CIM motors drawing uneven current

[anon28692278]

I have an andymark drive train (the one in the kit) that is powered by two cim motors on each side in a west coast style. The left motors are pulling more current (15 amps) than the right motors (14 amps). What could cause this, and how could I fix it?

[Skyehawk]

The simple stuff:

• With the bot off, do the sides turn by hand with the same amount of drag/friction?

• Are the connections between the motors and controllers all robust connections?

There will be some internal variation between CIMs (or any motor). This particular case seems excessive and I would bet one side of your drive train has more friction.

If the two sides feel the same in terms of friction by hand, swap the speed controllers (the motor controllers originally on left side now running the right side motors and vice versa) and see if the problem follows the controllers or stays on the same side.

[anon28692278]

The spark maxes would be the motor controller right? I want to make sure I do it right.

[Skyehawk]

Yeah, so the spark maxes (set up to run a brushed motor, like it sounds like you already have this working).

You would just plug the controller(s) that are normally for the right side into the left CIM(s) and vice versa.

Make sure you feel for consistent friction between the sides too. One test is to just push the bot directly from behind and see if it dramatically steers off course from a straight line.

[anon28692278]

If that doesn’t work, how could I limit the current to the left side so they both draw 14 amps?

[Peter_Johnson]

You don’t want to limit it. That just means you’ll need to limit the other side to even lower current, or you’ll turn. Uneven current simply means it’s taking more energy to have one side to keep up with the other to drive straight. Friction is the most common cause of this.

[SLAB-Mr.Thomas]

@Skyehawk

internal variation

To slightly misquote you, my team found that the forward and reverse directions of a single CIM were not identical.

And in your context of multiple motors, my team built a simple dynamometer to better select matching motors for the robot.

[Skyehawk]

@SLAB-Mr.Thomas

To slightly misquote you, my team found that the forward and reverse directions of a single CIM were not identical.

this too!

Robot curving with motors drawing different amounts of voltage.

One of the points of consideration I picked up years ago in designing drive-trains for autonomous operation was this: DC motors often draw different amounts of power when traveling forward vs backward. Since on set of your drive-train motors is most likely flipped to the other you will see different results driving forward vs backward. If you were seeing the same turning while driving forward and backward (i.e. your arc had the same center point for forward and backward) you could attribute this…

[zog]

The discussions about friction are the place to start - not just in bearings and the like but tensions of belts and chains can impact how much energy is needed to run an assembly at a given speed.

Weight distribution can also come into play; the heavier side tends to draw more current, but IMO, 1A difference right to left at full speed (I’m assuming that’s when this measurement is taking place) isn’t something that would bother me a lot (that’s an average of 500mA difference for each motor) given all of the other things involved in getting a robot doing robot things.

For independent drive systems like swerve, chassis deflection or out of level assembly can cause drive current differentials but that happens along a diagonal (larger current draws on the LF and RR modules for example). This comes from the fact that the robot weight is primarily being carried by two diagonally opposing corners which are also providing most of the driving force (the other two motors are sort of along for the ride to some extent). That’s not what you’re seeing but it is a thing.

I would also agree to not mess with this via software - it’s a mechanical issue (I’m considering the motors themselves as “mechanical”). If you need 15A on one side and 14A on the other to go straight, then do it so you can go straight.

[troy_dietz]

15 vs 14 amps is negligible. I’d do a cursory look around for anything off on assembly, then just send it.

[Skyehawk]

True, there is a certain healthy level of “well, this is how it is” and move on, some things aren’t a huge risk and aren’t worth chasing.

It’s also very possible things will wear in over an hour or two of run time and be almost equal.

[JamesCH95]

If memory serves

The directionality difference is intentional, and about 8%. The ‘forward’ direction is timed more effectively on purpose because the original design intent of the motor was in self-leveling RVs and powered 5th wheel ‘landing gear’ assemblies. In this situation the motor drives really hard in one direction and then is under very little load when retracting the mechanism.

In this case OP is observing a ~7% difference, almost exactly what we’d expect.

The AM datasheet has different CW and CCW duty cycles that reflect the asymmetric design of the CIM. The FRC version has the now-ubiquitous 8mm keyed output shaft, but the original 2000s era Atwood-flavor had a gear profile cut into it.

Exemplary mechanisms:

I feel so old…

[Skyehawk]

I honestly didn’t know it was possible to bake that into the motor (but I am really not an electrical person). It makes sense though, playing with the timing to make one more efficient is a neat trick and is really cool.

[FinIce]

That makes quite a bit of sense for why one direction on our RC bots always seems to be faster than the others, thanks for that!