Motor tuning issues

[cncsnw]

If the motor's no-load speed is 3700RPM on 140V, then it should be a little under 3200RPM on 121V. To apply any torque to speak of, the speed will have to be lower than that. You might be able to get to the 3000RPM required for 300 in/min, but 240 in/min or so would be a safer bet.

I recommend blocking the axis and dismounting the servo motor. That way you can:
1) Turn the ballscrew by hand and find out how much torque it takes to move up and down, and feel for tightness or binding spots.
2) Run the motor back and forth at various speeds, with no load on it, and see if you still get stalls.

[converterking]

I noticed your servo motor has a 90 vdc brake. You did check that you have 90 volts going to it?

[bencarlson]

I'll check the brake in a sec - I had to buy a $400 power supply JUST for that voltage, so I'm pretty sure! Haha!

I am uploading another video, (also I DM-ed you, am happy to pay for your assistance!) that shows things seem better after moving to 15amps on Z, and reducing max feed to 200(not that it has an effect yet).

I am able to get the Z to run the PID program a few times before kicking out - at 100% feed. Here's a photo of the waveform:

[bencarlson]

I just stumbled across this doc which explains motor tuning better - I was just 'shot in the dark' playing with the values before...

https://www.centroidcnc.com/dealersuppo ... ds/260.pdf

[bencarlson]

Here's the video:

[bencarlson]

Also, when it kicks out, it continues to give:

327 Fault: job cancelled
411 Z axis(3) full power without motion
SV_STALL ERROR Reported by CNC11!!!

[cncsnw]

You will need some Kv1: something on the order of 20 to 40.

If the axis is not balanced, you will need some Kg as well: maybe on the order of 10 to 20.

You can estimate Kv1 and Kg by using Slow, Continuous jog on the main PID Configuration screen (the one with live numeric values in table form, *before* you have pressed F1/PID to get into the table of PID gains and live tuning interface).

On the main PID Configuration screen, you can see PID output (percent current request) for each motor, in the "PID Out" column. Use Slow, Continuous jog mode to jog Z up, and estimate an average PID Out value. The number will be jumping all around, so the best you can do is a rough estimate of the average. That is good enough. Then jog Z back down, and estimate an average PID Out value for that.

Since you have Direction Reversal set on Z, on your machine Z+, head up, is negative motor movement; and Z-, head down, is positive motor movement. Therefore you would expect the average PID Out for an upward move to be negative. The average PID Out for a downward move would be positive if the axis were balanced. If it is so unbalanced that the motor has to hold it back (up) even when jogging down, then PID out for the downward move might still be negative; just closer to zero than before.

Your starting guess for Kg should be the mid-point (average) of the upward-going PID Out and the downward-going PID Out.

Your starting guess for Kv1 should be the difference between that mid-point, and either of the moving PID Out values.

For example, if the average PID Out in the upward move was -40, and the average PID Out in the downward move was +10, then your starting guesses would be -15 for Kg, and 25 for Kv1.

In CNC11 and CNC12, you may need to use a substantially higher value for Kv1 to make a dent in the error spike at the outset of each move. More Ka can help there as well.

[bencarlson]

I disconnected my z-axis motor and in the video show there is no binding of the ballscrew, but the z-motor tuning does the exact same thing as when it's connected! Looks like a motor issue maybe?

[converterking]

When you apply 90 volts to the brake, does the motor turn freely?

[bencarlson]

When you apply 90 volts to the brake, does the motor turn freely?

Yep, turns pretty easily!

-Ben