Class A Push Pull - output transformer specs

[systemtruck]

I’ve been learning more about class A push pull and they seem rare in the guitar amp world so am wanting to one day try converting an amp from AB to A.

Power transformer demands are pretty straight forward to figure out, but the output transformer seems like a challenge…

If we have some vintage output transformer on hand, is schematic / context the only way to approximate current capabilities of it?

These numbers below are extreme, but to keep the example math simple let’s say we have a push pull amp that runs its power tubes at 60% Class AB duty cycle and each tube is pulling 60mA. And let’s say we’re wanting to rebias that to Class A so 100% so then we want each tube to be able to pull 100mA. We’ll also need to redo the power transformer and rails, but aside from that, we want to be pulling an extra 40mA through each side of the primary of the output transformer.

Assuming we know little about a transformer’s specs other than the context and schematic it exists within, can we test or estimate the capabilities of those primary windings and take somewhat of a risk free attempt at the build?

DC resistance readings and design voltage

 

[CJ]

First thing you do is pick your tube type and how many,

Class A amps are almost always built with the lower wattage tubes, like the 6V6, 6BQ5, 6AQ5,

Why? Heat. And economics. A pair of 6L6 tubes running through a cathode resistor gives off enough heat to warm a small apartment. And the wattage goes down because you have to drop the plate voltage to keep them from melting. So a pair of 6V6 tubes running at 400 volts will approach the same wattage as a pair of 6L6 tubes running in class A. And remember that you have to quadruple the wattage to double the sound as the ear has a logarithmic response for safety reasons. So even if the 6L6 tubes put out a few more watts in Class A than a pair of 6V6 tubes, it will not seem that much louder, especially if the speaker cone area stays the same.

Tube manuals will tell you how to set up big tubes for Class A.

The transformer can be beefed up for Class A by choosing a size larger core or increasing the stack height of the lams used in the Class AB version. Then you bump the primary wire up a couple of notches, is run 30 AWG instead of 32 which gives you 100 ma instead of 62 ma for the #32. But the turns will go down because the wire is bigger. This us why you bump the core up a bit, so you end up with the same primary inductance. And since you running higher plate current with Class A, you also run the risk of having higher unbalanced DC on the core, so a bigger core will handle more DC flux.

Dagnal makes outputs for Marshall. They feature a large fire and small DC resistance, about 40 ohms, so they would probably handle EL 34's running in Class A all day.

Schumacher makes a robust output for the Twin and Super Reverb amps, they would probably be OK also.

The DeLuxe Reverb OPT gets hotter than a pistol when the SMP is cranking, so does the pwr xfnr, so that iron would make a poor choice for Class A.

Partridge probably makes the most over engineered OPT but they are big bucks and so old that they are ready to explode.

 

[abbey road d enfer]

I’ve been learning more about class A push pull and they seem rare in the guitar amp world so am wanting to one day try converting an amp from AB to A.

You have to realize that class-A is quite inefficient compared to class AB.
Your 50-watt class AB will become a 12-watt class A.
And some tubes are better suited to class-A than others. As CJ mentioned, 6BQ5/EL84 and 6V6 are favourites.

Power transformer demands are pretty straight forward to figure out, but the output transformer seems like a challenge…

If we have some vintage output transformer on hand, is schematic / context the only way to approximate current capabilities of it?

If you start with the notion that you need a xfmr that is rated for 4 times the expected power, there is some choice.

These numbers below are extreme, but to keep the example math simple let’s say we have a push pull amp that runs its power tubes at 60% Class AB duty cycle and each tube is pulling 60mA. And let’s say we’re wanting to rebias that to Class A so 100% so then we want each tube to be able to pull 100mA. We’ll also need to redo the power transformer and rails, but aside from that, we want to be pulling an extra 40mA through each side of the primary of the output transformer.

All this leads to the fact that the xfmr should be rated at, at least 2.8 times the power.
See how an original AC30 OT (30W class A) is about the same size as a Twin-Reverb's (85W class AB).

 

[systemtruck]

Thank you! That all makes perfect sense and yes I know the power loss is dramatic. I’m so curious about hearing a Class A push pull in person though, even if it’s just 8-10 Watts in the end. It seems like the best of both sonic worlds and worth a try.

I have a handful of old output transformers from old amps / radios / test equipment and basically am taking guesses about each one, wondering if one might be close to usable. Some are quite beefy compared to other transformers I see for the same PP tube setup. I suppose there’s a chance one might be overbuilt enough to handle a lot of extra current? Funny that 6BQ5 and 6V6 are the best candidates because that’s a lot of what I have.

