AMC drive tuning

No, +-10V is max

I made some progress. I verified motor rpm with an external laser tachometer to set the desired max rpm at 10v via the drive reference gain pot (pot3). For starters I'm limiting the RPM to 1500. If my math is right, with belt and screw reduction, that puts me at 200 ipm for xy and 166 ipm for the z. To have enough headroom, I set the max_output to 3.16 for xy and 2.63 for z.I can reduce that further if need be or bump up the pot setting. I'm working with output_min and max_limits of 3.33 for xy and was adjusting the output scale with FF1 set at 1.0. and P12. I was able to get the f error plot level for L/R movement at output_scale 3.37. Question is, is it acceptable to set the output scale value for tuning instead of FF1?
During all the PID testing/jogging, I have somehow lost the ability to keyboard jog after homing the machine. The machine thinks it's miss configured (probably is) and past the soft limits. It also did not zero out the coordinates. When I upgraded to 2.8, similar error would sometimes occur during shut down but wasn't an issues for jogging. Any ideas how to restore that?

Setting the output scale to what works with FF1=1.0 is I think the cleanest way to do this
since at that point your PID parameters are all in engineering units. A side benefit of doing
things this way is that metric and inch machines have the same PID parameters.

On 2.8 you use joint jogging before homing and axis jogging after homing so if you have a jog wheel
or similar you need to direct its output to both the axis and joint jog pins exported by motion and enable
jogging of both axis and joints.

Setting the output scale to what works with FF1=1.0 is I think the cleanest way to do this
since at that point your PID parameters are all in engineering units. A side benefit of doing
things this way is that metric and inch machines have the same PID parameters.

Thanks for confirming.

On 2.8 you use joint jogging before homing and axis jogging after homing so if you have a jog wheel
or similar you need to direct its output to both the axis and joint jog pins exported by motion and enable
jogging of both axis and joints.

This part was working with 2.8 prior to drive swap. The mpg is also disabled. Homing completes all 3 axis in order as before.
I'll reconfigure the homing part in the ini. I think the home value is the same as the soft limit. Could this have been masked by a large following error allowance previously?

The follow error limit should have no effect on softlimits except if the actual position is so far out of place and to exceed the limits

I changed the home positions by -.010" now it's fixed. I also had to turn back on the net x-vel-cmd => pid.x.command-deriv in the hal to regain operation of the mpg. Does that make sense?

I need you guys to look over the scope plot. First video is at 25 ipm. I need to reduce the spikes. I'm getting a .002" inch overshoot on the X, with a correction back to 0 measured with a dial indicator. At 5ipm and below it didn't overshoot. At 160 ipm, it returns to 0 and repeats but I could not gage overshoot. I hate to admit that I still lack the experience with halscope to fully grasp how the values translate into the real word. Keep in mind the encoder is on the machine table. Initial max error is at .050 and min error set at .010. Below that I get joint following errors for the X. Oddly, the y performs better with similar PID values with max error at .040 and min at .005.

And here with around 113ipm

I also had to turn back on the net x-vel-cmd => I also had to turn back on the net x-vel-cmd => pid.x.command-deriv in the hal to regain operation of the mpg. Does that make sense?
in the hal to regain operation of the mpg. Does that make sense?

No, I would make sure the pid.x.command-deriv is not connected to any signal in the hal file

The videos inserted report that they are private and and cannot be viewed

The videos inserted report that they are private and and cannot be viewed

Thanks for the heads up. It should be unlisted and visible now.

OK what is your acceleration? its seems quite quick. The initial spikes will be proportional to acceleration. They can be partially tuned with a very small amount of FF2