Wireless Packet Drop during competition

[boz336]

Wondering what could cause wireless packet drops during competition.
On a couple occasions we lost comms for about 30 seconds.
We talked to the FRC technical volunteers who did their best to help us diagnose, but we're not entirely sure and we have another competition next week. For reference we're using the new Robo Rio and a brand new DLink.

Their suggestions were:
1) Our mecanum drive system is drawing down the battery voltage to below the brown-out level. But this one was de-bunked after we went into the log file and looked at the voltage levels and confirmed we were above brown out.

2) Some "other" event caused the issue. Two theories were A) Wiring connection came loose (ethernet, or power) and B) our Dlink was not oriented correctly.

We re-secured our cabling, but had already duct taped all connections in place and they don't appear to move when the robot is moving around.
Our D-link isn't near any motors or other sources of EMI. Can the orientation of the D-Link make a huge difference? Vertical vs. Horizontal vs some random angle?

Thanks for suggestions!

[Daniel_LaFleur]

Originally Posted by boz336

Wondering what could cause wireless packet drops during competition.
On a couple occasions we lost comms for about 30 seconds.
We talked to the FRC technical volunteers who did their best to help us diagnose, but we're not entirely sure and we have another competition next week. For reference we're using the new Robo Rio and a brand new DLink.

Their suggestions were:
1) Our mecanum drive system is drawing down the battery voltage to below the brown-out level. But this one was de-bunked after we went into the log file and looked at the voltage levels and confirmed we were above brown out.

2) Some "other" event caused the issue. Two theories were A) Wiring connection came loose (ethernet, or power) and B) our Dlink was not oriented correctly.

We re-secured our cabling, but had already duct taped all connections in place and they don't appear to move when the robot is moving around.
Our D-link isn't near any motors or other sources of EMI. Can the orientation of the D-Link make a huge difference? Vertical vs. Horizontal vs some random angle?

Thanks for suggestions!

30 seconds sounds like a d-link reboot.
Are you using the original power cord that came with the d-link? I've heard that there are other connectors that look like they fit but will cause the d-link to loser power during vibration.

Orientation does little to the d-link, but location (deep inside the robot, surrounded by metal) may.

[boz336]

I'm pretty sure we are using the original D-Link connector, but will double check. Thanks for the suggestion.

I did look at some pictures of innards of the D-Link to see where the antennas are located. It looks like they use a dual antenna configuration for diversity... with the antennas in the corners of the enclosure.

https://q4rtmw.bay.livefilestore.com...902.jpg?psid=1

Ant 0 is in the upper left corner of the picture and Ant 1 is in the upper right. Both antennas are on the same side of the D-Link as the LEDs (I guess you'd call it the "front").

[MrRoboSteve]

30 seconds does sound like a radio reboot. Do you have a screen shot of the log from a match with the problem that we can look at?

Is the radio connected to the correct output on the VRM? Is there anything else connected to the VRM?

I would also mount the radio so that it is easy to see the lights from the side of the field, and let the FTA know about the problem before each match. If you have a problem, they can go look at the lights which could confirm a radio reboot.

[plnyyanks]

Originally Posted by MrRoboSteve

I would also mount the radio so that it is easy to see the lights from the side of the field, and let the FTA know about the problem before each match. If you have a problem, they can go look at the lights which could confirm a radio reboot.

Definitely do this - the field staff are very good at figuring out what's gone wrong, but that's hard to do if they can't see the status lights.

Like others have said, 30 seconds does seem like a radio reboot, and they often do have power issues. Have you checked to make sure your PDP fuses are pushed in all the way (and if you can get them out without hand tools, they're not in all the way)? Here are some pictures for comparison.

And again, take a look at the DS log - you should be able to clearly see where the incident occurred. What's the battery voltage right before you dropped? Does anything else look strange?

[rich2202]

It could also be a RoboRIO reboot. Check the RoboRIO power connections. Also check for debris in the connection channels (and MXP port). Shorts there might cause problems.

