Re: [CAD_CAM_EDM_DRO] Re: EDM spark control system

Larry Edington wrote:

> > A sensor (either resistor or current transformer) detects the start of the
> > discharge. A delay allows the discharge to continue for some adjustable
> > time. then, the electronic switch is turned off, and kept off for an
> > adjustable
> > time. then, the voltage is turned on again, and it waits for the next
> spark
> > to start.
>
> Easy enough. I assume we are talking about microsecond delays and not
> millisecond delays? Sounds like the more sparks per second we can get the
> better. Any idea what the commercial machine's delay ranges are?

Another response to this thread was 1-9 uS ON time, and 8-50 uS OFF time.
That could go from over 100 KHz down to 16 KHz, assuming the discharge
started immediately.

> > The drives will probably work, but I suspect there would be a need for
> much
> > software change, or addition of sensors, etc. Anyway, I think these
> discharges
> > are in the range of hundreds of amps. The MOSFET's series resistance will
> > lower the peak current, but one 47 A MOSFET may not handle it. I was
> thinking
> > of using about 4 similar sized units in parallel.
>
> Oh, I wasn't suggesting one 47A MOSFET would handle it. That's just what I
> have on my board now. But I can open the CAD package, rip out the stuff
> that's not needed for an EDM control and drop additional MOSFETs on the
> board in just a few minutes. Software changes aren't a problem as the core
> of the project is already done. The user interface code and display code is
> still useable as is. I'll just have to modify the menus. If the spark
> delays are between 1 and 255 us, or 1 and 255 ms, I don't even have to
> modify the delay code. If we need smaller than 1us resolution, I'll need to
> change that code.

Sounds good!

>
> > Close. What you want is to increase feed when there is no discharge, and
> > stop feed when the electrode shorts to the work. Within that range, you
> want
> > to modulate feed based on how many sparks / second there are.
>
> If it shorts to the work, would you output a short circuit signal which
> would cause the control package to back up?

Yes, this would be the best way. EMC doesn't really have a very good system
right now to do this. it would not be hard to modulate the feedrate based on
external conditions. With a bit more hacking on the GUI, you could probably
get the backup and try again to work, too.

> How would you suggest we signal a cnc control package like EMC from the
> control electronics? I can easily provde a spark rate output or other
> control signals. But I've never seen an EDM run. I do have some cuts to make
> in the future that will need to be done with a wire EDM. So I have a vested
> interest in making it work. Or is it impractical to use EMC or other
> standard CNC package to control the EDM?

I think EMC would be best, as we can change what we need to make it
work right. The GUI can already change the feedrate (with feedrate
override) during a programmed move. It gets a bit more complicated
to either have the GUI completely control the move, or to have it back
up a bit and then resume the move. but, i suspect there are ways we
could get the right behavior. As for the interfacing, sending signals through
the
parallel port sense bits would probably work. At the most basic, one for
number of sparks /second above some preset value, and another for short
circuit would work pretty well. When sparks/sec is showing above
threshold, then the feedrate is held. if sparks/sec drops, speed up.
If short, then stop a moment, and if the short doesn't clear, start backing up.

> > I have no idea what the
> > peak current is. 100 A would be expected - and the sparks produced sure
> > look like VERY short duration pulses in that range. 1000 Amps would not
> > surprise me a lot! I do have the gear to measure it, but it will take a
> bit of
> > gear to set it up. I will need my digital storage scope to capture the
> > discharge.
>
> I also have a couple of DSO's and a good current probe. So maybe what I
> should do is wire up an RC version like you did and take some measurements.
> I'll have to go find some caps though. All of my caps are either smaller
> aluminum electrolytics or surface mount caps. I do have several isolation
> transformers to keep it all safe.

I have junk dating back to the 60's 9and even a few bits earlier). I got some
big, oil-paper caps in the 1.5, 3, 6 and 12 uF 250 V size. these were just
perfect for the EDM. Some of these are actually Aluminized Mylar, some are
real paper, I think. All are very low ESR, ideal for this use.

Jon