caxis.comp - problems

smc.collins wrote:

4x quadratic counting
 

What specifically?  Do you see somewhere in the config files where I've not accounted for the 4x?  The encoder-in is 2500 lines, so I've set the scale to 10k.  And pulses-out doesn't use 4x.

I don't know enough about the system, I was asking if you accounted for it in your feedback from your encoder ? also, have you checked the drive settings ? I am sketching out the code for the hal component to feed spindle to t axis, I might have time in a week or so to try it on my lathe.

I did realize I need to scale the spindle encoder for counts/degree instead of counts/rev, but I was running open loop during the last round of testing.

The mismatch would account for the encoder velocity being off by a factor of 360, but not why the PID is clamping (or being clamped) by something.

Steps/degree are correct in the drive.  Otherwise it wouldn't actually move 180 degrees when a 180 command (or some other position) is given.

can you post your hal and INI ? I want to see what kinda code I need to sktech out. I am moving forward with my plan to route spindle speed to the axis from rpm to deg/sec.

They were attached a few posts ago.

found the hal, I need to think on this for a few days but I have a few ideas.

I think I figured it out, and this might be an LCNC bug.  Probably me though.

I think AXIS GUI is clamping the rotary axis speed.

No amount of fiddling with the [AXIS], [JOINT], or [TRAJ] values would have an effect on G0/G1 command speeds.  Even had the [DISPLAY] MAX_ANGULAR_VELOCITY = 3600.

I used PNCconf to cook up a quick config and compared the values to my own.  I had my axis/joint values set too low, but even after adjusting them up something was still clamping.

I then spotted this in [DISPLAY]:
MAX_LINEAR_VELOCITY = 3

Hmm... that's not sus at all.

Changed it to MAX_LINEAR_VELOCITY = 6, and now the rotary axis PID output is 6 during a G0.  Still too slow, but I think AXIS GUI is the culprit.

So why is AXIS GUI using MAX_LINEAR_VELOCITY value instead of MAX_ANGULAR_VELOCITY?

And maybe the C-axis isn't being clamped because...reasons?

Interesting development.

If I change the subspindle from W-axis to B-axis, LCNC stops using MAX_LINEAR velocity and starts using MAX_ANGULAR velocity.

Doesn't matter if the INI [JOINT] TYPE = ANGULAR.  If the axis is named anything other than A/B/C, either LCNC or AXIS GUI considers it a linear axis and applies the linear max vel ceiling.

Bug or feature?
AXIS or LCNC?

I don't feel like going through setting up a different GUI just to test.

So now I think I can clean things up and re-visit caxis.comp and see if I can get it sorted out.

Thanks Phill, that would save me some time editing a new config.

This appears not to be in the LCNC manual config/ini section?

EDIT - I wonder if there's a reason LCNC shouldn't default to the [JOINT] TYPE = N if the [AXIS] TYPE is not specified.

EDIT 2 - Now that I searched the code on GH, I see that this is an AXIS GUI thing and not LCNC.

Thx again