Robot Power Issues

[404'd]

During our competition at the 10,000 lakes regional, we experienced power issues only during the final rounds, after taking a heavy hit. We did notice that the signal light went out once, and we lost total control both times it stopped. During one of the outages a FMS guys told us our cRIO had been rebooting. Also during the same outage, the camera cycled power. The first time the power cut out, we regained control after a few seconds. (YouTube video http://www.youtube.com/watch?v=5m0em...el_video_title @1:22)The time the cRIO rebooted, we never regained control.

We brainstormed some and have some possible explanations for the outage, including main power cables, power board, and crio power connector/cable. If anybody has any ideas, feel free to post them.

Thanks,
Josh -967

[electron]

A good number of teams experienced this issue last year while traveling over the bumps on the field. (especially robots like ours that did not have a smooth ride over)

We found two things helped:
First: mount the cRIO on some kind of shock absorber. (we mounted all of our electronics on a few layers of anti-slip floor matting. it seemed to work fine)

Second: we noticed that our main breaker was acting very funny. replacing it with a brand new one seemed to help quite a bit. In fact, I personally believe that was our only issue.

Hope this helps!

[Liz Smith]

In that particular YouTube video that you posted I can say that I don't think you ever fully lost power, or wireless communications...

I can tell because the green light on the scoring table stayed green the entire time (save for one second where the ref walked between the light and the camera taking video), which tells me that all the robots on the field were still connected. If your CRio reboots, you will show a loss of communication with the field and that light would change color. How that helps? I'm not sure, but it could point to two completely separate non-related issues.

Any video of the other match?

[slijin]

Seeing as how the full safety inspection is supposed to include this, I doubt this is the case, but are your cRIO and D-Link both wired to their respective dedicated 24v/5v outputs?

[scottydoh]

Originally Posted by slijin

Seeing as how the full safety inspection is supposed to include this, I doubt this is the case, but are your cRIO and D-Link both wired to their respective dedicated 24v/5v outputs?

Thats exactly what I was going to suggest. A lot of teams seems to be having trouble with this whole powering the D-Link through the dedicated port thing...

[404'd]

Originally Posted by Liz Smith

In that particular YouTube video that you posted I can say that I don't think you ever fully lost power, or wireless communications...

I can tell because the green light on the scoring table stayed green the entire time (save for one second where the ref walked between the light and the camera taking video), which tells me that all the robots on the field were still connected. If your CRio reboots, you will show a loss of communication with the field and that light would change color. How that helps? I'm not sure, but it could point to two completely separate non-related issues.

Any video of the other match?

This video showed what happened the first time, only a short loss of communication. The second time, one of the FMS people came over and told us that our cRIO had rebooted.

One more thing, our cRIO and D-link AP power converter are plugged into the 24v plug-in on the power board. I don't think we are pulling to much current through that connector, because we never had any problems through qualification and elimination rounds. We'll be checking all connections at St. Louis to see if we can find anything obvious. A short could be a very possible thing after we got hit during the round.

[Kevin Sevcik]

Here's a highly relevant thread from last year:
http://www.chiefdelphi.com/forums/showthread.php?t=84725
I'd lean towards one of your motors having developed a short or a short to ground that's intermittently blowing the breaker to your cRIO, or possibly your digital sidecar. You're quite clearly not losing main power. If you'd lost all power, you'd be down for a good 45 seconds or more while you waited for the bridge to reboot and reconnect. I'd check resistance from all your motors to your robot frame and try giving your robot a 1 foot drop and see if you lose communications. If you do, start pulling breakers for motors till you find one that seems to fix things, and inspect all the wiring after that breaker.

[scottydoh]

Originally Posted by 404'd

This video showed what happened the first time, only a short loss of communication. The second time, one of the FMS people came over and told us that our cRIO had rebooted.

One more thing, our cRIO and D-link AP power converter are plugged into the 24v plug-in on the power board. I don't think we are pulling to much current through that connector, because we never had any problems through qualification and elimination rounds. We'll be checking all connections at St. Louis to see if we can find anything obvious. A short could be a very possible thing after we got hit during the round.

