Output Transformer Saturation

[lowell]

ok how would I go about putting DC current through the primary... so as to not cancel per the bias current? Just through 1 side? More specifically how exactly would I go about actually doing this?

[ThermionicScott]

One easy way would be to remove one of the tubes. Not taking any responsibility for damage though.

[lowell]

I don't see how removing a tube, thus NOT having ANY current through one side of the OT is going to ADD DC current through the primary. Is this because the bias current is no longer cancelled. I guess that's your reasoning now that I think about it. However, I do not want to change output class or output operation at all for that matter. Just interested in earlier OT saturation without using a different OT.

[teemuk]

...run a DC-current through it, with the AC-signal current "riding" on top.

How would a transformer with two primaries work, one primary being fed with the audio signal the second with a (possibly) user-variable DC voltage signal?

How does this saturation show itself, btw? I never had a chance to scope it and all the scope captures I've seen are related to using saturation in aid of switching applications. e.g.

It doesn't exactly look like something you'd want to add on to spice up your signal so I wonder if there's some other kind of saturation that takes place in audio applications. Is it, say, comparable to soft limiting?

[ThermionicScott]

Well, you'll need to change *something* to get it to happen. Use a smaller transformer, bigger tubes, and/or mismatch the bias somehow...

- Scott

[wizard333]

smaller transformer

That seems the most elegant solution.

[lowell]

The reason I'm delving into this is that I want to keep my amp in it's current configuration, and ideally, have a variable saturation switch/control. I don't want to have this be a permanent solution, as in a smaller OT. If it can't be done, so be it. If it can be, better.

[Old Tele man]

Yep - but it only saturates the polarity of the signal which reinforces the DC current.

...with PP OT, each half of input (pri) signal would be riding upon its DC-bias current, so together, the two DC-currents would sum at the OT centertap.

...would assume the astute tinkerer would "align" the DC-currents with the appropriate signal polarity, because both tube signals are electrically identical, but physically inverted, and re-aligned by the OT-primary windings and summed in the OT-secondary.

[ThermionicScott]

If you were making your own transformer from scratch, you could add an extra winding that would do nothing but carry a DC voltage...

[lowell]

So one might be able to use ultralinear taps for this?

[ThermionicScott]

...

[Old Tele man]

...using the 40-43% UL tap as a location to "inject" a controlled DC-current that would be summed and collected at the B+ center-tap is possible.

...however, the two controlled DC-current sources would have to be pretty close in tolerances...and simultaneously adjustable.

...basically, it's a dual implementation of a controllable version of the SE "DC-saturation problem" resulting from gradual magnetization of the core by the DC-current(s).

[R.G.]

...with PP OT, each half of input (pri) signal would be riding upon its DC-bias current, so together, the two DC-currents would sum at the OT centertap.

...would assume the astute tinkerer would "align" the DC-currents with the appropriate signal polarity, because both tube signals are electrically identical, but physically inverted, and re-aligned by the OT-primary windings and summed in the OT-secondary.

...using the 40-43% UL tap as a location to "inject" a controlled DC-current that would be summed and collected at the B+ center-tap is possible.

...however, the two controlled DC-current sources would have to be pretty close in tolerances...and simultaneously adjustable.

...basically, it's a dual implementation of a controllable version of the SE "DC-saturation problem" resulting from gradual magnetization of the core by the DC-current(s).

This is one of those things I actually thought about for awhile before posting. The reason I thought about it was that I spent about five years one year (!?) designing DC+/-AC carrying transformer/coupled inductors. I considered, but upon thought rejected the idea that one could inject DC one way into one half-primary, the other way into the other half-primary, and still get Class AB audio through it.

Here's the problem.
1. Ferromagnetic materials (which I'll from now on abbreviate as "iron") have a defining curve of flux density (denoted as "B") and magnetomotive force (H). When you flow current through a coil around some iron, the turns (N) times the current I make an MMF (H) happen in the iron. This MMF forces some of the magnetic domains to align in the iron, "conducting" the magnetic flux. The amount of flux density B which results from MMF H is graphed as a B-H curve. The B-H curve gives you a visual representation of what happens inside the material.
2. The B-H curves for materials other than ferromagnetic ones are sloped lines. The slope of the line is the "permeability" of the material, called "mu". For non "iron" materials, this slope does not bend, wiggle, etc. for any reasonable H.
3. The B-H curves for "iron" are not lines, but outlines of fat "S" curves. Force a small H into the iron; the B goes up some. For a certain amount of B, it's more or less linear. But eventually as you increase the H, the resulting B from each added unit of H is less and less. The "iron" is saturating with flux, and simply can't carry any more in the unique aligning-magnetic-domains stuff.
4. The B-H curves saturate both to the +H/+B end and to the -H/-B end. That is, you can saturate it to magnetic North, or you can saturate it to magnetic South.
5. Here's the kicker. You can't make iron saturate at *both* +B and -B at the same time. The same piece of iron can't have two values of flux at the same time.

