Cuts not repeatable, losing steps(FIXED) but will continue thread

All things related to the Centroid Acorn CNC Controller

Moderator: cnckeith

lavrgs
Posts: 486
Joined: Sat Aug 11, 2018 11:22 pm
Acorn CNC Controller: Yes
Plasma CNC Controller: No
AcornSix CNC Controller: No
Allin1DC CNC Controller: Yes
Hickory CNC Controller: No
Oak CNC controller: No
CNC Control System Serial Number: none
DC3IOB: No
CNC12: Yes
CNC11: No
CPU10 or CPU7: No
Location: Oregon

Re: Cuts not repeatable, losing steps(FIXED) but will continue thread

Post by lavrgs »

At a practical level I was unable to make parts that met print. Initially I used Intercon to mil a series of holes and then went back to cut a ring around those holes. I stopped the program after the first ring because it was not concentric. See the very first post in this thread. After lots of searching I found that SLODAT had a similar problem and was able to solve it by going through the DB25 instead of the terminal blocks. He was the first I found who defined the problem as losing steps. I went back into test mode and completed several runs on X and Y that moved 0>>-2>>2>>0. I did about 50 cycles and while the DRO said I was at Zero I used the DMMDRV app to monitor and it showed that I did not return to zero but was several thousand steps off...that represented the better part of 0.100. This same issue was exhibited in a set of nested parts that I made and when I laid them on top of each other the outsides did not match.
IMG_2460.jpg
The flat section was created because the program "ran out of material - i.e. it had overtravelled...Enter Kentonsj (Scott) to the rescue. He indicated he had similar issues losing steps and developed his differential board. TODAY, after having Scott's board installed for about a week I revisited making those parts, I was able to make four and they exactly fit on top of each other.
A7BA7C11-4ADF-4365-BAEB-07D16BF07ECE.JPG
. The testing I did last week after installing the board, using the DMMDRV app to monitor the steps after 50 cycles all axes returned to zero +/- 10 steps- usually better!!!

One more thing is that during the initial cyclic moves the lost steps would accumulate. However even a single move would not return to zero
Last edited by lavrgs on Tue Feb 08, 2022 10:37 am, edited 1 time in total.
martyscncgarage
Posts: 9912
Joined: Tue Mar 28, 2017 12:01 pm
Acorn CNC Controller: Yes
Allin1DC CNC Controller: Yes
Oak CNC controller: No
CNC Control System Serial Number: none
DC3IOB: No
CNC12: Yes
CNC11: Yes
CPU10 or CPU7: Yes
Location: Mesa, AZ

Re: Cuts not repeatable, losing steps(FIXED) but will continue thread

Post by martyscncgarage »

Thanks,
I would like to hear from Scott and/or Mike Richards on the technical side of things as well.

Marty
Reminder, for support please follow this post: viewtopic.php?f=20&t=383
We can't "SEE" what you see...
Mesa, AZ
slodat
Posts: 793
Joined: Thu Apr 12, 2018 11:16 pm
Acorn CNC Controller: Yes
Allin1DC CNC Controller: No
Oak CNC controller: No
CNC Control System Serial Number: none
DC3IOB: No
CNC11: No
CPU10 or CPU7: No

Re: Cuts not repeatable, losing steps(FIXED) but will continue thread

Post by slodat »

Marty, I did extensive testing on this 3-4 years ago. Posted ALL of my findings on here. Went through the same thing on my milling machine two years ago, because I incorrectly assumed it would have been fixed. Again, same tests, same solution. I have ran multiple hour test cycles on the mill and I'm never off more than ten encoder counts. Posted those results as well. I had attributed the issue entirely to DMM's DYN4 -> Acorn cables being connected to the screw terminal step/direction signals and using it's internal resistor to drop the +24vdc to the needed +5vdc. On my current lathe build (that I haven't had time to work on in a while) I bought Amphenol straight through cables and didn't even bother with the DMM cables. This will be the third machine I've built this way. For some reason, I could not get anyone's interest enough to solve the issue.

I do not use a differential driver, just the Acorn DB25 step/direction signals. Works very well. Again, I documented all of my work on here..
Richards
Posts: 692
Joined: Sun Sep 16, 2018 9:01 pm
Acorn CNC Controller: Yes
Allin1DC CNC Controller: No
Oak CNC controller: No
CNC Control System Serial Number: none
DC3IOB: No
CNC12: Yes
CNC11: No
CPU10 or CPU7: No
Location: South Jordan, UT

Re: Cuts not repeatable, losing steps(FIXED) but will continue thread

Post by Richards »

Losing steps can normally be blamed on two things: (1) Too much friction (steppers without feedback) and (2.) Electrical noise that interferes with a step pulse. Servo motors and closed loop steppers use internal encoders that when properly configured will detect missed steps due to friction. Electrical noise is an entirely different problem.

