[bill south]

I appear to have an issue with the x axis accuracy on my machine. I noticed some parts that have shifted in the X axis only by about .020". I have checked the backlash and it's almost nothing (.002"). I made a test cut on a piece of 1.5" round stock where I drilled a half inch hole in the center on my lathe then transferred it to the Pro. I dialed in the zero on the hole using my electronic edge probe and double checked it with a second probe. I then ran a simple profile cutting off .125 at .25 inches deep. The cut is off set to the right in the x axis .020". The cut is exactly 1.375" +/- about .002. which is very good for this machine. It is just off set from center.
I'll post a bad cell phone picture that may help show what the resulting cut is. It's almost like there is a setting in Mach that is consistently offsetting the cut by the .020. I've checked everything I know to check in Mach and all appears to be good. If I had a mechanical problem wiht the table, the roundness of the cut would be off (in my tiny little mind).
Any ideas will be greatly appreciated. I don't want to tear the machine down unless I really have to.
Attached is the Pix and the NC file I'm using for testing.
All help is appreciated.
Bill

[UniqueMachining]

Bill

Just a shot in the dark. Looks like BOBCAD code. I had this happen to me a while ago.
When aligning the stock using the stock wizard, somehow an offset was applied.
As it is only .020 it might not be noticeable on the screen.
I realize that if this is happening consistently, the above is unlikely. See Below.

As an aside your backlash should be sub .001. Sounds like some adjustments are in order.
I would check the X ball screw nut tightness and the X gibb adjustment.
Both are relatively easy to reach and do not require any dis-assembly.

Good luck
John

[Novakon]

One other possibility I have seen in a couple of circumstances before is the X axis coupler being loose. This would give you an oblong cut on the disk, but since you didn't indicate if it had a true overall diameter dimension in the X and Y direction, It could be this. A partial rotational slippage would give you this kind of error. It works out to about a 36 degree rotational slippage for the coupler if this is the case.

The backlash should be below .03mm or .0012". I think you can get this a little better also. GIBS and thrust bearing nut tightness are the prime causes for slight errors like this.

Regards,
John

[bill south]

Thanks Guys;
The fact that blows me away is the accuracy of the circular cuts. They are dead on at 1.375 (maybe .001 due to measurement error). In my mind, if there were any loose parts, they would impact the accuracy of the cut.
That being said, I'm going to inspect the X-axis mechanics and try to determine if the error is mechanical. I'm also going to repeat the test with code from Aspire instead of Bob. That will eliminate any software induced errors should there be any in the post.
Again, thanks for all the help. When I find the culprit, I'll post the corrective action for future use if needed.

Bill

[SCzEngrgGroup]

Since the hole is coming out round, I don't see any way this is a mechanical problem. Before tearing it apart, I would first do a test using hand-crafted code, to make sure the CAM isn't screwing you.

You post suggests you set X0Y0 to the center of the hole, but that is not what the code is doing. Where WAS the origin set?

Regards,
Ray L.

[bill south]

The origin is set at the hole x0 y0 using an electronic probe then verified using a different edge finder.
" I don't see any way this is a mechanical problem "
That is what has me baffled thus the reason for generating the same code using Aspire which has its own Post Processor. It almost has to be an offset of some sort that just needs correcting. Let the hunt continue!
Thanks for your input.
B

[SCzEngrgGroup]

The origin is set at the hole x0 y0 using an electronic probe then verified using a different edge finder.

Bill,

Are you saying X0Y0 is the center of the hole?? Because that is NOT what the g-code says....

Regards,
Ray L.

[bill south]

Hey Ray;
Yes, the origin was set up on the hole center by probing x then y to find center. I guess I don't quite understand what is causing any errors in the Gcode. I'll use another program to generate the code (unfortunately I never learned to code by hand) and try again. Something is most definitely out of whack. Oh yeah, I'm using bobcad ver. 26 for a cad/cam program. Maybe something in the post processor?
Thanks for taking the time to respond.
Bill

[SCzEngrgGroup]

Bill,

Hand-coding a quick and dirty circular interpolation with the hole center at X0 Y0 is trivial:

G0 Z0.1 (Set Z above top of part)
G0 Xr.rrr Y0 (Move X to hole radius at 3 o'clock position, r.rrr = HoleRadius - ToolRadius)
Ff (Set plunge feedrate)
G1 Z-0.1 (Plunge to Z-0.1)
Ff (Set cutting feedrate)
G2 Xr.rr Y0 I-r.rr J0 (Cut circle)
G0 Z1 (Retract)

Just put in appropriate numbers for r.rr, and plunge feedrate, depth and cutting feedrate, and see what you get

Regards,
Ray L.

[Kenny Duval]

If it's off in the G-Code could there not be a lead in and lead out position causing the issue?

[kvom]

G03 arcs are centered on X0Y0, so that looks OK. G02 arcs seem to be some sort of leadin. Did you check G54 value settings?

[Hirudin]

I haven't used any of them, but perhaps someone here can recommend a G-code editor that would plot the path and allow it to be measured. Of course, with such a small amount of code it could be plotted by hand in a CAD program as well. It looks like kvom may have done a little of that already.

[LeeWay]

I don't know much about Bobcad, but I know in Turbocad, you can easily shift the workplane if you are not careful. A slight shift in the work plane makes some very odd code. I hit plane by World often and then place on workplane for 2D stuff.

[MichaelHenry]

I haven't used any of them, but perhaps someone here can recommend a G-code editor that would plot the path and allow it to be measured. Of course, with such a small amount of code it could be plotted by hand in a CAD program as well. It looks like kvom may have done a little of that already.

G-Wizard Editor can handle that, can't it?

[SCzEngrgGroup]

G-Wizard Editor can handle that, can't it?

Simply examining the G-code will tell you exactly what it is doing - there are only 3-4 lines that really matter in this case.

Regards,
Ray L.

[kvom]

IJ mode for G2/G3 is very simple to discover the center of the arc. X of starting position + I value of G2/G3 = X coord of center. Same for Y.

[bill south]

Thanks Guys;
I'll do some more testing this weekend when time permits. Dang work got in the way this week!
B

[Hirudin]

G-Wizard Editor can handle that, can't it?

That's the impression I get too. If someone knows of an editor that can check g-code easily it might be useful for people who are unwilling or unable to scan the code unaided. Or to use with questionable code that is long or complicated.

It just seems like "here's a tool you can use to troubleshoot" would be better than "I used my big brain to check it out and it's perfect; take my word for it".

[MichaelHenry]

Simply examining the G-code will tell you exactly what it is doing - there are only 3-4 lines that really matter in this case.

Regards,
Ray L.

Just answering the question Ray.

[bhurts]

I don't know about newer bobcad but 17-21 would backplot g-code to cad. Also mastercam and alphacam are capable of doing so. Not sure about any others.

Ben