Thread created automatically to discuss a document in CD-Media.
Window Motor Lockups: Causes and Solutions by: ratdude747
This paper explains the causes of Denso window motor lockups and the procedures required to avoid them.
This is a tutorial that I wrote for FRC 2783 regarding the procedure required to legally remove the locking pins from window motors.
This paper covers the cases of Jaguar-related window motor lockups, the causes of left/right combination lockups, and the procedure required to avoid the lockups, including a pictorial tutorial on locking pin removal.
I’d like to note that there was previously at least 1 other pictorial guide on the pin removal procedure, as I learned the procedure from it last year. That said, when I was asked about the procedure this year, I could not locate it on CD-Media, so I decided to re-create and expand on it.
I only know of one robot where the jaguar-related lockup occurred, that being a ball kicker on 2783’s 2012 bot (why I this paper was wrote in the first place).
In that case, I was told that the intermittent issue stopped occurring after an emergency swap from serial-CAN to PWM. I am pretty sure that all the Jags on that bot were black Jags.
We had lockup issues this year on our hood. Our controls operators was able figure out a pattern to get it to take commands again after it would happen but it was an issue. I thought it was a myth for awhile as well but I can assure you that it does exist.
But locking pins would seem advantageous in such an application. The expected load is not likely enough to strip the gears and maintaining orientation of the module mechanically will reduce load on the motors.
Most Modules do not have that issue, as the wheels are evenly centered with the module’s axis of rotation. If the modules were more like casters, then there very well could be a backdriving issue.
We used window motors with the pins for steering at least one season. The only time it was a problem is when we slid into something with the wheel turned sideways. If the output of the gearbox is turned (like trying to pry the window down on a car), the pins would lock enough that the PID loop output couldn’t turn it. Steering into the lock would always unlock enough to recover so it wasn’t a big problem but since we’ve been removing them driving smoothness improved noticeably.
Thanks for getting this where more teams can find it.
Larry,
I believe the problem manifests itself when there is enough backlash in the system that the motor can vibrate with the output current waveform generated by the Jag. Systems such as steering in a crab drive system tends to damp any vibration that might occur. From those that have reported to me or here on CD it would seem that more than half of the teams that use the Jag/window motor combination experience the problem.
As reported, the difference between the left and right hand motors has more to do with the wire size the armature is wound with than actual performance. The lefthand motor getting significantly more use in standard operation then the right.
Nice job on the paper and the pictures it contains. This should allow more teams to overcome this small shortcoming by removing the pins. I would suggest you add the length of time this operation would normally take. Since it is so easy, more teams would take your advice and remove the pins. I would suggest that teams also keep the pins in a safe place so they can reuse them if needed.
Noted. Most of my information was based on random noting from threads over the past year or so. The sad part is that most of the details regarding the cause of the lockups is based off of observances…
IIRC most of my knowedge regarding that left-right issue was from one of your posts. Ironically, during the 2011 season, 1747, the team I was on before my family moved, asked me why their arm based on a left-right design was overheating so much… It was bewildering at the time; hence why I thought it was worth mentioning as a potential cause.
“Performance” may have been the wrong word; I was trying to imply that the motors weren’t quite compatible with eachother so to speak. Any suggestions on a better word or phrase to use?
TBH I never timed the process. I have only performed the procedure twice, both times as a student in 2011. The work on the motor was mostly done by a student that I was instructing. However, since I did do some of the work (as a demonstration), I was sure to make sure the motor used was a dead one; I wanted any motors that would see use worked on by the students.
Since I currently am not a permanent mentor (the photo shoot was by special request), I personally do not forsee myself getting a chance to time the job. Perhaps other teams that perform the procedure could post about how long it took them (I’d take a guess but in that shop, time flies).
I hadn’t thought about that… Noted. I may add that to a future revision.
It’s good that this came up again since some time has pasted. For the 2010 season, our team developed a swerve 4 wheel independent drive, 4 wheel independent steering continuous rotation bot. We used the Denso window motors for the steering. We had several problems. It wasn’t until the software issues were dealt with that we began to notice a problem with the window motors locking up and cutting out by the thermal protections built into the motor. We found out about the locking pins and removed them. Our steering motors were more responsive and ran cooler. Note that we were using them for positioning controlled by a PID. The window motors were driven by a Black Jag. Even with the pins removed we had intermittent shutdown issues. The problem was not full resolved until we replaced the black Jags with Victors. Since going with the Victors we have not had a problem. We still drive the bot to this day and have not had another incident. After this experience I would suggest that teams that are using window motors for a positioning application remove the locking pins and use a Victor not a Jag. For 2012 we used a window motor with the pins in for our bridge lowering mech driven by a Victor. No locking problems. The PTC thermal over load did shut down the motor when it was needed preventing the motor from releasing the magic smoke. The pin removal is not needed for ever application. However, There does seem to be a problem with the Jags. I think the high switching frequency of the Jag leads to self heating of the PTC. Also note that I believe the Jags for 2010 used locked anti phase control of the fets. The recent Jags I believe are low side switchers using sign magnitude modulation. Never tried a recent jag with a window motor. Ever team should Know of this issue so it can be dealt with if it arises. Also important since First has allowed other automotive units to be used this year. Many of the common units do have PTC protections built into the motors. The interaction of PTC’s and jags should be followed and notice given if problems arise.
Link to original thread on Chief Delphi. It’s rather long showing the round about discover of the problem. http://www.chiefdelphi.com/forums/showthread.php?t=83973
Larry,
I could see that using a right and left motor with the pins in place could lead to some locking of one or the other of the motors when used with either controller. However, I have also seen a number of teams use this configuration for arms without any reported difficulty. In those cases the teams intended that the anti-back drive keep their arm in place. They seemed happy with the performance.
In talking with John Novak who did some testing with these motors, it really does appear that the very low mass of the pins (I believe they are phenolic) can easily be set in motion by the switching frequency of the Jags when the motor is not actually moving.
Gary, I had not considered the self heating at 15kHz of the PTC as an issue. However, I think you may be right. If any teams can give us some more data on this, it would be appreciated.
I am going to guess that this operation takes about fifteen minutes per motor. Would that sound accurate?
I see. 1747 did have overheat issues as the result of mixing left and right motors on thier 2011 arm joint. They had locking pins in 1 of the two motors (it also happened wit 2 pinned motors). They ultimately resorted to attaching 2 victor fans to each motor to make the arm usable.
-Added a time estimate (15 minutes/ motor) and note about saving the old pins as suggested
-converted the JPEG compression to lossless
-Added my new personal Logo to the end.