[DARYL]

The encoders are mounted to the motors they are 1000 count per rev the motion board quads this input to 4000 counts per rev
I havent even tried to use the backlash comp yet it is set to 0 in setup
My program is
G90
G70
G01 X0 Y0 F10
G02 I-2.555 J0 -----at this point X readout .0006 Y readout .0006
G02 I-2.555 J0
G02 I-2.555 J0
G02 I-2.555 J0
G02 I-2.555 J0-----at this point x .003 y.0035
M30
after last move I should be at x0 y0 with in reason , even if I program a G00 or G01 X0 Y0 after the circle the accumulated error still remains , it's like the motion board thinks it's at the position but the readout indicates it is someware else also the table matches the readout position
it dosen't matter what the circle radius is I still get the error and it tends to accumulate each pass I have tested radius values of .5 to 5 in. I have tried this in each G2 and G3 same results

[keithorr]

I know you wrote the software multiplies the count from 1000 to 4000 per rev. So the encoders are set up for quadrature, not pulse in the setup.exe? I had an accumulative error when I set quadrature encoders as pulse on my system. Much more sensitive to noise.

What does .0006 equal as encoder tics? More than one, less than one? Is this an accumulative math error or a reading error?

All couplers tight? No shafts slipping?

Happen at all speeds, fast and slow?

G02 and G03, or just one direction?

When you say the "readout" is .0006, is that the number on the screen when it should be .0000, or is that a physical measurement on the machine?