Emphasis mine.

Wait a minute...Might just be a typo, but the CRio should be in the dedicated 24v port, and the 5v converted for the D-Link should be in the dedicated 12v port next to the one for the CRio.

[Captaindan]

sounds like you need to look into your wiring which should all be clearly labeled and easily traced to the crio by the way the crio is made to stand several g's worth of impact so its all about solid failsafe wiring

[Al Skierkiewicz]

404,
The stop at 1:22 follows a sharp turn. During turns, if your robot has significant friction, the current skyrockets. The Crio software monitors your battery voltage and when the voltage falls (or is pulled down) to 5.5 volts, the Crio output is disabled. If the battery falls to 4.5 volts for some time, the regulator in the PD will stop. When this occurs, the camera, radio and Crio all will reboot. The Crio takes anywhere from 20 seconds on up and the radio takes 52 seconds to reboot.
While I can't really see your drive, it appears to be standard tank with four wheels. In this type of drive, the motors are in near stall conditions in turns. Add several turns to each match and the battery is nearly exhausted near the end of the match. If you are using four CIMs, that is over 400 amps during the turn. Omni wheels are one way to eliminate this frictional loss.

[Nemo]

Al,

Our robot has a mecanum drive, so sharp turns should be less likely to cause low battery voltage. The possibility of momentary low voltage to the cRio is one of the possibilities to consider, though.

Thanks for all of the really helpful responses so far. This is already giving us some items to check off when we get to St. Louis.

Nemo

[404'd]

Originally Posted by scottydoh

Emphasis mine.

Wait a minute...Might just be a typo, but the CRio should be in the dedicated 24v port, and the 5v converted for the D-Link should be in the dedicated 12v port next to the one for the CRio.

Sorry, my error. The dlink is plugged into the dedicated 12v plug next to the 24v, where the cRIO is plugged in.

[Al Skierkiewicz]

Nemo,
It is possible that some of the rollers in the drive are sticking or dragging which gives the same effect. Did you ever try to crab (left or right) and also reset? I have had more than one team report their hardware on their mecanum wheels suddenly came loose.

[Jared Russell]

How many motors (and what kind) are on your bot? In Florida, we were running 4 CIMs, 1 FP, 3 RS-775s (that developed case shorts, providing low resistance paths between the battery leads), and the compressor, and at times we browned out the cRIO when everything was running at/near stall at once. The heavy current draw is even worse when you have poor or loose connections in the current path and/or BaneBots motor case shorts (adding to the resistance and therefore voltage drop).

Perhaps try watching your battery voltage if/when you can reproduce the problem. Plotting the battery voltage (with LabView or a Dashboard program) is even better. If you see it dropping (even momentarily) south of 5 volts, you risk causing a cRIO power cycle (the PD board requires ~4.5 volts to maintain nominal voltage to the cRIO).

[Jon Stratis]

A few other things to check:
- Make sure nothing inadvertently shorted to your robot frame. You said this started happening after a hard hit... it's possible that hit knocked something loose and short to the frame. After that, all it takes is one loose thing to cause a second short to the frame and momentarily cause a huge power drop.

- Check your main circuit breaker. There was a bad batch of them last year, and some may still be floating around. Turn on the robot and, while watching the LED's on the power distribution board, tap the red button on top of the main breaker. Not hard enough to turn it off, but don't go very soft either. If the lights on the PDB flicker, you probably have a bad main breaker and should try replacing it.

- Finally, try driving it around and recreating the issue during practice. Have someone watch the battery voltage readout on the driver station to see if it suddenly drops while performing any specific maneuvers. As Al has already said... issues in the drive train (loose chains, problems with the mecanum rollers, etc) can cause momentary binding that leads to a huge amount of stall current going to the CIM's, momentarily dropping your voltage below what's needed to keep the cRio powered.