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

Todd, Many thanks!! Attached is a set of traces of the "X" axes. File name is the setting from the SW3 settings/chart. I have not ponder these traces yet. The scale (the length of the "cruse" part of trace is longer) changes about 10 traces in. I started lower than the default setting, so the file name and SW3 setting are set the same. I am going to look at the traces and reflect tonight.
When we talked of "bracketing" of the adjustments/tuning, does it apply here? Look at 1 end of what is acceptable in the traces and go to other end of acceptable and pick a point in the middle?

Many thanks!!
 

Those look much nicer. Where did the bump in the analog output go?
From those I think I'd pick #10.

Todd, many thanks! These latest traces Pot 1, 3 & 4 are set mid-point, Pot 2 is 11 turns CW on both sets of .zip files, the earlier traces I am not shore of the Pot settings.
Should I consider getting 1 more set of only the 3 or 4 traces (with the settings for the widest & tallest trace, as to fill up o-scope screen) centered around the "best trace so far", as you have suggested trace #10? My thinking is that the 1st .zip fie and part of the 2nd .zip file are set with a narrow trace, and this suggestion makes all traces completely identical.

Many thanks

Todd, Thank you!! Attached are 10 traces, starting at 80 going to 120, from chart. File name 1 (the file name is long, but has a number at end) is 80 on chart for the "X" axes go to file name 5 (this is 120 on chart). File name 6 is "Y" axes and starting at 80 on chart, going to 120 on chart is file 10. Please review, these look the best.

Many thanks
 

I think you are done with the current loop tuning. Which trace do you think is best? I think probably set the drives SW3 switches to the 100ohm setting. Then reconnect the motors to the machine, set SW1-1,2,3,4 to the off position, unhook your o-scope from the current monitor pin and change it to the velocity monitor pin #15. Now it is time to try to tune the velocity loop in the drives again.

I think I would do it similarly to how you were tuning the current loop. Using the siggen test program. You may want to add back in the encoder settings for the axis you are testing so that you can look at the velocity feedback of the encoders in Halscope. Ask if you need help with this.

I think you will also need to change both the scale and frequency of the Siggen to make them more appropriate for a velocity command. It may need to be a lower voltage, and a slower frequency, depending on how fast the drive can accelerate the motor and axis. The frequency needs to be low enough for the machine to achieve a steady state move. That might be 10, 5 or even 1 hz. It might be a good idea to ask AMC for recommendations for good voltage levels and frequency to test at. (Other wise I think I might start with a 0.5v and 5hz signal and see what happens.)

Make sure that the machine is near the center of travel, only test 1 axis at a time, be prepared to quickly stop the test if things go wrong. You may need to adjust pot4 to stop drift. Then try to use pot 1 and 2 and SW2 switches to match the wave of the response to the square wave of the command, as closely as possible.

The response will be significantly more trapazoidal than the command's square wave. Just make it as good as you can get it. You will have to calculate the velocity feedback on the o-scope to know how fast it is going (the scaling will already be right in Halscope.) and figure out what your volts/speed of your command should be.

Once you have the velocity loop tuned in the drive, you can look at the slope of the velocity feedback of the transitions to calculate the max acceleration, and set the max acceleration in your Linuxcnc config to something like 75% of that.

Finally after all that you can open Linuxcnc again and tune the position loop there.

(PS I'm going camping for the weekend bye.)

Todd, Hope all your travels are filled with fun and relaxation. Many thanks for all of your help, I am very grateful. I will ponder your suggestions, have a safe trip.

Many thanks

Todd, Thanks for all of your help! Attached is a pic of the files with the encoder files included in the Halscope dir, just to make shore all is good.

As a review, at A-M-C drive (the P1 connectors on drive, change from Pin 8 to Pin 15 connecting channel 2 probe) as shown in this post (page 28): (15 Aug 2022 12:16 #249774); Pot 1, 3, & 4 turn to mid-point; Pot 2 is 11 CW (this, as per motor calculation requirement); SW1 (1 to 6 = OFF); Use the 100ohm setting (ref. A-M-C table on page 6 of 9 of the AB15A100 manual): motors are remounted on machine.
I have reached out to my A-M-C engineer, they out of the office until Monday (the 29th) with an email, I will wait until later in week to follow-up. I plan to use "auto" when ready.
I have not powered-up for the 1st time, but ready if the attached Halscope pic is correct.

Many thanks!
 

Todd, hope your travels were safe and fun for all. I have applied power & tried to get the "test" to send signal to motors, so I can try o-scope on the velocity loop tuning. The attached pic of the error is attached, I could not get to the siggen screen yet.

Many thanks
 

I'd strongly suggest testing only one drive at a time so be sure to comment out the Y axis analog output line while testing the X.

You had that error because you replaced the old analog output line without removing or commenting out the newer ones.
Here is a new test.hal that has both encoders set up for the X and Y and separate net lines for the X and Y analog outputs (with the one for the Y commented out.) 

When you want to switch to the Y axis simply remove the comment for the Y output line and add a comment (#) to the one for the X.

 

Also, according to the manual the velocity output pin (#15) on the drive is scaled at 22kHz/volt. That will most likely not be quadrature, so to compare with the encoder settings in Linuxcnc you'd need to divide by 4.

If you want your machine scaled so 120ipm is the max velocity, a 1V command should be 0.2 inches per second. Your encoder scale is 27500 edges/inch. So I think a 1v command to the drive should show 0.2*27500/4=1375hz, 1375/22000=0.0625v on the O-scope. (When every thing is tuned right (that or 0.25v if I'm wrong about the quadrature).)

If you are looking at the encoder feedback in Halscope, you'd look at the velocity feedback of the encoder which is in ips, so there you would be tuning for 0.2ips for a 1v command.