Parallel output/single ended

[dkevin]

I have acquired an old phonograph/instrument amplifier that I would like to convert to a guitar amp. It uses 2 6V6 tubes in push/pull design. I want to convert this amp into a single-ended, parallel output amplifier like the Gibson GA-9. I understand that I will need a replacement output transformer. My question deals with this choice.
I have heard that 6v6 tubes require a primary impedance of 5k to 8k. But driving the output tubes single ended and parallel changes the equation. Can you give me an understanding of what doubles and what diminishes and what stays the same when this change is made?
Thanks for your time!

[Chuck H]

To run two tubes in parallel you would half the load.

There may be other considerations. The amp your modding is AB1 so the PT may not be rated to handle full current, full time as it would be with a parallel class A amp. Not sure if the PT in that amp is up to it. There could be other power supply issues as well. Like a big ass choke (if the Gibson used one). Not sure, just making an observation.

Chuck

[R.G.]

I have heard that 6v6 tubes require a primary impedance of 5k to 8k. But driving the output tubes single ended and parallel changes the equation.

Sylvania technical manual for 6V6GTA says load resistance of 5K at 250V, 8.5K at 300V, single tube A1. It doesn't say anything about plate to plate in push pull. For parallel Class A1, those impedances would half.

By the way, they also mention that it produces 5.5W into 8.5K - at 12% distortion! I was impressed that they'd bother to quote power out at distortion that high.

[Chuck H]

they also mention that it produces 5.5W into 8.5K - at 12% distortion! I was impressed that they'd bother to quote power out at distortion that high.

They must have thought it sounded good. Ahead of their time

[dkevin]

Thanks for the help!
Along the same lines, what is the effect of paralleling output tubes (versus push/pull) on the size of the cathode resistor? I have pulled up a number of schematics using a parallel output section of two 6V6 tubes and the cathode resistor seems to be all over the map.

[tubeswell]

Thanks for the help!
Along the same lines, what is the effect of paralleling output tubes (versus push/pull) on the size of the cathode resistor? I have pulled up a number of schematics using a parallel output section of two 6V6 tubes and the cathode resistor seems to be all over the map.

Same as running two push-pull tubes off the same Rk. (You halve the Rk value) But you could run each tube off its own separate (470-500R) Rk. That way you could unplug one of the tubes at will and (by swapping the load resistance of the OT at the same time) run it on one tube

[Chuck H]

When running two tubes in parallel is there any reason not to wire the filaments in phase
opposition? This would cancel hum right?

Also, for half power you could just ground one of the grids. There was some debate here about whether this still requires a change to the primary impedance since the tube with the grounded grid isn't conducting. But it seems to me, for class A anyway, that the tube is already at max diss just sitting there so this should work without any impedance matching considerations and still achieve hum cancellation too, right?

Chuck

[dkevin]

I really don't have a desire to remove one of the tubes and run it as a single-output tube. I am just seeking to understand why the numbers seem to be all over the map.
While we're at it, is there any advantage to separating the output tubes and giving each of them an individual "470 ohm" resistor as opposed to a shared resistor?
One last area of concern; the Gibson GA-9 uses a field coil speaker. This shows up in the power rail as a "1000 ohm" resistor. The PT puts out 342vdc at the 5Y3 and flows first through the 1000 ohm resistor and then output transformer. The resulting voltage at the plates of the 6v6s is said to be 263vdc. Would you expect the voltage drop through this resistor and the OT to be this much? I wonder what would be the benefit of reducing 342vdc to 263vdc in this fashion? Would it provide a stiffer response when pushed?
Thanks for the help! The questions just keep on coming.

[Chuck H]

I really don't have a desire to remove one of the tubes and run it as a single-output tube. I am just seeking to understand why the numbers seem to be all over the map.

We have to compare apples to apples. Do the different amps your looking at all run the same plate voltage? Even if they all use the same tube it was common for different tube brands (even of the same type) to have different bias requirements and designs are often brand specific. Also, there is quite a bit of leeway in the strike zone for cathode bias because it is a "self bias" circuit.

While we're at it, is there any advantage to separating the output tubes and giving each of them an individual "470 ohm" resistor as opposed to a shared resistor?

The obvious reason to seperate the cathodes would be to control bias for each tube. Since your not in AB there is no need for tube matching, In fact it may be better for a guitar amp to intentionally mismatch the tubes for a more complex tone.

The non obvious reason would be to avoid phase cancellations due to non linearities between the two valves. A bypass capacitor makes this a moot point.

One last area of concern; the Gibson GA-9 uses a field coil speaker. This shows up in the power rail as a "1000 ohm" resistor. The PT puts out 342vdc at the 5Y3 and flows first through the 1000 ohm resistor and then output transformer. The resulting voltage at the plates of the 6v6s is said to be 263vdc. Would you expect the voltage drop through this resistor and the OT to be this much?

I would. I usually like to run a simulation on Duncan PSU before making a statement like this but considering the high resistance speculated by Gibson and the fact that ALL the amps current is pulled through it I would expect a substantial voltage drop.

I wonder what would be the benefit of reducing 342vdc to 263vdc in this fashion? Would it provide a stiffer response when pushed?

I don't think it was done with any benefit in mind. I think it's just a consequence of the circuit topology. Gibson tended toward more popular design trends in home audio than did Fender. At that time all the voltage being fed through a big choke was common in home audio. Gibson probably used a smaller than ideal choke (thus the high resistance) to keep the amp affordable. Ideally in class A there is no voltage sag because the tube never increases it's current. When pushed though the bias class can shift into AB (single ended?) and the tube can both cutoff and saturate. In this instance I would think that a higher plate voltage would give a little "stiffer" response.

Some of this is speculation. Corrections accepted.

Oh, BTW I was asking the above Q's for my own benefit but I thought they were appropriate to the thread, not hijacking.

Chuck

[dkevin]

I thank you, one and all, for your help. I will apply the lessons here to the building of a GA-9 clone from the donor amp.
As a parting question, what if I decided to eliminate the field coil resistor (of 1000ohms) and instead chose to use a different power transformer that would deliver the required plate voltage to the 6V6? Would you imagine an appreciable difference in the way the amp performed?
Oh well, as you can tell I have so many questions and there's so little time.
Thanks again!
dkevin

[melvin]

The speaker wouldn't work, of course. The choke, the "1000" resistor and the field coil speaker are all one and the same. The resistance of the coil in the electromagnet of the speaker is 1000 ohms and it acts as a choke. The choice of (field coil) speaker would determine the size of the choke.

The grid-grounding scheme is interesting, but I bet it would sound a whole lot different than if you shut one of the tubes off completely.

[tubeswell]

But you could replace the field coil speaker with an SE OT and a speaker

[dkevin]

Yeah, that was my thinking. I am not wild about field-coil speakers. So, I'm thinking SE OT in the 2.5K-4K range for 2 ea. 6V6's running in parallel at about 270vdc sharing a cathode resistor of 270 ohms.