Re: [CAD_CAM_EDM_DRO] EDM design

milespaul2004 wrote:

>Hi.
>Can anyone help.
>I have almost finished building a small spark erosion machine to use
>in my home Model Engineering workshop. I am have a finished but basic
>prototype working but I wish to modulate the voltage with a PWM
>signal and I am having some difficulty in getting output stage to
>work.
>I want to use an IRG4PC50KD IGBT to control the 70V DC 30A output.
>Radio Spares Technical said no driver was required but after several
>failures (I only have four IGBT's left)
>

IGBTs are very tricky devices. They CAN NOT be operated in the linear
region at all, even for a microsecond! If alowed to operate in the
linear region,
the spot on the transistor die that has the highest gain will hog all
the current,
and heat up. Because there is a positive temperature coefficient of gain,
this area will continue to hog current until it fails. Only when forced
into
deep saturation does the gain coefficient level out, and current sharing
across the die will occur. The only way to force the transistor into deep
saturation within one microsecond is to deliver HUGE current to the
gate. I have used the International Rectifier IR2110 and 2113 drivers
with IGBTs and FETs with excellent results, and recommend them
highly.

If you are trying to get linear operation from your circuit, you will
need to
switch to FET devices, which share current and withstand current spikes
much more gracefully than IGBTs. I'm wondering if the peak currents
of the spark action are forcing the IGBTs into the linear region despite
proper gate drive.

If you can be sure the current peaks are within the proper operating range
of the IGBTs, then the IR drivers should solve your problem.

Jon