Push/Pull Pair Removal Theory & Questions

[bob p]

4 pages and this thread still going on. Guys, really?!

Don't complain -- the free answers that you get here are worth every penny that you pay for them.

[Helmholtz]

That was at 1KHz. At 120 Hz it is 6.8H. The difference is probably due to core loss.

Why do you think this is low? It is perfect for the job it is intended for. Transformer design is compromise between various limitations. Keeping the inductance no higher than necessary means, for example, that the resistance can be kept down, lowering losses.

It makes no sense to measure OT primary inductivity with an LCR meter. The reason is that the ac permeability of steel cores strongly depends on flux. The inductivity needs to be measured at nominal (rated) current, repectively calculated from the voltage to current ratio. The small signal inductivity is typically too low by a factor of 10 to 15.

Never use an asymmetrical output tube arrangement in a PP amplifier, as PP-OTs are not designed for net DC flux. It is safe (regarding voltages) to pull 2 outer or inner tubes of a Fender Twin Reverb if you as well switch the primary of the PT from 110V to 127V or from 220V to 240V respectively. Generally a heater voltage of 10% below nominal is worse than 10% above for tube lifetime.

Sorry, if some of my info is obsolete/redundant, but I simply don't have the time to read all the posts.

[Mike Sulzer]

It makes no sense to measure OT primary inductivity with an LCR meter. The reason is that the ac permeability of steel cores strongly depends on flux. The inductivity needs to be measured at nominal (rated) current, repectively calculated from the voltage to current ratio. The small signal inductivity is typically too low by a factor of 10 to 15.

Never use an asymmetrical output tube arrangement in a PP amplifier, as PP-OTs are not designed for net DC flux. It is safe (regarding voltages) to pull 2 outer or inner tubes of a Fender Twin Reverb if you as well switch the primary of the PT from 110V to 127V or from 220V to 240V respectively. Generally a heater voltage of 10% below nominal is worse than 10% above for tube lifetime.

Sorry, if some of my info is obsolete/redundant, but I simply don't have the time to read all the posts.

If the inductance is much smaller at low flux levels, then you need to measure it at such low levels to assure that you have enough inductance to get the rated bass response, which is required at low as well as high signal levels. And what better to use than an LCR meter?

Of course this applies to PP-OTs , not SE-OTs, which have a high average current, and thus flux level.

I think one purpose of high signal level inductance measurements is to make sure that the transformer has a high inductance at its rated power, and thus does not saturate below it. So you need to use a low frequency.

I think it is quite complicated to fully characterize the inductance of either type PP or SE. One ought to make small signal ac measurements with dc current switched through a set of values covering the range of useful fluxes. Not so easy.

[loudthud]

The shop I'm doing repairs for just Hates Twin Reverbs. They can't give them away except to Steele Guitar players. They hired a guy who knows how to work on Steels so that brings in some Steele players. (He's out on tour now.) There might be a small niche market for modded Twins that use one channel as a buffer between the Steele and Volume pedal and possibly installing a 15 inch speaker (JBL ?).

[Helmholtz]

If the inductance is much smaller at low flux levels, then you need to measure it at such low levels to assure that you have enough inductance to get the rated bass response, which is required at low as well as high signal levels. And what better to use than an LCR meter?

Of course this applies to PP-OTs , not SE-OTs, which have a high average current, and thus flux level.

I think one purpose of high signal level inductance measurements is to make sure that the transformer has a high inductance at its rated power, and thus does not saturate below it. So you need to use a low frequency.

I think it is quite complicated to fully characterize the inductance of either type PP or SE. One ought to make small signal ac measurements with dc current switched through a set of values covering the range of useful fluxes. Not so easy.

You are right, bass response generally gets noticeably worse at very low power levels. (This doesn't seem to be a major issue in guitar amps, though.) Thus it may make some sense to check inductivity with an LCR meter at 100Hz also.

The large signal inductivity increases with AC flux density (current amplitude) up to the onset of core saturation from where it starts to decrease.
Inductivity decreases with DC flux density, though.

I used to measure primary inductivity with isolating mains transformers and a variac at 50Hz, voltage and current meters and a scope to control current distortion (core saturation). Voltages of up to 250Vrms are necessary, depending on the OT power rating.


A word regarding missmatching speaker impedances: Tube power amps are designed for an optimum Raa (reflected speaker impedance). If you cut Raa in half, plate signal currents and plate dissipation will increase. If you double Raa, screen currents and screen dissipation will be excessive at high output levels. Both effecte are likely to exceed the max. ratings of the tubes and thus reduce tube lifetime. Hot biasing makes things worse.

[bob p]

It is safe (regarding voltages) to pull 2 outer or inner tubes of a Fender Twin Reverb if you as well switch the primary of the PT from 110V to 127V or from 220V to 240V respectively. Generally a heater voltage of 10% below nominal is worse than 10% above for tube lifetime.

