Push/Pull Pair Removal Theory & Questions

[eschertron]

Let me offer this completely unproven assertion:
When trying to get less 'volume' out of a 100W amp by pulling two tubes and moving the impedance tap, the resulting loss of LF capacity in the OT can actually help the perception of a quieter amp.

[Mike Sulzer]

Agree with the theoretical concerns, differ on the practical ramifications.
I'd think this would be more of an issue with hifi than guitar amps, but point taken. It's still real and I imagine the tx designers have a specific primary z to build around.

Consider The Magnetics Components Classic Tone 40-18025, a fifty watt transformer. The impedance across the full primary is 3.6K (http://www.classictone.net/40-18025.pdf). I measure the primary inductance at 6.56H. The frequency f at which the inductive reactance equals the primary impedance is f = 3600./(2.*pi*6.56) = 87.34Hz. If you use this where you should be using a 7.2K primary, you have about 175Hz. This might matter; it depends on what you want.

[pdf64]

6.56H seems low?
The Hammond equivalent is spec'ed at >119H (tested at 240V 50Hz) http://www.hammondmfg.com/pdf/EDB1750N.pdf

[mikepukmel]

What a great discussion. So, the answer is . . .
Pop 2 tubes out (assuming you're popping the correct set out),
then pop in a different OT designed for the task.

[Chuck H]

What a great discussion. So, the answer is . . .
Pop 2 tubes out (assuming you're popping the correct set out),
then pop in a different OT designed for the task.

Or just pop out the tubes (in a four tube amp) and either double the actual speaker load impedance or half the impedance switch into the existing speaker load. Consequences not withstanding. Most OT's aren't so cheesy that you'll suffer noticeably for it at guitar and guitar speaker operating frequencies.

[Mike Sulzer]

6.56H seems low?
The Hammond equivalent is spec'ed at >119H (tested at 240V 50Hz) http://www.hammondmfg.com/pdf/EDB1750N.pdf

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.

[Enzo]

We can calculate and measure various things about a transformer, but in an amp, if we pull two of four output tubes, in general it works fine, as long as you adjust the speaker tap. Any little losses are easily compensated for if even noticed. In fact any number of commercial amps have a half power switch that does exactly that - disable two of four tubes.

[Chuck H]

Indeed! Pretty much all the four output tube amps I know of use a beefy OT with ample LF capability. That's sort of the point with four output tubes. And these amps are the only place you'll find a "half power" switch of that type. Something like a tweed Deluxe, that has a more marginally spec'd hunk of iron in it may audibly suffer for an impedance mismatch. But then it's not a four tube amp and it's not a head.

[g1]

Those amps that disable a pair for the 'half power' switch, do they even bother about changing impedance tap?

[Chuck H]

Not the ones I just looked at. I know there are others, because I've seen them along the the way in life. But a *oogle search reveals a lot of Ceriatone models (not sure if those should count?) and a couple of Mesa schematics. None compensate for impedance.

[Enzo]

And they don't because there are no tight specs in guitar amps.

[pdf64]

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.

I take your point but think it seems low as I'm used to seeing values x10 that, in Hammond's and other info.
And with a primary inductance in that range, it doesn't seem that the OT primary impedance would support the cab's bass resonance?
I'm likely missing something, but wouldn't that mean that the OT would end up dissipating the amp's output at the cab's bass resonance?

[Mike Sulzer]

I take your point but think it seems low as I'm used to seeing values x10 that, in Hammond's and other info.
And with a primary inductance in that range, it doesn't seem that the OT primary impedance would support the cab's bass resonance?
I'm likely missing something, but wouldn't that mean that the OT would end up dissipating the amp's output at the cab's bass resonance?

I found primary inductances only for Hammond's economy line (no audiophile bass):Hammond Mfg. - Universal Tube Output - Push-Pull Transformers - (125 Series). The values vary between 3 and 15H.

I think the voice coil would dissipate much more than the transformer; that is where the resistance is.

[pdf64]

I found primary inductances only for Hammond's economy line (no audiophile bass):Hammond Mfg. - Universal Tube Output - Push-Pull Transformers - (125 Series). The values vary between 3 and 15H...

Yes, that range is only specified (-1dB @ max rated power) down to 150Hz.
The inductance for the tube guitar amp 1750* OT range is noted on each individual info sheet, as per previously linked example.
All except the 18watt and AC15 replacements look to be >50H, -1dB 50 - 70Hz.

[bob p]

Those amps that disable a pair for the 'half power' switch, do they even bother about changing impedance tap?

Yes, sort of. Back in the early days power switching was done manually and a recommendation was made for the user to switch the load manually to avoid hard mismatch conditions.

In Mesa's original patented embodiment, the responsibility for switching power AND switching load were both tasks that had to be performed manually by the user. There was no electronic selection of these features in the first generation amps, and the user had to flip a switch and move the load wire to perform both changes. Insofar as these were both back-panel set-up selections they weren't the kind of selections that were intended to be switched while the amp was in play.

The original implementation -- when the first generation of adjustable power Mesa first came out with this 'patented' feature -- interrupted the cathodes on two of the four tubes and left the selection of the OT secondary impedance tap unswitched. The amps did have user-selectable secondary impedance taps (separate 4R and 8R jacks), and the responsibility for secondary matching was left to the user. The switching had to be done manually, by moving the load's cable from the 4R to the 8R jack on the back panel.

It's easy to look at the cathode switch not having extra poles that were directly connected as a speaker load switch and to assume that Mesa wasn't paying attention to changing the Z tap, but that's not quite the case. In fairness to Mesa, they DID tell users that they were supposed to change the load when power-switching and when switching tube types. There are specific instructions in some of the user manuals about needing to plug 8R loads into the 4R jacks. In this scenario the implementation of power switching was done manually, with the user flipping a switch, and the load switching was also done manually, with the user moving the load plug to a different jack.

So the answer to your question is a definite Yes, as even back in the early days Mesa specifically instructed people to perform the load matching task manually. I have to say though, Mesa's instructions about this weren't written all that well, but at least they spoke about the subject, even if people didn't understand it well after reading the instructions.


In later generations of amps some of these features became automated. Their flagship amps like the Road King were full of relays that were used for switching the power amp configuration. I don't have a schematic in front of me, but I'd like to say that in the later embodiments Mesa did do load switching in tandem with tube switching. Knowing the patent history of Mesa, they probably filed a patent application on the idea. Royalties may be the reason that other companies aren't doing it. It's cheaper for them to tolerate a mismatch than to pay royalties on the switching paradigm.