Victors behaving erratically when PWM is plugged in

[Sylphrena]

We have a mecanum drive with a CIM and a mini-CIM on each wheel, with their relevent Victors connected together with a y-cable.
When only one PWM cables are plugged in to the Victors, the robot behaves as expected. However, when two or three wheels are connected to the sidecar, the wheels will start jerking back and forth, even if no inputs are going to the wheels. This remains true even if we replace the Victors and the y-cables. This phenomenon occurs regardless of what order we plug in the Victors.

Why would this phenomenon occur and what could be done to fix it?

[Cash4587]

Is your side car being powered off of the PDB? I know last year my team plugged the digital sidecar into the camera port by accident.. We had problems with the robot going crazy as well.

[ArchosR4]

Have you checked the power connections to the sidecar? We had the same problem. The sheathing on the power to the sidecar was in the way and the wire was barely touching. Try resetting the connections on both sides of the power cables to the crio.

[dradel]

This probably isn't it, but electronics can act screwy if the voltage drops off. Check the battery voltage both at the battery and at the power distribution board. Something as simple as a loose ground can make stuff act a little nutty.

[kiettyyyy]

You may want to unassemble your digital sidecar to inspect the solder joints on the PWM connectors. I've had multiple instances where there was poor workmanship (i.e. no solder or cracked joints)

[Al Skierkiewicz]

I am going to suspect that you have rotation sensors on all wheels for your drive control. I also suspect you haven't calibrated the Victors. When all of them get going, the mis-calibration could cause the twitch you describe.

[Alan Anderson]

Are all three power LEDs on the Digital Sidecar brightly lit? Is the big orange Robot Signal Light blinking as expected?

The symptoms point very strongly to the Digital Sidecar not getting battery voltage on its power input. Enough current sneaks through the cable from the cRIO to let a couple of PWM outputs work (especially if you don't run them in reverse), but as soon as too many are connected they all get flaky.

[Sylphrena]

UPDATE: We replaced the y-cables with straight cables, and the problems disappeared. Now we're rapidly running out of space on the sidecar for more victors, but that's neither here nor there.

[Joe Ross]

Originally Posted by Sylphrena

UPDATE: We replaced the y-cables with straight cables, and the problems disappeared. Now we're rapidly running out of space on the sidecar for more victors, but that's neither here nor there.

I suspect that the root cause of the problem is the digital sidecar power, as others have suggested. By adding more motor controllers, you are increasing the amount of leakage current coming from the DB-37 cable which is making it work temporarily. If you remove the DB-37 cable, are all 3 LEDs brightly lit?

I would encourage you to not consider the problem solved until you can run with Y-cables.

[Alan Anderson]

Originally Posted by Sylphrena

UPDATE: We replaced the y-cables with straight cables, and the problems disappeared. Now we're rapidly running out of space on the sidecar for more victors, but that's neither here nor there.

As Joe says, it's likely that you just masked a power problem by opening more PWM channels. If you tell your robot to spin to the left (or right, depending on wiring), do the motors quit unexpectedly?

Seriously, look at the three power LEDs on the Digital Sidecar and make sure they're all on even with the connection to the cRIO removed. Verify that the power input of the Digital Sidecar is getting battery voltage (~12V) from a red/black pair of Wago connections on the Power Distribution Board, protected by a 20A snap action breaker.

[Al Skierkiewicz]

I would also like to add to the discussion... The three (not just two) LEDs need to be on and bright.

[nuttle]

A short in anything that uses 5V is another possibility -- causes similar issues with DSC power but the cause is different. If you have a meter, you might want to measure the voltage between GND and +5VDC on one of the sets of digital I/O pins. You can rig up a PWM cable to do this pretty easily.

[Sylphrena]

The LEDs on the sidecar are indeed brightly lit, and this problem persisted after we changed the sidecar, which is one of the reasons we're hesitant to accept it as a problem with the sidecar.

I thank you guys for your help--since we at least have a temporary fix, we haven't managed to find the time to search for something more permanent. I expect that we'll be able to spend more time on it today, as the electronics team hopefully () gets something of a pause in the action until our intake goes on the robot.