PDA

View Full Version : Any Method to Check If A Servo Needs to Be Replaced?



guoshihui
04-14-2016, 07:50 AM
Ok, I am new to robotics, and my robot is not quite taken care, so it falls a lot and sometimes over-heated.

So my question is, how could I determine if a servo needs to be replaced? Is there a definite rule or method to check if a servo is functioning properly?

Thanks ahead!

Shihui

jwatte
04-14-2016, 12:52 PM
If you can move it by hand, and it doesn't make horrible, scraping, clicking noises, the motor and gearbox is OK.
If it powers on, and does what it's supposed to when you tell it to, then the control mechanics are OK.

There is an intermediate mode for precision servos (like the MX series) where the motor is just damaged enough to cause further problems, but "seemingly works fine." For the MX series of servos, if they draw more than 100 mA while notionally idle (torque on, servo horn centered, no load,) then the motor has enough stickiness (from burning the coils/brushes) that it will start damaging itself, and possibly its neighbors on the bus, and should be replaced.

I imagine the same state can happen for other servos; the way you measure it is by looking at idle unloaded but active current for a "fresh" servo, and if the idle unloaded active current is a lot higher (2x or more) for the measured servo, then it's probably starting to go bad.

guoshihui
04-14-2016, 09:19 PM
If you can move it by hand, and it doesn't make horrible, scraping, clicking noises, the motor and gearbox is OK.
If it powers on, and does what it's supposed to when you tell it to, then the control mechanics are OK.

There is an intermediate mode for precision servos (like the MX series) where the motor is just damaged enough to cause further problems, but "seemingly works fine." For the MX series of servos, if they draw more than 100 mA while notionally idle (torque on, servo horn centered, no load,) then the motor has enough stickiness (from burning the coils/brushes) that it will start damaging itself, and possibly its neighbors on the bus, and should be replaced.

I imagine the same state can happen for other servos; the way you measure it is by looking at idle unloaded but active current for a "fresh" servo, and if the idle unloaded active current is a lot higher (2x or more) for the measured servo, then it's probably starting to go bad.

Thanks so much jwatte, this sounds a convenient solution, =)

Shihui