CIM Motor Failure

[sanddrag]

Yep, been there done that. Not a continuous duty motor, not even close.

[Richard Wallace]

Originally Posted by sanddrag

Yep, been there done that. Not a continuous duty motor, not even close.

The CIM motor data sheet indicates that the motor is designed for intermittent duty. Its specified test load is 64 oz-in running from a 12V supply, which corresponds to 4320 RPM on the motor's speed-torque curve -- the mechanical load is about 204 Watts and the motor is dissipating about 120 Watts as waste heat (in its armature windings and brushes) at this load.

Its specified test cycle (for endurance) is:

3 minutes at the above rated load, counterclockwise, followed by
2 seconds off, followed by
3 minutes at the above test load, clockwise, followed by
30 minutes off,
and then the cycle repeats, until the motor has run 1000 cycles.

So the motor is designed to run for about 100 hours at 200 Watts mechanical output from a 12V supply, with 16% duty (i.e., 30 minutes off for every 6 minutes running). Exceeding this load and duty will overheat the motor. Seems like that is what Al's team did.

[akoscielski3]

Originally Posted by Richard

The CIM motor data sheet indicates that the motor is designed for intermittent duty. Its specified test load is 64 oz-in running from a 12V supply, which corresponds to 4320 RPM on the motor's speed-torque curve -- the mechanical load is about 204 Watts and the motor is dissipating about 76 Watts as waste heat (in its armature windings and brushes) at this load.

Its specified test cycle (for endurance) is:

3 minutes at the above rated load, counterclockwise, followed by
2 seconds off, followed by
3 minutes at the above test load, clockwise, followed by
30 minutes off,
and then the cycle repeats, until the motor has run 1000 cycles.

So the motor is designed to run for about 100 hours at 200 Watts mechanical output from a 12V supply, with 16% duty (i.e., 30 minutes off for every 6 minutes running). Exceeding this load and duty will overheat the motor. Seems like that is what Al's team did.

In the case of it being a drive motor, would only turning it the one way (counter clockwise) and hardly ever turning it clockwise reduce the life of the motor?
I sure hope this wasnt a 2012 CIM.
Also do you usually reuse CIM's from previous bots (competition bots) or only remove them from the Practice bots? We have previously destroyed some of our robots and reused CIM's but in the last few years (2011, 2012) I wont let them touch the robots. Though our practice bot will be taken apart and reused for spare parts and probably prototyping in the years to come.

[Al Skierkiewicz]

Aaron,
The CIMs do have a slight bias that tends to let them run in one direction a little faster than the other but there should be no significant wear based on that alone. We have killed CIMs in the past with lengthy demos and heavy use in practice. The failure exhibited above is kind of a domino effect. As the motor builds excessive heat the insulation on the windings starts to break down. When that occurs the windings start to short. This causes the current to climb while the output power falls.

[Brandon Holley]

Al-

Does the failure of a CIM due to overuse/abuse have any key indicators we can note externally (ie: while still mounted to the robot)? Or is it a slow decrease of output power over time that eventually becomes noticeable.

-Brando

[Richard Wallace]

Originally Posted by Brandon Holley

Al-

Does the failure of a CIM due to overuse/abuse have any key indicators we can note externally (ie: while still mounted to the robot)? Or is it a slow decrease of output power over time that eventually becomes noticeable.

-Brando

Originally Posted by Al Skierkiewicz

... The failure exhibited above is kind of a domino effect. As the motor builds excessive heat the insulation on the windings starts to break down. When that occurs the windings start to short. This causes the current to climb while the output power falls.

Brandon, I know you asked Al, but let me try. Al's post above indicated a possible in-situ diagnostic, namely motor current. If the motor is drawing more current while powering the same load (i.e., while the robot is performing the same operation), then it is either hot or starting to fail, or both. Allowing the motor to cool for a while, say an hour, and then repeating the same robot operation while observing the current will tell you whether the motor was merely hot, or failing.

Jaguars with CAN provide one method of monitoring in-situ motor current. WildStang published another method many seasons ago.

[Brandon Holley]

Originally Posted by Richard

Brandon, I know you asked Al, but let me try. Al's post above indicated a possible in-situ diagnostic, namely motor current. If the motor is drawing more current while powering the same load (i.e., while the robot is performing the same operation), then it is either hot or starting to fail, or both. Allowing the motor to cool for a while, say an hour, and then repeating the same robot operation while observing the current will tell you whether the motor was merely hot, or failing.

Jaguars with CAN provide one method of monitoring in-situ motor current. WildStang published another method many seasons ago.

Thanks for the reply Richard.

The reason I ask, is we've had a weird issue regarding one of our drive sides all season.

We use 2 CIMs per side as most teams do, into a custom 2 speed gearbox. Even before ship date, we would monitor how warm the CIMs were getting during our testing. For some reason, one of the (4) CIMs always stayed cool to the touch. The partner motor to that one on the same drive side would become warmer than the 2 CIMs on the opposing drive side.

Thinking this motor just wasn't receiving power, we went through the debugging process, power the drive side with just 1 CIM and then alternating to make sure they both were "driving". We then switched the Victors that were powering them to ensure they were both "driving".

The motors always appeared to be functional, as in the robot would drive. However, during eliminations, or heavy testing, the robot would begin to pull slightly towards the "cold motor" side. We replaced the motors, and the issue has generally gone away.

I'm wondering if we were dealing with a CIM that was damaged from overuse/abuse or had a manufacturing defect.

