why the prevalence of push pull?

[Spookyman]

In preamp circuits, the anode output of one stage is typically fed to the grid of the next stage. OK.
Why can I not find a power amp circuit that does this? Every one I look at uses push pull. I know this is more efficient but surely somebody made an amp where the power amp had anode output of the first power valve fed to the grid of the next power valve, and no need for a phase inverter?
Or has large scale guitar amp manufacturing only ever gone down the push pull route? just curious.

[teemuk]

If I figured this out right it would still be a push-pull circuit but one of the power tubes would inherently handle the required phase inversion. Pretty cool idea, which at least according to quick SPICE simulation seems to work decently, albeit voltage gain of this configuration is not very admirable.

[jrfrond]

Well, single-ended Class A isn't push-pull, and there are PLENTY of these low-wattage jobbies around, especially nowadays.

Push-pull is the most efficient way of achieving higher power efficiently. A 50W, SE Class A guitar amp (it could ONLY be Class A), would require lots of iron in both the PT and OT and run hotter than hell. There ARE high-end hi-fi amps that use designs like this, but they are pricey and considered to be esoteric designs.

Push-pull, whether it is Class A, AB1 or AB2, still requires phase-splitting and inversion, no matter what the design, because previous stages are Class A single-ended.

[teemuk]

Something like this?

Also, those who were quick enough to read it, forget what I typed (and what was removed) about class-A. It took me seconds to realize how little sense it made.

[defaced]

Cascading (plate to grid) is used to alter the tone of a waveform. Since the preamp is used for shaping the tone, this makes since. Between each stage there are voltage dividers and caps that reduce the signal so the next stage isn't overloaded (voltage divider) and shape the tone (caps) - this causes distortion. A power amp is used to just make a signal very large with little distortion, so cascading doesn't make since for this purpose.

[Rob Mercure]

For non audio applications there have been lots of direct coupled amps that went straight plate to grid involving running each stage at higher power supply potentials from cathode to cathode. And for RF work there are Class C amps that only push part of the cycle. But other than push pull there is only paralleling lots of output tubes. There are other multi output device strategies other than push pull that "stack" devices but this is much more common with solid state and tricky with tubes as aging effects this more. And it is possible to have an AB single ended stage if you can accomodate the partially clipped waveform - if you run a Champ hard enough it can hit AB.

Rob

[loudthud]

I met a guy at the Dallas Guitar show some years back that patented such a design and did the patent search himself. He didn't give a schematic but just a general description in the patent. He used to advertise in Vintage Guitar Magazine but stopped a number of years ago. The company name was Kjl amplifiers out of Gretna Louisiana. The design got 30W out of a pair of 6L6's.

There was also an article in one of those 60's electronic mags (Electronic World?) about a stereo Hi Fi amp with one 12AX7 and four EL84's. One half of the 12AX7 was a gain stage and the EL84's were cathode coupled.

[R.G.]

In preamp circuits, the anode output of one stage is typically fed to the grid of the next stage. OK.
Why can I not find a power amp circuit that does this? Every one I look at uses push pull. I know this is more efficient but surely somebody made an amp where the power amp had anode output of the first power valve fed to the grid of the next power valve, and no need for a phase inverter?
Or has large scale guitar amp manufacturing only ever gone down the push pull route? just curious.

Class AB.

Anode to grid phase inversion works. It often needs some tweaking, but can be made to work as long as both tubes are always conducting current.

In an output stage, you get most of your efficiency in terms of power loss per power output by turning off one tube, then the other. They overlap in conduction only for a small part of every cycle, in the middle.

So they inherently can't drive one another for phase inversion if they alternately turn off.

[R.G.]

I met a guy at the Dallas Guitar show some years back that patented such a design and did the patent search himself. He didn't give a schematic but just a general description in the patent. He used to advertise in Vintage Guitar Magazine but stopped a number of years ago. The company name was Kjl amplifiers out of Gretna Louisiana. The design got 30W out of a pair of 6L6's.

There was also an article in one of those 60's electronic mags (Electronic World?) about a stereo Hi Fi amp with one 12AX7 and four EL84's. One half of the 12AX7 was a gain stage and the EL84's were cathode coupled.

