Can someone run my simple gcode to see if they can get a 1 inch round pocket within +/- 0.002"?
I can't get anything better than 0.980"

I used the mach3-inch post processor to generate this gcode.

Tool used was a flat endmill 0.25"

If someone can try this out on some scrap and let me know the results that would be very helpful and great. Thanks.

G00 G49 G40.1 G17 G80 G50 G90
G20
(2 1/2 Axis Pocketing)
M6 T1
M03 S2400
G00 Z0.1250
X-0.0089 Y-0.0607
G01 Z0.0500  F25.0
Z0.0250  F10.0
X-0.0239 Y-0.0577 Z0.0223  F2.7
X-0.0442 Y-0.0442 Z0.0180
X-0.0577 Y-0.0239 Z0.0137
X-0.0625 Y0.0000 Z0.0094
X-0.0577 Y0.0239 Z0.0051
X-0.0442 Y0.0442 Z0.0008
X-0.0239 Y0.0577 Z-0.0035
X0.0000 Y0.0625 Z-0.0078
X0.0239 Y0.0577 Z-0.0121
X0.0442 Y0.0442 Z-0.0164
X0.0577 Y0.0239 Z-0.0207
X0.0625 Y0.0000 Z-0.0250
G17
G02X-0.0625I-0.0625J0.0000 F15.0
X0.0625I0.0625J0.0000
G01 X0.1250
G03X-0.1250I-0.1250J0.0000
X0.1250I0.1250J0.0000
G01 X0.1875
G02X-0.1875I-0.1875J0.0000
X0.1875I0.1875J0.0000
G01 X0.2500
G03X-0.2500I-0.2500J0.0000
X0.2500I0.2500J0.0000
G01 X0.3125
G02X-0.3125I-0.3125J0.0000
X0.3125I0.3125J0.0000
G01 X0.3750
G03X-0.3750I-0.3750J0.0000
X0.3750I0.3750J0.0000
G01 X0.3529 Y-0.0383 Z0.0192  F2.7
Z0.0442  F25.0
G00 Z0.1250
M5 M9
M30

I don't think trying on a different machine would prove anything tbh.  Do you have ballscrews?  Backlash compensation on? The code looks correct.

Yes I have ball screws, and yes I have backlash turned on for 0.001" which I wouldn't think would be enough to yield a 0.1" undersized pocket. I even tried yesterday to lower my steps per rev from 1600 down to 200 to see what would happen and still, I get the same undersized round pockets. The strange thing is that when rechecking the motor calibration, it travels 9" spot on in both directions for the x and 4" in both directions for the y. Only when I'm cutting a round or square pocket do I get this annoying abnormality.

I ordered another controller card expected to be delivered today to see if maybe there is something off with the motion card. Because in addition to this issue, there were several times when I was jogging and mach3 got stuck in jog mode and would send pulses until the motors crashed. The Mach3 DRO was still continuing to send pulses even after the motors faulted. The only way to stop this runaway jogging was to unplug the motion card USB cable. Crazy huh?

Thanks for your response. It is very much appreciated.

You need to MEASURE the backlash and set up the compensation to match.

try this code:

G00 G49 G40.1 G17 G80 G50 G90 G61
G20
(2 1/2 Axis Pocketing)
M6 T1
M03 S2400
G00 Z0.1250
X-0.0089 Y-0.0607
G01 Z0.0500  F25.0
Z0.0250  F10.0
X-0.0239 Y-0.0577 Z0.0223  F2.7
X-0.0442 Y-0.0442 Z0.0180
X-0.0577 Y-0.0239 Z0.0137
X-0.0625 Y0.0000 Z0.0094
X-0.0577 Y0.0239 Z0.0051
X-0.0442 Y0.0442 Z0.0008
X-0.0239 Y0.0577 Z-0.0035
X0.0000 Y0.0625 Z-0.0078
X0.0239 Y0.0577 Z-0.0121
X0.0442 Y0.0442 Z-0.0164
X0.0577 Y0.0239 Z-0.0207
X0.0625 Y0.0000 Z-0.0250
G17
G02X-0.0625I-0.0625J0.0000 F15.0
X0.0625I0.0625J0.0000
G01 X0.1250
G03X-0.1250I-0.1250J0.0000
X0.1250I0.1250J0.0000
G01 X0.1875
G02X-0.1875I-0.1875J0.0000
X0.1875I0.1875J0.0000
G01 X0.2500
G03X-0.2500I-0.2500J0.0000
X0.2500I0.2500J0.0000
G01 X0.3125
G02X-0.3125I-0.3125J0.0000
X0.3125I0.3125J0.0000
G01 X0.3750
G03X-0.3750I-0.3750J0.0000
X0.3750I0.3750J0.0000
G01 X0.3529 Y-0.0383 Z0.0192  F2.7
Z0.0442  F25.0
G00 Z0.1250
M5 M9
M30


Thanks for the gcode. I tried the code you provided and I noticed it used both conventional as well as climb to cut the pocket. Well unfortunately the results were the same. It's still off by approx .008" at this point, which IS better than I was getting before with 0.100" undersize. It turns out it's not the code. I bought a new motion card, and the first thing I noticed with the new card was the the "thumping" I was getting every full rotation of the stepper motor, that is now gone. No more thumping. Second issue I had with the old card, was "runaway jogging" of the motors. Mach3 would jog to infinity until I removed the usb cable. For now that seems to have gone away also. Fingers crossed.

I have only calibrated the x/y using a ruler. I may be able to shave down that .008" shortcoming to .003" if I calibrate the x/y using a dial indicator.

Again, thanks for your help. Running this test help verify its not code related.

you can also easy calculate the calculate the step's/per.

example:

steppers have 200 pulses per revolution. ( 360 degrees divided by 1.8 degrees = 200 )
Micro stepping = 10. So 200 X 10=2000
Motor to screw gearing is 1-1. So 2000 X 1=2000
Screw lead is .200 of an inch.
screw will have to turn 5 rounds to move the nut 1".
1.0 divided by 0.200=5.
So 2000 X 5=10000.

UPDATE:

My undersized pockets issues are fixed. I replaced the motion card with a new one which solved my runaway jogging issues. I chnged my gcode to take less aggrssive cuts and slowed down my feed rates and that got me within 0.002" tollerance. Apperently the fast heavy cuts was causeing too much deflection possibly. I also greased my ways and ball screws with PTFE Super Lube and it made a big difference belive it or not. I was using bar/chain lube and it didnt seem to work at all. I kept feeling vibration in the table at a certian spot when jogging. After lubing the ball scews and all the ways, its as smooth as silk the full length of travel.