-Brando

[Al Skierkiewicz]

Brandon,
What you describe is usually caused by mis-calibration of the motor controllers such that one motor is doing all the work on that side. However. since you changed the motor and the problem corrected itself, the motor seems to be the obvious fault. There are certainly other issues that can arise in dual motor transmissions. Often these relate to mounting causing higher than normal frictions on one motor. Although the CIM motors are fairly well insulated (note the fiber insulation on the armature in the photos), it is possible to damage a winding in manufacture. All indications are that the cool motor was not performing the same amount of work as the warm motor.
Since the CIMs are sealed there is virtually no effective way to cool them. The majority of heat is built up in the armature. the only path for this heat to escape to the outside of the motor is through the end shafts, bearings and then to the end plates. Some heat can be radiated from the armature through the air to the magnet structure and case but this will produce very little cooling. As Richard stated above, these motors are designed for intermittent duty. Practicing for five minutes with a long cool down period is best. We are in training at the moment and require longer periods to achieve that result. The motor failure was the result of our aggressive schedule.

[Brandon Holley]

Originally Posted by Al Skierkiewicz

Brandon,
What you describe is usually caused by mis-calibration of the motor controllers such that one motor is doing all the work on that side.

Al- that was our thought as well, and recalibrating them was the first thing we did. Everytime it "acted up" we went back to that conclusion as well. It seems though, that was never the true cause (as you concluded).

Richard- Unfortunately, we didn't get to do any post-mortem on the motor because we swapped it at spare parts desk during one of our regionals for a fresh one. It would have been nice to know FOR SURE that the motor was the root cause, or to know what exactly with the motor was the root cause, however in the heat of competition, this was not a priority.

-Brando

[Ether]

...we didn't get to do any post-mortem on the motor because we swapped it at spare parts desk during one of our regionals for a fresh one. It would have been nice to know FOR SURE that the motor was the root cause, or to know what exactly with the motor was the root cause, however in the heat of competition, this was not a priority

What does the spare parts desk normally do with a CIM motor in that situation? Do they throw it away, or do they set it aside somewhere?

[Richard Wallace]

Originally Posted by Brandon Holley

..., we went through the debugging process, power the drive side with just 1 CIM and then alternating to make sure they both were "driving". We then switched the Victors that were powering them to ensure they were both "driving".

The motors always appeared to be functional, as in the robot would drive. However, during eliminations, or heavy testing, the robot would begin to pull slightly towards the "cold motor" side. We replaced the motors, and the issue has generally gone away.

This is an excellent example of debugging. Your problem was apparently solved by replacing the motor, so you now know that the PWM channel, the Victor, and the motor position on your gearbox are no longer suspect.

Did you measure the free current of the "bad" motor, after it was removed from the system? Often a bad motor will draw more current than normal when running free. If you don't have a convenient way to measure it, see me at IRI -- I have a test set you can use.

As Al pointed out above, it is difficult to cool CIMs. They don't have a flow-through path for cooling air, so the heat developed in their rotor assemblies stays trapped inside. The rotor heats up much faster than the case, so a hot case indicates a VERY hot rotor. In more quantitative terms, Al's "too hot for the finger" test indicates case temperature above 60 Celsius, and the pictures he posted indicate rotor insulation temperatures above 180 Celsius.

I've been a motor guy for a long time. CIM's are some of my favorite motors because they can really take the heat. But even a CIM can be overcooked. Thanks for the cautionary story, Al.

[akoscielski3]

Originally Posted by Al Skierkiewicz

Aaron,
The CIMs do have a slight bias that tends to let them run in one direction a little faster than the other but there should be no significant wear based on that alone. We have killed CIMs in the past with lengthy demos and heavy use in practice. The failure exhibited above is kind of a domino effect. As the motor builds excessive heat the insulation on the windings starts to break down. When that occurs the windings start to short. This causes the current to climb while the output power falls.

Is there anything we could do to prevent our motors from heating up easily, or to cool our motors down quickly? Like installing something on the motor, or doing something different in the electrical system, or even programming?

Second thing, what is CAN? I've heard of it, but i have no idea hwat difference it makes, and why teams would choose to use it. If you send me in the right direction to find out what it is, that would be good too

[Richard Wallace]

Originally Posted by akoscielski3

Second thing, what is CAN? I've heard of it, but i have no idea hwat difference it makes, and why teams would choose to use it. If you send me in the right direction to find out what it is, that would be good too

You'll get faster results by posting this one in a separate thread.

Several really good programming and control specialists haunt CD, but a thread titled "motor failure" might not attract their notice soon.

CAN is an acronym for "controller area network". In an FRC control system with Jaguars, it is a two-way communication channel that enables both commands to and feedback from the motor controllers.

[akoscielski3]

Originally Posted by Richard

You'll get faster results by posting this one in a separate thread.

Several really good programming and control specialists haunt CD, but a thread titled "motor failure" might not attract their notice soon.

CAN is an acronym for "controller area network". In an FRC control system with Jaguars, it is a two-way communication channel that enables both commands to and feedback from the motor controllers.

Okay thanks. Ill probably find something if i just search it

[Ether]

Originally Posted by akoscielski3

Okay thanks. Ill probably find something if i just search it

http://www.chiefdelphi.com/media/papers/2433 :

A tutorial on the hardware (electronics) aspect of CAN wiring using the black jaguar out of the 2011 (and later) KOP. Made from the experiences of a veteran wiring student who had no previous knowledge of CAN bus and had to ask a lot of questions and do some experimenting to make CAN work while keeping to a tight budget. Includes previously undocumented (as far as I know of) methods and quick tricks.

This has been updated several times since it was posted in January. It might be a good place to start since it was written by a student in the same situation as you (no prior knowledge of CAN).

(Be sure to read the thread as well as the papers)