Pid tuning 8i20.

So I'm finally there with the shenyang machine I've been retrofitting.  Currently boring a millimetre out of the spindle bore so I can fit a larger pull tube. As far as tuning the 8i20 cards I have a working set of pid values(I'll post a pic), I've read on the forum here that ff1 really dosent apply to torque mode drives.  But after much testing and going cross-eyed, I found that if I put around 15. Something in it it took away the (I geuss) wind up, as in as feeding along the error would creep up to say 1.2mm then hold there. Whereas with around 15 in ff1 this would remove the issue. But, at a higher than 3000mm/min speed, I  still get the same problem,  it dosent seem to scale properly.  Anyway any thoughts?the pic is before I put an ff1 value in. 

What is you servo thread rate? Torque mode tuning gets easier at higher thread rates

I've had it running at 3k or whatever-
Thinking of bumping it up now as I just turned off temp monitoring in the bios, (the fan runs constantly now) as i was intermittently getting real time errors if I left it unattended for hours.
I'll go to 3.5 and see.

If you were already running at 3 KHz, that's probably OK.

If you have such a huge following error are you running into the torque limits
or a stability issue?

(a ~1 mm error with a P term of 838 is 20X past saturation which suggests you are torque limited or not stable)

A halscope plot of a jog would help a lot here,

PCW wrote:

If you were already running at 3 KHz, that's probably OK.

If you have such a huge following error are you running into the torque limits
or a stability issue?

(a ~1 mm error with a P term of 838 is ~25X past saturation  at 30A fulscale which
suggests you are torque limited or not stable)

A halscope plot of a jog would help a lot here,
 

 

Probably torque limited, I will change the z drive to a 2:1 reduction belt drive, its 1:1 now, it just trips under rapid, I did this with x, and don't have any issues.

For future reference,  
I found after using halscope that my main problem seems that due to the servo motor being on the limit of what the 8i20 can handle the tuning has to be close otherwise it seems that the current gets jumpy and can be prone to running out of limit constraints.