Have others noticed this? Why doesn't the volume go down more? Maybe pulling one tube causes the B+ on the other tube to go up? (I haven't checked this theory, yet.)

(The radio I'm now fooling with is an E.H. Scott All-Wave 15 with a pair of 2A3s in push-pull. It has AVC, but that doesn't come off the final audio.)
Doug

T.

:Make sure that both tubes are putting out the same power (both have the same emissions). You should notice a change in tone quality, but may not notice a change in volume. I usually notice a slight change. It is most noticeable at higher volumes (distortion).
:
:T.

What yu have is a classic class "A" amplifier. Both tubes conduct for the full 360 degrees of the cycle. The advantage to this is that (over tubes in parallel) is that certain harmonic distortion is cancelled in the output transformer. If you were to measure the plate current of the tube, it would remain constant. All you do when you remove one of the output tubes of a push-pull class "A" amplifier is to make a single-ended class "A" amplifier.
If the tubes were run in pure class "B", one would conduct for one half cycle of the audio, and the other for the other half cycle. In between there is class "AB" where you let both tubes carry some current close to the crossover point. This only gets important when you're modulating the plate of a power amplifier tube, and yu need half of the transmitter output power to reach 100% modulation. For a 10 KW transmitter, you need 5 KW for 100% Mod. For that, you spend some time biasing your modulator tubes to save a buck on the power bill.
Lewis

rcw

:Lewis, thanks. I need to think about this - which at my advanced age may take a little while.
:Doug
::
::What yu have is a classic class "A" amplifier. Both tubes conduct for the full 360 degrees of the cycle. The advantage to this is that (over tubes in parallel) is that certain harmonic distortion is cancelled in the output transformer. If you were to measure the plate current of the tube, it would remain constant. All you do when you remove one of the output tubes of a push-pull class "A" amplifier is to make a single-ended class "A" amplifier.
::If the tubes were run in pure class "B", one would conduct for one half cycle of the audio, and the other for the other half cycle. In between there is class "AB" where you let both tubes carry some current close to the crossover point. This only gets important when you're modulating the plate of a power amplifier tube, and yu need half of the transmitter output power to reach 100% modulation. For a 10 KW transmitter, you need 5 KW for 100% Mod. For that, you spend some time biasing your modulator tubes to save a buck on the power bill.
::Lewis

::Lewis, thanks. I need to think about this - which at my advanced age may take a little while.
::Doug

Doug:
Izza matter of fact, on my first radio station, I put in a RCA BTA 1M transmitter, using 833 tubes as both the modulator and the final. These tube had an oval plate, with the long side of the oval towards the front of the transmitter. The normal operating state of these tubes was to be slightly red hot. When I made my daily transmitter inspection, I looked at the plates of those 833s, on the final tubes, a red spot about the size of a quarter was fine, indeed. If the spot was about fifty cent size, maintenance was necessary. The modulators, however, were class AB something, I forget, but the red hot spots would get larger and smaller with modulation. I could learn more about my transmitter looking at those 833s than I could from the meters. We could get pretty good hi-fi from that arrangement, but you had to check the bias on the tubes on a real regular basis, as with one tube carrying one half-cycle and the other tube carrying the other half cycle, a tube getting weak could push your distortion out of FCC limits in a New York second, whereas your class "A" just loses a little power and sounds the same with a weak tube.
Lewis

marv

:I've noticed this with several P-P radios and amps. Pull either one of the pair of output tubes, and the volume goes down a little but not greatly.
:
:Have others noticed this? Why doesn't the volume go down more? Maybe pulling one tube causes the B+ on the other tube to go up? (I haven't checked this theory, yet.)
:
:(The radio I'm now fooling with is an E.H. Scott All-Wave 15 with a pair of 2A3s in push-pull. It has AVC, but that doesn't come off the final audio.)
:Doug

I have recently moved, and haven't gotten my 'scope and other test equipment going - in fact, my bench isn't even set up yet.

But, Marv, your theory is very interesting.
Doug

:Doug,
:Have you o'scoped any of the output wave forms? Is it possible that a mirror image is appearing across the other half of the primary because of flux in the pri/sec from the operating tube? What if you loaded the plate connection to the transformer (50K or so) of the missing tube.
:
:marv
:
::I've noticed this with several P-P radios and amps. Pull either one of the pair of output tubes, and the volume goes down a little but not greatly.
::
::Have others noticed this? Why doesn't the volume go down more? Maybe pulling one tube causes the B+ on the other tube to go up? (I haven't checked this theory, yet.)
::
::(The radio I'm now fooling with is an E.H. Scott All-Wave 15 with a pair of 2A3s in push-pull. It has AVC, but that doesn't come off the final audio.)
::Doug

I would measure the voltage with both P-P output tubes installed, and then with one tube pulled. I can sure as heck detect 3 dB down with my DMM even if my ears can't.

