The AM Forum

I was talking with Steve and I just lost a FET catastrophically mid-transmission.  It's going to be a bugger to replace since it's in the middle.  One on the other phase looks cracked and I'll replace that one too.  It's on the end so it should be pretty easy to do.

Anyway, such is life with homebrew gear.

$6 repair

Jon

Hi Jon,

There was some type of crackling distortion on audio peaks suggesting either a parasitic or loss of drive under peaks or something.

Signal strength was good :D

Regards,  Steve

Yeah, I was pushing it pretty hard audio-wise.  With conservative audio drive I notice that my negative peaks hit just at 100% with my positive peaks hitting about 140-150-ish, according to the REA modulation meter.  Maybe my negative peak limiter can't keep up with such high audio drive.  My peak reading watt meter never left 2k while I was talking.  The shooting flame from the FET was a pretty good show though.  I wish that I had it on camera.

Anyway, I ordered 10 more FETs from DigiKey.  They're much cheaper than tubes, so experimentation is practical.  

Jon

Hey, there's another advantage of PWM (I just couldn't resist  ;D  ). 

The PWM has an overload shutdown, so if at any time the current rises faster than the voltage, the overload detect will pick that up and interrupt the pulse chain and shut everything down gracefully (and quickly - in under 100us).

Not sure on fets,  but on bipolar you need to ensure their is enough drive power to overcome the drop in impedance as the collector swings higher.

This is the reason on bipolar you need to modulate two cascaded stages,  ie...   Drive and final.   Otherwise,  the driver falls apart when it's asked to source more power for the next stage as it swings its collector.

This effect,  I've seen,  also can  cause anomalies with pos peaks vs negative peaks.

Not saying this is the problem, but I ran into this building my first xmitter.   Had a single stage bipolar.   Went two driver / final topology,  I can exceed 400 pct pos peaks.

My 500 Watt class e single ended mosfet I didn't notice e this on,  but it only requires like 2 watts of sinewave for saturation.

--Shane
KD6VXI

Was reading the mail on you guys when that happened. You were there, then crackle, crackle, exclamation!!!, you were gone.

Doug

After looking the situation over again, it's going to be easier to replace all 8 FETs than to squeeze the soldering iron in to replace two.  It'll be surgery, but I'll completely removed both drain busses for a do-over. 

Jon