[CEStanley]

Hi all you pros,

Thanks for your advise before yu give it

Here is my question:
I have been updating my Grizzly Mini Mill G8689 with a CNC Fusion kit. I had to Lap the X axis ways but the Y and Z are OK. So I was calibration the X axis and ask Mack 3 to move the table by .525 inch. The movement was something like 1.235 inches. After setting the steps per inch to over 10K, the requested .525 was .521 in the positive direction and -.515 in the negitive direction. I did this many times and even set backlask to .004 but that had no affect. .521 and -.515 remain the same. I bumped backlash to .04 and that did not correct the runs.
I went to the Y axis and had simular issues. Positive backlash was larger or smaller than negitive backlash.
When I got to the Z axis, a .525 movement was short by .01 in one direction and .018 in the other.

Anyone have any idea what is going on here?

I have a Gecko G540 driver and NEMA 23 200 oz/inch motors and an running Mach 3 on a newly built PC running Windows 7 32 bit.

[MrWild]

Your ways are loose and the table is cocking.
Lash in your lead screw/nut
Slop in your bearings.
You have a metric lead screw versus an inch lead screw..

My money is your gibs aren't as tight as they need to be, and the table is cocking back and forth as the moves are made.

[kawazuki]

If its a Fusion kit all the screws will be 5mm metric.
Your steps per inch will be 10160 (calculated) - this should get you in the ballpark for all measurable movement. Then use the Mach function 'axis calibration' under settings, do a commanded move that is measured ie. move X +1.000" - measure the resultant move. Tell Mach that and it will recalibrate the 10160 to correct it in the real world.
That doesn't overcome the backlash! So once the axis's are moving in one direction the correct distances, you need to measure how far the move in a backward direction. So go X +1.000" and it moved 1.000" then move X -1.000" and that measures 0.995" then that axis has 0.005" backlash. Drop these numbers in to the backlash compensation settings and actually turn backlash comp ON and you should be close to good. Worth checking out Hoss's videos on youtube on this subject as there are some tweaks that make BL Comp a lot nicer to use!

[CEStanley]

kawazuki, thank you for confirmation that 10K+ is the right setting and yes, it is coming in at about 10160 +/- a tiny bit (less than 2). I did check out Mr. Hoss' videos and I was following them last night.

MrWild, Thanks for the thoughts. I will check but I did not measure any movement the last time I checked so the gibs should be about as tight as they need to be.

Well, back to it and maybe I'll be a little better in the luck department today as it is a lot earlier.

[CEStanley]

kawazuki, you were right on the money. I am now down to .0005 inch which is the accuracy of the digital meter I have.:cheers:

[nmkranz]

If its a Fusion kit all the screws will be 5mm metric.
...

Where did you get this info?

Per the CNC Fusion page:

Standard ballscrew kits use 5/8" Thomson ballscrews with a .200 pitch.

[CEStanley]

Where did you get this info?

Per the CNC Fusion page:

I got it by calibrating my X, Y and Z axis. kawazuki just affirmed that it was a valid calibration. On my first pass of the calibration of all three axis I had the "steps per" set to 2000. When I requested a movement of .525 inch, The value generated by the auto cal came back above 10,000.