• The move to the new server is done. There are some software and database maintenance updates in process. This has us passing the hat around to help out. We appreciate any donations. Seriously, even a dollar helps. The payment page may be found here - https://www.audiokarma.org/support.html

Replacing VFETs with MOSFETs

I want to thank @PE9ZZ for his work. Doing the conversion on a trashed TA-4650, and I'm at the point where everything else works as intended, but am waiting the mosfets to arrive (prolly on Monday). Cheers! :thumbsup:
 
Damn. I got an oscillation issue with mosfets installed. Negative side of a test wave (1khz) shows oscillation hash already at 0.5Vpp output, both channels. That is without any bias. Bias alone will produce oscillation when high enough, without a signal. Investigating.
 
That is annoying. You may want to replace the drain resistors R333 and -4 with 10 ohm resistors of ¼ W so they blow up instead of your MOSFETs in case of destructive oscillation. Maybe Q304 needs a Miller-C. Put 10 pF across collector and base and see of that helps. I don't recall any oscillation but at the time (now 7 years ago) I also did not perform any stability tests like observing a square wave. That said, correct biasing is essential for proper operation and could kill the oscillations.
 
PE9ZZ said:
That is annoying. You may want to replace the drain resistors R333 and -4 with 10 ohm resistors of ¼ W so they blow up instead of your MOSFETs in case of destructive oscillation. Maybe Q304 needs a Miller-C. Put 10 pF across collector and base and see of that helps. I don't recall any oscillation but at the time (now 7 years ago) I also did not perform any stability tests like observing a square wave. That said, correct biasing is essential for proper operation and could kill the oscillations.
Click to expand...

It is not yet destructive at the bias and signal levels I'm testing. Naturally I won't crank it up until this is solved. Miller caps and various other compensation caps I tried made it only worse (lower oscillation threshold). Bigger gate resistors gave only marginal improvement. Might be a power supply issue, or something "simple" (wiring, relay etc.). Let's see.
 
No breakthrough yet. Interesting that I can wildly overbias either channel separately without a problem, but if I leave one at 50mv and start raising the other, when it reaches 10mv, instability ensues.
 
Without a load, and bias below 20mv, I can get the amp to clipping levels (about 80Vpp). When bias surges during clipping, the amp stays stable. When I try to adjust the bias over 20mv, without a load, instability kicks in. Fascinating and frustrating.
 
For the record, I'm using Exicon mosfets. While they are supposed to be Hitachi clones, the datasheets show 13-14% higher input capacitances. It could be the issue here.
 
Could be. 14% difference is not earth shattering though. I tried IRFP(9)240 with no stability issues (only thermal runaway).
 
Took me way too long to figure out that this isn't an oscillation issue at all. Oscillation is just a consequence of no bias. The real problem is the pow amp right channel, which goes into heavy conduction with bias and a signal. "Transistor protection" trs managed to mask this problem pretty well. Left channel is healthy.
 
JyriK said:
which goes into heavy conduction with bias and a signal. "Transistor protection" trs managed to mask this problem pretty well
Click to expand...
Huh? I have no idea what you mean by "heavy conduction". The whole raison d'être of a power amp is to conduct heavily to deliver enough current into the load to follow the input signal. So something goes wrong. What is it? The protection transistors cutting off too soon?
 
You must log in or register to reply here.
Back
Top Bottom