Hi Andy
if your path is a straight line
dont use linuxcnc , just hal
use steppers to advance
use an opamp window comparator
with a too high and a too low threshold
when your measured voltage is too high advance 1 step
else dont move, just wait
if too low, turn off the power , maybe retract
no linuxcnc needed
it cuts as fast as it can automatically
in most edm (any electrode material to steel workpiece ) the optimal gap
voltage is near 30V=
its almost 2 volts lower when roughing
and 2 volts higher when finishing
The discharge voltage Ua) varies with 'paarung' the pairing of material,
so for carbon wheel to wolfram, its will be even less,
i have seen stable cutting al low as 18V ( do not try to go lower )
a transistorized or FET generator is way better for MMR and wear.
to cut carbidem use a carbon of copper-graphite or copper tungsten wheel
edming carbide is mostly removing the binder.
using parrafinics is better than water but requres filtering and pumps
the nature of edm grinding is the 'best' flushing in all edm realms (
sink, orbiting, wire, hole drill )
because the tool to workpiece junction is continuosly new (like wedm,
but free on each side, unlike wedm )
so a bath of 'white spirits' may be enough
on AGie edm-grinders ( used mostly by sewing machine companies ) they
used thin ( low -visc) edm oil
there is no such thing as edm oil, its merely a layer in the cracker
tower that is removed and labeled (kerosene or edm fluid or axle grease)
get a copy of the OEL-HELD book on edm fluids
a higher supply voltge ( 80-100V=) makes it easier (and more accurate)
for a comparator to decide wether to adance , hold still, or retract.
a use of adaptive feed wull never apprach the needed jittery advance of
edm, adaptive feed never reacts negatively, it can only reacts 'less-ly'
the motion techniques for mills is not the motion for edm.
I did a lot of work for Union Special on such machines
http://www.unionspecial.com/
are you using a lathe spindle as the edm spindle?
the plans that Garret Visser link to at MIT are a good idea
previous 'the garden of edm' was good but remains unfinished
https://cscott.net/Projects/FabClass/final/edesign1.html
oh, splash/spray guards are neccesary, high rpm is not
hth
I'm kneedeep in rpi-gpio driver here . is **cp same as &array[] ??
hahaha i have to write a lot of test code to see how the original code
worked.
best o luck
tomp
On 1/31/21 8:28 PM, andy pugh wrote:
I have started to play around with EDM. The intention is to make an
EDM "grinder" to profile tool steel and possibly carbide cutting tools
using CNC-turned profiles. At the moment I am using a cobbled-together
XY table with a couple of STMBL drives. One of the STMBL analogue
inputs is used to measure the current through the gap. (ie using the
voltage across the resistor that a capacitor is charged through). Then
a PID controller tries to achieve a target current by adjusting the
adaptive feed pin. I think that it shows promise, but only partly
works. I am using a 50R resistor and a 1000uF capacitor. I suspect
that this is too high on both counts. I am working at 40V and it seems
that the tool welds to the work too readily. I suspect that I would
get different results if I controlled to the gap voltage, rather than
charging resistor voltage. And probably better still with some sort of
signal processing on that voltage. What combination of R, C and V
would be typical for a servo-controlled eroder? I imagine it might be
different to a "doorbell" style one. Current set up:
https://youtu.be/nxpmEFnmK-A
_______________________________________________
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users