Another DYN4 and Acorn Losing Steps Issue

[cnckeith]

hello!

- if you reference a document (guide from DMM?) please post it.

- Are you using the DMM cable? or did you make your own? if you made your own is the cable twisted pair with shield and ground? if so, how is it connected to the acorn and dyn4, supply good photos and even a sketch of how and where you connected the drain. show good photos of your cabinet wiring, make a photo album with good hi res images with that iphone and post a link here to the album.

I can help you debug, but you must cooperate and use the setup guide i attached in my first post.

[slodat]

Walkaround of the machine, showing wiring, and a detailed problem description. I posted this same video in the first post of this thread in an effort to provide as much relevant information as possible.



Yes, I am using the DMM DB25 cable. Since the cable came from DMM, I asked them how to use it. They sent me this DMM DYN4 setup document:

DYN4 AC Servo Drive to Centroid Acorn - 5040426.pdf

(390.85 KiB) Downloaded 123 times

I also have a detailed thread with every part number I used in my servo conversion , linked in my signature of each post.

Video of successful bench test:



Video showing the axis moving well in both directions directly from DMM's drive software:



Control cabinet photos:


I'm now headed to the machine to change the counts/rev and repeat the tests from the first video. I'll post my results. The reason I used 16384 is for resolution..

[slodat]

The only deviation from the setup guide you posted and my original configuration was the GEAR=4096. I changed the gear to 2048 and steps/rev to 8192. The problem persists. Granted, there's half as many pulses, so the problem is less dramatic, but I am still losing encoder counts.

G0 moves with encoder position for reference:

Code: Select all

g0x59		-6284769
g0x0		 -439025
g0x59		-6285145   376 counts
g0x0		 -439452   427 counts

[cnckeith]

ok..great.. now we are on a base line...

did you try at 200 khz?

does it only happen on one axis?

[cnckeith]

and for reference..

[slodat]

I lose less counts at 200khz.

It’s happening on the Y axis also. This video shows it happening on Y with the higher resolution. I’m waiting to sort out the lost counts on the Y before I do any huge speed testing on the Y because the racking it caused on the gantry creeps me out.

[cnckeith]

ok that is a big hint.

i would un bolt the axis motors from the machine until you get this sorted. put position indicators on the motor shafts so you can visually see shaft returning to an exact position.

a common debug technique when one axis has a problem and another does not is to "swap" drive cables with another axis. un bundle and place all drive cables as far away from any other wires as possible, and move the problem motor to another axis that doesn't have an issue. swap at drive end of cable and swap at acorn end of cable and test each case.. does the problem follow the cable or the drive?

and describe how the electrical cabinet is grounded.

[Richards]

When you're in the middle of a frustrating problem, it's hard to step back far enough to see what works properly and what still needs to be fixed. I spent the entire day yesterday chasing down a communication problem on something totally unrelated to Acorn. The basic problem was that I thought I understood the communication protocol, particularly since my first test worked. I relied on the results of that first test all day long and got nowhere. It wasn't until I re-opened the manual and re-checked each parameter that I realized that I had been doing things wrong all day long.

The hard part is admitting my error, after all, I've been involved in almost every aspect of computing since 1976 when I bought my first IMSAI kit. Since then, I've designed and built dozens of complex process control computers, some that had hundreds of I/O points and precise timing requirements. Those years of experience didn't make me immune to not understanding something new.

Just as Keith said, check everything. I would go even a little further than that. Even if something seems to be working, check it by the book anyway. If I had done that yesterday, I could have solved the problem in 30 minutes instead of wasting an entire day.

[slodat]

Up to now I’ve used the encoder absolute position from DMM software as my position indicator to tell me I’m not returning to the same commanded position (0 and 59” in the previous example). I have swapped cable/drive/motor combos between x,y,y2. At low enough speeds 50-100ipm I don’t lose counts. I am losing encoder counts on all three motors, in the many different combinations. The higher the output frequency from Acorn, the more encoder counts I lose, on all three axis. On the Y this results in racking of the gantry. I have no indications of electrical noise. Just losing steps between Acorn and the drives.

There are several threads on here where guys switch from the screw terminal outputs to the db25. Some go further and use a differential driver between the db25 and DYN4. When they do, they stop losing encoder counts. I’m wondering if the problem is the 2k pull down resistor in the DMM DB25 cable shares between the step and direction circuits not pulling down enough for both? Or the 24v screw terminal connections vs 5v db25 connections?

I sincerely appreciate the help, I obviously need it.

Thank you!

Steven

[slodat]

Richards,

That’s sound advice. I have gone over everything, from zero a few times. What I’ve come to, each time, is I’m losing encoder counts when speeds start to increase. The higher the speed, more counts.

Lots of talk on here about using the 5v signals from the db25 have me thinking the screw terminals work great on machines with lower feedrates.

Again, I sincerely appreciate the help and input. I obviously have something I need to get sorted out.

Steven

P.S. I’m not frustrated. I diagnose, troubleshoot and repair high voltage generator and transmission protective relaying for a living. This is just part of the process- getting my arse kicked while I figure out what’s going on