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

It is unlikely I and D will be much help but it may not hurt to try, but you are a long way from that point.

So far I have been very concerned about the apparent delay of response of the motor to input changes of the drive's command. So far, the delay has been too much to allow adequate P to maintain a respectable following error.

Todd thank you for all the help!! I need to be clear in my mind as to the settings allowed to be used & if there is a preferred sequence to get good traces. I have tried to duplicate the trace you reference #245143. I agree the trace units were not uniform, now I make shore all the same values in the traces.

Is it ok to adjust POT1 & POT 3, with the FF1 = 1.0 all other = 0.0? Making shore what POT was turned & how many turns & in what direction. Still keeping eye on the traces. Doing this for each SW3-7,8&9. Each new SW3 setting requires starting fresh again with POT 1 then POT 3 so I will start with factory settings & have 4 sets of out comes. I will only post pics that have value, even though I saved pics as I go.

Many thanks!

Yes, I would suggest making all drive adjustments with Linuxcnc sending an open loop signal to the drive (FF1=1 and all other PID settings 0).

 Todd, Many thanks for your help. I have been working on the traces. Please help me be clear on this suggestion from a few posts (#246021) earlier:

"Adjust your gains for the Halscope signals so that 1inch/sec is the same number of ticks for each signal you are reading."
This says to me make all units the same in all the traces I do, so it is uniform. Am I missing anything here? The pics are not what I was expecting.
The attached pics show the way the trace moves. This is sw3-7 with the pot1 set to 8.5 & pot3 going from 7.0 to 10.0 turns. If I change the sw3 settings to other like 7&8, very same progression of the trace & very similar values set for pot 3. The file name is the settings. All units are 200m/unit.

Edit: This is how I have gathered this set of pics. Power off set sw3-7 (for example) reset to factory settings the pot 1 & 3. Power up increase (CW) pot1 until motor gets nosey, back off 3 turns. Start on pot3, turn CCW until motor does not jerk. It smooths out to what is in the pics until the trace flattens out completely. The white trace never gets close to the green trace.

Many thanks, your suggestions/thoughts are welcomed!!

What are you seeing?

What I meant was that the velocity-cmd trace (red), PID-output (green), and encoder-vel-feedback(violet) should all be using the same scale (200m/dev for example). The units for all of these should be in inches/sec assuming that your encoder scale is correct. So at 200m/dev one vertical line in Halscope will be equal to 0.2inches/sec (or 12ipm) for each of those traces that are in the same units.

The following error is only in inches, so it's scale is unimportant and does not compare with the others. You only need to know the scale of it to determine how large the following error actually is.

Todd, thank you for ALL your help!! Attached are pics with same units & few changes to POT 1 & others. The file name are the values that are not seen in the screen, example: sw7pot18.5pot37.0.png =sw3-7=on;pot1=8.5 turns;pot2=7.0turns.
The pics in this post are a re-do, very close to the post #246098 (2 post up), that include a trace of 6.0 turns on pot 1. The motor would jerk, any lower than shown.

Many thanks!

You are moving in the right direction turning down pot 3, probably need to go 2-5 more turns. You need to get that purple line up to or slightly above the commanded vel.

Are you sure that your encoder scale is correct? (Is 1" of movement on the Linuxcnc DRO actually 1" on the machine?)
Check this with the servos disabled, moving the axis by hand.

Todd, Thanks
Yesterday when I made a lot of traces (and saved on usb) starting low, doing 1 turn at a time to pot3 (and to pot1) the motor would jerk until pot3's turns did not jerk the motor. Always adjusted pot1 first, then pot3, 1 turn at a time. Attached are the lowest (when it would not jerk motor) of the sw3-dip switch. As I see the traces, the distance gets larger between the purple line (shown in white) and green or red traces. Only the lowest pic attached.
Many thanks!

The last pic (file name starts with swf= factory setting for sw3). The 1st pic is todays file starts with 7sw (sw3-7) this was adjusted until the last 1/8 turn was saved as you see, I have a 2nd pic as well.

If turning down pot3 below 6 turns causes a jerk, what does the trace look like at 6 turns? (leave pot1 at 12)

What does the trace look like when the jerk occurs when pot3 is turned slightly past the point when the jerk starts?

It is possible you are running into the limits of what that drive can do with your motor and the power supply you have feeding it. In other words the drive can't spin the motor fast enough. (Was the old setup capable of these speeds?)

(Edit)
PS,  Adjusting pot3 shouldn't cause any kind of jerking, it should only be adjusting how fast the drive tries to turn the motor for a given command voltage.  Basically it just scales the speed of the command.  However if you are running into the performance limits of the system all bets are off.