Now I have to find my audio tone generator among all the boxes and debris from my move. (I do have my DMM in captivity.)

(This P-P question originally arose after setting up my E.H. Scott receiver in the "new" house - it is one of my main sets in the place.)
Doug

Doug,
You would be surprised how little difference 3dB makes to your ears. At AT&T, we routinely identified a group of 60 circuits by placing a 75 Ohm dummy load across the input, thus dropping the power 3dB, which showed up very well on a DB meter, but nobody ever complained, as they never noticed it.
Lewis

Here was the set-up. I inserted an 420-Hz audio signal into the phono input. The AC input voltage was kept comstant during the experiment.

There are two 2A3 output tubes in P-P: a Tung-Sol and a JAN (unknown manufacturer, but possibly RCA). With my Fluke DMM, I measured the AC output voltage at the speaker terminals. Here are the results:

Both tubes installed: 95.0 mV output
Tung-Sol pulled: 74.8 mV
JAN pulled: 79.5 mV

It seems that the Tung-Sol is slightly stronger than the JAN. For comparison purposes, let's use the average single-tube output = (74.8 + 79.5)/2 = 77.2 mV, and compute the power drop to the speakers:

20 log (77.2/95.0) = 20 x (-0.0901) = -1.8dB.

If the power had dropped by half with one tube pulled, we would have expected -3dB. So, it would appear that something is happening in the amplifier circuit that partially compensates for the pulling of one P-P tube. Once I get my bench back in operation, I can explore this further - possibly the B+ is going up with one tube pulled or maybe something like Marv's theory.
Doug

::Marv, what about just using my DMM to measure the voltage across the speaker with a tone generator inputing the phono connection on the receiver? (I did this once before when I was testing the audio frequency response of the receiver.)
::
::I would measure the voltage with both P-P output tubes installed, and then with one tube pulled. I can sure as heck detect 3 dB down with my DMM even if my ears can't.
::
::Now I have to find my audio tone generator among all the boxes and debris from my move. (I do have my DMM in captivity.)
::
::(This P-P question originally arose after setting up my E.H. Scott receiver in the "new" house - it is one of my main sets in the place.)
::Doug
:
:Doug,
:You would be surprised how little difference 3dB makes to your ears. At AT&T, we routinely identified a group of 60 circuits by placing a 75 Ohm dummy load across the input, thus dropping the power 3dB, which showed up very well on a DB meter, but nobody ever complained, as they never noticed it.
:Lewis

Doug:
How 'bout the negative feedback? Dontcha think that may keep the amplifier from showing the expected result?
Disconnect the feedback and see what happens then.

Lewis

This 15-tube set has a signal-strength meter that derives its input from the AVC line. When I pull one of the p-p tubes, the signal-strength meter doesn't even twitch. (Since the AVC comes off before the phono connection, this makes sense.)

But, yes, there is some subtle negative feedback creeping into the picture - for example, if pulling one of the p-p tubes were to cause the B+ voltage to go up, that would be, in effect, negative feedback (sort of). But until I can really "get under the hood," all I can do is speculate.
Doug

:
:Doug:
:How 'bout the negative feedback? Dontcha think that may keep the amplifier from showing the expected result?
:Disconnect the feedback and see what happens then.
:
:Lewis
:

Doug:
You are right, (now that I've seen a schematic), there is no feedback as was used in later years, from the voice coil back to the first AF. No, I wasn't confusing the AVC, that is totally different. In class "A" amplifiers, the current flows through the tubes for 360 degrees. This is a lot like having the tubes in parallel, except you can reduce distortion by having them work out of phase, and combining them in the output transformer. But, yu are right,it would seem that you would get 3 dB by removing one output tube, without having a lot of distortion. Your math looks right, even using 20 times log (P1/P2), instead of 10 as one would do for power. Hmmmmmmmm.....back to the old thinking cap, I guess. Talk to you later.

Lewis

But, I think your original point is very valid. 2-3 dB is almost impossible to hear unless you compare them instantaneously. I can understand that cutting a telephone's audio by 3 dB would be pretty negligible.

For sure, a p-p amplifier isn't anywhere near twice "as good" as a single-ended amp using the same output tube. In fact, the difference is rather minor.