Wasn't he the guy that did single ended by turning off the signal to one of the output tubes? You know, use the whole OT iron by putting a constant current source on one side of the primary to reset the iron and prevent saturation so you can use a smaller tranny for class A single ended?

That works OK. I did it years ago with a power MOSFET for the CC side. Maybe I should have patented that.

[Steve Conner]

The technique Spookyman is talking about would probably count as self-split push-pull. A Google search for this will turn up a few examples, including Mark Lavelle's Decimator homebrew amp:

http://www.harmonicappliances.com/de...Decimator.html

As others have stated, the coupling is usually done by connecting the two cathodes together with an unbypassed cathode resistor to ground, rather than connecting one tube's grid to the plate of the other.

It's not commonly used because it gives less undistorted power than you would get from driving both tube grids. The reason as RG mentions, is that it's limited to pure Class-A: once the driven tube has cut off, it can't drive the other one any more. OTOH, even "Class-A" hot cathode biased amps can transition into Class-B on large signals.

FWIW, if you run a Champ hard enough it can hit Class-C, never mind AB It basically becomes a switched-mode converter pumping out a square wave to the speaker. So maybe that should be Class-D...

[Rob Mercure]

Steve,

Let me rephrase that - it's common for a Champ when driven as hard as the stock circuitry allows to hit AB as it clips asymmetrically.

And as long as we're looking at PP without a "normal' PI there is the circuit that derives the opposite phase from the "first" output tube's screen grid.

See attached schemo

Rob

[jrfrond]

Steve,

Let me rephrase that - it's common for a Champ when driven as hard as the stock circuitry allows to hit AB as it clips asymmetrically.

And as long as we're looking at PP without a "normal' PI there is the circuit that derives the opposite phase from the "first" output tube's screen grid.

See attached schemo

Rob

That is usually reffered to as the "Philco" circuit. Phase inversion is still required, but accomplished within the power tubes. A neat system for a reduced parts count (originally used on Philco table radios, and I believe they MAY have had a patent on it, though I am not sure).

Looking at all of this aesthetically, there are MANY ways to amplify audio. What works for high-fidelity amps probably won't work well for guitar amps, and vice-versa. The criteria are different. BOTH will amplify the guitar signal, but most high-fidelity designs utilize low-gain stages and multiple feedback-loop schemes, all designed to keep distortion down. Unless you are a serious purist, this doesn't work for guitar, as it destroys the feel of the amp. Turntables and CD Players don't care about feel, and audiophiles are not jazzed by early breakup in their tube amplifiers.

What it all comes down to is that the guitar amps that most of the world knows and loves are relatively simple affairs, with tried-and-true designs that work. For the most part, it is unnecessary to reinvent the wheel, but it's always good to try. Hey, until Music Man came along, we (or at least I) had never seen a cathode-drive power amp in a guitar amp, and those worked out well. Guess Leo didn't want to plagarize his original designs.....like everyone else!

[gmoon]

... but surely somebody made an amp where the power amp had anode output of the first power valve fed to the grid of the next power valve, and no need for a phase inverter?...

In simpler terms: So why don't guitar amps cascade the power amp section?

The preamp sections are primarily concerned with boosting the voltage of the very weak pickup signal (along with shaping the signal.) The more stages, the more signal boost, clipping, etc.--whatever the builder's goal.

But the power amp's primary function is as a current amplifier--i.e., creating the power necessary to move that speaker cone around.

Power tubes have a well-defined maximum power dissipation, as well as a decently high input impedance (like most valves.) Even a low-current 12AX7 can drive a power pentode. Feeding a high-current signal into the grid of a power tube won't raise the power output of the second tube beyond the max of either tube.... So why use a second tube to get the exact same power output?

That's where all the other topologies (parallel SE, push pull, etc.) come in...

[Gregg]

Maybe I should have patented that.

I'm afraid you're a bit late. Randall Smith did it already. As usual...
Check out Mesa's new Express series amps (patent 7,173,488):

http://www.mesaboogie.com/Product_In...ss/Express.htm

[R.G.]

I'm afraid you're a bit late. Randall Smith did it already. As usual...
Check out Mesa's new Express series amps (patent 7,173,488):

No doubt.

Didn't Randall Smith also invent the internet?

No, wait. That was Al Gore...

I posted that either on my web site or a net forum years ago.