Unbalacing a PP output transformer?

[dshortt9]

I have read that some mfgrs - VAC, RMJ - unbalance the output transformer on purpose to increase detail. Does this work and is it then as good as a SET? How much of an imbalance works? 10 mil or so?

[unclejed613]

seems like it would add even harmonic distortion (or maybe that's what the advertisers mean by "detail")..... that's the whole reason for using matched tubes and balanced output xfmrs, to minimize distortion...... unbalanced xfmrs might sound good in a guitar amp, but for hi-fi, it's a source of distortion.

[HollowState]

Humm....I've never heard of that. Seems to me like it would only increase distortion.

Victor

[tubelab.com]

I read about this trick several years ago, and even did several experiments. The theory goes like this:

Every transformer (or electromagnetic device) has two non linear regions. One is near zero current, and the other is the saturation region. Look up "BH curves". We all know about saturation in a transformer, and the resulting ugly sounding distortion. There is another, much smaller non linear region where some of the initial energy applied to the tansformer is used up magnetizing the core without generating any current in the secondary. This leads to a "dead zone" around the zero net current region.

This effect does not come into play in a SE amp since the DC current flowing through the transformer primary biases the transformer right into the center of its linear region. The transformer in a P-P amp ideally has no net magnetizing force at idle, so the non linear region occupies the center of the important "first watt" zone. When a P-P amp is used with efficient speakers the effect can be a loss of detail and a "dull sound".

It is possible to move the "dead zone" away from the "first watt" region by applying a significant offset in bias current of the output tubes. The tubes must be biased hot enough so that the tube with the lowest current does not generate distortion. This technique works best with class A P-P amp designs. A significant DC offset requires a gap in the transformer core, which then reduces the primary inductance, requiring a much larger transformer.

My experiments used an 80 watt guitar amp transformer ( I have a lot of them) similar to the one I used in the 300Beast (30 watt P-P). I restacked the core to include a gap, which was optimized to allow the transformer to work in a SE amp.

I tested the modified transformer with the idea of building an asymetrical P-P amp for HiFi purposes, and believe that the idea has some merit for lower powered P-P amps (I was making 10 watts). I came to the conclusion that it would just be easier to build an SE amp.

These experiments quickly lead to the idea that the asymetrical P-P amp had some serious use as a guitar amp. It is possible to use two different output tubes, at two different bias currents, to build an amp that sounds clean up to a point, then transitions to nasty abruptly without needing to be extremely loud. Much more experimenting is needed here.

[dshortt9]

Thanks, that's exactly what I wanted to know. So if one wants to use PP with an efficient speaker, perhaps one with the minimum power needed for the required volume might be better than a high powered one. And SE is the better way in theory at least.

[BudP]

Or, you can use a PP OPT that does not suffer from hysteresis induced zero crossing problems. They do exsist.

Either amorphous core or clever E/I core construction, take your pick and pay the price.

Even guitar OPT's are available, that perform in this fashion, ask Andy Marshall at THD, or check out the new Gibson Amplifiers. Or, for that matter, the new Sylken amps.

Bud

[ray_moth]

Do you know if a toroidal PP OPT would be any better in this respect? I'm thinking of Plitrons.

[jlsem]

If you balance the ac signal but have unbalanced DC current in the output transformer, distortion could decrease.

John

[BudP]

Don't put unbalanced DC into a toroidal OPT unless you know for certain that it has a gap cut into the core ring.

In E/I and C core OPT's you can add DC bias for reduced distortion. Chicago Transformer, Peerless and others did this by deliberately unbalancing the DC resistance in the primary halves. It also messes with the tubes and is not good when the system is called upon for maximum performance.

Better to put a gap in the core and flatten the permeability hump that way instead. That is what DC off set and gaps do and it has a direct relationship to how much distortion you have under 400 Hz. Commercial core does not provide anything but a ferrous focusing window for the B Field event, above 400 Hz, and the increased permeability in the core, as frequency drops, just interferes with the antenna event and creates distortion.

Only problem with either scheme is that you will eat inductance, which is what gives you the reactive load match that provides low frequencies in the first place. Only free lunch is no saturation induced reminance in the core so no zero crossing distortion in PP and no permitivity defined "settling" time on the back half of a signal in a SE OPT either. No way out of inductance equals distortion either.

Bud

[jlsem]

Toroidal transformers generally have an intrinsic distributed gap which allows some leeway for DC current mismatch. Inductances in push-pull type OPT's are generally high enough that some may be sacrificed by adding a slight gap which makes the transformer more useful.

John