versa). Does the problem move with the controllers?
–Check connectors. We’ve seen similar things with loose electrical connections from PDP->controller and controller->Motor. That being said, it’s very odd that the connections for both motors would simultaneously become loose.
–Talon SRX’s do retain settings over power cycle. Have you tried doing a factory reset on both and seeing if that resolves the issue?

A few others have mentioned brownout conditions, which don’t appear to be happening here (all motors should shut off simultaneously, and you should see flashing red on the battery voltage on the driver station, and it should be logged).

Does the driver station ever spew errors about CAN messages and the PDP/controllers? What is your current CAN bus load? Do changes in either of the above correlate to the issue you’re seeing?

Would you be willing to post your code as well? Although it seems unlikely, there may be some clues in the software implementation.

I think your issue is not CAN because the CAN bus would have to affect both sides. I recommend checking the connections from your motor controllers to your motors on the affected side to see if there are any issues. We had a similar problem on our intake which ended up being anderson powerpole connecters being put together wrong where any jostling of the robot would cause that side to lose connection. I hope this helps.

I also second what others said on the breakers for those motors. You should definitely replace these because they might be bad. (We have had experience with that as well as the loose connections)

This. As the breaker pops certain parts are distorted and/or damaged slightly and the breaker is never the same afterwards. Once we identify that a breaker has popped we throw it out because our testing has shown an extreme likelihood that it will pop prematurely in the future.

Also this.

The Talon and Victor will not function below 6V. Your main system voltage is dropping precipitously low during parts of the log, increasing the chances of a motor controller seeing <6V. You may consider checking all of the wiring related to those drive motors for loose or sub-par connections (quick disconnect terminals wear out quickly, for example). And, of course, testing and charging your batteries.

Having said that, the ‘off’ time looks about right for tripping breakers, so I would start there.

How big is your drop center?

Regarding CAN, this is not true. It is entirely possible to lose only the portion of the can bus that is downstream from a problem. It will not necessarily affect all the can hardware.

That said, you would also be seeing can errors on the driver station error window if you were losing devices.

Yup, partially what I was going for at least. Even with a perfect electrical system, there’s improbable but technically possible scenarios where the messages for one side of the drivetrain are getting dropped. See how Device ID and message arbitration are related. If the bus load is near 100%, it is more likely possible this could be at play.

Can you provide the mechanical information on your drive train? What are the gear ratios used for low & high gear? From your robot photo it looks like you are doing a WCD with traction wheels in the center, omni’s on the front - are omni’s also on the back?

It seems like you’re pulling a lot of current so I’m just curious what your setup is like mechanically.

I did not know that, thanks for the information. This makes me think that the OP’s problem is CAN related, in which case I recommend soldering all of the connections if not already done.

The zero voltage on the talon coincides with a longer low on the battery. That tells me that the breaker is probably tripping on the talon. With the victor in follow mode it will also shut off.

I’d look at that gearbox assembly to see if there are tight spots or misalignment in that side, as that would cause high current draw (which will show as lower battery voltage).

Also, does this only happen with wheels on the ground, or does it also happen with wheels off the ground?

It’s rarely the case that a breaker’s innards become warped. The typical “damage” is pitting on the actual electrical contacts due to a tiny bit of arcing when the breaker opens. That decreases the quality of the contact, increasing its resistance, leading to a bit of heating during normal operation. Since they are thermal breakers, that primes them to open sooner or at a lower current than they are rated for.

The bottom line is that a breaker that has tripped a few times is indeed likely to trip again when it really shouldn’t.