SE transformer, air gap?

[gtrrich]

What happens if I use a push pull OT in a SE Class A design? I have read that the single ended trans will be specially made, with an "Air Gap", but I have also heard of some folks using a push pull transformer (no gap), and only using half of it, center tap to one end, or just not using the CT. If you have the right impedances and wattage, is it ok to do this? Thanks!

[Chuck H]

If the impedance is correct it will "work", but it won't perform very well. SE OT's have a lot more mass as well as an air gap. Both serve to idealize the OT for SE use. A regular PP OT forced into SE operation will saturate too quickly and suffer poor fidelity. And the more you push the amp, the worse the OT will perform.

Chuck

[Steve Conner]

It will indeed saturate and sound like poop.

A SE OT has an air gap to prevent saturation. But the air gap reduces the magnetizing inductance, so more iron and more turns are needed to bring it back up to a reasonable level.

[R.G.]

The comments you've received are correct. There is a fundamental difference in the operating conditions between P-P and SE. In P-P, the transformer operates at effectively zero DC current through the primary, since the two sides take turns going about as far as each other, adding up to a net zero. In SE, the DC current through the primary must be half of the peak signal current for the primary, or the transformer will forever be running into saturation on one polarity of signal peak.

The amount of DC current offset that a transformer can withstand is composed of two parts - that part that offsets the iron magnetically, and that part that creates a magnetic field in air gaps. The magnetic "goodness" of iron is tens of thousands of times (literally; that's a representative number for iron's permeability, not something I made up as a figure of speech) better than air. But iron saturates and air doesn't.

Mother Nature has rigged things so that iron is expensive and air is almost free. So the trick in designing a P-P transformer is to design one with a minimal air gap because the currents sum to zero, or very nearly. That lets us get the maximum signal swing with the least iron. Air gaps only make things worse for P-P transformers because they lower the magnetic qualities and that forces us to use big amounts of iron and more turns (pounds/kg) of copper to get the performance we want, and that costs money.

As a side benefit, in P-P mode, you are pulling the transformer core (in a magnetic field sense) from zero/balanced toward positive saturation on one polarity of the input signal, and toward negative saturation on the other polarity of signal. This lets you use the full positive-to-negative swing of the magnetic field in the iron for transforming signal.

All this taken together means that P-P transformers are the cheapest way to build an OT for a specific power output and frequency range. The price you pay to get this is that you must use two (or paired) output devices and the circuit must be a bit more complicated to drive the push-pull nature of the output swing. Generally, you save so much on the transformer cost that the extra parts are paid for, and you get the deluxe benefit of much higher power output per pound of transformer. This paragraph pretty much puts in a nutshell why most tube amps are P-P.

SE operation however, requires that there be a standing current of half the peak output current in the transformer all the time, on average. With only one output device to make the transformer primary current swing up and down, the iron must be "biased" to one half of its saturation magnetic field, and can swing only up to max and down to zero. Comparatively, the swing is one quarter of the available swing of the same amount of iron used in P-P because you only get to use one polarity of the M-field in the iron, and then only half of that because you have to do positive and negative swings inside the half you can use.

Just that requirement means you have to use at least four times as much iron in an SE transformer than you would in a P-P. But it gets worse. Relying on just the iron to hold you in the middle of the magnetic field operation is not very usable. The only good way to ensure you won't saturate is to add an air gap to give you some saturation immunity. Adding the air gap means that the magnetic field has a much bigger range of swing before saturation makes things go to pot; but it also lowers the good magnetic properties of the volume of iron that the magnetic field is in by that factor of tens of thousands for the portion of the magnetic path that the air is in. Add in air gap, you lose primary inductance, and have to make the iron core bigger again to get back the low frequency response that losing primary inductance cost you.

Overall, a good rule of thumb is that an SE OT needs to be about ten times the mass of a P-P for the same output power for normal operation.

With all that as background, enter the amp-boo-tweekers and those poor unfortunates afflicted with BUMS (Blind Urge to Mod Syndrome).

"Kewl!" they say, "All you gots to do to get one of those nifty, keen-O SE amps I read about over on TweekTilYouDie.com is to pull out one output tube and rebias the remaining on until you get full signal swing. How kewl is that?" So a tube gets pulled, bias is tweaked, and the sound is ... um, different... and distorted, but - wait - yeah, yeah, it's distorted in a good way (probably) so what I did was really good, right? "All right! Now all I have to do is post this hyper-secret mod to TweekTilYouDie.com and I'll get millions of approval markers."

It is possible to use a P-P transformer in an SE circuit. There are two ways.
1. Carefully set it up to use a bias current that puts the iron at about half it's M-field maximum. This *might* take some math and some careful measuring to do. Then run the amp at about 1/10th of the power it would otherwise do, lowering the power supply B+ voltage as needed to make this work out. (Note here that we haven't even scratched on the power considerations in the power tubes and power supply for converting from P-P class AB to class A.)
2. Yank out the signal drive to one side of the output tube pairs. Now the OT still sits at nearly zero magnetic field offset because there's the non-signal tube running an offsetting DC bias to keep things balanced magnetically. And you can start dinking with the bias because you need to up the standing DC current to the working output tube to get signal swing in both directions, not just one. That requires upping the DC bias to both tubes to keep the transformer magnetically balanced, and so you eventually work your way to a DC bias current that would otherwise halfway saturate the OT, but is balanced by the current in the inactive tube. You have to do this because of the output tube needing to be able to turn off as much as it turns on to get signal swing. So the tube is working as hard as it is in a SE setup. So is the inactive tube, which is resetting the transformer iron to near zero.

We're just about there. If you get all that of #2 done, you will notice a funny smell as the radiated heat starts melting the transformer, wooden enclosure, plastic parts, etc. That's because the power supply voltage is big enough that setting the output tubes at half of full current all the time massively overheats the tubes. It also pours more DC current, not signal current, through the OT windings designed for P-P DC offset, which is much, much smaller. So the OT starts to overheat from the DC bias conditions. To correct this, you have to turn the power supply voltage down to keep things from overheating.

After several passes of all of this, you can work your way into a situation where you can, in fact, use a P-P output trannie in an SE fashion.

... but you find out that the output power you get is about 1/10th of what you'd get with the same transformer used in P-P again. Mother Nature just insists on having her little jokes, doesn't she?

[MWJB]

Not going to argue with any of the statements posted so far, but this case can be somewhat overstated.

If you don't have a SE OT at all, but do have a PP OT on hand that will handle the plate current and will also give you an acceptable impedance match...then you may as well try it. If it truly sounds bad, then buy the SE OT that you knew you needed anyway.

Plenty of guys have recorded & gigged with PP amps running with a power tube pulled.

Fender's factory OT fix for the tweed SE Princetons, was to replace with a PP Princeton OT, CT capped off, 8ohm load. I have an amp configured like this and have had the opportunity to A/B test against SE amps up to a couple of grand in price...the 5F2A in question is still my yardstick for SE guitar amp tone.

If you're after headroom & plenty of cleans, then buy the right tool for the job...but then you wouldn't build a SE guitar amp in the first place.