Reason for capacitive coupling in vintage push-pull stage

[Elvee]

Hi,

I was wondering why, in pre-complementary days, the all-PNP push-pull stages always incorporated a capacitive coupling between the phase splitter transistor and the lower output transistor.

I see no definite advantage to this arrangement, on the contrary:
-It requires an additional electrolytic
-It is inefficient, as the base resistor of the output devices consumes drive, but does not contribute to the bias current of the driver.
-It could cause nasty dynamic effects, because under large signal condions, the capacitor will charge through the BE diode of the output transistor.

The pic shows a typical such stage; C2 is the capacitor in question.

So, why not get rid of C2, R1 and R6, and adjust the value of R4 to the new conditions?

[Fogonazo]

If you look at the diagram you will see that is powered by a single rail, on the union of the output transistors there is a voltage of Vcc / 2.
It is necessary to isolate the tension of the speaker, but otherwise he would be sending this DC and burn.
The effect of placing a capacitor in series with the speaker is compensated by the feedback network

[Elvee]

Originally Posted by Fogonazo

If you look at the diagram you will see that is powered by a single rail, on the union of the output transistors there is a voltage of Vcc / 2.
It is necessary to isolate the tension of the speaker, but otherwise he would be sending this DC and burn.
The effect of placing a capacitor in series with the speaker is compensated by the feedback network

I am talking of C2, not C1.

[Elvee]

Up!!!!!!

[jechentau]

maybe to let the lower output transistor work as ccs, thus having a more stable idle current? or equivalent to make it able to work with a "window" of DC on the input,
with which it still functions. (this suggests the input is DC coupled to the previous stage, for some reason, perhaps global thermal stability or DC or so)

[analog_sa]

Originally Posted by Elvee

So, why not get rid of C2, R1 and R6, and adjust the value of R4 to the new conditions?

Germanium transistors with lots of thermal issues?

[Elvee]

Originally Posted by jechentau

maybe to let the lower output transistor work as ccs, thus having a more stable idle current? or equivalent to make it able to work with a "window" of DC on the input,
with which it still functions. (this suggests the input is DC coupled to the previous stage, for some reason, perhaps global thermal stability or DC or so)

It is a class B, not class A.
And anyway, it doesnt change a lot of things: in one case the bias is provided
by R6, in the other it would come from R7 via Q3.

Germanium transistors with lots of thermal issues?

Yes, I didnt represent it, but R1 is made up of a NTC + resistor.

But this composite could be adjusted in value and in "mix" to give an identical Iq and stability.

[Conrad Hoffman]

We're smarter now and it's just a matter of style and complexity. Back then designers tended to treat each active device as a block, getting the DC bias where they wanted it, and isolating the whole thing with caps. Today we have a lot more experience with direct coupled circuits and know the benefits (and pitfalls). I always think of my SWTPC Tiger amp, the first direct coupled circuit I worked on. When something failed, typically everything failed, as it was all tied together. One quickly learned to check every single device and to limit current on start up until all was well. You almost never have that problem with circuits using a lot of coupling caps to isolate the stages- failures are invariably local.

[Elvee]

Originally Posted by Conrad Hoffman

We're smarter now and it's just a matter of style and complexity. Back then designers tended to treat each active device as a block, getting the DC bias where they wanted it, and isolating the whole thing with caps. Today we have a lot more experience with direct coupled circuits and know the benefits (and pitfalls). I always think of my SWTPC Tiger amp, the first direct coupled circuit I worked on. When something failed, typically everything failed, as it was all tied together. One quickly learned to check every single device and to limit current on start up until all was well. You almost never have that problem with circuits using a lot of coupling caps to isolate the stages- failures are invariably local.

Yeah, it could be the reason. It's something I hadnt considered, as our forefathers are supposed to be so much smarter, but it is a possibility.

[jechentau]

well, it definately stabilizes the idle current more (heat q3, what happens?)