1
$\begingroup$

I am working with a ball screw mechanism driven by a brushless dc motor with an encoder mounted on the load side. We are using a PID position controller with tuning settings by optimization however the ball screw keeps oscillating around the reference position. We are thinking it is caused by the backlash and we accept the precision reduction caused by it. The only thing we would like to have is a non-oscillating response. Is there any trick/technique or any source that we can get an idea of?

Thanks in advance,

$\endgroup$

2 Answers 2

3
$\begingroup$

Backlash is only an issue when you change directions, because that's when you'll open the space in the gear train. If you're using PID control, consider an overdamped response instead of a typical "critically damped" response that overshoots. An overdamped response should have no overshoot and thus no directional change. You'll have a slower response, but should eliminate the oscillation/chatter.

$\endgroup$
3
  • $\begingroup$ Hi @Chuck thanks for the answer. This is what we are doing now but as backlash changes over the length, its effect changes. Do you know any book as a reference that points out practical applications for ball screw control ? $\endgroup$ Commented Dec 9, 2021 at 4:03
  • $\begingroup$ @controlsHeaven no, sorry. $\endgroup$
    – Chuck
    Commented Dec 9, 2021 at 14:34
  • 1
    $\begingroup$ @controlsHeaven can't you somehow measure the backlash as a function of position on the screw and provide the data in tabular form to the controller? And by "provide" I mean keep the PID unchanged but add some sort of compensation in the feedforward loop. $\endgroup$
    – Vorac
    Commented Dec 11, 2021 at 4:37
1
$\begingroup$

Backlash is not limited to just when you change directions. You can open the space in the gear train by slowing the motor down faster than the load would slow down on its own and yet both the motor and load will still be moving in the same direction.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.