Acorn Trials and Tribulations - PROBLEM SOLVED

All things related to the Centroid Acorn CNC Controller

Moderator: cnckeith

Post Reply
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

Acorn Trials and Tribulations - PROBLEM SOLVED

Post by cncman172 » Mon Nov 27, 2017 2:41 pm

Team,

Well I finally removed all the parallel port stuff and Mach3 on my machine and installed the Acorn. I am having some issues with accuracy, no exactly sure what is happening. I am a newbee with the Acorn but have been using Mach3/Mach4, and KmotionCNC for many years.


This photo is all the old stuff removed, the mill was working perfect, as I used it to cut many aluminum parts always within .001, so went Mach3 did not go off any do something unexpected it worked great. The machine has ground ballscrews and nice nuts and linear guides, and panasonic servos.
IMG_1884.JPG
Now after getting all the motors working and homing all working I did some quick checks to see if the calculations for movement were correct. In my case the encoders are 2000 p/r and the 10mm ballscrews made it 2.54 as the multiplier. I put those in and cut a piece of plastic. The Y axis is moving on each pass probably .015 making that part oblong. You can see the steps in the plastic, never had this issue previously.
IMG_1882.JPG
Then I attached a long linear scale to check the distance for Y using travel of 10" and got 10.004" with a modified multiplier. On the X axis it was spot on with the standard calculation of 2000 x 2.54, so I modified the multiplier to 2.538784 using Marty's spreadsheet. Still not doing things correct. I also checked for loose coupling, nut, thrust bearing, and all were good. I also greased the ballscrew and made sure it was very clean.
IMG_1883.JPG
When I tell it to go out to 10" on Y and then tell it to return to Y=0 it comes back to zero using a dial indicator and the long linear scale, so does not appear to be backlash.

The next issue was a simple chunk of gcode to drill some holes. It works fine but the very first hole never gets drilled. I goes to that position and then just moves down to about z=.1" and then moves to the next hole and all the rest of the holes drill correctly, very strange. Works fine in backplot.

(MATERIAL - ALUMINUM INCH - 2024)
( T1 | 1/8 DRILL | H1 )
( T2 | 1/8 FLAT ENDMILL | H2 )
N100 G20
N102 G0 G17 G40 G49 G80 G90
N104 T1 M6
N106 G0 G90 G54 X-2.2818 Y-1.3174 S6000 M3
N108 G43 H1 Z.1
N110 G99 G83 Z-.235 R.1 Q.1 F4.11
N112 X-1.319 Y-1.319
N114 X1.319
N116 X2.2818 Y-1.3174
N118 X-1.319 Y1.319
N120 X1.319
N122 X0. Y2.6347
N124 G80
N126 M5
N128 G91 G28 Z0.
N130 M01
N132 T2 M6

The next issue is learning to use the Centroid software. I run the code it comes to a tool change, I have to change tools manually. The machine pauses and asks the user to install the next tool. I do this and it says press cycle start to continue. When I press this is starts moving the Z axis down towards the table. In my cause I do no have tool holders just the collect in the spindle. I never saw the opportunity to reset the Z axis zero. To get past this issue I just posted the second tool in another file. I am sure there must be something simple did not see this in the manual or a video. Any help here would be much appreciated.


Russ
Last edited by cncman172 on Sat Dec 09, 2017 9:22 pm, edited 1 time in total.

cnckeith
Posts: 514
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 » Mon Nov 27, 2017 2:56 pm

since you are not using the Offset Tool Library.
this is one way to measure each tool each time you load it.
Load first tool in gcode progrfam. got to part zero menu, F1 setup, F1 Part and set the z0 location with that tool #
run the gcode for that tool.
when the g code prompts you to load the next tool press ESC
Load the next tool. got to part zero menu, F1 setup, F1 Part and set the z0 location with that tool
press F4 RUN, F2 Search , CNC12 will automatically fill in the Gcode line that you pressed ESC on, press F10 to accept
Press cycle start and the Gcode program will be run from the place you left off.

sounds like a lot of steps, but once you do it a few times its quite easy.
but the best way is to setup the Tool Offset Library so you only every have to set a Z zero part position once and its good for all the other tools which eliminates the process described above.

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 » Mon Nov 27, 2017 4:37 pm

Keith,
I followed some of what you said. On a router which has just a collet and collet nut there is no way to pre-measure those tools and put them in the offset tool library. Why? Because when you use a router and it comes to a tool change you loosen the nut slide in another tool and tighten the nut. The depth the tool slides into the collet will vary, so that is why in Mach3, Mach4, and KmotionCNC and others allow you to set the tool height when you are required to change tools manually.

I tried to hit the ESC when the tool change command executed and told me to change tool. However, when i did this it said user cancelled program?

