Re: Aug-30-99 EMC release set free

Ian,

I have noticed the .0001 discrepency in xEMC, but never paid enough
attention to it to notice if it only occurs on a single axis. I have been
running in commanded view and last night I switched to actual and the
discrepency went away.

Regarding a 4 axis EMC setup, the current idiots guide is you can't do it as
the g-code interpreter is strictly 3 axis. Fred said this is one of the
things in the works.

Tim
[Denver, CO]

> From: "Ian W. Wright" <ian@...>
>
> Hi Fred,
>
> I tried both commanded and actual and get the same result although, as
> my machine will use stepper motors I would expect to use commanded
> values until I get a DRO set up. The reason I mentioned it is that it
> only seems to affect the 'x' axis and happens
> even when there is nothing connected to the port. If the resolution is
> reduced to 0.0100 per step or less, it seems to correct itself but, for
> anything finer than this it seems I would have to home it then change
> resolution to 0.0001 and jog the axis to exact zero manually. As the
> other axes don't exhibit the same effect I thought it was probably a
> minor software problem but it may, of course, just be an anomaly with my
> machine. I expect other users could comment on this. Also, is it normal
> for the digits on the screen to noticeably flicker as a g-code script
> runs? I assume it is a factor of the refresh rate but wonder if it could
> be stabilised in future developments. I noticed your comments on 4-axis+
> use of EMC in another post and didn't understand a word of it! Since I
> will need at least 4 axes, is there any chance you could do a kind of
> short 'idiots guide' to setting this up please - I'm sure others on the
> list would also be very grateful for this. Thanks,
>
> Ian
>
> Fred Proctor wrote:
> >
> > Ian W. Wright wrote:
> >
> > > When I try to home the axes in Jog(Manual) mode,
> > > the 'x' axis almost always fails to zero exactly, i.e. it finishes on
> > > 0.0001 or -0.0001(sometimes -0.0002). The other axes zero exactly.

Also

> > > I have noticed that, if I jog the 'x' axis in 0.001 increments it goes
> > > 0.0100, 0.0110, 0.0120 etc up but 0.0109, 0.0099, 0.0089 etc down. I
> > > don't see this as being a particular problem and when running G-code
> > > scripts in step mode, they always seem to finish on the right figures
>
> > Are you displaying commanded or actual coordinates? Normally you display
> > actual coordinates so you can see where the machine actually is. You can
> > select which numbers are displayed from one of the xemc menus. If you
> > are displaying actual values, you may simply be off a bit, and
> > increasing the P gain would help. If you are displaying commanded
> > values, then this is a software problem. I haven't seen this myself but
> > it could be roundoff in the display code.