Deadband Gains

I have an ethernet ultralite running Yaskawa sigma 5 drives/motors (torque mode). I am trying to reduce buzzing/vibration by putting in some deadband settings, but they do not seem to work. I am following the manual and I have roughly 10 counts of following error while sitting idle and buzzing. So I set Ixx64 to -12 and Ixx65 to 640. I would expect this to reduce my buzzing when I am sitting, but it seems to make no change. I am not using the extended servo algorithm. Is my problem in the yaskawa? If I reduce Ixx30 by 20%, the resonance goes away, but the dynamic performance is much poorer.

 

Thanks,

 

Adam

The buzzing may be due to quantization noise. If this is the case, please try to decrease your velocity feedback gain and/or reducing the servo frequency.

I have the same controller with sigma 5 motors. The sigma 5 motors have a signature sound. If you use the autotune in the amp you will end up with the same buzz. You also get it with the Yaskawa controller. I was able to reduce this noise in the Delta Tau tuning to a point where you don't notice it. My approach was to lower the proportional gain and increase the Derivative and Velocity Feed Forward gains. I found with all the feedback on the sigma 5 motors proportional gain seems to make them unstable. Here is a typical tuning set i am using with good results. This should give you an idea of how the gains typically scale in my experience with Sigma 5 motors.

 

I130=20000

I131=25000

I132=25200

I133=1000

I134=1

I135=25000

acarder,

 

What is your clock settings on the Ultralite? More specifically, I6800..6802, I8 and I10.

Also, MI992 and MI998 on ACC-85M?

acarder,

 

What is your clock settings on the Ultralite? More specifically, I6800..6802, I8 and I10.

Also, MI992 and MI998 on ACC-85M?

 

I6800 = 5895

I6801 = 0

I6802 = 4

I8 = 2

I10 = 4193066

MI992 = 5000

MI998 = 4

Can you report on what is the frequency of the buzzing? Can you look at the following error in the frequency domain, i.e. FFT of the following error?

 

Dead band in your case is likely two wrongs trying to make a right.

 

If the buzzing is truly much higer frequency than the servo bandwidth and is happening primarily at one frequency, then you will likely get better performance by introducing a notch filter at the buzz frequency.

 

On the flip side, if the buzzing is due to a relatively low frequency mechanical resonance then the notch will just kill your phase margin and you won't really get much of a benefit. In that case you will likely have to go with lower gain.

acarder,

 

In addition to tahoe brian's suggestions, I recommend increasing the phase and servo rates on the machine. This will allow you to respond much faster with lower gains. Since Yaskawa Sigma V drive and ACC-85M can support 8 kHz servo frequency, I would suggest the following settings for the Ultralite and ACC-85M:

 

Servo freq.: 8 kHz

Phase freq: 16 kHz (This will allow the ring to run twice as fast so the single ring delay in propagation of data is canceled)

On Ultralite:

I6800=3684

I6801=0

I6802=1

I10=1048320

 

On ACC-85M:

MI992=3125

MI998=1

 

Notice that the servo loop gains should be "re-tuned" to take advantage of the new servo frequency. As a rule of tomb, if we have quadrupled our servo frequency:

- Kp (Ix30) stays intact.

- Kd (Ix31) is divided by 4

- Kvff(Ix32) is divided by 4

- Ki (Ix33) is multiplied by 4

- Kaff (Ix35) is divided by 16.