Would this be a matter of slow gradual burnout and then inevitable collapse over a bunch of usage? Or is this one of those things where it’ll just fry the moment you turn it on? For example, if I wired up a beefy OT with 6V6’s in Class A instead of Push Pull (with all rails / PT factors considered and adjusted and i understand about lowering the rail) but otherwise the same circuit, will I just be waiting to see if the OT gets increasingly warm and then hot? And if it stays cool then basically we know it’s fine?

From what I gather, since it’s running full cycle at all times, volume of signal won’t increase the operating temp of tubes and transformer, so as long as it stabilizes cool enough then that’s it?

 

[Murdock]

I have a somewhat related question so I hope it’s ok to post here.
Is there a standard how the manufacturers stated the specs back in the days?
Like the Stancor A-3250 for example.
Catalogue states 20k primary impedance and max 15ma current.
Is that all for one winding/tube?
Or for both halves/both PP tubes together?

 

[emrr]

Every broadcast monitor amp like BA-4 is a PP class A, BC-2B monitor amp, etc. There's a ton of PP 6L6 class A 10W amps out there.

 

[systemtruck]

I have a somewhat related question so I hope it’s ok to post here.
Is there a standard how the manufacturers stated the specs back in the days?
Like the Stancor A-3250 for example.
Catalogue states 20k primary impedance and max 15ma current.
Is that all for one winding/tube?
Or for both halves/both PP tubes together?

Every broadcast monitor amp like BA-4 is a PP class A, BC-2B monitor amp, etc. There's a ton of PP 6L6 class A 10W amps out there.

My own personal guess is that 15ma is the max DC current across the entire primary winding. And if that’s correct then it’s 7.5mA each half of primary.
Looking at the specs that are printed on that Stancor, it advertises it as Single Ended OR Push Pull. Since the end user of a single ended design is always focused on the primary current capabilities when choosing transformers, I’m guessing that this 15mA rating is meant for them to see. But the other folks here have actual knowledge, not me!!

But this all leaves me confused now, I have to admit…

I thought one of the major perks of a push pull format is that we don’t have to worry so much about DC current ratings because it gets canceled out by the two legs in opposite phases.

So… why are we now trying to figure out if a PP transformer can handle DC current in the context of Class A current draw as opposed to Class AB current draw? Aren’t they still “canceling each other out” even though they’re running at significantly higher draws individually?

 

[abbey road d enfer]

My own personal guess is that 15ma is the max DC current across the entire primary winding. And if that’s correct then it’s 7.5mA each half of primary.

15mA across the whole primary is also 15mA across each half primary.

I thought one of the major perks of a push pull format is that we don’t have to worry so much about DC current ratings because it gets canceled out by the two legs in opposite phases.

The 15mA rating is governed by the heat generated by Joule effect. Even if it cancelled out magnetically, the Joule effect would be the same.
What actually cancels out in a push-pull stage is the magnetizing current.

So… why are we now trying to figure out if a PP transformer can handle DC current in the context of Class A current draw as opposed to Class AB current draw? Aren’t they still “canceling each other out” even though they’re running at significantly higher draws individually?

They don't cancel out regarding heat elevation.

 

[systemtruck]

15mA across the whole primary is also 15mA across each half primary.

The 15mA rating is governed by the heat generated by Joule effect. Even if it cancelled out magnetically, the Joule effect would be the same.
What actually cancels out in a push-pull stage is the magnetizing current.

They don't cancel out regarding heat elevation.

Oh right, 15mA everywhere on the primary winding. That’s a Keirschoff’s Law thing, correct?

Ok gotcha, so the self cancelling in the magnetic domain is what lets us not need to gap the transformer, hence the convenience of a PP transformer build. BUT, we still have DC current ratings to abide by. Roger that.

So if we have a PP amp that’s using a pair of 6v6s, and they’re pulling some amount of current, if we want to switch to Class A and it’s running constantly at 100% dissipation we then need to reassess the current maximum of the transformer on the primary and make sure the increase is something the primary can absorb. And it probably ISNT, unless said transformer happens to be way overbuilt for the AB design.

 

[Murdock]

15mA across the whole primary is also 15mA across each half primary.

The 15mA rating is governed by the heat generated by Joule effect. Even if it cancelled out magnetically, the Joule effect would be the same.
What actually cancels out in a push-pull stage is the magnetizing current.

They don't cancel out regarding heat elevation.

I’m confused…
But each of the two PP tube can only be biased up to 7.5ma, right?

 

[CJ]

That Stancor is one if those "one size fits all" transformers designed to handle single ended or push-pull operation. It is a cheap way to cover both markets with an off the shelf solution to both circuits.
So the 15 ma is for single ended operation.

It is a gapped core design with a center tap. So the core will not saturate in SE mode but primary inductance will be quite a bit lower due to the gapped core.

