Re: emc for wire edm control

[Bob Simon <bsimon-at-qualmach.com>]

>
>I can see this approach also.  It could be implemented in either real time or
>at the interface level.  At the interface level we would just have to scale
>the signal from 0 -100 and put a set command in the tkemc update loop.
>
>Again at an interface level, if we wrote a separate tickle process like tkio
>we could run that loop much faster if we needed greater response that 0.2 or
>0.1 seconds.  If we needed 0.001 we'd need to get into the real-time
>processes.

Wire EDM isn't much different than a sinker. Here's some things I know 
about sinkers;

Early sinkers that tried using servos and steppers on the ram were very 
unsuccessful because servo response was too slow to handle the changing 
conditions in the spark gap. The inertia of the ram and motor drive was 
just too great for economical electric servos of that era. Most machines 
used fast, powerful servo hydraulic systems to overcome the inertia 
problems. You've heard of Moog valves...

The erosion speed/voltage controls available on most machines were purely 
analog in nature. One knob adjusted the cutting gap and another adjusted 
the ram speed. A more technical way of looking at it was the cutting gap 
knob adjusted the minimum gap voltage and the speed knob adjusted the servo 
gain.

The servo response on a hydraulic system is only as fast as the frequency 
response of the Moog valve. I'm not sure, but I think it's in the low to 
middle audio range. Obviously, electric servo rates could be much higher 
with modern motors, and amplifiers.

On to wire;

I think the slower pace of the older wire machines (1-2 cu.in./min.) would 
allow stopping in the cut until the short cleared.  Faster machines would 
probably have to reverse. I would guess that the reversal issue is were the 
first speed improvements were made in the history of wire machines.

I don't know what would be the easiest method to handle the servo gap/speed 
(voltage/gain) issue. Here's some thoughts...

Speed is an established parameter of CNC motion through programing and feed 
rate overide. We only need to tackle the gap condition issue. We need to 
control the motion proportional to the gap voltage to be effective. As the 
gap voltage approaches the preset value, the feed rate approaches 0. 
Reverse motion would begin if the voltage drops below the preset value. I 
think you could get away with a constant programmed reverse feed rate. It 
just needs to back up until the gap voltage is restored. I see an analog to 
digital converter in our future. Can you read the game port? I think they 
convert to 8 bits accuracy or so, maybe 12.

Not knowing much about EMC yet, I guessing EMC could handle the gap voltage 
preset control (knob) in the GUI much as it does with feed rate overide.?

-Bob