3551
General Mach Discussion / Re: Spindle/ C Axis using CSMIO/IP-A
« on: October 16, 2018, 02:17:41 PM »
Hi Mick,
the encoder resolution of 2048 lines/rev is the simulated encoder.
The servo encoder and the servo drive have much finer resolution that that. I think the servo encoder is 17 absolute and therefor has a potential
resolution of 170,000 counts per rev.
You could theoretically increase the simulated encoder resolution up to the same 17 bit. Given that the CSMIO closes the loop then the simulated encoder
resolution is the limiting figure. The downside is that the CSMIO has limited signaling speed, and in any event signaling over 500kHz between controller
and servo drive is DIFFICULT at best there is therefore a practical limit to the resolution beyond which the signaling speed limits the rotational speed.
The second issue is the 'zero error window' My servo has a programmable window, 4 encoder counts (out of 8000) by default where the servo drive
determines that the actual position is close enough to the commanded position that it will not try to close the error any further. You can reduce the
zero error window however the servo will 'hunt' when its close to its commanded position, sometimes you will hear it as a buzz. I use the servo drive
to close the loop and therefore use the servo drives manufacturers programming facility to do this. In your case you will have to use the 'zero error
window' feature or its equivalent, if it has it, in the CSMIO.
My experience is that for practical purposes that a resolution of 0.18 degrees or 10.8 arc minutes is realistic even if greater resolutions are theoretically
possible. Thereafter gear reduction is the only feasible way forward. Consider a harmonic drive.
Craig
the encoder resolution of 2048 lines/rev is the simulated encoder.
The servo encoder and the servo drive have much finer resolution that that. I think the servo encoder is 17 absolute and therefor has a potential
resolution of 170,000 counts per rev.
You could theoretically increase the simulated encoder resolution up to the same 17 bit. Given that the CSMIO closes the loop then the simulated encoder
resolution is the limiting figure. The downside is that the CSMIO has limited signaling speed, and in any event signaling over 500kHz between controller
and servo drive is DIFFICULT at best there is therefore a practical limit to the resolution beyond which the signaling speed limits the rotational speed.
The second issue is the 'zero error window' My servo has a programmable window, 4 encoder counts (out of 8000) by default where the servo drive
determines that the actual position is close enough to the commanded position that it will not try to close the error any further. You can reduce the
zero error window however the servo will 'hunt' when its close to its commanded position, sometimes you will hear it as a buzz. I use the servo drive
to close the loop and therefore use the servo drives manufacturers programming facility to do this. In your case you will have to use the 'zero error
window' feature or its equivalent, if it has it, in the CSMIO.
My experience is that for practical purposes that a resolution of 0.18 degrees or 10.8 arc minutes is realistic even if greater resolutions are theoretically
possible. Thereafter gear reduction is the only feasible way forward. Consider a harmonic drive.
Craig