But there is something about this issue that deserves more study. I need to find out for sure why pulling one p-p tube drops only 1.9 dB, not 3 dB.
Doug

:
:Doug:
:You are right, (now that I've seen a schematic), there is no feedback as was used in later years, from the voice coil back to the first AF. No, I wasn't confusing the AVC, that is totally different. In class "A" amplifiers, the current flows through the tubes for 360 degrees. This is a lot like having the tubes in parallel, except you can reduce distortion by having them work out of phase, and combining them in the output transformer. But, yu are right,it would seem that you would get 3 dB by removing one output tube, without having a lot of distortion. Your math looks right, even using 20 times log (P1/P2), instead of 10 as one would do for power. Hmmmmmmmm.....back to the old thinking cap, I guess. Talk to you later.
:
:Lewis

Doug,

Is the choke shown on the schematic from the centertap of the filament winding for the 2A3s with a 2 mike cap in parallel to chassis/ground/B-) part of the speaker or separate?

Radiodoc
**********

You are referring to the 780-ohm choke, I think?

Look at the plug and socket. Everything to the right WAS part of the speaker, presumably. But I don't have the original speaker, nor did I ever have it. So, I have built a box that duplicates the circuit shown for the speaker and am using a permanent-magnet speaker. The two chokes shown, one of which is the field coil and the other the 780-ohm choke, I have replaced with resistors.
Doug

:
:Doug,
:
:Is the choke shown on the schematic from the centertap of the filament winding for the 2A3s with a 2 mike cap in parallel to chassis/ground/B-) part of the speaker or separate?
:
:Radiodoc
:**********
:

Yep that's the one. Not being able to physically see the speaker, I just wondered if the choke was wound on the same core as the field coil or if it was a choke mounted separately on the speaker proper. With a cap in parallel would create a low frequency tuned circuit. Wouldn't it be interesting to go back in time and discuss with the engineer(s) why a choke instead of a resistor was decided to be used there?

Radiodoc
**********

:Doc, you're looking at the power section schematic shown at the top half of page 15-27, correct?
:http://www.nostalgiaair.org/PagesByModel/599/M0016599.pdf
:
:You are referring to the 780-ohm choke, I think?
:
:Look at the plug and socket. Everything to the right WAS part of the speaker, presumably. But I don't have the original speaker, nor did I ever have it. So, I have built a box that duplicates the circuit shown for the speaker and am using a permanent-magnet speaker. The two chokes shown, one of which is the field coil and the other the 780-ohm choke, I have replaced with resistors.
:Doug
:
::
::Doug,
::
::Is the choke shown on the schematic from the centertap of the filament winding for the 2A3s with a 2 mike cap in parallel to chassis/ground/B-) part of the speaker or separate?
::
::Radiodoc
::**********
::

I wonder if it was essentially just a wire-wound resistor with no particular purpose for the inductance? I honestly can't think of a purpose for a choke at that location.

Doc, you said that you'd like to go back in time and talk to the Scott engineers. If you hurry, you won't need a time machine. Murray Clay served as E.H. Scott's chief engineer, 1934-38. His initials are in the title block of the schematic for my All Wave 15. The last I heard, he was alive, at age 100, and still lived in Chicago (where Scott Radio Labs was located).

After Clay left Scott, Marvin Hobbs served as Scott's chief engineer, 1939-47. Hobbs also is still alive, at least as of a few years ago when the second edition of his history of Scott was published.
Doug

:Doug,
:
:Yep that's the one. Not being able to physically see the speaker, I just wondered if the choke was wound on the same core as the field coil or if it was a choke mounted separately on the speaker proper. With a cap in parallel would create a low frequency tuned circuit. Wouldn't it be interesting to go back in time and discuss with the engineer(s) why a choke instead of a resistor was decided to be used there?
:
:Radiodoc
:**********
:
:
::Doc, you're looking at the power section schematic shown at the top half of page 15-27, correct?
::http://www.nostalgiaair.org/PagesByModel/599/M0016599.pdf
::
::You are referring to the 780-ohm choke, I think?
::
::Look at the plug and socket. Everything to the right WAS part of the speaker, presumably. But I don't have the original speaker, nor did I ever have it. So, I have built a box that duplicates the circuit shown for the speaker and am using a permanent-magnet speaker. The two chokes shown, one of which is the field coil and the other the 780-ohm choke, I have replaced with resistors.
::Doug
::
:::
:::Doug,
:::
:::Is the choke shown on the schematic from the centertap of the filament winding for the 2A3s with a 2 mike cap in parallel to chassis/ground/B-) part of the speaker or separate?
:::
:::Radiodoc
:::**********
:::

