Acorn Trials and Tribulations - PROBLEM SOLVED

All things related to the Centroid Acorn CNC Controller

Moderator: cnckeith

diycncscott
Posts: 634
Joined: Thu Mar 04, 2010 4:03 pm

Re: Acorn Trials and Tribulations

Post by diycncscott » Tue Nov 28, 2017 4:20 pm

I'm not sure that we haven't reached the limit of what can accurately be measured with your setup. The max error you report is ~.004" and you indicate that there is no mechanical connection between the scale and the Z axis.

In the tests, except for the last 3 readings when the control cable from the acorn to your diff board was swapped, your average error is less than .5 thous (~.000409") across all measurements.

Try this. My experience here, using lasers that measure in the sub millionths, show that an axis will coast into the lash band even at slow rates (10"/min). We use a laser (and ball bar) to map every axis on all of our 5 axis, precision racing machines. Early on, we were seeing some inconsistencies in the laser measurements which I was able to prove where being cause by inertia driven "coasting" of the table into the lash band -even on a very heavy, box way machine with less than .0005" lash. I created a command that "creeps" into the starting position and ending position.

Use this approach and command for all measurements I'll use X axis in the example (all units imperial):
Do not jog the machine or reverse direction at any point in the commands listed below. If a reversal is needed, start over at step one. You can write a program or use MDI to command the moves. HINT - You can use the up/down arrow keys to access the MDI history to greatly reduce typing.

1. Move to -.05 on the axis. G0 x-.05

2. Move to X0 at a VERY slow rate: G1 X0 F.25

3. Set axis dro to 0 G92 X0

4. Command move to X9.9 G1 X9.9 F10

5. Command move to X10 at a VERY slow rate. G1 X10 F.25

6. Record DRO and scale readout

7. Lather, rinse and repeat with different feed rates.

cncman172
Posts: 94
Joined: Thu Nov 02, 2017 4:07 pm
Acorn CNC Controller: Yes
Allin1DC CNC Controller: No
Oak CNC controller: No
MPU11 & GPIO4D -w/ 3rd Party Drives: No
DC3IOB: No
CNC11: No
CPU10 or CPU7: No
System Serial Number: 985DAD429EE9-0926170118
Location: Westerville, Ohio

Re: Acorn Trials and Tribulations

Post by cncman172 » Tue Nov 28, 2017 4:31 pm

Keith,

I ran the text again using the Mitutoyo 12" digital caliper. That is kind of hard to hold and align perfect but here are my readings after running it five times. I did move the axis prior to doing the test to ensure no backlash would impact the settings by G1 Y.5 F5, then reset the indicator and the Y axis in Centroid software.

G1 Y10 F20 10.0085
10.0135
10.0025
10.0040
10.0020

I also checked to ensure no AC wiring was near the encoder or motor leads going to the amplifiers. Normally if that happens I get an error on the amplifier. It appears to overshoot in each case in the best case by only .002".

Russ

cncman172
Posts: 94
Joined: Thu Nov 02, 2017 4:07 pm
Acorn CNC Controller: Yes
Allin1DC CNC Controller: No
Oak CNC controller: No
MPU11 & GPIO4D -w/ 3rd Party Drives: No
DC3IOB: No
CNC11: No
CPU10 or CPU7: No
System Serial Number: 985DAD429EE9-0926170118
Location: Westerville, Ohio

Re: Acorn Trials and Tribulations

Post by cncman172 » Tue Nov 28, 2017 4:58 pm

Keith,
Sorry missed you previous post will do those moves again with the coasting adjustment of going almost there and then change the feedrate to F2.5 and see if it make any difference. When doing the testing on the Mitutoyo a very slight angle on the caliper could cause several thousands variance. In my case I had three readings that were .020, .025, and .040 over.

Anyway with things that close it is hard to believe that when I cut that plastic I was seeing steps greater than .015 causing the part to come out oval in the Y direction. The X measures correct.

I might try and cut the part again and see if anything changes.

You are correct I am getting to the limits of my ability to measure things to a higher degree of accuracy due to my indicator limits.

Russ

cnckeith
Posts: 520
Joined: Wed Mar 03, 2010 4:23 pm
Acorn CNC Controller: Yes
Allin1DC CNC Controller: Yes
Oak CNC controller: Yes
MPU11 & GPIO4D -w/ 3rd Party Drives: Yes
DC3IOB: Yes
CNC11: Yes
CPU10 or CPU7: Yes
System Serial Number: none
Contact:

Re: Acorn Trials and Tribulations

Post by cnckeith » Wed Nov 29, 2017 9:35 am

when you swapped the connections at the Acorn board between a good axis and the problem axis and reran measurement tests what was the result?
did the problem follow the cable swap or stay with the original axis?

cncman172
Posts: 94
Joined: Thu Nov 02, 2017 4:07 pm
Acorn CNC Controller: Yes
Allin1DC CNC Controller: No
Oak CNC controller: No
MPU11 & GPIO4D -w/ 3rd Party Drives: No
DC3IOB: No
CNC11: No
CPU10 or CPU7: No
System Serial Number: 985DAD429EE9-0926170118
Location: Westerville, Ohio

Re: Acorn Trials and Tribulations

Post by cncman172 » Wed Nov 29, 2017 3:39 pm

Well I decided to test the machine again after doing the various tests on accurate distance since there were only off 2-4 thousands on the Y and perfect on the X. Clearly something is very wrong, not exactly sure what is happening.
IMG_1890.JPG
In this photo you can see how the part continues to step away from the intended path. I generated the gcode so it was only taking .015 per pass which is nothing, but the more loops around the part the farther off the part gets.
IMG_1892.JPG
In this picture you can see the drawing below the plastic and see just how far off the part gets as it gets cut. Even the very first holes do not line up with the drawing perfect. If you look closely you can see these holes do not even end up in the center of the pockets.
IMG_1889.zip
(3.22 MiB) Downloaded 5 times
Here is a short video of the part being cut the movement is very smooth, but you can watch as it gets farther and farther on each pass.

