What are the similarities and differences of push-pull and single-ended OTs?

[deci belle]

I know some people use p/p OTs in SE applications, with patents to prove it, but why wouldn't/couldn't I normally do that (without needing to copy a patented wiring circuit)?

Is it a matter of too low primary-winding impedances that it isn't generally desirable to use a p/p in a SE application?

What other considerations typically require specifically p/p or SE OTs for their design applications?

[Enzo]

A patent doesn't prevent you from duplicating a circuit. The patent is there to help protect its owner from other manufacturers building it for profit.

The similarity is that they are both transformers.

The difference is that the SE part has a gap in the iron because of the continuous DC current through it.

[loudthud]

There was a thread a couple of years ago where a single ended amp was built and it made almost no sound. After much trial and error trying to figure out what the problem was, it was discovered that a push pull transformer was used. It was saturated by the DC current and had very little output.

Link to thread: http://music-electronics-forum.com/t16311/

[deci belle]

Enzo said:

The difference is that the SE part has a gap in the iron because of the continuous DC current through it.

Gap?

[g1]

Related thread:
http://music-electronics-forum.com/t15200/

[R.G.]

The differences are profound. A push-pull transformer is designed for a zero DC current level in its primaries. That lets the internal magnetic field be pushed all the way to saturation one way and then all the way to the negative of that in the other direction, so you get an internal magnetic "swing" of 2*S if you think of "S" as the saturation flux density - all the magnetic swing the iron can do.

A single ended OT is used so that current only flows one way through the primary winding. The current through it can never go negative because it's a single supply, so the magnetic field can only go from 0 to S in one direction. To keep this from distorting, you have to "bias" the core at 1/2 the saturation flux density, S.

So the same amount of iron can support swings of +/-S (i.e. 2*S) in a P-P transformer, and 0 to S in a single ended one. You only get half the magnetic field swing in an SE transformer.

This causes a lot of side effects. You need to design the SE transformer so it can carry half the maximum current in the primary *all the time*, not alternating two halves like in a P-P. So you need bigger wire, usually. You usually have to introduce an air gap in the core, to help with it withstanding the 1/2 S magnetic offset and being stable there. The air gap reduces the inductance of the primary windings, so for the same bass response, you have to use more turns.

More turns of bigger wire means the wire won't fit, so you have to go get bigger iron laminations to get enough window space, and that often changes the gap, then turns, then windings, then iron again,

Well, you can see where this is headed. A P-P transformer has to be fairly delicately balanced at average zero magnetic level. The bias on opposing primaries is in opposite directions, so it cancels inside the iron. An SE OT is deliberately unbalanced off to half the maximum to one side, and you pay for this in bigger iron and more turns.

There are tricks you can play, using a P-P core and playing single ended on one direction of the magnetic field and offsetting the core back to zero with an equal-and-opposite DC current in the other primary to keep the AC excitation single ended but the net magnetization zero. There are a few patents on this, but it's really right out of magnetic amplifier theory from the first half of the 1900s.

[deci belle]

Thank you for mentioning the "gap", Enzo, that R.G. expanded on hugely— Yay! I'll check out the links, g1 and loudthud!!

[mcgruff]

Push-pull also requires an extra tube (or half a tube) in the role of phase inverter. This is the part which splits the signal to create the pushes and pulls.

There are some inherent advantages for push-pull in terms of noise cancellation but I'm afraid I forget the details.

[guitician]

In a PP tranny the B+ ripple is canceled in each 180 degree out of phase primary side of CT, when balanced that is. I recently used a Hammond 125A in a SE 6AQ5A amp, and it worked, but was distorted in a way that wasn't too appealing. I replaced it with a PP transformer about half the size and was pleasantly surprised at how good it sounds, that saturation thing I guess

[loudthud]

In a PP tranny the B+ ripple is canceled in each 180 degree out of phase primary side of CT, when balanced that is. I recently used a Hammond 125A in a SE 6AQ5A amp, and it worked, but was distorted in a way that wasn't too appealing. I replaced it with a PP transformer about half the size and was pleasantly surprised at how good it sounds, that saturation thing I guess

Wait a minute! You put a push pull transformer in a Single Ended amp and it sounded better?

[guitician]

Opps, that was an SE type, not PP.