Effects of flux reversal in PF OPTS?

Jim R. · 4644

0 Members and 1 Guest are viewing this topic.

Offline Jim R.

  • Hero Member
  • *****
    • Posts: 2194
  • Blind Bottlehead
on: October 22, 2009, 06:57:11 AM
Don't want to start a holy war or anything like that, just want to learn the facts about what goes on in the PF circuit.  Some people claim that because the flux reverses in the core of a PF output transformer, the sound is degraded from what happens in a typical direct or series feed transformer.  At least that is my understanding as to what the PF critics are saying.  I understand the basic principle of splitting off the function of the plate loading choke where you can have higher inductance and where the capacitance is not critical, and then the OPT can be designed to best suit the sonics, have no DC offset meaning it can be unggapped, etc.

So, what's the tradeoff and why does it seem to make much less difference than the critics claim?  Is it that flux densities are generally a lot lower, that there is no gap, or what?  The only real potential disadvantage I see is the large cap, but there are some really nice caps around these days.

Looking to be enlightened,

Jim

Jim Rebman -- recovering audiophile

Equitech balanced power; uRendu, USB processor -> Musette DAC -> 5670 tube buffer -> Finale Audio F138 FFX -> Cain and Cain Abbys near-field).

s.e.x. 2.1 under construction.  Want list: Stereomour II

All ICs homemade (speaker and power next)


Offline Paul Joppa

  • Global Moderator
  • Hero Member
  • *****
    • Posts: 5751
Reply #1 on: October 22, 2009, 08:45:33 AM
I've posted about this before, here and in other places, but not recently.

The criticism you describe is based on an incorrect understanding of how magnetic flux works in core materials. Really, it's that simple.

Magnetic materials have a range of flux levels over which their small-signal behavior is pretty consistent; there is no magical zero-flux point. The places where the behavior deviates are the extremes where the core approaches saturation. The fluctuation due to small signal variations anywhere in the linear region looks pretty much the same no matter what the mean flux level is.

Incidentally, the push-pull guys have the opposite myth - that the amp sounds much better if the DC currents are perfectly balanced and there is zero net average flux. Both myths are equally persistent and are completely incompatible with each other.

I'll take this opportunity to talk about the parafeed capacitor issue as well, since it is even more persistent. In parallel feed, the signal current flows through the tube, the parafeed capacitor, the output transformer, and back to the tube. It is isolated from the power supply by the plate choke. In series feed the signal current flows through the tube, the output transformer, and the power supply, before returning to the tube. That means that the power supply (mostly the final capacitor) is in the signal current loop. The power supply capacitor is usually a large electrolytic, and is unlikely to sound any better than the parafeed cap.

Paul Joppa


Offline Jim R.

  • Hero Member
  • *****
    • Posts: 2194
  • Blind Bottlehead
Reply #2 on: October 22, 2009, 08:59:05 AM
Paul,

Once again thank you for the great explanation.  I looked around and did not find any of your previous posts, so now at least it is here and available for others to see as well.

Indeed, I'd rather have the signal going through almost any film or PIO cap instead of an electrolytic, but a good one is the next best thing to nothing at all.

Thank you again for busting this persistent myth that I encounter fairly often when the subject of PF vs. DF comes up.

-- Jim

Jim Rebman -- recovering audiophile

Equitech balanced power; uRendu, USB processor -> Musette DAC -> 5670 tube buffer -> Finale Audio F138 FFX -> Cain and Cain Abbys near-field).

s.e.x. 2.1 under construction.  Want list: Stereomour II

All ICs homemade (speaker and power next)