Push/Pull vs two tubes parallel (Single Ended)

[martin manning]

Thanks for the reference John; I'll have a look when I get home. FWIW, I think the .pdf I posted above addresses the question very succinctly. It's not the voltage swing at all (SE and PP have the same swing for a given B+, as pdf64 pointed out), it's the current.

[schaublin65]

Hi Martin,

it's always the three basic laws of electricity isn't it?

V=IR

W=IV

Don't put your hand in there without draining the filter caps:shock:

Take care....

John

[matt h]

(deleted)

[martin manning]

Current swing. By varying the load you can get any current swing and power absorbed that you want for a given voltage swing. However the power delivered to the load has to be within the dissipation limits of the device delivering it, and that sets the minimum load and maximum power. Operating in Class A at 100% dissipation with zero signal power is a very inefficient way to use the power handling capability of a tube, so the power that can be delivered to the load is much reduced. Current and voltage are interrelated, and my analysis accounts for both; the current is implicit.

[schaublin65]

Hi,

as I far as I understand it the reason we typically see more than twice the power from push pull operation is down to the operating conditions push pull allows us use. Each valve is swinging its voltage for only half the time so we can run it harder.

If we were to run the two tubes at the same operating point as a single ended class A tube we would only get twice the power.

The design centre values typically take this into account and lead us towards a set of operating conditions that take advantage of push pull efficiency in this regard.

Just thinking....

Take care..

John

[matt h]

(deleted)

[jazbo8]

We are really making it harder than it should be, if the OP just spend a minute or two reading Merlin's pages or RDH and plot out the load lines for SE and PP, he can have all the facts he is looking for, using simple algebra. The reason why class AB PP has higher Po than PSE, has already been covered by Matt, Martin, tubeswell and myself here and elsewhere, perhaps not in exactly the same way, but we all said pretty much the same thing...

[54Goldtop]

John thank you - Reading your recommended section now - I thank all of you that have and continue to discuss this thread

[pdf64]

Interesting that he concludes from loadline analysis that most class AB amps are more appropriately viewed as being class A.
As opposed (kind of) to Aiken's view that many amps regarded as class A are in reality class AB!


..Wink I am not sure how you reach that interpretation..

See p19 of the loadlines document, regarding pentodes in PP, he notes that cut off is pretty sloppy and for most AB amps, some plate current is always flowing.

[jazbo8]

and for most AB amps, some plate current is always flowing.

I see how you got that idea, but it still has to be viewed in context and perhaps better worded - instead of saying "the tubes are really operated in Class A for all but the highest power outputs...", it would be more technically correct to say "the tubes are really operated in Class B for all but the lowest power outputs..." which is really what happens in most PP guitar amps.

[martin manning]

There is a lot of subtlety in this question, and I don't agree that a few minutes study can provide anything like a complete understanding of the nuances. The RDH4 article cited above does not address the question, and Merlin's pages really only hint at the answer. IMO posters here have struggled to find a rational way to compare the two configurations and settle on what should be taken as the operating limits.

There is no doubt that the max plate dissipation of the tubes should be utilized to maximize the power output from each configuration, but within that there are options with respect to selecting the plate voltage and load.

Beginning with a constant plate voltage assumption is a useful way to get some understanding, and as I noted above the output power can be doubled at the same Va and load line Pa max because the load can be reduced. That leaves the question of how to utilize the dissipation capacity made available by the fact that in push-pull each tube is in cutoff half of the time.

Reducing the load (halving it) is one option if a constant plate voltage assumption is to be maintained. If allowed, another possibility is increasing the plate voltage (by the square root of two). Either way, the load line Pa max is increased to 200% of the tube's Pa max, and the output power is doubled again, increasing it to four times the power of two tubes in parallel Class A.

[schaublin65]

Hi

I cited p570 of RDH4 to the OP because it starts to deal with PSE topology there.

Earlier in the chapter the concept of plate circuit efficiency is introduced.

By p589 the method of determining maximum power for class B leads to the conclusion at the top of p590 that six times the power of class A operation is possible. This being a result of greater plate circuit efficiency.

I hope this helps....

Take care

John

[Tony Bones]

and for most AB amps, some plate current is always flowing.

I see how you got that idea, but it still has to be viewed in context and perhaps better worded - instead of saying "the tubes are really operated in Class A for all but the highest power outputs...", it would be more technically correct to say "the tubes are really operated in Class B for all but the lowest power outputs..." which is really what happens in most PP guitar amps.

Classes B and AB are concepts based around the idea of each tube shuts off for up to 50% of a cycle. The problem is that tubes don't shut off that way. They shut off more slowly as the grid is driven more negative. It's very nonlinear. True Class B is, for all practical purposes, impossible to achieve. (That's where both tubes are biased just barely at cutoff at idle but the slightest signal causes current to flow.)

It's all very soft with no clear transition. Even Class A to Class AB doesn't have a clear transition as the tubes don't really shut off, they just "slow down", first just a little then more convincingly. It might be argued that if the tubes never shut off then it's all Class A. But at what point do we agree that they really have shut off if they only asymptotically approach zero current?

This all becomes more relevant to power output when we realize that it affects the load that each tube sees. PP transformers are spec'ed 'plate-to-plate'. An output transformer might be wound so that 8 ohms across the secondary appears as 5k ohms across the entire primary from one end (plate) to the other. We're told that with pure Class A each tube sees 1/2 of that or 2.5k while with pure Class B each tube sees 1/4 (1.25k). What do they see in Class AB? (Notice that 1/4 the entire plate-to-plate load requires that one tube is cut off completely during each 1/2 cycle so that that half of the transformer primary is an essentially an open circuit.)

It's complicated! I think that Steve Bench has a good grasp though it's been quite a few years since I've read what he has to say about it.

[matt h]

(deleted)

[54Goldtop]

I started this thread with the hopes of coming up with a consensus, a summary statement that can be shared with confidence and hopefully understandable (to some degree) to the lay person - What do you think guys? Doable?