At this point to validate my machine is still mechanically good, I am going to disconnect the Acorn and hook the Parallel Port and Mach3 and cut the same part, my machine has never been off like this so something is off significantly. Keith you mentioned noise, which is always suspect but on my servo system when I get noise the program normally crashes or the amplifiers get errors neither which has happened.


Russ
Attachments
IMG_1891.JPG

diycncscott
Posts: 634
Joined: Thu Mar 04, 2010 4:03 pm

Re: Acorn Trials and Tribulations

Post by diycncscott » Wed Nov 29, 2017 5:00 pm

Please see http://www.centroidcnc.com/downloads/ce ... manual.pdf for the correct use of G83.

You will need to specify an XY location on the G83 line as shown in the example in the manual or specify the current position in the list of holes to drill if you wish to drill at the position at which the G83 is issued.

From the manual:
G83 X0 Y0 R.1 Z-2 Q.5 ; drill 2" deep hole in 0.5" steps

Without the entire program I can't comment on what is wrong with the arc.

I've already shown that the Y axis variance is isolated to something in your Y axis control cable, diff board, drive or motor. If you perform the tests as I have prescribed and provide the data I've requested, I may be able to help you further.

cncman172
Posts: 94
Joined: Thu Nov 02, 2017 4:07 pm
Acorn CNC Controller: Yes
Allin1DC CNC Controller: No
Oak CNC controller: No
MPU11 & GPIO4D -w/ 3rd Party Drives: No
DC3IOB: No
CNC11: No
CPU10 or CPU7: No
System Serial Number: 985DAD429EE9-0926170118
Location: Westerville, Ohio

Re: Acorn Trials and Tribulations

Post by cncman172 » Wed Nov 29, 2017 7:15 pm

Scott,

As I suggested I was going to unplug the control cables and connect the parallel port and Mach3 to determine if there was any hardware issue on my machine.
IMG_1895.JPG
This is the same exact part, same exact machine, same exact power supply the only thing changed was the brain, meaning Acorn and the Centroid software, with the except I used three control cables to the servos but those are also identical. You will note the part comes out exactly perfect all sizes are right on the money, no skew at all.
IMG_1893.zip
(2.1 MiB) Downloaded 1 time
Here is a short movie of the machine cutting the same exact part just using Mach3 and the Parallel port.

As I ponder this issue it hit me perhaps the pulse width or rate leaving the Acorn is not allowing the Panasonic drivers to see all the pulses. This has nothing to do with calibration as you saw I could run my tests for 10" and was extremely close, and perfect on the X axis. I will pull up the specs on the Panasonic and compare them to the specs on the Acorn for pulse rate and pulse width. These were adjustable on Mach3, Mach4, ESS, and KmotionCNC, but I did not see anything in the documentation on those type adjustments on Acorn.

At least I know the hardware is still rock solid and I was able to get that Security Camera job out tonight.


Russ

cncman172
Posts: 94
Joined: Thu Nov 02, 2017 4:07 pm
Acorn CNC Controller: Yes
Allin1DC CNC Controller: No
Oak CNC controller: No
MPU11 & GPIO4D -w/ 3rd Party Drives: No
DC3IOB: No
CNC11: No
CPU10 or CPU7: No
System Serial Number: 985DAD429EE9-0926170118
Location: Westerville, Ohio

Re: Acorn Trials and Tribulations

Post by cncman172 » Wed Nov 29, 2017 7:17 pm

Scott,

I forgot to add that gcode does drill the first hole correctly on Mach2 and KmotionCNC, I tested it and it all works exactly as listed. It also simulates perfect. Will review the section you mentioned in the manual.

Russ

cnckeith
Posts: 520
Joined: Wed Mar 03, 2010 4:23 pm
Acorn CNC Controller: Yes
Allin1DC CNC Controller: Yes
Oak CNC controller: Yes
MPU11 & GPIO4D -w/ 3rd Party Drives: Yes
DC3IOB: Yes
CNC11: Yes
CPU10 or CPU7: Yes
System Serial Number: none
Contact:

Re: Acorn Trials and Tribulations

Post by cnckeith » Wed Nov 29, 2017 7:26 pm

Mach3 and others are renowned for making up there own way for a g code to function.
Centroid CNC software follows the industry standard G code as close as possible while others just decide on a different way to do it.
So, Scott is correct please read up on the code and how it works in the Centroid operators manual here. http://www.centroidcnc.com/downloads/ce ... manual.pdf there are good examples of each G and M code and how it works. Chapter 12 g codes and Chapter 13 m codes.

DICKEYBIRD
Posts: 252
Joined: Sat Jul 08, 2017 7:38 am
Acorn CNC Controller: Yes
Allin1DC CNC Controller: No
Oak CNC controller: No
MPU11 & GPIO4D -w/ 3rd Party Drives: No
DC3IOB: No
CNC11: No
CPU10 or CPU7: No
System Serial Number: n/a yet
Location: Collierville, TN USA

Re: Acorn Trials and Tribulations

Post by DICKEYBIRD » Wed Nov 29, 2017 7:51 pm

Just curious Russ, looks like a simple part, can you do it in Intercon & compare its posted code to what worked prior to Acorn? It may just be something real simple you can change in the PP to make your machine happy with Acorn. That's what I'm getting ready to do with my lathe & hope it turns out to be simple. :)
Milton in Collierville, TN

"Accuracy is the sum total of your compensating mistakes."

Post Reply

Who is online

Users browsing this forum: GraemeP, Tapburrbit and 1 guest