Set up using AMC AB15A100 drives, brushed DC motor Prototrak Plus with Encoders

Todd, The attached traces are 1/16 of turn lower, both at the same settings, if I try lower Pot 3 than 4.0 turns, the motor will jerk.

Will I, set the "Y" axes with the same settings as this or start a new tuning from scratch? My thinking is I will set "Y" to these POT 1, 2, 3 & 4 (I understand pot4 is for the drift control) settings; before starting the tuning of drive to Linuxcnc with P, FF1, FF2, DEADBAND etc.

Are these any better, I see the over travel at the beginning of the trace.

Many thanks

Thanks, Todd, Pot3 will not go any lower than 4.0. The attached pic is Pot 1 at 11.813 turns (this is 1/16 turn more, and is all it would accept) & shows some of the spike is lower. I will try any suggestions, this was all I think can do to make trace better.

Many thanks

I don't understand what is causing it to "jerk".

Perhaps we need to experiment with some of the dip switches. But we need to limit changes to 1 thing at a time to try to understand their effects.
What happens with all settings the same as above, but switching SW2-1 on? What does the trace look like? Will that allow turning down pot3 further or does it make it worse?

Edit: I am hopeful that the SW2 switches (increasing the integration time) may be the answer.

Todd, This is a pic of sw2-1=on, I have not compared this trace to others. Units are uniform at 50m/div & all other settings the same.

many thanks!

Todd, This is better, sw2-1 & 2 = on. See pic! I turn power off so it takes time to reset up screen.

many thanks!!

Todd, this is the last of the settings from sw2-. One of the pics is sw2-1,2 & 3=on & the other id sw2-1,2,3, & 4=on. I have not studied them. One of the files has the 3 missing in the name (should be named sw2-1,2, &3 on pot1 11.813 pot3 4.0 feed 20.0) is does have sw2-1,2,3 = on.
Many thanks!!

Ok, I think the last image with 1234 is for sure over dampened. The 2nd image might be the best of the bunch, The better question is does any of those settings allow you to turn down pot 3 any more to try to raise up the encoder velocity trace to the proper level?

Also I don't really like the little wart that is appearing on the acceleration portion of the trace, in the last couple of images. I wonder if that is due to not having the current loop tuned properly?

Many thanks, Todd, the trace is looking best yet!! Thank you!! The file name is shorten,example: (s12pot1 11.687pot3 3.312) = sw2-1&2=on; pot1 is 11 11/16 CCW; pot3 is 3 5/16 CW; feed 20.0"/min. I have not tried pot3 any lower of turn, to make the trace rise further, because the overshoot. Notice that pot1 has less turn (from11 13/16 to 11 11/16) this took some of the vibration out of the motor. Now motor has real faint vibration if hand is on motor. I will try the last 3 pic (1/16 turn between pics). I am seeking a second opinion about the traces before I go further.

Many thanks

Looking much better. Ok, now try with SW2-3 only on, then SW2-2,3 on, and finally SW2-1,2,3 on. The idea is to try to bracket the response and minimize the overshoot without degrading the following. If SW2-1,2,3 still has overshoot, move up the chart in the manual to the next higher capacitance switch setting... After finding the best switch settings, go back and fine tune pots 1 and 3. Then try a jog at a higher feed rate, see how fast you can go before things fall apart. (You may now be able to jog it at 60ipm without a problem.)

Once you've done the above, then you will be able to tune the PID in Linuxcnc (Keep FF1=1 approximately, some fine tuning is acceptable)!

You can and probably should use the settings you find for X as a starting point for tuning Y. But you may find the final settings are different because the mass the Y-axis is moving is not the same as the X-axis.

PS; A small Halscope tip: Slide the trigger position slider up about 3/4 of the way to the top, so that you can also see some of the steady state after the end of the move. This will let you see how well the servo is settling after the move. (You might also need to readjust the horizontal position slider to recenter.)