(Quasi)complementary JFET power amplfier

[R-K Rønningstad]

I accidentally posted this in the F4 group buy, and Jacco helpfully suggested a re-post under the Pass Labs section, so here we go:

More power from JFET Post #323
The Lovoltech JFET device does not have a P-channel counterpart, so a N/P push-pull output stage is out of the question.
A quasi-complementary version should in theory at least, be possible with the N-channel.
To me it seems much of the idea behind the F3 and the ZV9 is to operate the JFET at a special part of its transfer curve, that is, where it resembles a triode.
My question is then, a quasi-complemantary amplifier with the JFET might "spoil" this concept, and there might be any reason for it left. Maybe even not a N/P-complementary amplifier with an imaginary P-channel JFET would give the F3 sound?
Any comments from preferably the Zen-master himself? ;-)

Regds
Rolv-Karsten

[Zen Mod]

Originally posted by R-K Rønningstad
I accidentally posted this in the F4 group buy, and Jacco helpfully suggested a re-post under the Pass Labs section, so here we go:

More power from JFET Post #323
The Lovoltech JFET device does not have a P-channel counterpart, so a N/P push-pull output stage is out of the question.
A quasi-complementary version should in theory at least, be possible with the N-channel.
To me it seems much of the idea behind the F3 and the ZV9 is to operate the JFET at a special part of its transfer curve, that is, where it resembles a triode.
My question is then, a quasi-complemantary amplifier with the JFET might "spoil" this concept, and there might be any reason for it left. Maybe even not a N/P-complementary amplifier with an imaginary P-channel JFET would give the F3 sound?
Any comments from preferably the Zen-master himself? ;-)

Regds
Rolv-Karsten

I'm the Zen .....
.......master .... pretty questionable , but also definitely not the one you asked

just take jfet +cascoded mosfet as combo ;

put them up and down (one up,one down) ,and you have quasi complementary , with all issues solved .

[Nelson Pass]

Originally posted by R-K Rønningstad
The Lovoltech JFET device does not have a P-channel counterpart, so a N/P push-pull output stage is out of the question.
A quasi-complementary version should in theory at least, be possible with the N-channel.
To me it seems much of the idea behind the F3 and the ZV9 is to operate the JFET at a special part of its transfer curve, that is, where it resembles a triode.
My question is then, a quasi-complemantary amplifier with the JFET might "spoil" this concept, and there might be any reason for it left. Maybe even not a N/P-complementary amplifier with an imaginary P-channel JFET would give the F3 sound?

I am currently tinkering with such a thing, and it's certainly true
that it would be nice if there are a P channel power device. It's
possible, but the last estimate for development was way over
any budget I would consider.

So you can go quasi-comp, or you could do a JLH topology, or
a circlotron, or a transformer coupled split drive. The first is
probably the most practical candidate if you ever want to
consider Class AB operation.

[GRollins]

I have done some preliminary work on a quasi amp. Time constraints have pretty much got me stonkered at the moment. Nelson will--no surprise--probably get five JFET amps out in the next week. I'm having trouble getting to the bench to work on the two I currently have in mind. I ain't picky...either would be nice.

Grey

[R-K Rønningstad]

Thank you for your replies.
But do you think some/all of the F3 magic will be lost in a quasi-complementary JFET amp? I mean, going to Class AB means operating the Lovoltech not on the "triode-like" part of the transfer curve?

Best regards
RK

[EUVL]

> So you can go quasi-comp, or you could do a JLH topology, or a circlotron, or a transformer coupled split drive.

Something like this ?

LU1014 with MOSFET Cascode, a la Zen V9.
The gate resistor for the JFET might not be necessarily.


Patrick


PS I don't have my CAD with me, so this is a bit quick & dirty as far as picture quality goes.

[Joe Berry]

Originally posted by EUVL
> LU1014 with MOSFET Cascode, a la Zen V9.

If you remove the output series resistors, feed the CCS circuits from the opposite rails, and replace the zeners with adjustable shunt regulators so you can tweak the gain and balance, I bet it will work fine.

[Joe Berry]

Sorry, bad word choice above. Please replace "gain" with "operating point."

[flg]

Well, there is a patent for using a follower drive stage powered by the opposite rails and I use the opposite rail for my current sources so we know that works... I haven't had much time to digest the rest of it... The output series resistors??? Hmmm, I'm not sure about that. You must however supply a return path for the input signal to gnd, without unfloating the supplies and I think those are the resistors you are speaking of... If however you use F.B. that can do the job without the resistors. Just a few thoughts...

[Joe Berry]

Originally posted by flg
output series resistors???

I was referring to the two horizontal resistors on the far right in Patrick's diagram.