Odd Shaped Bore Holes

[ALSjosh]

Yes,I have run these exact toothpaths with the same tooling. Same feeds and speeds and same endmill.

[cncsnw]

I must be missing something. I downloaded the "op2_top.nc" file and graphed it, and it looks sort of like a circular pocket cleanout ca. 7.5" in diameter.

Where are the 2.5mm diameter holes described and shown in your first post?

I suspect you are running into an arc-chording issue, exacerbated by your comparatively low resolution (4000 counts/rev and only 1.7 revs/inch). That may be distorting small circular arcs. If so, it is likely you will be able to see this on the F8/Graph preview in CNC12.

[ALSjosh]

@cncsw I am not sure what you are seeing but it should look like this. I downloaded the NC file I provided in my previous post and it is showing correctly. This job has 10 small holes to be bored and a contour cutout.

[cncsnw]

Can you post a couple screen shots of the CNC12 graph of the NC file: one with F9/Zoom All, and one zoomed in so that one of the small holes takes up at least a quarter of the screen?

[cncsnw]

When I graph it, it looks like this.

[suntravel]

Looks ok for me.

But I still think it is a mechanical problem.

You really have no backlash on the 20mm pitch screws?

Uwe

[ALSjosh]

Zoomed out it looks just like Uwe's screenshot. Everything looks as it should. Zoomed in, it looks like the toothpath from Fusion. I will measure the backlash tomorrow but considering nothing with the ballscrew has changed, I still am leaning towards a software issue. In the video I posted to the Onedrive folder, you can watch the coupler at the stepper motor and see the jerking action. Is the standard Fusion post processor using arcs (IJK values) for Centroid what everyone else is using?

[ShawnM]

You may have your turns ratio dialed in now but they are outside the "sweet spot" for a router. Rather low actually at 1.26 and 1.27. What is the pitch of your ball screws? I'm guessing you don't have any gear reduction. I notice in the report you posted that you have no backlash value defined. It's HIGHLY unlikely you don't have any backlash. Have you run a backlash program and checked?

[suntravel]

There is no software issue in CNC12 causing your problems.

with 20mm pitch and 4000 Steps the shortest move possible is 0,0002 "

Bore 2 is not with G2/3 but very short vectors

If your machine is at Z -0,0011 and should go to Z -0,0014 it is not possible, Z-0,0013 or Z-0,0015 are the next possible positions.

Do not use short vectors for a machine with a coarse resolution and look after the backlash.

[ALSjosh]

I will be measuring and applying backlash compensation today. By "software issue" I dont mean a problem with CNC12, rather a setting this is incorrect or not optimal in either CNC12 or Fusion. What do you reccomend for the steps per rev and to avoid short vectors? I noticed this setting in Fusion?



Also, what is the explanation for being able to run this program without this issue on the stock controller but not on my new Centroid controller, other than incorrect settings?