Hi Jeremy,
After seeing that cheap harbor freight plasma cutter, I was thinking
how hard is it to make an inverter plasma cutter? So, I've decided to
build one. I don't expect anything close to a real plasma cutter, this
is mainly an "educational" project.
The torch, of course, I would buy (I'm not that crazy). But the power
supply is just a 5-20 khz squarewave H bridge into a high frequency
transformer. Output voltage is about 150 volts. The secondary of the
transformer is loosly coupled to the primary, as the secondary voltage
drops considerably while cutting. Secondary current is around 30 amps.
For the controller, I was considering using a PIC microcontroller (my
programming is better than my soldering). Current adjustment would be
nice, but initially I will just have an adjustable 5-20 khz squarewave
output. The mains will be rectified and smoothed with a electrolytic.
An H-bridge IGBT module (600 volts, 25 amps) will be pulsed from the
PIC. My question is, what type of transformer should I use? Is there
anything else I should consider? Any good books?
Getting a project like this to work is harder than it looks.
I'll say right now:
* I AM NOT AN EXPERT. Don't take my advice very seriously - I'm just
hoping to point you in the direction of reasonable looking
documentation.
* SAFETY FIRST. Be very careful with high voltage supplies. Keep one
hand in your pocket
. Be particularly careful with charged
capacitors. It's possible to get a nasty shock even with the power off!
Ensure that you are not the only one around while you work on the thing.
I attempted to build a high current H-bridge as part of my undergrad
thesis. My report (and some photos) are available here:
http://www.biccard.com/alan/thesis
You might consider reading the application notes I have referenced in the
bibligraphy - particularly those from International Rectifier.
The inductance of the leads holding the bridge together can result in some
nasty transients when you turn the H-bridge switches on or off fast. These
transients can destroy the switching devices and/or the gate drivers.
You will need to be extremely careful with layout. Use high power modules
instead of paralleled discretes where possible.
Use google to search sci.electronics.design for some good posts on snubber
networks by Win Hill. Read them all twice.
You will need a fast storage oscilloscope and a variable DC supply to
drive the bridge with while you test it.
When something fails, there will be a big bang and it will be very hard to
determine what failed first
. The best approach I know of is to test
the bridge VERY cautiously. I'd use a procedure something like the one
outlined below.
1. Use a light resistive load to test the bridge for the first time.
2. Keep the bridge supply voltage at zero, and look at the gate drives to
the IGBTs. Check that bridge phasing is correct, that the gates are not
being overdriven. Look for overshoot at the edges of the pulses (due
to wiring inductance oscillating with the gate capacitance). FETs and
IGBTs can be easily damaged by exceeding Vgs(max) or Vge(max). Be aware
that your oscilloscope probe has capacitance of its own.
2. Slowly increase the bridge supply voltage. Look at the
drain-source/emitter-collector voltages at the edges of the square pulses.
Zoom right in to the rising and falling edges. Look carefully for
very short spikes which would exceed the maximum Vds / Vce. Exceeding Vds
even very briefly can write off FETs. Dunno what the situation is with
IGBTs.
3. Look out for capacitive coupling from the drain/emitter to the gate. It
is possible for this capacitance to turn on the switching devices briefly
at inappropriate times (possibly with catastrophic results). For this
reason, do not to use gate resistors which are too large.
There was an interesting discussion on this group about the purpose of
gate resistors for fast switching of power FETs. There seems to be some
disagreement about why we put them there. Maybe they appease the Gods :>
I'm sure people more knowledgeable than I will post good advice on this
topic. I think the best advice I ever read on line-powered switching PSUs
was in the Art of Electronics (by Win Hill, Paul Horowitz)- "Buy it."
Best of luck - and let us know how it goes!
Alan