We normally blame ourselves when we have electrical noise because we assume that the motor driver is not noise sensitive, i.e., that it will function properly in an industrial environment if standard practices are followed (shielded cables, proper cable routing, etc.) However, a driver's electronics may not be as robust as we think. When several users who have followed generally accepted wiring practices give conflicting reports about missing steps, then we might justifiably wonder whether the driver should be blamed.

I only have one DMM DYN4 driver/servo. I've only used it on the test bench, but it has worked as expected, i.e., the encoder readout showed that a cut would be withing 0.001" over a distance of 6". I'll never be able to afford a mill that can hold tolerances tighter than that and I could never afford to run temperature control so that a precise mill could hold 10ths.

So let's jump in and look at how a differential driver can help. First of all, we should understand a standard single-ended signal, like signals used by the Acorn and most stepper and servo drivers. A single-ended signal uses GND as its reference point. A 5VDC signal will be about 5V above GND. A 24VDC signal will be about 24V above GND. A single-ended signal is usually very reliable because chip designers designed chips to respond to two separate ranges of values. For instance, the old TTL (74xx series) allow a LOW signal to be anywhere from 0V to 0.7V and a HIGH signal to be anywhere from 2.2V to 5V. Anything above 0.7V and below 2.2V that lasts long enough to be detected is not allowed. Take a quick look at the Acorn spec manual, page 7 to see the acceptable range of voltages. You'll note that LOW is 0.0 to 0.44V and HIGH is 3.5V with a 5VDC VCC supply.

A differential signal uses two opposing voltages, we'll call them A and B. If A is HIGH, then B is low. If B is high than A is LOW. Because of the A and B signals, GND is not used to carry return current. It most cases it is only connected to the cable's shield. So, how does an A and a B signal stop electrical noise? If some noise gets on the A signal, an equal amount of noise will be seen on the B signal so the DIFFERENCE in voltage between the A and the B signals remains the same. The noise is cancelled out. It's very effective. Look at the SATA connector between your computer's hard drive and the mother board. Differential signaling allows very high reliability and very fast speeds. The SATA connection replaced the older single-ended signal IDE ribbon cable.

Usually, differential drivers are paired with differential receivers. The driver splits the signal into an A and a B component. The receiver turns the A and the B components back into a single-ended signal. Usually a twisted-pair cable separates the driver and the receiver. The DMM DYN4 driver can accept a signal from a differential driver. It does not require you to use a differential receiver to reconstruct the signal. On the other hand, the Teknic manual warns against using just a differential driver. The voltage level between an A and B signal and the current being carried by that signal is too low to reliably trigger the Teknic ClearPath's step and direction inputs. Adding a differential receiver near the ClearPath motor would allow you to use differential signals between the Acorn and ClearPath.

This post is already much too long, but I wanted to try to explain enough that everyone could see why a differential driver works well with DMM DYN4 drivers. On my test bench, when I rigged up a differential driver, I had NO lost steps. NONE. If I ever put my DMM DYN4 driver/servo into production, I will use a differential driver to feed signals to it.
-Mike Richards
martyscncgarage
Posts: 9912
Joined: Tue Mar 28, 2017 12:01 pm
Acorn CNC Controller: Yes
Allin1DC CNC Controller: Yes
Oak CNC controller: No
CNC Control System Serial Number: none
DC3IOB: No
CNC12: Yes
CNC11: Yes
CPU10 or CPU7: Yes
Location: Mesa, AZ

Re: Cuts not repeatable, losing steps(FIXED) but will continue thread

Post by martyscncgarage »

slodat wrote: Tue Feb 08, 2022 12:06 am Marty, I did extensive testing on this 3-4 years ago. Posted ALL of my findings on here. Went through the same thing on my milling machine two years ago, because I incorrectly assumed it would have been fixed. Again, same tests, same solution. I have ran multiple hour test cycles on the mill and I'm never off more than ten encoder counts. Posted those results as well. I had attributed the issue entirely to DMM's DYN4 -> Acorn cables being connected to the screw terminal step/direction signals and using it's internal resistor to drop the +24vdc to the needed +5vdc. On my current lathe build (that I haven't had time to work on in a while) I bought Amphenol straight through cables and didn't even bother with the DMM cables. This will be the third machine I've built this way. For some reason, I could not get anyone's interest enough to solve the issue.