You *can* make two H's from two different coils and force them into the iron. The iron happily sums up the two H's, including direction. So if you have two windings (like the primaries of a PP transformer) and force in equal H's from opposite directions (either by different current directions or different winding directions; both work) then you *can* make the H add to 0 and cancel, leaving no net flux from the additive H's.

But again: you can't make the iron saturate at *two different* B values at the same time. Mother Nature is very good at keeping the "iron" from getting confused about this.

The implication of this is what I wrote: if you shove an "iron" core off toward saturation one way, and impress an AC signal on it, whether in the same coil or another one, the H's from the DC and AC components add in the iron. If the DC has the "iron" pushed near to saturation on one side, the AC polarity that helps the DC toward the closest saturation will show saturation effects, such as not coupling well to any secondary winding. The AC polarity that pushes the "iron" away from saturation will just be not saturated, and will maintain its coupling to any secondaries.

The idea that since we use two currents to add to zero in two half primaries in OTs, we must be able to use two different currents somehow to get variable saturation is seductive. But in the normal case, the two DC offsets add to a single value - that being zero if everything is set up well. And in the two other currents case, they will still add to a single value of B in the "iron".

The saturation people talk about in OTs is - I think - the soft, primarily third order distortion that happens when you nearly symmetrically drive a core into saturation. In that case, the peaks of the signal are softly compressed by the core saturation. You can't make smaller volt*time signals compress/distort that same way with DC in the core because you can't move the operation of the same piece of iron to *both* a lower B+ saturation and a higher B- saturation at the same time.

If you could, this would amount to being able to actively adjust the maximum flux density of the core material. If you discover a way to do that, I'll drive you to the patent office myself. That would be a *very* useful thing to have. Unfortunately, that seems to be buried in the quantum mechanics of unpaired electron spins, as far as we can tell.

But I didn't stop there. We played a lot with more complicated cores and magnetic structures too. One thing we had was a three-leg core which was gapped on only the two outside legs by our machine shop, who thought we were crazy. By putting primaries and secondaries on all three limbs, we were able to get DC output from either the center/transformer connection or one of the two outer legs continuously, which was previously thought to be impossible; we were getting DC through a transformer with no copper connection between primary and secondary. We did file patents on that one. It was a combination of flyback transformer and variable DC chopper formats of switching power supplies, so we immediately wanted to call it a fly-chopper.

That name got shot down.

We did magamps, pulse transformers, saturating transistor drives, and some others I can't talk about. Pretty much, we flogged every magnetic structure we could think of. One of the guys in that group was the one that put me onto asymmetrical distortion in the Crybaby inductor. Which is what you get when you push one magnetic core off center too far.

So I thought some more before posting. If I can't saturate both directions at the same time in the same piece of iron, why not two different pieces of iron? One saturates positive, the other negative, and then we combine those.. er, somehow... about then I started thinking about crossover distortion issues much worse than we get in class AB amps.

The idea of just offsetting the iron one way and getting nice symmetrical distortion is appealing - but I can't see any way it could happen.

[raiken]

The saturation people talk about in OTs is - I think - the soft, primarily third order distortion that happens when you nearly symmetrically drive a core into saturation. .

I think the saturation people talk about in OTs is generally a bunch of misinformed hype. I doubt saturation is as prevalent as people think, as true core saturation can be a very ugly thing.

On a related topic, I have played around a bit with SE OTs, injecting a DC current into them with my trusty HP6181C variable current source, while simultaneously applying an AC-coupled frequency sweep. You can watch the low frequency response drop off neatly as you increase the DC current through the primary.

[lowell]

Wow... RG you've got lots of great experience w/ this. Wish I had access to a machine shop and all that knowledge. It may take me a while to fully grasp your post.

Raiken, so it can be done in SE operation?