High Gain Phase Inverter - MOSFET Follies 8-)

[R.G.]

I did some work with some Zetex high voltage MOSFETs. These are a bit unique in that they are TO-92/E-line packages, same size as your garden variety small plastic signal transistor. They are 0.7W, 450V devices, and their input capacitance (Ciss) is only 55pF, unusually low for a high voltage MOSFET.

That's low enough in fact that a 69K rp 12AX7 plate will drive it flat to over 100kHz.

So I messed with this thing as a LTP. I put 100K plate loads on each side, used a 22K common source resistor and 1M bias resistors to a 68V zener bias point.

The result? It's beautifully balanced, produces high drive, and more importantly, has enough output that you could shape the "clipping" nature of the thing to be as tube like as you want to mess with. This thing will swing 200V peak to peak on each side of the LTP with a 350V supply.

And did I mention it had high gain?

It was so high that I had to put a monster divider in front of it to get it down to a signal level on the output tube plates that was not just all square waves all the time.

I modelled it up and the sim says that it has an open loop gain of a bit over 80db... That's enough to just plug your guitar right into the phase inverter and still overdrive the output tubes.

So I put some source degeneration into the LTP, 1K on each source before the 22K source resistor. That got it back down to only 46db.

For anyone wanting to play with the MOSFET Heresies, these devices are good fodder. The Ciss is low enough to let you get full audio band response from them in an environment which 12AX7 tubes work in. They're too linear to actually be tubes, of course, and they need enhancement biasing instead of depletion biasing, but other than that, they're a really useful adjunct to tube stages.

I've messed with the cascode setup a bit - depletion mode JFET running into the source of the MOSFET - and that's interesting as well, but I don't have any firm circuits yet.

Another thing I have to go mess with is current buffering a 12AX7. I have a couple of ideas on how to make a ZVN0540A multiply the current out of a 12AX7 by a fixed amount. That would let us use a "12AX7" as an output tube.

Sigh. So many circuits and so little time...

[Ray Ivers]

R.G.,

80dB gain! No worries implementing FX returns, reverb, etc. with this thing, eh? And what a killer input stage for a 'hybrid Zen-amp push-pull' kinda thing - two gain blocks and done.

I wonder how two depletion-mode devices would work in a 9-pin replacement 'super-PI module'? I could see people being interested in such a thing... and also a 12AX7 output stage!

Ray

[R.G.]

I don't know of any depletion mode devices that will drop right into the circuit.

These devices have an unusually high transconductance, low Ciss, and handy packaging.

The biasing is easy, but just not depletion mode. You have to tie the gates up at some voltage that then sets the right current through the source resistor and drain resistors, and it needs to be large compared to the variation in Vth.

I could probably have gotten away with maybe 25 or 30V for a zener bias point.

[Steve Conner]

That's pretty cool R.G.! I like the "Too linear to be a tube" part especially What's the part number of the device you used?

BTW, I noticed you mentined current multiplying. I've been puzzling for years about how to use a solid-state output stage to boost the power of a tube amp, while still letting the tubes react to the speaker impedance. I wrote this years ago

http://www.scopeboy.com/boots.html

and I finally figured out about 2 weeks ago how to make a suitable output stage with unity gain, low distortion, and soft clipping. So now I can have some real MOSFET heresies like a 400W bass amp powered by a single OTL 6V6. I ain't telling how though! :P

[Gregg]

I think he mentioned ZVN0540A. Unfortunately either not available or available only in bulk.

[Steve Conner]

Hmm, you're right, the ZVN0540 seems quite hard to get. The usual suspects like DigiKey, Farnell or Mouser don't seem to carry it. Where did you get yours RG?

Hey, I just thought, if I can't get any can I substitute a 12AX7?

[TDS]

Steve, I've just seen them listed on the Farnell site for £1.54 p/no 9525424

TDS

[Alex/Tubewonder]

Originally Posted by Steve Conner

That's pretty cool R.G.! I like the "Too linear to be a tube" part especially What's the part number of the device you used?

BTW, I noticed you mentined current multiplying. I've been puzzling for years about how to use a solid-state output stage to boost the power of a tube amp, while still letting the tubes react to the speaker impedance. I wrote this years ago

http://www.scopeboy.com/boots.html

and I finally figured out about 2 weeks ago how to make a suitable output stage with unity gain, low distortion, and soft clipping. So now I can have some real MOSFET heresies like a 400W bass amp powered by a single OTL 6V6. I ain't telling how though! :P

Current dumping? OTL to drive it - what about "the sound is in the iron" aspect?

[Steve Conner]

Well, think of it as an experiment to see whether the sound is "in the iron" or not. I could get more ferric goodness by using a choke in the plate circuit of the tube output stage. If I wanted to boost a PP amp though, I can't see any way round using an output transformer.

I'm not telling whether it uses the Quad current dumping principle or not. :P

[R.G.]

Try a ZVN0545 in the TO-92 package. They're quite similar. Digikey has them here for USD$1.50 each.

[Alex/Tubewonder]

Originally Posted by Steve Conner

I'm not telling whether it uses the Quad current dumping principle or not. :P

I'm not telling either, your secret is safe