Re: [CAD_CAM_EDM_DRO] Re: H bridge progress

Lee Studley wrote:

> Another important feature that is overlooked initially is the
> fast acting over-current protection circuit. The constant
> off-time or a syncrounous-off time circuit is described in the
> application notes of the L2xx bridge part. Its a good
> starting point. Also at the voltages you guys are playing with,
> IGBT's may need come into play.

IGBTs are very tricky. They need to be driven HARD, because any
time you let the IGBT go into the linear mode, current hogging will
occur, and the device will be destroyed. You need a very clean and
'firm' drive of the gate waveforms to get the transistor through the
linear region faster than emitter current can build. (You also need
the reverse, cutting the transistor off quickly, and allowing the parasitic
inductances to be supplied by ultrafast recovery diodes.) At 80 volts
or so, the IGBT is not a good choice. Over 200 V it starts to look
very good, and above 400 V it really shines. An IGBT running at rated
current will have an emitter-collector drop of 2 to 2.5 V. A low voltage
FET can come pretty close to that, and beat it on switching losses.
A 600 V FET can't beat the IGBT unless you use a 30 A FET at 5 A.
Typical 600 V FETs have a drop of 10 - 15 V at rated current, and the
power dissipation becomes real significant.

>
> Discontinuities in the brush commutation on bigger DC servo motors
> can add a impulse noise that can do strange things to a H-bridge. I
> made a servo drive that I was testing to 180vdc. It worked fine at
> voltages up to 80-90vdc, but then as the motor got loaded, you
> could see massive ringing spikes(that varied with motor RPM that
> seemed to rise quickly in proportion to the current especially
> if the motor was near being stalled and above 90vdc. Since
> this was my first drive design, I had overlooked the current foldback
> features needed to keep the MosFETs in the safe operation region.
> Needless to say I popped a few. The international rectifier appnotes
> have some good tips on reducing the overshoots and protecting the
> bridge drivers. It awfully dull reading however.

I use a very expensive 2-pole output filter on my PWM servo amp.
The inductors are some of the most expensive parts on the board.
But, it has a number of advantages, not just the reduction of radiated
RF from the PWM carrier. It slows the rise time of fault currents so
that the fault logic can respond in time, even to a fault to ground.
It also prevents brush noise and armature inductance from getting back
into the FETs and causing excessive transients. It must be working, as
this (my 3rd generation, at least) hasn't popped a transistor in 2 years!

Having nearly 90 V dropped across a stalled motor sure develops a lot
of inductive energy in the armature windings. I can't imagine the motor
can take that for long!

Jon