Now that can’t be a good sign: I am no electrical engineer, but when a motor gets hot, that is not a good thing. What could be the possible source of this? Now they seem to be hooked on all right, but one gets hot while the other one stays cold to the touch. Something must be short circuited. And while I was testing the motors today, I saw sparks jump out from the corner of my eye (I was staring at the computer screen, so I do not know the cause of the spark) It is not software that is giving it a problem, I am giving them the same voltage. Would it be the Black Jaguars? Only the black jaguar one seems to heat up, that was the case for the other motor that burnt out. Seems like the most obvious, but I would assume the newer ones were more “stable” than the beige jaguars.
Where are you using the motors? Is one getting more load than the other? I see no reason for one of them to heat up, and the other not: It is impossible, given the amount of force, for one window motor to drive another one, due to the worm gear and anti-backdrive mechanism.
Of course, thats assuming that they are both hooked to the same shaft. Again, what are they used for?
No load, just on a table hooked up to the Jaguars using CAN. Giving both the same voltage. The only difference is that one is the beige and the other is a black jaguar.
my former team, 1747, was reporting to me that they have been having the same issue on their arm’s 2 window motors
apparently they only removed the locking pins from one side of the arm. the other motor has the pins… i wonder if that did it
did you remove the pins first?
Jags have issues with the windows motor locking mechanism due to higher output switching frequency (15kHz vs 120Hz). You can legally remove the locking pins by <R47-C>. However, since resistance to backdrive is about the only reason to use window motors IMHO (unless you are out of other motors), I suggest you use Victors to control them.
Is the motor casing hot, or is the transmission casing hot? Neither one is good but being a slight more specific in our observations might help towards troubleshooting this problem.
And I agree that using a Jag is a bit overkill (capabilities, space, weight, parts…) for the window motor, a victor or spike should work just fine!
Good luck!
Petrie
Oh, great, thank God there are people that agree with me. Another programmer and I have been going at it with each other since day one. I told him “lets use Victors, they are reliable, more robust and teams rarely have problems with them.” Keep in mind he is a rookie. His argument is that it is new technology and that using CAN would make us go to championships… Now I explain to him that no one cares how the motors are controlled or programmed, the only thing that would get us to championships is the performance on the field.
I see no real advantage of using CANs… I just need to manually get a closed loop working.
It is not that the Jag is overkill, the Jag combined with the window motor locking pins is a serious problem, one has to go. The motor is stalling because the locking pins do not always release when the motor vibrates at 15kHz (output frequency of the Jag) but do release at 120Hz (output frequency of the Victor).
[EDIT to reply to David]
The real problem is the windows motors were not designed to be run with any speed controller (Victor or Jag), so a spike relay is the best solution if you can do without variable speed control. They were designed to be run with a relay (on/off switch) like on your car window. As a safety feature (to prevent window related injuries to human appendages), they are designed not to run at stall by using positive temperature coefficient (PTC) resistor (which is what is getting really hot). Reconsider your application of these motors. Window motors are heavy, break easy (plastic output strips out), require an odd complier to interface to a shaft or a gear, and are the weakest motors in the kit (barring servos). Their only positive is that they are the only motor with a worm drive gearbox, and due to this Q&A ruling you can’t use the window motor gearbox with a better motor (seem to be the only gearbox that is considered integral to the motor). I have had first hand experience with several component using window motors, and basically every one has been unsatisfactory, often to the point where we redesigned with a different motor. I understand some teams have had success applications for them, particularly in steering swerve modules.
There are numerous advantages to Jags and CAN. Jags (PWM or CAN) are the only speed controller you can switch between brake/coast on the fly (we run auto in brake and teleop in coast) and their input to output function is more linear (easier to program control loops IMHO), increased update rate. CAN advantages are multiple control modes (include closed loop for speed and position), increased feedback (including reading encoders without using limited CRIO resources), daisy chain wiring with locking connectors (PWMs tend to come loose without clips). I use Jags for fine control (i.e PID) of certain components (i.e. drive & arms).
Victors have their advantages, too. They are smaller, lightweight and most veteran teams have a large stock of them. I use them for components that don’t require precise control (i.e. rollers, winches).
No speed controller is going to punch your ticket to Championship. They each have different features and you need to find what works for you.
