Spice model assistance

[VE3ELQ]

Trying to learn how to use LTSpice to model an new modulator output circuit design.  Having limited success with the learning curve and getting a bit frustrated. Perhaps some of you folks who do this stuff could help.  The circuit is a Half Bridge using 2 IRFP4332 FETs with 170V DC input assuming a perfect DC source ( no R or L).  It drives a 4 pole butterworth inductive input filter L1 170uhy,  L2 220uhy, C1 1.15ufd, and C2 .47ufd into a 16 ohm load. Pulse width goes from 5% to 100%.

What I want to know most is the pulsed Drain current of the FETs, a parameter that I cant measure. Any help much appreciated.

73s  Nigel

[KQ6F]

I tried setting up your model but ran into a problem I've had before.  There is something in the SPICE model from Infineon for that transistor that LTSpice doesn't like and consequently the model won't run.

So I changed to the IPP600N25N3 MOSFET that is included in the LTSpice directory of parts.  It's not as beefy a part as the IRFP4332 but gives a pretty good idea of what is going on.

There is a problem running the model assuming zero DC source impedance.  The problem stems from the recovery time of the drain-source diode in the OFF device - the device that is turning on sees it as a momentary short circuit.  The peak current will vary depending upon how rapidly the active device is turning on.  In my model I assumed a 25ns rise time and that resulted in 900 amp peak currents!!  Slowing down the rise times to 50ns helped but not much.  The diode recovery time for the IRFP4332 IS 190ns typical while that for the other part is 127ns so they are roughly equivalent, at least as far as that's concerned.

Changing the DC source Z from zero to 0.5 ohm lowered the peak current to a more reasonable 300 amps.  The pulse widths in both cases are 5ns or less so it may not be a really big deal.

Ignoring the huge pulse currents, the current in the active device during ON time is about 10 amps which is consistent for a 16 ohm load.

I can send you the LTSpice model if you'd like.

[VE3ELQ]

Rod,
Thanks for modeling this and for your detailed analysis. I could not get it to run at all but this is my first attempt with Spice.

I ran up against lower FET diode recovery time in real life and cured it by slowing down the turn on of the upper FET with a gate R working into the Gate C. Same with the turn off delay of the upper FET which required even more gate R on the lower one. Its somewhat counter productive as it slows down the switching speed yet still provides good results. There are Bridge driver chips that provide for adjustable turn on/off delays thus maintaining the fast switching speed which I may explore further.

The 10A agrees closely with the working model now bench running at 40% pulse width at carrier giving 68V to the RF deck at 5.4A.  But I plan to push the modulator much harder with a different RF deck using 22A FETs so wanted to explore the modulator FET limits first.  These IRFP4332s are fantastic, they are running cold at 144 khz even with the gates slowed down a bit.

Yes please FW the file to play with and possibly post it, you may have to ne-name it to .doc if thats allowed.

73s Nigel

[VE3ELQ]

While reading the specs for the LM7301 RR I/O op amp came across an interesting circuit for reading current into a load. This could be used in a half bridge on the lower FET, source to ground, to measure pulse current with very little affect on the circuit due to its potential very low shunt resistance. It could be easily calibrated against a good ammeter under steady state DC conditions and with a 4 mhz BW should produce a descent pulsed scope trace at say 10A = 1V on the output referenced to ground. When I get time will build one on a small PCB and see how it works.  This could be a valuable tool for these measurements.
See article 8 figure 27

73s  Nigel
http://www.ti.com/lit/ds/symlink/lm7301.pdf