That is a good point. There is a difficulty in exercising that sort of plan on this side of the Atlantic, because we don't have the multi-tap primary PT that are common in countries that run on 220-240V mains. Here in the USA we get "domestic" PT that don't have any selectable primary taps. Although it is possible to do what you suggest if someone intentionally chooses an "export" PT for their "domestic" builds (which will be more expensive ... I actually do this with the Hammond 300 series iron), the vast majority of people don't have that option with "domestic" iron.

This situation has improved somewhat, as in recent years Hammond has provided alternate primary taps on their 200 series "domestic" PT, allowing people to chose between a 115V and a 125V tap. For decades we never had this option, as Hammonds were designed in the era of 115V mains and were never updated in spite of the mains voltage increasing to 125V. At least now we have a 125V mains option on the PT primary, but to really make things work the way you suggest just isn't an option for us. We'd need a higher mains option that we don't have. In order to implement your suggestion, someone would have to take the counter-intuitive step of designing the amp to operate on 125V mains using the PT 115V tap, and allowing the switch to the 125V tap when pulling tubes. Most people would never consider doing that.

I like the 300 series because if offers even more ways to do this: the suitable mains options for me are 100 / 110 / 120. As you've noted, taking steps to keep the heater voltage in-spec is a necessary step.

[bob p]

The shop I'm doing repairs for just Hates Twin Reverbs. They can't give them away except to Steele Guitar players. They hired a guy who knows how to work on Steels so that brings in some Steele players. (He's out on tour now.) There might be a small niche market for modded Twins that use one channel as a buffer between the Steele and Volume pedal and possibly installing a 15 inch speaker (JBL ?).

The 15" Twin Reverb -- wasn't that called a Vibrosonic Reverb during the 70s?

I *LOVE* Twin Reverbs. Great amp as a pedal platform. They're definitely not popular anymore -- now you can buy a big amp like a TR with premium speakers for a fraction of the price of a Deluxe Reverb.

[Helmholtz]

....or use a variac and adjust for nominal internal voltages.

[bob p]

^ practical on the bench, but not on stage.

[Helmholtz]

To be fair, in some more conservative designs pulling two tubes and doubling speaker impedance works just fine. While anode and heater voltages may rise by 5% above nominal, also the bias voltage magnitude rises, thus decreasing idle currents. This helps a lot. The most critical thing to check is screen voltages and dissipation.

[g1]

Generally a heater voltage of 10% below nominal is worse than 10% above for tube lifetime.

My understanding was that the opposite were true?

[bob p]

Smart guys don't go near 10% in either direction.

[Helmholtz]

Yes, IIRC too high filament temperature is bad, but too low oxide cathode temperature is even worse. Effects like cathode poisoning and cathode hot spots causing sputtering of cathode material can irreversably reduce emissivity.

[pdf64]

To be fair, in some more conservative designs pulling two tubes and doubling speaker impedance works just fine. While anode and heater voltages may rise by 5% above nominal, also the bias voltage magnitude rises, thus decreasing idle currents. This helps a lot...

Idle conditions are a consideration certainly but I think it's the full load conditions that are likely to be most significant; as the lightly loaded power supplies will have somewhat less sag, and so with the g2 node maintaining a higher voltage, the plate currents will tend to be higher, potentially resulting in overdissipation of the plate and g2.
The reduced sag may tend to result in a more sterile dynamic response, and the higher loaded voltages may tend to produce a hardsher tone.
So the amp will tend to sound worse when overdriven; whereas a key reason (though not in this case the OP's) for reducing the amp's power may often be to get power amp overdrive at lower power levels.

...What modification can I do to facilitate the fact I WILL be running them with two tubes for the conceivable future...

Perhaps the appropriate mod is to reduce your expectation of your amps sounding as they're intended to?
To mod the amps, I suggest to replace the transformers and power supplies for those suitable for the equivilant 2 tube model.

...It's like going to an ice cream shop and them saying...

"come back when you can afford an ice cream"

[Helmholtz]

Idle conditions are a consideration certainly but I think it's the full load conditions that are likely to be most significant; as the lightly loaded power supplies will have somewhat less sag, and so with the g2 node maintaining a higher voltage, the plate currents will tend to be higher, potentially resulting in overdissipation of the plate and g2.
The reduced sag may tend to result in a more sterile dynamic response, and the higher loaded voltages may tend to produce a hardsher tone.
So the amp will tend to sound worse when overdriven; whereas a key reason (though not in this case the OP's) for reducing the amp's power may often be to get power amp overdrive at lower power levels.


Perhaps the appropriate mod is to reduce your expectation of your amps sounding as they're intended to?
To mod the amps, I suggest to replace the transformers and power supplies for those suitable for the equivilant 2 tube model.


"come back when you can afford an ice cream"

Without concrete numbers this will remain speculative. It is easy to restore original sag behaviour with a (switchable) power resistor or choke wired in series with the PT primary.