[Jarren Harkema]

This occurred for us a few times during early playoff matches. Looking at the logs, voltage drops and packet losses were occuring at the same time. We never lost control for more than a few seconds though, and FMS showed communication with the robot the entire time. It seemed to occur when we accidentally drove our lift down on top of another tote, tipping the robot up and drawing a lot of current. After adjusting driver performance to eliminate the chance of that, we never lost an abnormal amount of packets. It was then tested while tethered in queuing, and we could cause packet loss by stalling our lift motor. We still aren't sure why this is causing packet loss.

[plnyyanks]

Originally Posted by Jarren Harkema

This occurred for us a few times during early playoff matches. Looking at the logs, voltage drops and packet losses were occuring at the same time. We never lost control for more than a few seconds though, and FMS showed communication with the robot the entire time. It seemed to occur when we accidentally drove our lift down on top of another tote, tipping the robot up and drawing a lot of current. After adjusting driver performance to eliminate the chance of that, we never lost an abnormal amount of packets. It was then tested while tethered in queuing, and we could cause packet loss by stalling our lift motor. We still aren't sure why this is causing packet loss.

Check your PDP fuses

[Shu]

Originally Posted by Jarren Harkema

This occurred for us a few times during early playoff matches. Looking at the logs, voltage drops and packet losses were occuring at the same time. We never lost control for more than a few seconds though, and FMS showed communication with the robot the entire time. It seemed to occur when we accidentally drove our lift down on top of another tote, tipping the robot up and drawing a lot of current. After adjusting driver performance to eliminate the chance of that, we never lost an abnormal amount of packets. It was then tested while tethered in queuing, and we could cause packet loss by stalling our lift motor. We still aren't sure why this is causing packet loss.

We had the same thing happen in our first semi-final match causing us to drop a container and wrecking our elevator when the arm came crashing down on top of it. It was interesting that they chose to replay 2 of the quarterfinal matches but not ours. Fortuneatly for us, we were able to do a quick repair and score enough points in our next 2 matches to get to the finals.

[Jpatterson1710]

Originally Posted by Shu

It was interesting that they chose to replay 2 of the quarterfinal matches but not ours.

FTAs typically would need three or more robots to have connection/lost packet issues before calling a rematch. One or two robots having issues very rarely is caused by field or FMS fault. Radio Brown Outs are a huge issue for many teams, and while frustrating, it is by no means a reason to replay a match.

[boz336]

Thanks for all the comments and suggestions.
We had ONLY the Dlink on the VRM radio power rail.... so we're pretty sure that wasn't an issue. However, we did realize too late that our orange status light was not connected correctly. It was plugged into the VRM instead of the Roborio...so the FRC volunteers were'nt catching it as it was happening.

We will certainly check the fuses and other connections very carefully. Also, we're planning on backing off the PWM level to the drive motors to maybe help prevent possible brownouts

Fortunately, we were able to scrape together enough money for a second regional... so maybe these suggestions will put us in a better spot next week. I didn't think to grab a screenshot of the log before bagging the robot, but may try on Thursday at check in. If so, I'll post it then.

Thanks a bunch for all the good feedback!

[boz336]

https://drive.google.com/drive/mobil...FWjlTYlNROF E

Here is a screenshot. It doesn't look like we have any big battery dips... or even packets lost.... but we were definitely dead for several seconds. Solid orange indicator light on the robot rather than blinking like it normally does. Apparently we haven't fixed the robot yet.

[Maldridge422]

Originally Posted by boz336

https://drive.google.com/drive/mobil...FWjlTYlNROF E

Here is a screenshot. It doesn't look like we have any big battery dips... or even packets lost.... but we were definitely dead for several seconds. Solid orange indicator light. Apparently we haven't fized the robot yet.

Screenshot link isn't working for me

[boz336]

Please try this link.

Thanks!

[boz336]

You can see the high cpu% in the 15 second auto period but I dont see any obcious problems after that. It doeant even look like were logging packet drops...but were DEFINITELY not able to control the robot.



Thanks!