I do not use a differential driver, just the Acorn DB25 step/direction signals. Works very well. Again, I documented all of my work on here..
Yes. I remember.
I am interested to hear about the differential driver board Scott created.
Marty
Reminder, for support please follow this post: viewtopic.php?f=20&t=383
We can't "SEE" what you see...
Mesa, AZ
lavrgs
Posts: 486
Joined: Sat Aug 11, 2018 11:22 pm
Acorn CNC Controller: Yes
Plasma CNC Controller: No
AcornSix CNC Controller: No
Allin1DC CNC Controller: Yes
Hickory CNC Controller: No
Oak CNC controller: No
CNC Control System Serial Number: none
DC3IOB: No
CNC12: Yes
CNC11: No
CPU10 or CPU7: No
Location: Oregon

Re: Cuts not repeatable, I'm losing steps

Post by lavrgs »

kentonsj wrote: Thu Jan 27, 2022 3:49 pm
Richards wrote: Thu Jan 27, 2022 2:51 pm I'm interested in learning more about the Kentonsj differential board. I have not yet built a differential prototype for my DMM DNY4 750W driver/servo.
My board uses two TI AM26C31 quadruple differential line drivers to convert the TTL logic from the DB25 to differential. I also added some breakout so that I could combine the enable signals from the screw terminals and take the BKO output from the Z axis drive and faults signals in one neat package. Schematic_Acorn Differential Version 2_2022-01-12.pdf
This board is really just a 4 channel differential converter, it would work with any step/direction drive that accepts differential inputs. I use a DB9 connector and then solder a DB25 connector on the other end to match the DYN4 pins. Here's a wiring diagram of how it interfaces with the Acorn.
Acorn to DYN4 Differential Driver - Terminal Wiring (2).pdf
Here's the cable soldering/pinout for the custom cables.
SJ Thermy-AcornDMM Cable.pdf
This is what it looks like in my cabinet. Each cable runs to the DYN4, no DB25 breakouts and all the signal wires are shielded.
IMG_20190316_113214.jpg

I made and sold a tested board to lavrgs in hopes that it would help him eliminate lost step issues. It seems to have helped him. Keith, Marty and others on the forum have demonstrated that a differential converter is not required. Keith's most recent test from last week proves it isn't required. I mainly made the board to clean up the wiring in the panel, in the process, I designed this board with help from this forum. I thought I may as well take the opportunity to change to differential outputs and make the signal chain more robust. I'm not looking to go into business selling these boards, but I will sell tested boards to interested parties. I can sell you an un-populated board, or a tested board. I have one more design tweak to make on the next revision, I placed the DB9 connectors very close together and only the thinnest of DB9 connectors will work, a little bit more breathing room on that would have been a good idea. I was also dreaming about a re-design to make it strictly a 4 channel converter with no BKO DYN4 support, which would let me use a shielded CAT cable for all the connections, change to all surface mount devices and have a circuit board shop populate the board so I could lower the cost. When you hand solder the through-hole board and all the cables, it's a couple of hours' worth of work. 9 pins are required if you use the DYN BKO output, 8 without. I think a lot of people just use the acorn brake output, I used the DMM one to save outputs for a future tool changer, this was all designed before the ethernet board was available. DM if you'd like more information.
Scott
Marty It's already been posted. If you need more info you should PM Scott. The original post on 1/27 has attachments that detail the board and cables
martyscncgarage
Posts: 9912
Joined: Tue Mar 28, 2017 12:01 pm
Acorn CNC Controller: Yes
Allin1DC CNC Controller: Yes
Oak CNC controller: No
CNC Control System Serial Number: none
DC3IOB: No
CNC12: Yes
CNC11: Yes
CPU10 or CPU7: Yes
Location: Mesa, AZ

Re: Cuts not repeatable, losing steps(FIXED) but will continue thread

Post by martyscncgarage »

I was looking for a summarization.
Thanks anyway
Reminder, for support please follow this post: viewtopic.php?f=20&t=383
We can't "SEE" what you see...
Mesa, AZ
Richards
Posts: 692
Joined: Sun Sep 16, 2018 9:01 pm
Acorn CNC Controller: Yes
Allin1DC CNC Controller: No
Oak CNC controller: No
CNC Control System Serial Number: none
DC3IOB: No
CNC12: Yes
CNC11: No
CPU10 or CPU7: No
Location: South Jordan, UT

Re: Cuts not repeatable, losing steps(FIXED) but will continue thread

Post by Richards »

