Spark Eroder pulse generator

[shortbus=]

Your welcome RB. No our machine didn't spin the electrode, but from other jobs that we did spin an electrode I don't think it would make that much difference.

I made a mistake in the frequency of the pulses, went back and read some of my references and they should have been - 2kHz to around 1MHz.

Another problem with the 3D EDM "milling" is that you need a way of deciding which of the axis is the one to retract on when it shorts out.

And using copper for an electrode needs all spark parameters to be correct. Real EDM quality graphite (Poco or similar) is the best. This graphite has a much higher melting point than copper and is easier to machine. Copper has a real problem with burrs when machining.

 

[Mr RB]

Thanks again for the excellent info.

 

[bsparky]

hi there first of all very interesting subject hope you don't mind me asking a few ?'s the resistors in dougy83 circuit r1 varies between 0.1 - 0.01 are these a special type ? there is no part reference in spice list for it & also the l1 47u inductor ?.
the on time's i have read about relate to micro ( or millionth's of a second ) as opposed to milli seconds , on time from 50 to 500 micro sec's off time from 10 to 100 micro sec's . the scope shows on times based on milli sec's ?

i have a friend who has a machine ( home made ) but its based on banks of resistors to limit the current on psu & has around 15 amp output & the heat it generates is immense. so you may be able to see my interest in dougy83 circuit design.

best regards

 

[dougy83]

the resistors in dougy83 circuit r1 varies between 0.1 - 0.01 are these a special type ?

Nothing special, though they obviously have to be able to handle the peak current. This may mean that they end up being wire shunts.

there is no part reference in spice list for it & also the l1 47u inductor ?.

Again, inductor has to handle the peak current.

the on time's i have read about relate to micro ( or millionth's of a second ) as opposed to milli seconds , on time from 50 to 500 micro sec's off time from 10 to 100 micro sec's . the scope shows on times based on milli sec's ?

The circuit shown has a current rise time limited by the value of the inductor. It takes 17us to go from 0 to 15A (with the assumed 1 ohm work-piece load). The fall time is over 100us. It would be possible to decrease the rise and fall times by adding extra switching components and a higher voltage reservoir. If you need microsecond pulses with square edges, then the circuit shown previously is not for you.

You could perhaps utilise a class-G amplifier type setup, where a number of input voltage taps provide the power to the linear current source. This will reduce the power dissipation immensely while keeping the benefits of a linear driver.

 

[bsparky]

hi many thanks for that dougy83 i will look into a class g amp setup.


best regards