Problems with circles

[bbergami]

Hi,
I recently completed my first router which I intend to use on wood and aluminum.
I've adjusted my x and y axis to travel 1" within .001 on a dial indicator and to travel say 12" and return to the same spot.
Today I tried to pocket a 4.25" circle, .25" deep using a .125 end mill taking .050" per pass. I noticed on the first pass that diagonally opposite 0,0 there was material left between each trip around the circle and the closer to 0,0 each trip around overlapped removing all the material. On each pass afterwards the circle would shift towards 0,0 maybe .020 or .030". Basically I'm getting a circle fatter at one side.

Any suggestions on where to start with this would be appreciated.
Thanks
Bert

[IJ.]

Same pitch on X and Y leadscrews?

[wjbzone]

Have you checked your backlash?

[bbergami]

Both x and y are 2 turns per inch. I did have to adjust them slightly different in the software (DeskCNC) to achieve 1" of travel when I told it to jog 1".

[bbergami]

I probably need to check the backlash a little closer. When I was adjusting it was returning to the same spot.

[IJ.]

When you're cutting and watching the readouts is it going to exactly the correct reading at the diagonal of 00? (just thinking if it's a mechanical thing or not or maybe a software setting)

[bbergami]

The readout shows it to be where it's supposed to be. I was wondering if I was losing steps somewhere. But I put some pretty big motors on it (680 in/oz).

[kong]

You could try cutting the same code at a very slow speed - as slow as you can wait for. If the error is the same, then it shouldn't be lost steps due to loss of torque.

[bbergami]

I did try it at very low speed. 2 ipm. I also tried it fast. Same thing. It does seem that it accumulates the more times around the circle. I'm wondering if my settings are a problem. Originally I had it set at 3200 steps/inch. When I jogged 1" I was actually getting something like .991 in one direction and .982 in the other. I adjusted my steps per inch upward on both so it jogged 1". They are slightly different on each axis. But it was cutting the circles off at the original 3200 spi too.

[kong]

Surely your steps per inch should be the same for each axis?

[bbergami]

That was my thought. But I wasn't getting even movement on each axis. Is there something else I should be looking at ?

[High Seas]

You could elminate RFI (from the router) as a suspect cause by putting a marker (felt tip/pencil/etc) in the collet and try running again without the router on. Like maybe one axis cable is too close to the router and not the other?