Strange noise on acceleration curve during motor tuning

[the1depot]

I am upgrading an Anilam control to the all in one unit. I am seeing a sinewave riding on the acceleration curve on the tuning scope in the PID screen. The motors are also emanating a sound. I have the report attached and a picture of the waveform. I was running at 30 IPM with 0,5S pause. The encoder is the original 1000PPR. Also The original power supply is being used. I ordered a new Capacitor and rectifier board from Centroid as the Cap is the original and the ripple was about 8%

Any Ideas would be greatly appreciated.

[tblough]

You drive is oscillating. See the Centroid Technical Bulletin below:
https://www.centroidcnc.com/dealersuppo ... ds/260.pdf

[ashesman]

Usually high frequency control oscillation is caused by too much proportional gain or not enough derivative gain. Usually a good PID tuning method is to start will really low Kp and 0 Kd and 0 Ki, low enough Kp that response is sluggish. Then start increasing Kp bit ny bit until you start seeing signs of oscillation. Then add a bit of Kd to settle it down. Keep increasing Kp and Kd alternately until you can not get them any higher and then back them both down a bit. Next introduce a bit of Ki to remove the steady state error. Be careful, too much Ki can cause big slow oscillations. Usually Ki is a very small number compared to Kp. Kd is usually 2 to 8 times bigger than Kp.

To think of it in simple terms:
Kp is how much the control tries to react to a difference between target and actual position (referred to as the error)
Kd is how much the control reacts to a CHANGE in the error.
Ki is how much the control will keep increasing its output incrementally over time to reduce steady state error

But, obviously the tech doc mentioned above is the best place to start...

[ashesman]

Also, starting with slower acceleration times (like 0.5s) would make tuning easier. Then works towards faster acceleration.

[the1depot]

Thank you for the help. I will try this tomorrow.