OT: Do we always need a phase inverter in class AB amps ?

[fikri]

Almost all guitar amps uses a phase inverter circuit and the others uses some kind of power tubes driver (?). what is the differences between them ?

[aaronkessman]

in my extremely limited knowledge it's the PI that almost *Defines* the AB-ness of the amp. that is, the PI sort of splits the signal in two, inverting the phase of one side to drive the two or more power tubes in class AB. So you dont need a PI in a single power tube setup as in: http://www.ampwares.com/ffg/schem/champ_5e1_schem.gif

my guess is that if you *did* try to use a PI, your phases would cancel and you'd get a very weak sound.

[Brett Clark]

Almost any push-pull tube amp has a phase inverter of some kind, whether the amp is class A, AB, or B (or even class C in the case of RF amps). There are some odd exceptions, such as the Circlotron design.

The reason is that one tube (or tube set) must operate in mirror-image fashion to the other tube (or set of tubes, in an amp with 4 or more output tubes). That is, when one side is swinging positive, the other side must swing towards ground (speaking of the plate voltages).  So, for the plate voltage to move in opposite phase, the grid must be driven in opposite phase. This is the job of the phase inverter.

In a transistor amp, you have the option of just using output devices of the opposite polarity. Tubes only come in one polarity: "N-channel", as the charge carriers are electrons. Solid state devices may have either electrons (N-channel) or "holes" (P-channel) as charge carriers. There is no vacuum-state equivalent of a hole carrier. I guess a "positron tube" could exist, but antimatter is kind of expensive  :lol:

[lovekraft0]

While most commercial push-pull tube amps have an explicit PI, there are several self-split designs that work pretty well. The trade-off is in power/efficiency vs simplicity and lower parts count, so it's not particularly attractive to the manufacturers, but it's a DIY staple - Doug Hammond's Firefly springs to mind immediately.

Class (A, AB, B) is about quiescent bias current in the output tubes, and has nothing to do with the phase inverter.

[Sir H C]

For tube amps you need some way to invert the phase for the two drive sides in a push-pull amplifier.  I guess you could have single ended A-B but you would have a lot of distortion in the system.

Since transistors have complimentary devices, you do not need a phase inverter/splitter.

[lovekraft0]

How's this?

Two triodes set up as a differential pair, push-pull output, no phase inverter in sight (unless you're going to consider the power amp itself as a phase inverter). And it sounds great!! Thanks, Doug!!  

[Transmogrifox]

lovecraft0:

That's what many amps use as an inverter, essentially, but instead of coupling directly to the OP transformer, they take each end off a resistor and each goes to one  polarity of the push/pull stage.

If you did this with power tubes, it would essentially be a class A amp since you would have to bias the power tubes so hot in order to get any respectable current on the negative swinging cycle, and to avoid the distortion that would result from the point where the one tube turns off, and the extra gain is then reduced to that of the tube with a resistor in the cathode.

Not to mention you would need a pretty heavy duty resistor to withstand the required bias currents to get much power out of that configuration.

All the same, I can imagine using that configuration may be interesting in a lower power application since indeed it does induce soft clipping on the positive half cycle and hard clipping on the negative half-cycle (w/reference to the input).

Hopefully that's useful info to somebody.

[lovekraft0]

Yeah, it does resemble a standard LTP, doesn't it? And your caveats about bias currents and low power are absolutely correct, but as I said, the trade-off is efficiency vs simplicity, and it really does sound good. There's also a similar architecture that drives the second power tube from the screen grid of the first, but I don't have it handy just now. My point was that while their inherent disadvantages will probably insure that we won't be seeing any self-split Fenders or Marshalls, there is more than one way to flay that feline, and since this is DIY territory, they're worth looking at (especially since they do sound very nice when pushed - Doug has some samples linked from his Firefly page, as I remember). If I can find that alternate self-split pentode setup, I'll post it for anybody who's interested.
<EDIT>FWIW, here's the article that started most of the DIY interest in self- split power amps: Compact Hi-Fi Power Amplifier, Electronics World, June 1961 - courtesy of Triode Electronics

[lovekraft0]

As promised, the screen fed self-split (as described in RDH4):

It turns out that Magnatone actually used this topology in a production amp.
Magnatone Melodier 109 (courtesy of Stephen Keller)

[fikri]

Thanks for all the replies. I have seen a Peavey Classic 50 schem and Orange graphic MK2, are they push pull type of amps ?

[aaronkessman]

I can say with 100% confidence without even looking at the schematics that they probably are

[Doug_H]

The cathode fed self-split circuit (triode article, firefly, etc) is class A. It's efficiency is very poor. In fact I kind of wonder if it would be louder just going single-ended with a big cathode bypass cap. Never compared that.

Doug

[fikri]

Well, i just wanted to know the diferences, the advantages and so on, about using the PI and a power tube driver before we hit the power tubes section.