Doug:
I think that choke - now resistor - is providing your negative feedback to the audio section that prevents you from losing 3dB when you pull one output tube. I have forgotten exactly how it works, I will look it up and see what I can find.
Lewis

:
:::
:::You are referring to the 780-ohm choke, I think?
:::
:::Look at the plug and socket. Everything to the right WAS part of the speaker, presumably. But I don't have the original speaker, nor did I ever have it. So, I have built a box that duplicates the circuit shown for the speaker and am using a permanent-magnet speaker. The two chokes shown, one of which is the field coil and the other the 780-ohm choke, I have replaced with resistors.
:::Doug
:
:
:Doug:
:I think that choke - now resistor - is providing your negative feedback to the audio section that prevents you from losing 3dB when you pull one output tube. I have forgotten exactly how it works, I will look it up and see what I can find.
:Lewis

Doug:
What I am trying to say is that the same process that gives you negative tube bias gives you negative audio feedback. When you pull one tube, you change bias and the feedback, so your output will not change 3 dB.
Lewis

:Lewis: yes, see Terry's posting down below. The 780-ohm thing is a common cathode resistor to provide self biasing of the 2A3s.
:Doug
:
:

Well, I will frankly admit I don't know for sure, I am just throwing ideas around and seeing if we can agree or disagree, just like I have done for years with problems with things that don't do like they are supposed to. Two mF would be about what 75 Ohms at 1Khz? That might provide a little feedback, heck I am just throwing ideas up in the air.
Lewis

marv

:Marv, what about just using my DMM to measure the voltage across the speaker with a tone generator inputing the phono connection on the receiver? (I did this once before when I was testing the audio frequency response of the receiver.)
:
:I would measure the voltage with both P-P output tubes installed, and then with one tube pulled. I can sure as heck detect 3 dB down with my DMM even if my ears can't.
:
:Now I have to find my audio tone generator among all the boxes and debris from my move. (I do have my DMM in captivity.)
:
:(This P-P question originally arose after setting up my E.H. Scott receiver in the "new" house - it is one of my main sets in the place.)
:Doug

Another thing that possibly could be going on here is speaker efficiency. Maybe at 1.9 dB down, the speaker is slightly more efficient? I don't know.

But I'm several weeks away from having my bench back in operation, which will allow more thorough testing. When we moved, the actual, physical, wooden bench stayed with the old house, because it was built into the basement. I'm not going to set up a temporary bench on a card table! ;>)

Maybe somebody out there in RadioLand with a p-p set would like to carry on this noble experiment while I'm out of business?
Doug

:Doug,
:As the old saying goes, "a picture is worth a thousand words". Using a 2 channel o'scope vs the DMM would allow comparative analysis of the push/pull activity in the time domain, and cause and effect determination. Lewis, supposedly 1db is the least amount of change the human ear can detect (normally), and -3db is mathmatically the half power point, but not necessarily the way the human ear decerns it on a logarithmic scale. Air density and background noise in the listening environs might even enter into the equation.
:
:marv
:
:OK, I found my audio tone generator, and performed a little experiment with my P-P E.H. Scott receiver.
:
:Here was the set-up. I inserted an 420-Hz audio signal into the phono input. The AC input voltage was kept comstant during the experiment.
:
:There are two 2A3 output tubes in P-P: a Tung-Sol and a JAN (unknown manufacturer, but possibly RCA). With my Fluke DMM, I measured the AC output voltage at the speaker terminals. Here are the results:
:
:Both tubes installed: 95.0 mV output
:Tung-Sol pulled: 74.8 mV
:JAN pulled: 79.5 mV
:
:It seems that the Tung-Sol is slightly stronger than the JAN. For comparison purposes, let's use the average single-tube output = (74.8 + 79.5)/2 = 77.2 mV, and compute the power drop to the speakers:
:
:20 log (77.2/95.0) = 20 x (-0.0901) = -1.8dB.
:
:If the power had dropped by half with one tube pulled, we would have expected -3dB. So, it would appear that something is happening in the amplifier circuit that partially compensates for the pulling of one P-P tube. Once I get my bench back in operation, I can explore this further - possibly the B+ is going up with one tube pulled or maybe something like Marv's theory.
:Doug

:The push pull 2A3s are biased for class AB1 via a common 780 ohm bias resistor in the filament circuit. When you pull one tube the bias changes to class A and the current in the remaining tube goes up from about 40 ma to about 60 ma. The B+ will go up somewhat and the tube will be in a higher transconductance region so power will not drop by half but depending on individual tube characteristics will drop by a db or two. All this assumes a constant audio input signal voltage.