Losing connection when ramping over defenses

[Kevin Kolodziej]

We are seeing the same behavior, with both the competition and practice robots. Both robots worked well for the first couple of minutes and now any forceful contact with the defenses causes a loss of comms. We have double checked the fuses, ethernet cables, all connections, main breaker, battery terminals, etc. and cannot find anything questionable. We have wiggled and jiggled all wires and even done drop tests of the robot on all sides and cannot replicate the problem - only when physically driving. I took a high speed video (240fps) of the control board during impact and what I see is this:

Robot impacts the defense, about 10 frames later (about .05s) the RoboRio power light turns orange for a few frames, then back to green for a few frames, then goes dark for about 10-12 frames, then goes back to green.

So it seems to be a power issue to the RoboRio, at least in our case. Lights on all the Talons behaved as normal during this and lights on the router behaved as normal during this.

We will continue to investigate...

[RufflesRidge]

If you're sure the fuses are good and the roboRIO power wiring in both ends is good and you're battery, main breaker, and PDP terminals are secure I would open up your roboRIO and vacuum out any debris.

[Al Skierkiewicz]

Kevin,
Did you perform a tug test on the Robo Rio power cables? I have seen this when power wiring is just a little too short in the push type connectors. Others have reported damaged power connectors that appear to be OK but in fact are partially depressed allowing a pretty good connection but not a perfect connection. The only sure sign is to look down into the terminal with a bright light and watch for the contact to return to fully closed after pushing on the actuator.

[BigJ]

We tried pulling on every wire, but we'll try some more tomorrow morning

We'll also look at the power connectors, and check to see if there is any debris in the RoboRios (this happens on both our practice and competition bots, which is concerning).

The only consistent thing we could see in our DS logs that was bad (as far as I know, which is not a lot ) is that a bunch of "5V brownouts" happen. We aren't really seeing the battery voltage on the DS go that low at all, and they also seem to pop up when driving normally, so I am unsure if that is a normal thing to see happen every so often while driving. It might also be worth noting that we don't actually have any 5V stuff wired up to the roborio at the time.

[Al Skierkiewicz]

Josh,
Kevin can give me call tomorrow morning. We are trying to get our robot together (big surprise I know) but I should be available. I am pretty sure he has my cell phone #.

[Alan Anderson]

Originally Posted by BigJ

The only consistent thing we could see in our DS logs that was bad (as far as I know, which is not a lot ) is that a bunch of "5V brownouts" happen. We aren't really seeing the battery voltage on the DS go that low at all, and they also seem to pop up when driving normally, so I am unsure if that is a normal thing to see happen every so often while driving. It might also be worth noting that we don't actually have any 5V stuff wired up to the roborio at the time.

Seeing "a bunch of 5V brownouts" suggests that you have debris shorting the 5 volt rail on the RoboRIO. It might not even be on the inside of the housing. Look very closely with a bright light at the recessed areas around the PWM, DIO, and Analog pins. You might find a nearly invisible strand of wire, or a crumb of metal, or a thin film of aluminum dust, that sporadically connects the 5V and GND pins together.

[Chris Mounts]

For what it's worth we discovered a few weeks back that hitting some of the defenses, for example the rock wall, can cause a hard stall depending on angle of approach, tires, etc.

We had made a mistake in putting 8" numismatics on the stock KOP chassis with only slightly adjusted gear ratios. According to the JVN calculator Excel we would be at over 90 amps per motor.

The power draw was so high in this situation that we observed arcing across the primary breaker polls. This lead to many brown outs most typically the radio dropping.

What a lot of you are describing sounds very much like what we experienced. Make sure you've run the numbers on your drive train.

[BigJ]

We tried everything suggested, cleaned out the roboRios, etc. We were about to call Al when another mentor tried yanking the connection to the main breaker pretty hard.

After uninsulating it turns out our connection was loose there on both robots, but not loose enough to disconnect with our normal tug test. the shock of running into defenses must have provided strong enough force to jostle it and momentarily disconnect.

After tightening and reinsulating we went over rockwall/ramparts roughly 10 times without problem where before we would disconnect on the first or second try.

Thanks for the help everyone!