I will run down and test your instructions outlined below. Part of what I am dealing with is learning a new system. For example zeroing the G54, I now follow the multi-step process and have that all down now.

The bigger issue right now is trying to understand how the Y axis is so far off when cutting something simple like a piece of plastic. Tools are brand new and very sharp so nothing to do with drill due to dull cutters.

Other interesting thing is the X axis is spot on using 2000 P/R from the encoder and 2.54 for a 10mm ball screw. In Mach3/4 my steps per unit were 5080, which is exactly the same 2000 x 2.54 = 5080.

10mm ballscrew = .39369" 1/.39369 = 2.54 turns/inch

I also attached a dial indicator to the Y axis and can power down the machine and rotate the axis by hand the dial indicator tracks perfect when you switch directions, I have never had backlash on this machine since it has ground ball screws, by NSK.

The feedrate when cutting the plastic was 15 ipm so pretty slow. So missing steps on a servo motor would cause an error pretty quick, no errors are reported by the amplifier. I will run some more tests but very strange.

On the gcode example if you run this little chuck of gcode you will see it goes to the first hole position and never starts the peck drilling. It moves down to .1" and then immediately moves to the next hole and drill all remaining holes correctly. Is this potential a bug or some unsupported gcode, no errors are generated in centroid software.

More to follow. Thanks for the help.

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 » Mon Nov 27, 2017 5:41 pm

Keith,
Following your guidance I was able to do a tool change and reset zero on the second tool with no issue. I am playing with acceleration which could be causing the issue. My X value was .3 and my Y value was .5 seconds, which seems to suggest Y has longer to accelerate, I changed them both to .3 seconds. Will try to increase and see if that makes any difference. I am also going to clean the Y ballscrew very good, I can see some chatter when I move it back and forth. I removed the grease and replaced it with fresh but there could be some aluminum flakes in the balls from cutting aluminum since they are not 100% protected. The ballscrews do have shields but thin aluminum flakes could get pulled it with grease. Anyway checking everything to get the precision back to where it has been for years.

Russ

cnckeith
Posts: 514
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 » Mon Nov 27, 2017 5:49 pm

cncman172 wrote:
Mon Nov 27, 2017 4:37 pm


I tried to hit the ESC when the tool change command executed and told me to change tool. However, when i did this it said user cancelled program? << yes you are cancelling the G code program at the tool change, go to the part setup screen and reset the part Z zero with the new tool, then restart program at the tool change. this is basic simple way to do it. you can also custom edit the M6 code (the code the tells the control what to do when a tool change is commanded, we are adding this custom M6 that tells the control to "remeasure at every tool change" to the next release of software, you'll be able to select whether or not you want to remeasure at every tool change and how and where you want to do that) >>>

I will run down and test your instructions outlined below. Part of what I am dealing with is learning a new system. For example zeroing the G54, I now follow the multi-step process and have that all down now. << you can also use MDI mode and G92 to set zero locations.. read up on MDI and G92 in the operator manual.>>>

The bigger issue right now is trying to understand how the Y axis is so far off when cutting something simple like a piece of plastic. Tools are brand new and very sharp so nothing to do with drill due to dull cutters.

Other interesting thing is the X axis is spot on using 2000 P/R from the encoder and 2.54 for a 10mm ball screw. In Mach3/4 my steps per unit were 5080, which is exactly the same 2000 x 2.54 = 5080. <<< You will have to please attach a report for me to comment on this...>>>
viewtopic.php?f=60&t=1043

10mm ballscrew = .39369" 1/.39369 = 2.54 turns/inch

I also attached a dial indicator to the Y axis and can power down the machine and rotate the axis by hand the dial indicator tracks perfect when you switch directions, I have never had backlash on this machine since it has ground ball screws, by NSK.

The feedrate when cutting the plastic was 15 ipm so pretty slow. So missing steps on a servo motor would cause an error pretty quick, no errors are reported by the amplifier. I will run some more tests but very strange.

On the gcode example if you run this little chuck of gcode you will see it goes to the first hole position and never starts the peck drilling. It moves down to .1" and then immediately moves to the next hole and drill all remaining holes correctly. Is this potential a bug or some unsupported gcode, no errors are generated in Centroid software. <<< where did this program come from? please re create the same program using Centroid's Intercon conversational software included with CNC12 and take a look at how Intercon creates G code to do the same thing. you may have to modify the post processor of the cad/cam system you are using to match the Intercon generated G code.>>>

More to follow. Thanks for the help.

