Society of Robots - Robot Forum
General Misc => Misc => Topic started by: madsci1016 on August 15, 2010, 02:57:12 PM
-
Man, between experience with other tech companies and the horror stories floating around here, I'm starting to assume the worse.
It seems that one of the $40 encoded motors I bought has a defect and is internally binding. Explains why it's been so hard to get steady results during my hours of PID tuning.
I finally got the clue and plotted PWM duty cycle being sent to each motor while both maintain the same speed. Got
(http://www.billporter.info/wp-content/uploads//2010/08/motordefect.png)
Removed the motors and one is definitely harder to turn by hand then the other.
Pololu's website gives no clue to there defective return policy, and that makes me nervous.
I wrote a nice email to to them explaining everything and asking for a replacement, so we'll see how it goes. Anyone ever dealt with them in this regard?
-
Pololu has given me decent tech support . . . and they occasionally (on rare occasion) post on SoR. But I've never done a return.
It seems that one of the $40 encoded motors I bought has a defect and is internally binding. Explains why it's been so hard to get steady results during my hours of PID tuning.
I've noticed those motors on their site . . . they're cheap, but to me they look like the 'made in china' type. That type sucks in terms of efficiency/quality.
For example the Tseng Neng motors, popular because of price, but horrible in terms of efficiency/quality. And if you put a screw in too tight when mounting the motor, the gearing box will jam.
If you've only used cheap motors, you should consider giving an expensive quality motor a try - you'll be amazed at the difference. Cleaner, quieter, stronger.
btw, you didn't label the axes or even title your chart, so I can't make sense of it :P
-
btw, you didn't label the axes or even title your chart, so I can't make sense of it :P
plotted PWM duty cycle being sent to each motor
so it's PWM duty cycle, and it's not real important to know the detail, just that there is a significant difference between the two motors. You can see the oscillation in the faulty motor, makes me think there's a bad spot on a gear or a place during rotation that it rubs.
but if you must (and can't take an educated guess :P ), it's 8bit PWM duty cycle on Y, and an arbitrary time scale on X.
I was careful with screws, and originally had Lynxmotion motors that seem to be much better quality, with nice US digital encoders, but they were not fast enough. Do you know of an equivalent 400+ RPM gearbox motor of high quality that has encoder shafts?
-
Maxon sells whatever you need. But it'll cost ya :P
Oh, and shouldn't your chart plot PWM duty cycle to either output velocity or torque?
-
Oh, and shouldn't your chart plot PWM duty cycle to either output velocity or torque?
each motor while both maintain the same speed.
I could, but relatively I think I get my point across by having them at the same speed, but one is requiring twice the power of the other. If I remember right, each wheel was set to run 1.5m/s, and i forget what rpm that calculates to. Both wheels were achieving set speed, and the robot was going straight.
-
oh!
So you have a feedback system to adjust the PWM duty cycle to stabilize velocity, now it makes sense :P
-
Yup, a full blown PID loop at 37 Hz.
Sorry If I didn't make that clear.
-
Looks like Pololu is going to step up to the plate. Got an email that they were going to ship me another motor.
-
New motor tested
(http://www.billporter.info/wp-content/uploads//2010/08/motor2.png)
Now that's a pretty thing.
Overall happy with Pololu customer service, they sent me a new motor fairly painlessly. They are asking that I send back the failed motor on my own dime, but I'm ok with that.