Losing robot comms over bumps

Hey friends,

Our robot seems to not appreciate going back and forth over the bumps on the field. We have tried diagnosing it ourselves and can not seem to get it to reliable hold a connection going over a piece of metal the same size as the one on the field. We lost communication and robot code for about 35 seconds on DS before we are able to re-enable and drive again.

Here’s what prints to the DS console

Warning  44002  Ping Results: link-bad, DS radio(.4)-bad, robot radio(.1)-GOOD, roboRIO(.2)-bad, FMS-bad FRC: Driver Station ping status has changed.  Driver Station

We have tug tested wires from PDP to roborio and have changed multiple things about them. The fuses on the PDP have been replaced to ensure there is not an issue there. Let me know if there is anything else you would like to know and I can provide a DS log file if needed. We have also had this issue last year when sending our robot off level two of the hab platform. Thanks!

Are the fuses to the roboRIO pushed all the way in to the PDP? It is surprisingly easy to push them most of the way in, to the point where a strong jostle intermittently breaks the connection, but it otherwise works.

Here’s a presentation that covers common electrical problems that can lead to the issue your describe: Five Fatal Mistakes: Why your robot is dead on the field

Particularly have you checked all your 6AWG terminals to ensure they are tight? (PDP, main breaker, and all of your batteries)

1 Like

Hot glue anything that seems like it might become loose. It’s easy enough to pull apart if you really need to replace it, but strong enough to prevent those little VRM wires from falling out.

1 Like

See I’ve heard that a lot and we pushed it in as far as we possibly can with our hands, is there some tool we can use to better push it in because it’s not going anymore with just pushing it in with all of our strength

We just redid all of our batteries at the beginning of build season and I tightened down the terminals on the PDP and the ones on the breaker aren’t moving either. I did notice a little give in the PDP terminals and I tighten it down after making this post but I would be very shocked if that magically fixed all of our communication issues, but I’ll know for sure Wednesday.

From what I have seen, usually when you loose connection for about 30 seconds after being bumped it is because the radio is rebooting because of a loose barrel connector. I would recommend powering your radio with power over Ethernet if you are not already.

2 Likes

Next time u do it watch to see if the rsl turns off, if this turns off u have a problem with power before the Rio. If it stays solid the whole time my bet is the barrel connector is coming loose in the radio. You can try and use a POE injector cable to limit this.

3 Likes

If you’re still have disconnection issues at your next event, make sure to get your local CSA to look over your wiring. They’ve likely been around the block so they know what to check for. And it’s a lot easier for them to diagnose the problem in person than for us online.

1 Like

Thanks for the tip, I believe the RSL is turning off and we’re already using the PoE and barrel jack for the radio.

The error message was very informative and explicit.
The radio (GOOD) is fine and stays powered, but the RIO (bad) is losing power.
It’s not a total loss of power which would also reboot the radio, but there is enough of a voltage dip occurring somewhere in the chain of electrical connections from the battery to the rio to cause the rio to reboot.

We had a similar issue and found that our ethernet cable connecting the radio and RoboRio was becoming momentarily disconnected when going over bumps or breaking quickly. Properly securing the ethernet cable, RoboRIO and radio fixed the issue for us.

If you’re still having issues, a log file may be helpful to some folks on the forums. I also found this WPILib docs article to be very helpful when trying to decipher the logs.

possibly a bad Ethernet cable/connection, Bad connection between the PDP and the RoboRIO (power cables to be specific), or the barrel connector is falling out. make sure you connection isn’t tight and straining the cable. I would recommend a POE for the radio instead of using the barrel connector as we have found barrel connector to fall out without some hot glue. make sure that your RoboRIO has a good connection with your PDP and isn’t turning off due to a loose connection at ether ends of the power cable. If you think its the power cable for the RSL, you could simply hot glue it onto the RIO although I wouldn’t advise it.

It’s mostly likely a loose connector, but when you abuse your robot over aggressive surfaces, other things can happen, too. During Stronghold we ended up with something coming loose inside our PDP, which we discovered using GM’s longstanding “shake test”. (If your electronics rattle when you shake them, replace them.)

So do what everyone else said first–seat the connections and breakers really well, hot glue them in, make sure you don’t have any shorts from stray strands of wire bridging the weidmuller connectors, etc, etc–and if none of that works, try swapping your PDP.

Last year we had a similar issue and found the issue to be our breaker connections being loose. If you haven’t already make sure those bolts are securely connecting your wires.

Something that hasn’t been mentioned yet here is the “breaker test”. I have seen a handful of breakers over the years that will break connection when being jostled too hard. Ensure your robot is on and tap the top of the red button and look for lights to go out/rebooting issues. If the connection breaks with just a small amount of force, throw the breaker away and get a new one.

1 Like

You can rule out the first and third thing by reading the status message in the OP, where the radio connection is good / has power. A loose Ethernet cable wouldn’t take the full amount of time it takes for the roboRIO to reset to reconnect.

In addition to the things to check shown in the excellent presentation linked by @forbes, use a bright flashlight and look at the wires going into the various Weidmuller connectors from various angles to look for anything shiny. It is common for there to be a strand or two that don’t go down into the hole. These can short to the adjacent wire.

This topic was automatically closed 365 days after the last reply. New replies are no longer allowed.