Hi all,
We attached four encoder kits (AMT-103) to four CIM motors, however only one of them stays on, unless you tug with some pressure. The others slide off if you gently pull on them.
Do you have any suggestions on what to do?
Thanks!
2 Likes
Make sure you assembled them properly. Also, make sure you are using the “blue” insert to get a good friction fit on the motor.
The AMT 103 is designed for the plastic base plate to be secured to the motor (step 4 of assembly). Unfortunately, the largest bolt circle on the base plate is 1" diameter, and the CIM mounts are 2" apart. (Reference the figures at the bottom of page 6 of the spec sheet.) To use this encoder with a CIM, you’ll need to get creative. Two ideas: mount the encoder base plate to the plate to which you mount the motor, or 3d print a larger base plate similar to the AMT 102’s wide plate (but even wider).
If you’re not planning this for FRC, you might consider disassembling the CIMs and tap a pair of holes in the shaft plate that match holes in the encoder mount plate and be sure to use short enough screws, but this is not an FRC permissible motor mod per 2020’s R28.
Don’t mount your encoders directly on your motors; that puts them in front of the gear backlash, which can either be not a problem or a huge problem, depending on how much backlash you have.
As a rule, try to measure positions as close to the end effector of your mechanism as possible.
4 Likes
What FRC situations have you seen where backlash has been detrimental to the function of a mechanism? We have actually had very good luck measuring everything from the motor.
Our team has been largely sticking with Neos for most things, so most of our speed/position feedback is right on the motor. We have implemented a Ramsete Controller into our drivetrain with good success using this method of feedback.
We also have run PID position control of an arm with a quite large reduction, and a fair amount of backlash. Wasn’t a big deal because gravity tended to favor one side of the backlash, so we just measured from there.
I am also under the impression that some swerve teams are using the built in encoder in there brushless motors to steer the module to keep all of their control loops on the motor controllers.
1 Like
Lots of places - drives and arms are the most common, though. Anywhere you have loose belts or a sloppy gearbox, the backlash is going to hurt your controllability.
2 Likes
Haven’t seen it much in our drivetrain in past years (Falcons - 2020 and NEO’s - 2019), however we also aren’t trying anything too advanced (got close to path planning in 2020, and both those years we started with basic, drive X number of inches based on encoder feedback).
We have had “issues” with backlash when working with arms in the past though. The further back in the geartrain our feedback sensor was, the harder it was to tune our control loops and maintain position. Was it so detrimental that it cost us our season? No. However it would make me think twice in the future as to where we place our sensors, with the goal being to have them as close to the final output shaft as reasonably possible.
1 Like
While I generally agree that you should place your encoders after as much reduction as possible in order to avoid backlash (and that’s a good engineering principle to learn) - it’s not particularly helpful to answering the OP’s question. The OP already has these CIM-mountable encoders procured, and is looking for a while to make use of them.
1 Like
You can mount AMT-103s in a large number of ways, and the CIM mounting is difficult to do for reasons that @GeeTwo covered above.
1 Like
Excellent point! So back to OP, in the future, try mounting as close to the final stage/output as possible. For this instance though, have a look at something like this. I’ve used similar custom mounts that we’ve 3d printed for these encoders in the past (albeit, for non FRC applications) and it works great!.
The above link is designed for the 102 flavour of encoder, but I would imagine a similar mount could be cooked up for the 103 series.
2 Likes
I’m not sure I fully understand the problem. The plastic base should be mounted to the motor. The best way to mount the AMT encoders, no matter where you are mounting them, is VHB (Very High Bond double stick tape), just make sure you use the centering device as you install it.
The adapter should lock on to the shaft fairly securely but it doesn’t need to transfer much torque. The adapter is supposed to be a slip fit in the encoder and it not intended to retain the encoder itself.
1 Like
My team, 3512 uses AMT 103’s for our drivetrain encoders.
We secure them directly to a turned down hex shaft with a 3D printed mount:
Here’s a drawing of the turned down axel:
20-DRT-005.PDF (148.0 KB)
3 Likes
Don’t quote me on this but I think the 103 works with this from a hole mounting standpoint the issue becomes the direction the 103 wires come out. With my design they would interfere with the mounting plates that cim was on but if you can tailor the plate to accommodate the wires then that could work.
You’re absolutely correct here. It’s been a while since I’ve worked with these encoders, and now that you say that and I look more closely at the differences, that would be an issue that needs designing around…
This topic was automatically closed 365 days after the last reply. New replies are no longer allowed.