And, just to add some more noise:
The window motors are, by design, intermittent use devices. They aren’t intended to be run for more then few seconds at a time. They will of course run much longer then that, but heat will accumulate. Just running with no load (other then the transmission) will warm them up. Running with any significant load will cause them to get very warm to the touch.
Luckily, like most of the automotive sourced motors in the past, they are pretty hardy. They won’t typically fry from ‘normal’ FIRST use. If things get overly warm, an internal thermal protection circuit will open and the motor will stop long before it’s damaged.
So, whether you are running them with a relay or speed controller, be aware that some heat is to be expected.
So having a closed loop system that constantly compensates is a bad thing for the motors?
In 2009, our team used an Ackermann steering system where we had the front wheels controlled by Denso motors. They got pretty hot because of the constant use. We ended up using heat sinks (about 1.5"x2"x.75") and fans (similar in size to the victor fans) to cool off the motors. We also used some thermal paste (arctic silver brand) between the motor and heat sink.
Also, removing the pins can cut down on the heat depending on what the motors are being used for.
I’ve heard that this can be fixed as simply as switching the Jag to Coast instead of Brake.
Not necessarily. Certainly it’s done frequently in FIRST without harm. Whether it works depends on the average load applied and the terms of the the PID loop. I couldn’t possibly hazard a guess as to whether your use is kosher without seeing it.
The simple test is: does it work? The motors PTC circuit will open before the motor is damaged, so that’s the budget your are working within. Over the course of two minutes does the motor heat to the point of the PTC opening up? If so, adjust the loop to reduce over controlling of the motor or reduce the load. When you can make it through the match with everything still working then you’re fine.
The nature of the PTC is that it acts like a breaker more then a temperature sensor. A hot motor ‘preloads’ the PTC, but it’s really intended to protect the motor from stall current, not over use (they can only run for a few seconds before stalling in a car, after all). It’s pretty rare to just plain overheat the motor.
If you’re really pushing it on match turnaround, and starting matches with hot motors, you can always spray them down with coolant. It won’t do the motors lifespan any great favors, but how long do they really need to work for anyways?
A couple more software tips:
-
I run a deadband at 0.4 on our window motors. The machine physically could not move outside of that range, so running them there wouldn’t do anything except heat them up.
-
I run an anti-death function that kills the output if the sensor does not change enough, and it has been at least 700ms since a state change. This prevents the motors from stalling when they cannot reach their target (We sometimes drive the target past the mechanical limits, so it will hit the limit for certain, since some of the design requires it to go as far as possible).
-
You shouldn’t get much (if any) fiddling by the PID loop if it is tuned right., since the motor can’t fall, any power it is applying is to move somewhere (not to hold its position).
-
We run all Victors, no Jaguars at all.
We had black jaguars for our arm, and it would cause the arm to lock up. It wouldn’t come out of the lock unless we put full power into the window motors which sent it slamming down or flying up. During and after the lock ups, the motors would get VERYVERYVERY HOT. After some research we switched to victors, and they worked better but for some reason when we would bring the arm down it would keep going when the joystick wasn’t being touched. Putting the victors in break mode fixed this. Now the worst that happens, the motors might get warm after running them for awhile. We were driving around using the arm for a good half hour and didn’t have a problem
We were running 2 window motors to power it
Hope this helps!
Guys,
There needs to be a little more info given before jumping to conclusions. David said he had the motors sitting on the table. David, the left hand and right hand window motors are not identical so merely programming for the same speed does not make the motors run the same. While initially it seemed that the window motors suffered problems with the Jags, it was not always the case. A lot has to do with the application. The locking pins are not a new phenomenon but have been around for a long time. John Novak reported opening a 2001 motor and finding them. They (the locking pins) are very small and light and could interact with the switching frequency except when at full throttle.
It is possible that the higher switching frequency of the Jags may cause additional losses in the armature. These are called ‘eddy current’ and can cause heating. However, I have not seen a discussion on this topic and have no data that either controller would cause heating.
David, if the motors were coupled together and one was doing much more work, it could/would get hotter. Meaning if one motor was slightly faster, it would be pulling the other motor along.
CAN will not get you to Einstein. We have a long history of success and don’t use the Jags so therefore we don’t use CAN. What gets you to a win is a robot that doesn’t break, can perform a variety of tasks, is driven by students who have practiced like crazy and a strategy team that collects the needed data for alliance selection.
An old musician riddle is “How do you get to Carnegie Hall? Practice, practice practice.”