Russ

martyscncgarage
Posts: 720
Joined: Tue Mar 28, 2017 12:01 pm
Acorn CNC Controller: Yes
Allin1DC CNC Controller: Yes
Oak CNC controller: No
MPU11 & GPIO4D -w/ 3rd Party Drives: No
DC3IOB: No
CNC11: Yes
CPU10 or CPU7: No
System Serial Number: none
Location: Mesa, AZ

Re: Acorn Trials and Tribulations

Post by martyscncgarage » Mon Nov 27, 2017 6:14 pm

Russ
to be clear,
You did linear moves on X, Y & Z and checked them.
You mentioned you modified Y:
"Then I attached a long linear scale to check the distance for Y using travel of 10" and got 10.004" with a modified multiplier. On the X axis it was spot on with the standard calculation of 2000 x 2.54, so I modified the multiplier to 2.538784 using Marty's spreadsheet. Still not doing things correct."

Did that improve Y's accuracy?

Your also mentioned step over, That is only in the Y axis direction? I see steps in your part. If it were an open loop stepper, I would say you were loosing steps. DId you double check all electrical connections from Acorn to your Servo drives?

Could you temporarily swap X&Y axis to the servo amps and see if the problem follows?

Marty
"Jack of all trades, master of none!"
Mesa, AZ

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 » Mon Nov 27, 2017 6:30 pm

Marty,
YES, I tested all axis and the only the Y axis did not come out perfect. These are NSK ballscrews so they are ground and have no backlash. I did discover when I tried to do the moves at higher feedrate the error would get worse. This seems to point to acceleration being a potential issue.

I can attach a dial indicator and rotate the axis by hand unpowered and the dial indicator follows perfect, never a delay in the needle on the indicator when changing directions. I also can send the axis say 10" and then used the MDI command like G1 Y0 F30 and it comes back to zero.

At feedrate F15, and distance of 10.0 and it would go 10.004" with an adjusted ratio
At feedrate F50 and distance of 10.0 and it would go 9.45" which again suggests perhaps velocity or acceleration but velocity is very slow with those feedrates.

Keep in mind all of these same servos and drives worked perfect in Mach3, I took a working system and just replaced the brain.

I am also taking the ballscrew out for close inspection to ensure there is no binding.

Russ

martyscncgarage
Posts: 720
Joined: Tue Mar 28, 2017 12:01 pm
Acorn CNC Controller: Yes
Allin1DC CNC Controller: Yes
Oak CNC controller: No
MPU11 & GPIO4D -w/ 3rd Party Drives: No
DC3IOB: No
CNC11: Yes
CPU10 or CPU7: No
System Serial Number: none
Location: Mesa, AZ

Re: Acorn Trials and Tribulations

Post by martyscncgarage » Mon Nov 27, 2017 6:47 pm

I would like to see you swap the X & Y axis at Acorn and see if it now does the same thing with the Actual X axis. (Your axis will be swapped on the machine for the test) What we want to see if now the actual Machine X axis now having the problem or does the problem stay with Y.

Marty
"Jack of all trades, master of none!"
Mesa, AZ

diycncscott
Posts: 631
Joined: Thu Mar 04, 2010 4:03 pm

Re: Acorn Trials and Tribulations

Post by diycncscott » Tue Nov 28, 2017 7:37 am

Please ALWAYS post a report.zip when asking for help. F7-Utility->F7-Report

Without the report, it's all just speculation.

Along with the report, let's start with:

1. Since your X & Y appear to be identically mechanically configured, go back to your original configuration (ie.. 2000 & 2.54 on both axes and accels at .5) - switch ONLY the X & Y control signals at the ACORN.

For each axis - starting at -.5, command a move to axis 0 at .5"/min. Confirm axis is displaying 0 or set the axis to 0.
Now, command a move to 10 at 10"/min (yes, I know it's slow but humor me.)

What does the scale read?

Does the problem move to the X? If not, the problem is in your wiring, amplifier and/or motor.

If problem does move to X, don't change anything. Describe your findings and send a report.zip.

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 10:14 am

Keith, Marty

As requested... On the results I expect a small amount of variation of a few thousands since there is no way to mechanically connect the spindle to the long linear scale, so I am pushing the display along the scale using a round rod installed in the spindle.

Russ

Reports.zip attached.
Experiments on accuracy
CNC12-D Mill V4.07 Dev Test, Rev 8

Encoders = 2000, Ratio = 2.54 on both axis

G1 Y10 F10 > Reading on long linear scale 10.014
G1 Y10 F30 10.012
G1 Y10 F80 10.011
G0 Y10 10.008

G1 X10 F10 10.003
G1 X10 F30 10.001
G1 X10 F80 10.020

NOW, Reversing Control Cables on X, Y

G1 X10 F10 10.020
G1 X10 F30 10.016
G1 X10 F80 10.013

G1 Y10 F10 10.058
G1 Y10 F30 10.089
G1 Y10 F80 10.111

Post Reply

Who is online

Users browsing this forum: No registered users and 1 guest