We are having an issue with our robot working great for about a minute before the battery starts to drain rapidly. It continues to drop to about 11 volts at around 3 minutes of hard defense before it starts to jitter like seen in this video.
Also shown in the video is that nothing else is able to be controlled while it is happening. The motor controllers run green yellow red rapidly when this is occurring. We have changed out the breakers and checked all the wiring connections and are still not able to figure it out.
You are drawing enough current when you are trying to turn that you are dropping the battery voltage to the point where you are browning out. 3 minutes of hard defense might very well do it, especially if you scrub when you turn your drive train and spend a time pushing without spinning your wheels (stalled).
Make sure you don’t have any mechanical binding in your drivetrain, make sure your gears are greased, sprockets / belts aligned and tensioned, and all your bearings are good. Double check you’ve got drive from each one of your motors individually (just pull breakers one at a time to make sure all your drivetrain cims are working). One way or another, if you’re starting with a fully charged battery you’re draining it quickly.
We found with the Andy-Mark kitbot wheels the center wheel wore fast enough that by the end of a competition all 6 wheels were rubbing the carpet at once causing a lot of scrub when we turned.
Would you mind posting the Driver Station Log from the video (or any instance of this happening)? The logged battery voltage is something I would like to take a look at.
My first instinct is to check the battery and how many Amp-hours it will still hold since it has been used. A brand new battery is generally at 17 Amp-hours. The lower the A-Hr, the faster the voltage on a battery will drop. A full battery dropping to 11V after 3 minutes is not unheard of, but not common either.
Does anyone see the RSL (Robot Signal Light) flash in the video? I am not seeing it flashing or are momentarily blind (or both). The jitter could be a low battery undervolting the sidecar, but without the Log file, it is just speculation.
Any FTA’s, FTAA’s, CSA’s, or anyone else are more than welcome to throw out other ideas.
I can try and get the logs on thursday when I get to the robot again and try and upload them. The rsl is not blinking because of a broken connector which is on the list of things to fix. I do know that when it does jitter the voltage goes down to around 4.5, which I could see messing with the sidecar. I will check to make sure all the cims are working properly as well.
Good to hear you still have the logs. If the battery voltage is dropping down to ~4.5V during the jitter, I am quite confident that that is most of the problem. If the voltage to the sidecar drops below a certain point (I think 5V), the sidecar will not be able to generate the PWM signals correctly.
I agree with what others have said. I suspect that the log file will show frequent drop and recover of the battery voltage. The drop is due to large draw by the motors at stall. The recovery is because the cRIO FPGA cuts the outputs in order to avoid reboots by critical systems such as the radio and cRIO.
Verify that the battery is good, verify it is fully charged, and see what the voltage drop is when you initiate a turn. If it isn’t hard, swap the wheels or place new ones on the center. I can’t see much on the robot, but I assume it has dropped center wheels.
You should also check all of the battery power connections. Check the lugs on the power distribution board and each of the batteries. You could have loose connections that causes heat. How warm are the battery cables after you turn off the robot?
If the battery voltage drops to 8 or below you will run into other issues. We were told by a FTA that cRIO will crash which may not be recoverable. Check for pinched PWM wires, if moving parts pinch and short for short duration, the circuit breaker may not get activated, but battery drains out.
Related to the above, can you check the drive motor circuit breakers to see if they are tripping? You should hear an audible clicking. Also, the PDB has a red LED which should light up when this occurs.
This is not a classic demonstration of battery voltage brownout but it may be complicated by a frame short somewhere. When you are listing your battery voltage are you telling us the dashboard reading or are you measuring the battery directly? I would check all wiring first, starting with the power to the DSC and the multi-conductor between DSC and cRio. The rapid on/off that appears in the video is likely caused by a circuit breaker rapidly resetting. That is usually caused by a short or a partial or intermittent short. Coupled with a rapid battery discharge, a pinched cable (power to the DSC) is the likely culprit. If you have any servos, check for a pinch in those PWM cables since they are fed from the DSC. If it is a short, then something is getting warm. Carefully feel around all wiring for some thing is hotter than the others.
What is your wheel setup? I would expect this issue with 4 high traction wheels in that configuration. If you have omni’s on front or back, I would not expect this behavior due to the drivetrain configuration.
Edited for clarity:
**If you do not have omni’s on one set of wheels, you will stall your drivetrain motors, likely pulling too much current and tripping the breakers.
Don’t discount a mechanical issue as a result of an electrical symptom.
If this occurs after three minutes of heavy drive motor usage, you could likely have a dead battery. If you have a Battery Beak, check the condition of the battery when it exhibits these symptoms. Hopefully, it will point you in the right direction.
This is very strange, and I don’t know the cause other than low battery, but something I know is that we would have the same problem from time to time, and I’m pretty sure the robot is enabling and disabling at the same time while this happens, as whenever this would occur in a match, our pneumatics would retract and extend with the twitches, if they were extended beforehand. Also, our LEDs would flicker greatly, which were directly wired to the PD board at 20 amps. I don’t know if this would help, but just contributing if possible.
A cRIO crash due to low voltage or otherwise should not brick the cRIO. Every time you open the Anderson switch to “turn off” the robot, you are removing power from the cRIO. It isn’t aware you were going to do that, and it runs out of juice and has an abnormal shutdown – a crash.
When voltage drops below a threshold, the cRIO disables outputs. In its status packet back to the DS, it explains that it cannot run outputs because it is disabled. This is similar to what happens with a watchdog timer. The DS it telling the robot it can be enabled. The robot is saying it chose not to be. I don’t believe rebooting the DS has any effect on this. I don’t believe it hurts anything, but I don’t believe that it fixes issues with a low voltage battery in anyway.
Please post the logs and that will give some data to confirm our speculations. Or it will hopefully point in another direction.
I will post logs ether tonight or tomorrow morning due to the fact i don’t have the robot or driver’s station at the moment. I am mostly thinking it is a short like many people have said. I will check to see what gets really warm when we run it today and see if it is in fact the power cables for the DSC.
The battery leads do get somewhat warm as well as all 4 Drive cims, The main breaker does not get hot as well, this is why i am pretty confident is just a short somewhere.
Thanks for all the help.
We run a 8 motor drivetrain (6cim 2minicim) and this happens whenever our battery is low (Which can easily happen after 3min hard defense). I don’t believe it is any problem with your robot or wiring, instead the battery at low levels is simply unable to provide enough current for 4 cims to run at full from stop, especially when turning and running a compressor. When you attempt to draw to much current from a low battery, the voltage can drop so low that your robot stops running commands. (I believe the minimum is around 5 volts). The constant starting and stopping is due to the motors drawing to much current, losing control because of low voltage, stopping, regaining control, then repeating. If you just try slowly accelerating you should be fine, but you should just change battery.
watch the driver station - if your ever dropping below 5 volts you’re probably at low battery. (or you’re stalling motors, inefficient drive base, etc…)
It’s unlikely because you mentioned it happening after 3min of hard defense, but this can also happen because of communication interference when to many wifi networks are present. You’ll see the robot enable and disable erratically on the driver station. In this case one fix is to switch to 5ghz wifi.
hope this helps.