F4 power amplifier

[luvdunhill]

Originally posted by GRollins



Ooops! My mistake. That's not a dingus, that's a whatsit.

Grey

wait, hold the BOMs! I just noticed that there is a 0.4% and a 0.5% version. Anyone know which sounds better? The most expensive TO92 one I can find is the Microsemi-IPG TL431BCLP that is a 0.4%, anyone know if this might add a bit of prat compared to the cheaper Zetex ZTL431BCSTZ?

:bs:

[luvdunhill]

Originally posted by Babowana




What is dingus while I could guess what the whatshit is . . . ?

I think something got lost in the translation there

[Nicola]

I would be interested to learn about the difference in sound between the F4 and the F1 if any of you have had the opportunity to compare them

Kind regards

[cviller]

I would be interested to learn about the difference in sound between the F4 and the F1 if any of you have had the opportunity to compare them

If you had read the thread you would know that "some guy" from Lowther-America has compared them already. It is pointless to keep asking questions already answered in this thread - it is too long already!

[GRollins]

Originally posted by cviller


If you had read the thread you would know that...
It is pointless to keep asking questions already answered in this thread - it is too long already!

This is the point at which the Aleph-X thread transitioned from technical discussion to space-filling. Ironic, but predictable. People are too lazy to read, so they ask questions that have already been answered, thereby making the thread even longer, which made it even more intimidating for those who didn't want to read all that went before, so they asked the same question, thereby making the thread even longer...
Don't get me wrong, I'm pleased the circuit was received as well as it was, but human nature will eventually reveal itself.

Grey

[Kashmire]

Thanks, GRollins

Two questions about the F4 ...

(1) what is the C3/R25 and C4/R24 network? At first glance, it looks to be power supply filtering downstream of R5 and R22. However, what is the function of R24 and R25? Altering the values don't seem to change frequency response much, nor the distortion spectra.

(2) is a servo loop (crude attempt attached) desirable? Or is a little bit of offset actually good? How much does the DC offset drift during warmup of a cold amplifier?

I spent some time simulating the F4 and it is pure genius. According to simulation (which isn't perfect, but usually tells the general truth), the distortion spectra is carefully contained in very low harmonics, almost fully concentrated on the even-order harmonics. The distortion-to-power relationship is pretty dramatic (i.e. distortion increases with increasing output voltage). All considered, quite an amazing design.

As soon as I clear some other projects, I'm certainly going to solder a F4 together, without a doubt. (my wife already thinks there are too many amps in the basement, why not add a couple more?)

Image attached: F4 with (crude) DC servo loop

[GRollins]

I don't have a schematic handy to answer the question about the caps. I'll try to look at it later.
Why do you feel that a DC servo is necessary? It's a power follower, pure and simple. The main difference between the F4 and other follower designs you may have seen is that the F4 incorporates a driver stage. It's a pretty stable circuit. Adding a servo strikes me as a rather arbitrary addition to the circuit.

Grey

[Nelson Pass]

Originally posted by Kashmire
(1) what is the C3/R25 and C4/R24 network? At first glance, it looks to be power supply filtering downstream of R5 and R22. However, what is the function of R24 and R25? Altering the values don't seem to change frequency response much, nor the distortion spectra.

(2) is a servo loop (crude attempt attached) desirable? Or is a little bit of offset actually good? How much does the DC offset drift during warmup of a cold amplifier?

C3/R25 and C4/R24 are bootstrap networks for the front end.
They slightly improved the maximum output and bandwidth, but
their primary purpose is to protect the JFETs from over-voltage.

[Kashmire]

DC servo to null the output DC offset due to potentiometer drift, natural changes during the warmup phase, etc. Just a consideration. I'm not familiar enough with solid-state push-pull (all of my prior push-pull projects are tubed) to be comfortable with direct coupled and a little DC offset. Tube gear isn't direct coupled and doesn't have DC offset; it's an unfamiliar byproduct of solid state that a DC servo can correct for. I see many highly-regard push-pull designs (Borbely, D’Agostino, Levinson) use servo loops, although Pass certainly falls into the upper echelon of highly-regarded designers.

[nobody special]

How would you implement a servo when there is no voltage gain in the circuit?