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

Remember at the 200u resolution, in Halscope each vertical graduation is the equivalent of about 6 of your encoder's edges. (Not just the lines but the edges of the lines (each line having 2).  

Todd, many thanks! Attached a pic of traces in the p=800 range. I started the opposite direction, seems more consistence trace. You will see that the start of the cruse

Todd, thank you. Referencing your post 07 Sep 2022 19:39 #251433 at the bottom it talks about starting with "enough "P" to get some stable movement that is 60-90% of commanded move, with FF1 set to 0.0". This is comparing ch1 (axis.0.joint-vel-cmd) to ch4 (hm2_5i25.0.encoder.00velocity) correct? If so, a value under 10 is required for "P", I have printed out your #251433, to reference, and tried all the numbers (by 10.0 increments) to 850 & the ch2 error will go from hi to low as the "P" goes lower.
This I over looked, but pondering more about this, if the error in this step is large it will follow all the way through. I was thinking the "bracket" approach to this 1st step and prove to myself all is correct.
Channel 4 as the value of "P" increases, generally this trace oscillates more as you increase "P".
Many thanks!

Not sure I follow what your trying to say.

Have you tried incrementing larger D values to see what happens?

With P=800, try D=1, 2, 3, 4. Does the oscillation get smaller? If you don't notice any or much change, try D=10 or 20. What kind of change in the trace happens with the larger values?

Todd, Thanks!! What I was trying to say is that ch2 & ch4 are always on top of each other, (never separate, with a section of screen between) so 60-90% from where? I think it is 100% with different amounts of oscillation as I go up in values. I was thinking (I know, their I go again, thinking again) that the trace may of not been low enough and missed the 60-90% "sweet spot" instead went all the way to 100%, and your suggestion was specific in the #251433 post.
The attached traces start with P=800, the units are ch 3 & 5 = 500m/div; ch 1 & 4 = 200m/div & ch2 = 500u/div. Notice all other settings are 0.0.
Many thanks!

When you have all of the feed forwards (mainly FF1) set to zero this is expected. In this situation the output command is driven only by the error. With every thing set to zero except for P, the out put to the drive is simply the error multiplied by P, so the larger the error the larger the output command to the drive.

Adding more D is certainly improving the oscillation. I think you should definitely try adding more. (You should also put the FF1 and FF2 settings back to where they were at, about 2.1 and 0.06.)

Todd, thanks! Attached is best trace, the units are ch 3 & 5 = 500m/div; ch 1 & 4 = 200m/div & ch2 = 500u/div. The settings are on the screen. What do you think?
Many thanks!

Yes that looks pretty good. Might be able to improve the cruse error with some more little tweeks of FF1 and FF2. But otherwise that is probably about as good as your going to get it.

Also put the deadband back to where we had it. (That should quiet down the dithering when the servo is at rest.)

Once your done be sure to save and back up your settings.

Todd, thanks! The attached pic is the best trace with settings for "X" axes. Now for "Y" axes, start with the "x" settings?

Many thanks!