Re: [CAD_CAM_EDM_DRO] Re: setting up the CNC mill

My controller uses the opto-isolators too. And they are the problem as you
say in your case.
I tested the same G code program with demo version of DesKNC for DOS and the
program and the machine work OK.
But when I use TurboCNC the error is here again. I don't blame program but
the opto-isolators or something with that.

In my case the problem is not in loosing step or steps.
Don't know how to explain but something like that is happening in my mill.
In each reversal from - to + (or + to - depends how is defined + direction
LOW or HIGH state) the first step in + direction is in - direction and all
other are correct in + direction.

For example 6 step in - and 6 step in +:

go-; go-; go-; go-; go-; go-; go+(but it still goes -); go+; go+; go; go+;
go+;

correct
| > > > > > > |
| < < < < < < |

wrong, what happening in my mill
| > > > > > > |
| < < < < < > |

Maybe the problem goes away if direction bit is send in one programming line
earlier then step bit.
But the problem will go away if I put away opto-isolators and I'll do that
just right now.
And I will still use the TurboCNC. It's really good.

Boris

> Boris:
>
> Do you mean that if you execute G01 X-1 and G01 X+1 once it returns to
> almost 1.000 but is off by some small delta, but if you repeat this many
> times it drifts away from 1.000 a little more with each execution? If it
> doesn't return to the same point after each loop it's more than a
> backlash problem.
>
> If this is the case it may be losing a step with each reversal + to - or
> - to +. Jon Elson identified a similar lost step in my machine. I was
> using EMC with an opto-isolator which inverted the signal.
>
> You might run this loop many times, say 100, and measure the error in
> returning to zero. If this is 100 times a single step (mine is .00025")
> you're probably losing one step each loop. If double it could be two
> steps, etc.
>
> I had the same error in both X and Y and it was consistently one step
> for each reversal in one direction. I tried a different controller and
> the error went away. I then took out the opto-isolator and the error
> went away with EMC. It was probably some wierd timing problem in EMC but
> once I got it working I didn't try to isolate the problem further.
>
> If this seems to match your problem try this. If not, never mind. :-)
>
> Hugh Currin
> Klamath Falls, OR