Southwest Industries Prototrak DPM conversion

[martyscncgarage]

The noise you are hearing is dithering between encoder counts. I would imagine it can be turned out. A video clip showing the motor and sound would be helpful along with the other information requested.

[CRM]

I did not get a report with baseline numbers as far as I remember; that would have been about 8 years ago though. I ended up having to shelve the project for a couple years just as my servos were returned with the new encoders installed so it is entirely possible that I was given information that I didn't realize I would need later and didn't take the proper care to file it accordingly.
Once I got back to the conversion, built the electronics cabinet and had the hardware in place, I did the initial testing and determined that the rapids were significantly slower than the original control was capable of (less than 100 ipm vs. 150 ipm). I disassembled the original drive and studied the layout; SWI was direct rectifying the current. I talked this over with Keith and said I would likely be OK to do the same; but understandably, he would not be liable for servo damage. I understood and accepted the risk. Once I converted over to direct rectification, the rapids returned to pre-conversion speeds so I am confident it will be fine. See the attached photos of the motor nameplate and the SWI servo board.
I've read elsewhere that it is now preferred to make a Google album of all images in my build thread; I will consolidate them there in the near future.

SWI servo drive.jpg (54.86 KiB) Viewed 2877 times

The label reads: 160V 2.97 amps 2200 RPM

[martyscncgarage]

Well, if you have 110V it gets close to the 160V rating.
You are not using the old servo drive anymore correct? You are using the All in One DC?
I would have imagined Scott sent a file with baseline.
DO post a new fresh report when you can please. That will tell what is in there.

[CRM]

Marty, you are correct; NOT using the old servo drive. The original Magnetek servos are upfitted with 8000 count encoders and wired directly to the All in One DC. With a digital multimeter, I get a 120 VAC reading at my mains to the mill control (and isn't that the standard for residential single phase power?), so that's were the uncertainty about direct rectification came in.
I will post a report and get some video within the next couple days. Please stand by...

On a different, but not totally unrelated note:
I am too far down my current retrofit path, but now that the Acorn exists (didn't when I first started this odyssey) I wonder if it would be possible to use it in conjunction with the original SWI drive/motor assembly? That would make an incredibly affordable conversion for other people. The trick would be determining the signals (step, direction?) that the SWI computer module sends to the drives, and on what pins. Hmmm...

[martyscncgarage]

When you take 120vac x 1.414 you get 169.68VDC. it may be ok. Maybe a good tune will resolve. Marc or Tom may chime in. You *might* be able employ Marc for a remote tuning session, but I'm not volunteering him.
The report will be helpful.

I can't answer to the SWI drives. Do they take step and direction input? I'd be surprised if they did.

[CRM]

Well, so much for getting the servos tuned.
Went to run a report as requested by popular demand. Then decided to try changing Kd value after re-reading my tuning documentation and notes. Almost gave up and ready to restore the original value (3.000) after lowering it by .25 several times, but when I lowered it one more time to 1.000 the servo I was working with (Z axis) got much quieter when holding steady position. Cool! Finally making progress. Ran the PID collection moves and watched the scope. While trying to balance the following error overshoot I accidentally entered 0.40000 instead of 0.00400 in the Ki field, instantaneously sending the Z axis into violent oscillation. Before I could react to hit the e-stop, I heard a snap and saw a flash from the electronics enclosure. I powered down and checked fuses and breakers, finding the glass fuse for axis 3 (Z axis) on the ALLIN1DC board blown and the Capboard AC supply 15A breaker tripped. I left the fuse out and the breaker tripped, then tried to power up. LED 1 on the board (the numeric display) failed to light and the PC was unable to ping the ALLIN1DC board. Close inspection of the board found no visible damage except a blown transistor chip at location C91, near the axis 3 output terminals.

But I doubt that is the only thing damaged. I've got an email off to Centroid support to assess my options.

FWIW, here was the report I had ran just prior to blowing up the ALLIN1DC.

report_0301110110_2019-11-29_16-09-04.zip

(5.04 MiB) Downloaded 127 times

The build part of this retrofit was pretty fun. The servo tuning part has been incredibly frustrating...and now expensive.

[martyscncgarage]

Well, so much for getting the servos tuned.
Went to run a report as requested by popular demand. Then decided to try changing Kd value after re-reading my tuning documentation and notes. Almost gave up and ready to restore the original value (3.000) after lowering it by .25 several times, but when I lowered it one more time to 1.000 the servo I was working with (Z axis) got much quieter when holding steady position. Cool! Finally making progress. Ran the PID collection moves and watched the scope. While trying to balance the following error overshoot I accidentally entered 0.40000 instead of 0.00400 in the Ki field, instantaneously sending the Z axis into violent oscillation. Before I could react to hit the e-stop, I heard a snap and saw a flash from the electronics enclosure. I powered down and checked fuses and breakers, finding the glass fuse for axis 3 (Z axis) on the ALLIN1DC board blown and the Capboard AC supply 15A breaker tripped. I left the fuse out and the breaker tripped, then tried to power up. LED 1 on the board (the numeric display) failed to light and the PC was unable to ping the ALLIN1DC board. Close inspection of the board found no visible damage except a blown transistor chip at location C91, near the axis 3 output terminals. 20191129_174552.jpg But I doubt that is the only thing damaged. I've got an email off to Centroid support to assess my options.

FWIW, here was the report I had ran just prior to blowing up the ALLIN1DC.report_0301110110_2019-11-29_16-09-04.zip

The build part of this retrofit was pretty fun. The servo tuning part has been incredibly frustrating...and now expensive.

Sorry to hear that happened. Hopefully Centroid will be able to repair it.
There is a bit of black magic experience to tune DC servos.
Marty

[CRM]

I have video, but can't seem to post it up. It's in .mp4 format. Am I doing something wrong (likely) or have to convert to a different format?

[martyscncgarage]

I have video, but can't seem to post it up. It's in .mp4 format. Am I doing something wrong (likely) or have to convert to a different format?

You have to upload videos to YouTube and the share a link here