Here's a schematic that I used to connect the Acorn to a single DMM DYN2 driver. The schematic is based on Scott's PDF (listed in his posts above). I used a 2-channel UA9368CP chip instead of the 4-channel chip that Scott used. I used a DB25M breakout board on the Acorn and another DB25M breakout board on the DYN4 driver. I used the +5VDC and GND from the power supply that came with the Acorn to power the UA9368CP chip. I used +24VDC from that power supply to power pin 16 of the DB25M breakout board on the DYN4 driver and to power the Acorn. I clicked the Active LOW Enable box in the DMM software so that the DYN4 drive is enabled when the Acorn pulls the Enable signal low. The Acorn H2 Enable-1 signal is connected to DB25M pin 15 to enable the DYN4 drive. Pin 2 on the UA9368CP chip is channel 1 input, which is connected to Acorn DB25M pin 2 (step). Pin 3 on the UA9368CP chip is channel 2 input, which is connected to Acorn DB25M pin 3 (direction). Pin 8 on the UA9368CP chip is output (+) on channel 1. It is connected to DYN4 DB25M pin 11 (step +). Pin 7 on the UA9368CP chip is output (-) on channel 1. It is connected to DYN4 DB25M pin 23 (step -). Pin 6 on the UA9368CP chip is output (+) on channel 2. It is connected to DYN4 DB25M pin 22 (direction +). Pin 5 on the UA9368CP chip is output (-) on channel 2. It is connected to DYN4 DB25M pin 10 (direction -).

.
DYN4_DIFF.png
-Mike Richards
martyscncgarage
Posts: 9912
Joined: Tue Mar 28, 2017 12:01 pm
Acorn CNC Controller: Yes
Allin1DC CNC Controller: Yes
Oak CNC controller: No
CNC Control System Serial Number: none
DC3IOB: No
CNC12: Yes
CNC11: Yes
CPU10 or CPU7: Yes
Location: Mesa, AZ

Re: Cuts not repeatable, losing steps(FIXED) but will continue thread

Post by martyscncgarage »

Richards wrote: Wed Feb 09, 2022 2:28 pm Here's a schematic that I used to connect the Acorn to a single DMM DYN2 driver. The schematic is based on Scott's PDF (listed in his posts above). I used a 2-channel UA9368CP chip instead of the 4-channel chip that Scott used. I used a DB25M breakout board on the Acorn and another DB25M breakout board on the DYN4 driver. I used the +5VDC and GND from the power supply that came with the Acorn to power the UA9368CP chip. I used +24VDC from that power supply to power pin 16 of the DB25M breakout board on the DYN4 driver and to power the Acorn. I clicked the Active LOW Enable box in the DMM software so that the DYN4 drive is enabled when the Acorn pulls the Enable signal low. The Acorn H2 Enable-1 signal is connected to DB25M pin 15 to enable the DYN4 drive. Pin 2 on the UA9368CP chip is channel 1 input, which is connected to Acorn DB25M pin 2 (step). Pin 3 on the UA9368CP chip is channel 2 input, which is connected to Acorn DB25M pin 3 (direction). Pin 8 on the UA9368CP chip is output (+) on channel 1. It is connected to DYN4 DB25M pin 11 (step +). Pin 7 on the UA9368CP chip is output (-) on channel 1. It is connected to DYN4 DB25M pin 23 (step -). Pin 6 on the UA9368CP chip is output (+) on channel 2. It is connected to DYN4 DB25M pin 22 (direction +). Pin 5 on the UA9368CP chip is output (-) on channel 2. It is connected to DYN4 DB25M pin 10 (direction -).

.DYN4_DIFF.png
So you found that this was necessary even with the DB25 output from Acorn?
Reminder, for support please follow this post: viewtopic.php?f=20&t=383
We can't "SEE" what you see...
Mesa, AZ
Richards
Posts: 692
Joined: Sun Sep 16, 2018 9:01 pm
Acorn CNC Controller: Yes
Allin1DC CNC Controller: No
Oak CNC controller: No
CNC Control System Serial Number: none
DC3IOB: No
CNC12: Yes
CNC11: No
CPU10 or CPU7: No
Location: South Jordan, UT

Re: Cuts not repeatable, losing steps(FIXED) but will continue thread

Post by Richards »

Marty,
The DYN4 drive/servo still showed lost steps when using the DB25 signals from the Acorn. The number of lost steps (as reported by DMM software) was minor, less than 50 encoder counts when the total move was 65536 encoder counts. With the differential driver, there were no lost steps. The differential driver chip costs less than $1.50 from Digi-Key. The problem is building boards when there is no great demand for a solution. I just hacked things together using a white punch-wire type proto board. If I needed a permanent solution for a single machine, I would use a pre-drilled proto board and kludge something together. Taking the time to design a real board and having prototypes made isn't feasible unless there is a market for several hundred boards - otherwise the cost is prohibitive. DMM's DYN5 drive may be immune to lost steps making a differential driver unnecessary.

I doubt that there is any problem with the Acorn board. I've never noticed missed steps with DM542 stepper drivers or with Teknic ClearPath servos.
-Mike Richards
Post Reply