However, sometimes cheap sounds good as the Fender Tweed DeLuxe used a similar transformer, a gapped core used in push pull. Inductance was about 40 Henrys IIRC. It was the weird one with the speaker common soldered to the frame which gets grounded to the chassis when you bolt it down.

Class A output tubes have a very high idle or quiescent current. So when you take a break you want to hit that standby switch.

Power transformers are designed for different types of rectifier circuits. A full wave bridge will mean that the secondary has current running through it at all times. With a center tapped full wave circuit, one half of the transformer secondary gets a break when its diode us not conducting. So the duty cycle will be 50% on the secondary winds. All this means is a heavier wire will be used on the full wave power transformers.

 

[abbey road d enfer]

I’m confused…
But each of the two PP tube can only be biased up to 7.5ma, right?

Current does not pass twice in a winding. 15mA in eaach half is 15mA total.

 

[merlin]

I think you're over thinking it. Any [power output] transformer that can handle a pair of tubes in Class-AB will also be able to handle them in Class-A. The idle current will be higher, the peak current will be lower, six o' one, half o' dozen of the other. Transformer won't care, it's not a fragile little flower.
The main difference is that you will want a lower rail voltage to allow you to bias at higher current without redplating the tubes. Don't try to start off with a rail of a bajillion volts like your last project! About 250Vdc (i.e. 200Vac transformer) is perfect for 6V6, EL84 etc. Maybe a few volts more if you're going for cathode bias.

 

[abbey road d enfer]

I think you're over thinking it. Any transformer that can handle a pair of tubes in Class-AB will also be able to handle them in Class-A. The idle current will be higher, the peak current will be lower, six o' one, half o' dozen of the other. Transformer won't care, it's not a fragile little flower.

Actually, the example chosen by Murdock is inadequate, since it's a quite low power xfmr (typically driven SE with a 12BH7), so yes, the 15mA limitation is to be taken seriously, I guess.
BTW, I'm not sure this particular xfmr is usable in a push-pull, since I don't see a mention of a center-tapped primary, when the secondary is clearly ibentified as CT.

 

[merlin]

Yeah I don't know why he picked such an esoteric, line-level transformer (it does have a CT though). I was talking about ordinary push-pull power output iron.

 

[Murdock]

Current does not pass twice in a winding. 15mA in eaach half is 15mA total.

Is it because of the way push pull works? The current is drawn alternately between the two tubes?
I simulated a PP circuit in spice where two tubes each have around 12mA cathode current.
When inserting a 100 ohm resistor between B+ and CT of the OT it states a current of 24mA.
But when looking at the two transformer primary halves one states 12mA and the other -12mA.
So would a PP OT with 12.5mA current rating be safe to use in that circuit?

 

[merlin]

So would a PP OT with 12.5mA current rating be safe to use in that circuit?

There is less than 12.5mA in any part of the winding. So yes it would be safe.

 

[abbey road d enfer]

Is it because of the way push pull works? The current is drawn alternately between the two tubes?

No.
It's because of the way current pentrates and exits nodes.
You have a node where current enters, supplied by the positive rail, and exits via two branches, each drawing a similar current.
Check Kirchoff.

I simulated a PP circuit in spice where two tubes each have around 12mA cathode current.
When inserting a 100 ohm resistor between B+ and CT of the OT it states a current of 24mA.
But when looking at the two transformer primary halves one states 12mA and the other -12mA.

That's correct.

So would a PP OT with 12.5mA current rating be safe to use in that circuit?

Yes.

 

[Cqwet Dbdfte]

For guitar amps bias "class A" makes little sense as it is not a music re-producer, but a "color" amp. Anything that comes out is the true original.
Acoustic guitar being an exception, I would think
If you want louder, mainly the speaker should be the center of attention. Efficiencies vary wildly.
Look closely at the dB/W/m numbers, or some use 2.86V instead of W.
Most of the energy in music is in the low frequency range, say under 200Hz, and the low frequency energy that determine output transformer size.
Happily, low frequency energy can be handled with switch mode amps, which are low cost, and can be crossed over at some suitable frequency.

 

[systemtruck]

For fear of this thread steering itself into yet another Class A vs Class AB debate, I’ll just state that I’m personally aware of the power loss and fully embrace it. And I personally have been getting more and more curious about what ‘amplifiers that don’t chop waveforms and put them back together’ sound like in the end, and also am now interested in seeing what that sounds like with the added benefit of Push Pull ripple and/or noise cancellation. The power boost from PP Class A vs single ended Class A is just a perk and not at all my own focus and reason for posting this question about using PP transformers in Class A. I just want to see what PP Class A sounds like in person.