WHy this PI design?

[eschertron]

I stumbled across an early Fender deluxe 5B3 schem http://www.thetubestore.com/lib/thet...-Schematic.pdf

and am wondering about the 'opposite' grid on the PI. I see that it is fed by tapping off the grid leak for the 'first' 6V6 power tube, and comparing numbers I see the 6SC7 tube in the PI needs a gain of about 36 to balance the PI grid signals.

So is this a clever design that got lost to history, or is it a parts-count-reducer's nightmare? I'd think a pot in place of the 7k divider resistor would allow for tweaking in the PI balance to taste. So, does this PI design work? I don't think I've seen this anywhere else. Education, please

[catalin gramada]

It is a paraphase, pretty popular design and variants. Ampeg used a lot.
RDH4,p.524-526

[eschertron]

It is a paraphase, pretty popular design and variants. Ampeg used a lot.
RDH4,p.524-526

OK, I have heard (read) of that type of PI. I don't think I've seen it drawn that way, if I've seen it at all. I'll take a look at some Ampeg schems.
Are there reasons why it is not more popular? It looks as easy -or easier- to tweak the PI balance on this than the LTP. The available output headroom is greater, correct? (no long tail)
It looks like NFB can be applied to this design with no problems.

[soundmasterg]

Merlin has info about the paraphase (used in many Ampegs, older tweed Fenders, Supros, etc.) and the floating paraphase (Used in Vox AC100, some other amps too) on his website. Good reading so check it out.

The paraphase types have a lot of gain and decent current drive but aren't balanced as well as some other phase inverters since the second stage gets amplified by the first stage, tapped off, and amplified again by the second stage whereas the first stage has only one stage of gain contributing. The LTP in comparison has decent gain, decent current capability, and a distinct sound that most guitarists like since it was used so widely. The cathodyne has good balance until it overdrives, but less gain than the other types.

Greg

[catalin gramada]

There are a lot of variants. The principle is the same. I did it once in one of my project I just remember a terible awfull overdrive when clip. and not so easy to balance aso. I don.t know if this is a reason for its unpopularity but give a lot a character in sound for sure

[eschertron]

There are a lot of variants. The principle is the same. I did it once in one of my project I just remember a terible awfull overdrive when clip. and not so easy to balance aso. I don.t know if this is a reason for its unpopularity but give a lot a character in sound for sure

D'oh! I didn't think about what happens when the first side clips! So better suited to a 'high headroom' design, to be sure. I'll try it if I ever build a Hi-Fi amp
Thanks for the info, one of the reasons this place is so valuable!

[Leo_Gnardo]

One of the hassles with paraphase in older amps is, the resistors drift and then it's no longer balanced as designed. That can make an amp sound awful, gain minimized or exaggerated in an imbalanced way. It pays to check the resistors that are involved & replace with proper values if they've drifted.

[Ken Moon]

The so-called Tweedle Dee Deluxe adds a PI balance pot to the traditional 5E3 circuit:

[pdf64]

My finding is that the biggest nasty with paraphase is when they share a cathode resistor with the preceding stage, as was commonly done, presumably for economy's sake.
In the linear range, the balanced signals at the cathodes cancel each other out, so a bypass cap might apparently be optional / not really necessary.
But if it's not fully/properly bypassed, it facilitates a positive feedback loop that can result in blips of parasitic oscillation when overdriven, ie when the signals at the cathodes are no longer balanced.
Even though it gets kinda 'masked' by subsequent power tube grid clipping, an unpleasant artefact is created.
If feasible, separating the cathodes is best, but if not, ensure there's a fresh, low ESR bypass cap whose value is more than just adequate.

[Dave H]

The so-called Tweedle Dee Deluxe adds a PI balance pot to the traditional 5E3 circuit:

How can that ever be balanced with a 110k plate resistor when the PI balance max resistance to ground is only 99k?