I was wondering about the use of a bypass cap in push-pull...
That RCA-28x that I'm working has 25L6 output tubes fed from 6sq7's ...and it doesn't use a cathode bypass cap.
So I thought I'd try experimentally putting one in to see what it does.
Now I have oticed that normally in a regular AA5 with a 50L6 output tube or something like a 35L6 in a 6-tuber fed by a 12sq7 .. as soon as I put a bypass cap across the cathode resistor the volume ALWAYS jumps up. But on this RCA28x push-pull there was no noticeable audio increase or difference at all putting a 25uf cap across that 68 ohm cathode resistor.

I see on other amplifier schematics that cathode bypass caps are sometimes used in push pull amplifiers.
.. so..
Is there something different about push-pull cathode biasing that I'm over looking that caused me NOT to hear any audible difference when adding that cap?
Is there... something in the biasing or circuit design that precludes that noticable volume boost with a cathode bypass cap?

When one is pushing, the other is pulling, and if both tubes are synchronized properly, the cathode resistor shouldn't see much, if any, voltage variation. A condenser will do little here. I noticed this a long time ago in my Crosley 1117, and more recently in my Majestic 180.

A condenser is always a bonus, but isn't really necessary. If each tube has its own resistor, on the other hand, a condenser on each resistor may help.

T.

Or, rather, since tubes only allow current in one direction, push-pull isn't really the correct term. When one is pushing, the other isn't. As one increases in conduction, the other one dies down, since it's receiving the signal out of phase. Neither one pulls, though.

Hi Peter

In theory Thomas is correct. A cathode bypass capacitor isn't needed. In actual practice you will find that a capacitor bypassing the cathode resistor still increases volume and helps low frequency response. Some of this is due to tubes not being matched.

Norm

:Or, rather, since tubes only allow current in one direction, push-pull isn't really the correct term. When one is pushing, the other isn't. As one increases in conduction, the other one dies down, since it's receiving the signal out of phase. Neither one pulls, though.

Vintage tube-type, push-pull amps, e.g. my stereo Dynaco ST-70, have a small-value cathode resistor which is unbypassed. The purpose of the resistor is not to provide grid bias (which is fixed) - but rather to provide a convenient way to measure cathode current. The cathode current needs to be measured when setting the fixed grid bias.

In the ST-70, each pair of p-p output tubes share a precision 15.6-ohm cathode resistor, so when the grid bias is correct, there will be 1.56V DC measured across the cathode resistor.

OK, so what's my point? I tend to think that bypassing the cathode resistor in this case might goof-up setting the fixed grid bias? Comments?
Doug

:
:Hi Peter
:
: In theory Thomas is correct. A cathode bypass capacitor isn't needed. In actual practice you will find that a capacitor bypassing the cathode resistor still increases volume and helps low frequency response. Some of this is due to tubes not being matched.
:

A condenser across the resistor you are talking about won't ruin bias settings. If you take a meter reading there, and there are any voltage fluctuations, they will be more sluggish with a large value condenser.

In Peter's and my case, though, the resistor we are talking about is definitely used for bias. My two push-pull radios must have some fairly well matched tubes, because I cannot discern a difference with or without a condenser. However, with one tube pulled, I can, but that's because it's going back to single ended output.

T.

In provididng negative bias to the tube grid, you are also providing negative feedback. So, when you put a capacitor, you (depending on the cap) you bypass some of the feedback and increase the gain, and alter the frequency response. That's as far as I have dug into it, someone might see how it's different for class A, AB, and B amplifiers. My first radio station used two class B modulators, biased a little on the AB side, and with the cathode negative feedback we could get down to .75% distortion at 100% modulation with new tubes. As the tubes aged, they had to be balanced and the bias carefully set, or the distortion would go through the roof.We had a TRF monitor back at the studio, and with a night's sending tones through the transmitter and tweaking the tuning, you could get a durned good sound through the monitor speakers. Pretty good for a AM station.
Lewis

:Vintage tube-type, push-pull amps, e.g. my stereo Dynaco ST-70, have a small-value cathode resistor which is unbypassed. The purpose of the resistor is not to provide grid bias (which is fixed) - but rather to provide a convenient way to measure cathode current. The cathode current needs to be measured when setting the fixed grid bias.
:
:In the ST-70, each pair of p-p output tubes share a precision 15.6-ohm cathode resistor, so when the grid bias is correct, there will be 1.56V DC measured across the cathode resistor.
:
:OK, so what's my point? I tend to think that bypassing the cathode resistor in this case might goof-up setting the fixed grid bias? Comments?
:Doug
:
::
::Hi Peter
::
:: In theory Thomas is correct. A cathode bypass capacitor isn't needed. In actual practice you will find that a capacitor bypassing the cathode resistor still increases volume and helps low frequency response. Some of this is due to tubes not being matched.
::
:

Negative feedback only occurs in the bias circuit naturally if the bias is established by the use of a cathode bias resistor. The resistor makes the cathode more positive than B-, which effectively makes the grids more negative than the cathode if they are connected to B-. No other bias method typically employed will allow negative feedback to occur. The negative 'feedback,' occurs because of loads imposed upon the cathode bias resistor. When the grid shifts less negatively because of a certain part of a signal imposed upon it, combined with its negative bias, the tube conducts more. This puts an extra load on the cathode bias resistor, because the cathode is drawing more current. What this also does is make the cathode a bit more positive, since the resistor only supplies so much negative energy, and the cathode wants more. Having the cathode shift more positive is the same as shifting the grids more negative, or at least has the same result. The differene between the two is the same, regardless of external voltages and conditions. Since the grid will be shifting a bit more negative, its less negative condition will be slightly cancelled out. This is why a single ended tube has less gain without a cathode bias resistor bypass condenser. What the condenser does is it retains a charge, and reduces variations in the voltage across the cathode resistor.

In a push-pull circuit, one tube is conducting while the other is not. As a signal wave passes through center, both tubes conduct half of what they would or would not if the wave was at its peak or trough. The two tubes balance eachother, since they are doing the exact opposite of eachother. If one of the tubes conducts slightly better than the other, this balance may be thrown off slightly. Under ideal circumstances, voltage fluctuation will not occur at the cathode bias resistor, if the resistor serves both tubes.

Thomas