Re: [CAD_CAM_EDM_DRO] Re: emc with camtronics more help please

bgroover1 wrote:

> Good suggestion, but it happends on all 3 axis and I've checked the
> pulleys and hardware. I don't change direction when I'm testing the
> steps to take backlash out of the picture. I run the machine down to
> where you can hear each step and it makes no difference.

So, when you command the move, does it click or not? If yes,
then it is mechanical. If not, then it is the inch -> mm -> inch
conversion. I certainly don't have such a problem on my servo
system, I can make individual moves down to .0001" with no trouble
at all. My X and Y axes have .00005" encoder resolution, the Z
is .000025". I can't get reliable table movement in .0001" steps,
due to stick-slip friction, but the motors and leadscrews do make
the move pretty reliably. (These numbers are not typos.)

> I'm thinking
> this is either an emc setup problem or a not enough computer
> problem. This setup works reliably with dos and windows demo
> software.

I can't believe it is insufficient computer speed. If you were losing
steps at the high end, or just couldn't reach a rapid feed rate you
were happy with, I could understand blaming the computer. But,
in the slow range, the computer speed is not an issue. Can you
really make .001" moves with complete reliability of table motion?
You don't mention your machine type, so I can't tell whether it is
reasonable to expect that or not.

The real test is if positions are repeatable. If you can make these
moves, observe positions being skipped, and then return to the
same point, you just have roundoff in positioning. If you CAN'T
return to the same position, then steps are actually being lost,
which is a completely different problem. One of the ways this
happens is violating the rules for setup and hold time on the
change of the direction signal into the stepper driver.

Jon