Re: stepper edm

Graham,
I'm more interested in your progressing than this geometry stuff,

I just think the cutting area is smaller when orbiting,
and that orbiting in XY has less active area than a fully seated electrode.

I understand your point that 'on the way down' the cutting is all edge cutting with plunge.
BTW: That kind of cut is a very good test of a machine ( a 'shave').

>
> --- In CAD_CAM_EDM_DRO@yahoogroups.com, tomp-tag <tomp-tag@...> wrote:
>
> > The slightly smaller size is still like a 2.002" dia ball in a
> >2.004" cavity, the interfacial area ( cutting area ) is small (in
> >theory a point ), and whatever that area is, that is the 'leading
> >edge' you speak of.
> > You always have a 'leading edge'.
>

Me too, an XY orbit, cutting on the ZX and ZY walls (if the z is fixed)
The whole face does not touch, in a cylinder only a line is in contact
But practicly in EDM an area of inch fraction and the height of the cavity is active.
We've all watched the orbit sparking go round and round
and never saw sparks all the way around. That's the leading edge I spoke of.

> I am refering to CNC orbiting, i.e. in the finishing stage the
> electrode is lowered into the cavity, the z is fixed and the electrode
> moved in the x,y plane. The small amount of material still to be
> removed is removed by the whole face, the wear is shared. In a simple
> sinking operation the edges of the finisher does all the cutting. In
> at least the example of a blind straight sided cavity I was thinking of.
>
> > and electrodes ( tools) made inexpensively,
> > then multiple electrode are faster due to a larger cutting area
> > ( shell mills cut faster than engraving tools )
>

A full contacting cut (sparks all around) is like a shell mill,
a small area approaches an engraving cutter ( the wear is concentrated not distributed)
Your view of edge cuttin is correct, the edge gets beat up because everything happens along the edge.
Even worse with corners.(than edges)
sorry if it's a bad analogy

> I really don't understand how an orbiting electrode could be
> considered an engraving tool compared to a shell mill.

Yes, 'wiggling' during a rough cut helps plunging a rough cut.
It 's main adnatage (to me) was to remove the 'slime' debris that collects on the vertical surfaces.
Only a few 'tenths' wiggle is desired, I find that wiggling a few thou slows me down!
(I do test trials on competing machines )
The reduction of this slime (spanking the the dirt off ) also increases the flow, i agree.

> Take a
> cylindrical electrode made slightly undersize undersize, (to account
> for the gap) then set it making a cavity, the surface area presented
> to the work is that of the end of the electrode, it is not smaller
> because it is orbiting, it is only slightly smaller because it is
> well, slightly smaller. If it lets you up the power because of better
> flushing then that's even better.
>
> > The number is relative, and the problem of multiple electrode
> > accuracy is a given element in the work equation.
> > Standard technologies from manufacturers will suggest up to 7.
> > ( ie: Mitsubishi & Agie, carbide < 1.2 um finishes )
>

> I don't really get your point.

More tools are suggested by professionals as the finish requirement increases.
You have to use more tools anyways.
So the idea of 1 tool one job isnt applicable across all jobs,
(and IMO not in many jobs.>)

> I think I've pretty much got basic 2-axis interpolated moves with
> trapizoidal velocity profiles programmed up in the propeller.

great stuff, did you handle the 'sqr root of 2 ' problem,
( when bresenham says both axis must be stepped to stay on circumference)
where the velocity requested must be reduced when 2 axis interpolate?
If not, just div the requested vel by 1.4

> I'm doing it with a single processor, it would be easier with multiple
> ones but I think the speeds should be good enough.

Are you thinking of a time period (step delay) that is analagous to the gap
(or difference between actual and desired gap ) ?

> For gap control
> use the code would need to be altered and effectively clocked by the
> gap conditions rather than the processor clock, I think :)
>
> Cheers,
>
> Graham
>