[andrewmb]

Hi All,

I am using the following code to cut a circular pocket (16mm diameter) in aluminum. To the best of my knowledge, this code should work.

Now here is my problem: when I actually cut the pocket, it measures out as 16.2mm, which is hugely over. I cannot for the life of me figure out why this is happening. I am thinking this is a mechanical problem with my vise or the collet (ER32 with CAT40 holder) allows major tool runout?

I would really appreciate any input that you have as my only two ideas don't really make sense and I'm a bit of a n00b.

G-Code snippet follows (Dimensions are inches, tool is 6mm EM ~.2362"D. No machine cutter comp. Flood coolant.):

N48 G90 X.9514 Y-.4331
N49 G43 Z.1 H01
N50 G01 Z-.0591 F3.75
N51 X.9711 F7.5
N52 G03 X.9318 R-.0197 F30.
N53 X.9711 R-.0197
N54 G01 X1.0302
N55 G03 X.8727 R-.0787
N56 X1.0302 R-.0787
N57 G01 X1.0892
N58 G03 X.8137 R-.1378
N59 X1.0892 R-.1378
N60 G01 X1.1483
N61 G03 X.7546 R-.1969
N62 X1.1483 R-.1969

[andrewmb]

An update:

Because I think this has to do with my vise, I checked the hole again. It is elliptical. 16.3mm along the y axis (parallel to jaw travel) and 16.02mm along the x axis (parallel to jaw face).

Is this due to some tolerance in my vise? Am I not clamping this material correctly? (it is a small piece of .25 thick solid bar. I'm using parallels and clamping in my vise.

As always, thanks for your advice

[awerby]

This sounds more like backlash than any error involving your G-code, vise or spindle. What kind of machine are you running this code on? Have you checked for backlash in the Y axis?

Andrew Werby
ComputerSculpture.com — Home Page for Discount Hardware & Software

[andrewmb]

I'm running it on a sharp SV2412S vertical mill. I don't think that there would be any backlash since the machine is brand new.

How would I check for that?

[awerby]

run the suspect axis in one direction, to a measurable point, and call that zero. You can use a dial indicator parallel to the direction of motion to do the measurements, since there's not a lot of distance involved. Then command a move in the other direction, say 10mm. Does the commanded position agree with the position you've achieved? Run it back to what should be the zero point. Does it arrive exactly where you started, or is there a discrepancy? If either of these measurements don't agree with the positions you've asked it to go to, then you've found your error, and it's either backlash or some other mechanical problem, like a loose pulley.

Andrew Werby
ComputerSculpture.com — Home Page for Discount Hardware & Software