Digital Side Car Error

[Tom Bottiglieri]

Hi all,

We are having some weird issues with our practice DSC, and hoping someone might have some relevant advice.

The side car will drop into a strange state after a duration of use. In this error state, the PWM outputs die, or work sporadically, but with an incorrect map of input to output. The relay 5-8 red and green leds turn on, as well.

I first found this error after driving the robot full speed into another trailer. As soon as it hit the obstacle, everything died. After troubleshooting for a while, we swapped the DSC to our backup and the problem seemed to go away, leading me to believe there was a hardware issue induced by the large shock.

After a little more testing, the 2nd side car reverted to the same (not working) state. No shock this time.

Interestingly enough I am still able to read encoder values off the DSC even though I can not output to the motors.

I have not swapped back to the original DSC yet, and the robot is still a sitting duck. I also swapped the DIO module in the cRIO before swapping to the second side car.

Any ideas?

[JohnBoucher]

This was causing us problems.

http://www.chiefdelphi.com/forums/sh...ad.php?t=73891

[Mike Betts]

Tom,

Check the 12V power input to the DSC.

Read this thread for more background.

Mike

[Tom Bottiglieri]

Originally Posted by Mike Betts

Tom,

Check the 12V power input to the DSC.

Read this thread for more background.

Mike

I will double check the wiring tonight. I am 99% sure everything is wired up correctly, but I will double check just in case. We stole some cabling, with attached wago connectors, from the competition robot before shipping. That robot had been working no problem for a few weeks. Maybe we missed something in the transplant.

The reason this seems strange though is the behavior of the relay LEDs. They only go on once everything goes down hill. I have no idea if there is any kind of built in diagnostics, but if there was, this seems like it would be it.

[Mike Betts]

Tom,

This has happened to a few teams. In one case (as well as my addled mind can recall), the 12 volts was not hooked up at all but leakage current sneaked in via the DB37 and allowed the board to function until you went to full reverse on one of the joysticks, then chaos followed...

If 12V looks good, disconnect the DB37 and look again.

Mike

[Tom Bottiglieri]

So I tested all of the power out tonight. Everything was wired correctly. I ran the diagnostic window and got a red light for "FPGA and COMM OK". Motors still didn't work.

We did some digging to find that the screw holding the anderson connector to the robot could sometimes cause a short between the power lines and the frame (when we wiggled the wires). Not sure how this is possible, but it happened. The cRIO was mounted to a bracket that was dangerously close to a frame member. There may have enough shock to cause both to short at the same time, but this seems highly unlikely. Beyond that all electrical components are isolated.

So I ask someone with a bit more knowledge of the guts of the DSC/DIO modules/FPGA - Is there any way the situation explained above could fry only the PWM driving circuitry on the DSC? I assumed the PWM signal came directly from the FPGA. Not sure if there is any amplification on board the DSC that could have gotten whacked out.

Thanks in advance.

[Alan Anderson]

Originally Posted by Tom Bottiglieri

So I tested all of the power out tonight. Everything was wired correctly.

Did you test the voltage at the Digital Sidecar, or did you just verify the wiring? It's not hard for a WAGO connection to bite down on the insulation if the wire isn't stripped the proper amount.

[Russ Beavis]

My vote's with Alan on this one. Please confirm that, when powered up, all 3 power LEDS are fully lit on the Digital Sidecar (all 3 LEDs next to the power input). If you're not providing 12V from the PD (typically due to clamped insulation at the PD's WAGOs per Alan's comments), the 5V power LED will be dimly lit (due to phantom power from the 37-pin cable) but the 6V and 12V LEDs will be VERY dimly lit or off.

If you've got a chassis isolation issue (you should check that too), you may have found a new path for phantom power. If all 3 power LEDs on the Digital Sidecar are lit, remove the breaker in the PD that provides DSC power and see whether the power LEDs are still lit. If so, you're sneaking power into the DSC via some other path (maybe a combination of 37-pin cable and chassis isolation). If the LEDs do turn off when you remove the breaker, you may still have a chassis isolation issue but it's hard for me to imagine how that current path could be closed. You might need to start disconnecting digital I/O devices and Spikes, Vics, Jags to find the culprit.

Russ

[Tom Bottiglieri]

Yes, we checked voltage at each component. We held them in the air, we put them on rubber pads. We tried every combination of possible power input. There is no power ghosting going on here, just a straight feed from the battery.

We ripped apart the entire system. With just a battery, cRIO, PD, and DSC, both DSCs turned on directly into this state. We could nothing with them, and every couple of minutes they drive motors at full speed for about .1s. This is with no connection to a DS, or a disabled DS.

[Russ Beavis]

What are you using for motor drives? Does it happen for every motor driven via the DSC? Does this happen somewhat consistently?

Does this happen when the PWM cables are disconnected from the DSC? If so, maybe your motor drive has a fault such that it's spontaneously energizing its H-bridge every once in a while (not unlike the reported Jaguar failures).

Russ

[Tom Bottiglieri]

Originally Posted by Russ Beavis

What are you using for motor drives? Does it happen for every motor driven via the DSC? Does this happen somewhat consistently?

Does this happen when the PWM cables are disconnected from the DSC? If so, maybe your motor drive has a fault such that it's spontaneously energizing its H-bridge every once in a while (not unlike the reported Jaguar failures).

Russ

For motor drives (I'm assuming you are speaking of the down stream speed controllers) we are using a mix of Jaguars and Victors. The problem is apparent on all ports. We have 7 motor controllers on the robot now, so we have only tested between 0-7 plugged in, but I assume it would be the same with more.

The random bursts have happened on a variety of motors. Sometimes only one will drive, sometimes they will all drive. I have not seen any kind of pattern to when these events occur, they just happen after a long period of the robot being powered on.

We have got our hands on a fresh DSC. We have double checked everything else in the system for isolation, and should be testing the DSC tonight. I will let you know the results.


Also, can anyone elaborate on what the "FPGA and COMMS OK" indicator on the FRC Diagnostic Window is representing?

[Tom Bottiglieri]

Quick update:

Swapped the DSC tonight. Everything seems to work fine.

[PhilBot]

Hi.

Before assuming that you have a bad sidecar, we ran into a problem yesterday at the DC regional of the DB37 cable vibrating out of the DSC (or maybe it was that collision in auton).
This causes what looks like a progressive failue, that started with the encoders and worked it's way to PWM's.

This cable is not anchored at the DSC end, but it can be.

Take a cRIO plug-in module that you are not using (Spare analog or digital) and remove the small hex posts that the db37 screws into.

Fit these posts to the DSC, (at each end of the db37 connector) and then re-install the cable.

This makes for a solid connection.

Phil.