F4 power amplifier

[juma]

Hi there,
I'm glad that with F4 Mr. Pass gave us a circuits that betters the Aleph (I believe Steen when he says so - there is probably no circuit in this forum that he did'nt build and/or critically listened to).

The thing that keeps nagging at me is that we have been repeatedly told over the years that push-pull topology is something unnatural and that it should be avoided. Zen 5 was an exception with no significant echo in community, but it was a circuit with a voltage gain (MOSFETS in common source connection).

Suddenly there is a plain complementary buffer that surpasses all the single-ended designs we had in the past (Alephs, Zens, Firstwatts).
I enjoy my Aleph amp very much (although I build it out of sheer curiosity) and I will build the F4 for the same reason, but I'm very confused with this sudden change of stance towards push-pull.

Please Mr. Pass, can you shed some light on this ?

P.S. I don't believe that you didn't try the circuit in figure 1. from Zen 5 article in that time.

[MiiB]

This is exactly the circuit mr. Pass has put to use in F4. The bias is provided with the resistor network arround the TL431 and the DC Shift (ofset) of the signal is fed through the coupleing caps. This DC shift is normally done by taking the driving siglnal outside a resistor pair. thus reducing the possible voltagesving of the amplifer. Personly i think this is one of the unique elements in this design. We only need a Jfet pair that can handle a larger voltage sving. So let's invent a 100 V jfet pair, so we can make a +50 V version.
Attached is a revised input circuit with 24 dB gain and changed jfets so voltage can be raised a little.

[Nelson Pass]

Originally posted by juma
The thing that keeps nagging at me is that we have been repeatedly told over the years that push-pull topology is something unnatural and that it should be avoided. Zen 5 was an exception with no significant echo in community, but it was a circuit with a voltage gain (MOSFETS in common source connection).

Suddenly there is a plain complementary buffer that surpasses all the single-ended designs we had in the past (Alephs, Zens, Firstwatts).
I enjoy my Aleph amp very much (although I build it out of sheer curiosity) and I will build the F4 for the same reason, but I'm very confused with this sudden change of stance towards push-pull.

Please Mr. Pass, can you shed some light on this ?

Well, I don't think I told you those things. I like single-ended
circuits for their simplicity. Others have explained the appeal
of SE circuits by speculating that 2nd harmonic sounds less
offensive than 3rd.

The Zen and F series have been exploratory from my viewpoint,
and one of the benefits of this has been the conclusion that
people don't necessarily prefer 2nd over 3rd harmonic, in fact
many times people prefer 3rd harmonic.

I offer as examples the common preferences of the BOSOZ
over the BOZ, and the sound of the F1 over the F2. None have
negative feedback, but the former circuits are "balanced single-
ended" and the latter purely single ended. The former have
3rd harmonic, and the latter 2nd harmonic, otherwise they
are very similar in parts and execution.

What to make of this? First we should note that "balanced SE"
circuits have intrinsically less distortion through cancellation of
2nd harmonic, usually about 1/5 the distortion. This might
be adequate to explain the F1/F2 preferences, but less likely
to explain the BOSOZ/BOZ as the BOZ still has about 1/10
the distortion of an F1.

So I continue to go about my business, designing up circuits
and listening to them. If I ever discover the real secret to
this, I'll never tell.

Originally posted by AD
I also don't understand why on earth did you move this post away to the Trade area?

Perhaps it was intended to stimulate availability of the boards.

[Netlist]

An F4 board run is likely inevitable. We can only be glad Jason started the fire and giving him a separate thread will surely encourage others to make it happen. Any discussion of board production or sale can now take place there. I’ll gladly move it to group buys if needed.

/Hugo

[gl]

Ah yes, Captain Flints! Noyo harbour, Fort Bragg - best fish and chips for miles around. Nice view too. A long drive from TSR but well worth it.

Cheers,
Graeme

[Netlist]

Carpenter, good idea.
I updated the post:
http://www.diyaudio.com/forums/showt...60#post1174260

/Hugo

[Malotron]

Ah yes, Captain Flints!

Went there for my thirteenth BD... wrestled a burger about the size of my head!

We only need a Jfet pair that can handle a larger voltage sving.

Have you tried the 2sj105/2sk330 pair? 50vdgs

-Mal

[GRollins]

The problem with using higher voltage JFETs is that you run into heat dissipation problems if you don't back off on the current. The problem with backing off on current is that distortion climbs because you have trouble driving the cumulative Gate capacitance of the outputs.
What we need are good TO-220 case JFETs that can take more than 3-400mW of heat.
And while I'm at it, I'd like to request some nice TO-247 JFETs, too.
Silly of me, I know, but I've always been a dreamer.

Grey

P.S.: I gave Hint #1 for modifying the F4 elsewhere. I now drop...
Hint #2: Parallel lower Idss devices for the input complementary follower. Don't have BL? Use two or three GR, or whatever.

[Malotron]

Thanks GR. UUPS man just drove home the point too with a big bag of jfets: the 2sk330 vs 2sj170
-Mal

[GRollins]

Yeah, and people keep saying Japanese parts are easy to get here...
(Same thing I've run into--you have no say in whether you get GR, BL, or VI...[of course, the VI are analogous to the Nonesuch])
You may find that the parallel JFETs have their own capacitance increase. It's workable. See Hint #1.

Grey