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

Todd, Thanks! After re-reading the top of page 6 of 9 in A-M-C manual (see attached pic) it says:
"A larger resistor value will increase the proportional gain, and therefore create a faster response time".
My thinking is to leave the SW2 as is (factory .047) & start with SW3 (factory 20) moving up the chart to increase the resistor value. SW3 has units that are in resistance units, (SW2 & SW3-6 to 10 are ALL capacity units) All of the trace pics so far are at the factory setting.

Many thanks

The SW2 switches are for tuning the velocity loop in the drive. They are completely disabled and bypassed while you have the drive set to current mode.

SW3 1-5 are the proportional gain and are going to control how fast the current rises (make the rise and fall more vertical. (Concentrate on this first it is where you will see the most improvement.)

SW3 6-9 are the integral. The larger the capacitance the more rounded the leading corner of the square wave will be. Use this to tame a sharp spike of overshoot at the leading edge.

SW3 10 will disable the integral, bypassing all of the current loop capacitors. Use this for observing only the effects of the proportional part of the loop, or if having all 6-9 on is still too much capacitance. (In the chart it is listed at the right technically this should probably be at the low/left end not the high as printed.)

PS:  It is possible the overshoot you see in your last scope image may not be from needing more integral (capacitance) and may be a side effect of needing more proportional gain (resistance).  Such effects are mentioned in AMC's current loop tuning tutorial paper.

Todd, thank you!! I am to be looking at channel 2 (the blue trace) correct? The only setting that I adjusted was the SW3-1 to 5. The higher I went after the attached pics, the more jagged the trace became. The setting referencing the file name (the name is hard to edit in the o-scope) as follows: The files attached are numbered (the last number in file name).I have other pics, this is the best traces, with the "test demo" key in "~/linuxcnc/configs/test/test.demo" at the terminal screen.Many thanks!Edit:Read the file name, they are not in order. These are the values/settings for the pic: The files attached are numbered (the last number in file name is ref to the value)  only (SW3 setting/unit) were adjusted. pic30=80 (sw3 settings) pic31 70; pic 32 = 90; pic33 = 100; pic34 = 110; pic 35 = 50; pic 36 = 60; pic 37 = 60(do over, but spike is left or mid). I will only post 4 of the set listed, all have same scale, both channels are 500 units. The spike floats around.Many thanks 

Well that is certainly a vast improvement over where you started. I don't think I can see any significant improvement between any of those images.

What do the "jagged" traces look like? If they have better rise time than these, they still might be an improvement. That is if it is possible then smooth out the jaggedness by increasing the capacitance with SW3 6-9. If they have the same rise time then don't bother.

So if that is as good as you can get it, I would be inclined to go with the lowest setting that gave the above response.

Even still you may yet want to experiment a little with SW3 6-9 and 10. Give it a try with the next lower and higher capacitance and see what effect they have and maybe even a try with SW3-10 switched on just to see what happens. Otherwise if you can't get further improvement, I'd say your done tuning the current loop.

Once done tuning the current loop, turn the velocity loop back on, and reattach the motor to the machine, and attempt to tune the velocity loop again (the pots and SW2). Hopefully it goes better.

You edited the post while I was typing. I can only see 4 of the images right now (31, 32, 34, and 35). Of those I think I would pick 34.

(The forum might be limiting how many you can attach to a single post.)

Todd, Many thanks! I got lots (at least 1 pic at low resolution for each of the SW3 settings) of pics, did not want to hog the forum. When you talk about the "rise time" this is the angle of the ramp to plateau, all look the same to me? Or something else? The jagged shows in this pic 27 = 310 (1 from last in chart) and pic20= 240 and pic9= 120. These are lower resolution traces, are they good enough to show the important part of trace? To do the higher resolution traces, you need to be quick at the scope, (because 2 sets of things need attention to) so the motor is not on too long. To try a different set of adjustments it requires to re-power the cnc PS and start at the "terminal" screen, (but I copy and pasted the command after the 1st time) to take next set of traces.
If I need to take "hi res traces", that ok, if this will work & now is the time to move back to the Pot 1 & 3 adjustments, I am ok with that. Or if need a bit more of SW3 6-9 and 10 and then re-visit SW3 1-5, I am ok with that.

Many thanks!

 

I'd say that Pic9 is the best of those (and maybe the best overall). The other two definitely look like they are showing overshoot from too much P.

Yes when I talk about rise time, I'm referring to the angle of the ramp up to the plateau.

I think you are probably done with the current loop tuning. If you look at the trace images in the AMC app notes, theirs often weren't any better. And they mention that the drive's voltage and the motors inductance are often the limiting factor there.

You can probably set the other drive/motor combo with the same settings and be good. But if you want it won't hurt to test it to. To change the test.demo to send the signal to the other drive, just edit the hal file to change which analog out the command goes to.

Todd, thank you! As I look at my notes & look at the USB stick, I need to be confidant that all is correct, there appears the o-scope saved 2 files with the same name (as I see in the "files" the only difference is the small icon before the file's name is only difference, the images look the same. In addition to the group of files that are uniform in the "scale of the units" currently, stops at pic 34=110(chart units) that pic is in prior post. I am going to try the 120 & 130, with uniform scale set large to full screen, this will give me setting from 50 to 130, a 2nd review of these settings.
I am wondering, how would you consider motor placement? The motor that the scope is connected to now is the "newer" to look at, of the 2 motors. The axes we are tuning is the "x" axes (the long travel, with less weight to move) the other motor (just looks older, brushes look same in both) has always been in the "x" axes, and the "newer" motor was at the "y" axes with the short travel, but with more weight (carrying the saddle&table&vise) vs. just table&vise,(no part considered). I am considering changing motors, and tune this, the newer of the 2 motors to the "y" axes, thinking the new motor is better suited for the heaver of the 2 loads. May go though from 50 to 130, on other motor with this drive. I am open to suggestions.

Many thanks!

Edit: Attached are 1 pic, the same settings as Pic9 (in prior post) but scale set to 500 for Channel 1 & 2. Pic9 (setting=120) this attached set 120 also, but fill screen. The last part of the file name is the original file name as pic number. but I renamed the beginning of a small group of files to reflect the settings, so on USB stick it stands out. many thanks

It probably makes little difference, so long as they're both in good working order.