Choosing an output transformer

[dsavitsk]

Some background reading http://www.jensen-transformers.com/a...rs Chapter.pdf

[BudP]

...and the divergent results from different winders for the same design, and ...

Boy and how true that is. I finally began letting the secondary turns / wire size for one layer, of 1/4 of the total required CMA (circular mils per Amp Moon) for the 4 ohm secondary wire, with CE 6500 rated narrower winding offsets from the primary (3mm), determine my basic construction layout. Doing this dropped the deviance to below 1% across a 100 pc run. Makes the OPT's a bit bigger than typical commercial size, but I am not out for large volumes, AT ALL. For DIY designs I relax even further from commercial standards and allow low DCR to drive the design layout.

Moon, I have only scratched the surface here. There are actually philosophical, musical concerns to take into account. And then there are production concerns, dielectric materials, matching coil and core permittivity, how much of a dielectric circuit to build into the coil and then we can talk about core...

Bud

[Johan Potgieter]

Originally Posted by moonbird

the ancient greeks had a wonderful word for this 'techne" (teknay) ... there is the scientific theory and technology -- and then there is the artful application of it ... making xformers seems to be a great example of this idea.

but comon you other gurus -- don't be bashful here - there is this heavy stuff -- very interesting -- but many great amp makers do their thing without using any of this stuff -- lets hear how you do your black magic too!!!

While others are compiling their comments ...

Moonbird,

As a word of relief perhaps, I would caution against overly much dedication of this art business to something that is really quite normal science. In this business it is always wise to take a sober look at the science behind a product before giving a logical thing an aura of black magic or whatever. Yes, there is quite a wide field to cover before total control is achieved. One need to know quite a bit about power tubes and NFB stability theory (basics of which is found in the analyses of a mathematician called Nyquist), but the width of knowledge should not lead to confusion with complexity. In the end all will come together quite nicely, if one starts at the beginning and progresses step by step without skipping any.

You will realise that your initial questions (if not said so then) asked questions about both transformer design and tube technology. That is why I pleaded for some reading to be posted by those having such information (mine is mainly in books); it will save lengthy explanations here. Concerning the design of a p.p. output transformer not too much remains ... in the sense that one usually have certain parameters provided, and just have to fit them into the empty spaces.

See my next post somewhat later - I must unfortunately go off now.

[TerryO]

Originally Posted by Johan Potgieter

While others are compiling their comments ...

Moonbird,

As a word of relief perhaps, I would caution against overly much dedication of this art business to something that is really quite normal science. In this business it is always wise to take a sober look at the science behind a product before giving a logical thing an aura of black magic or whatever. Yes, there is quite a wide field to cover before total control is achieved. One need to know quite a bit about power tubes and NFB stability theory (basics of which is found in the analyses of a mathematician called Nyquist), but the width of knowledge should not lead to confusion with complexity. In the end all will come together quite nicely, if one starts at the beginning and progresses step by step without skipping any.

You will realise that your initial questions (if not said so then) asked questions about both transformer design and tube technology. That is why I pleaded for some reading to be posted by those having such information (mine is mainly in books); it will save lengthy explanations here. Concerning the design of a p.p. output transformer not too much remains ... in the sense that one usually have certain parameters provided, and just have to fit them into the empty spaces.


See my next post somewhat later - I must unfortunately go off now.

Johan,
I have a friend that teaches electronics, that is about the sharpest guy I've ever met and he's a pure science guy. All electronics can be known through Applied Physics, etc., but he acknowledged that the exception is transformer design which seem to defy the normal rules and ends up depending on the experience and skill of the winder. It was like pulling teeth to get that out of him, but that was about two years ago, so maybe things have changed.

Do you have a spread sheet that will enable us to to plug in the values needed when the design has progressed to the point where one usually has "certain parameters provided, and just have to fit them into the empty spaces?"

The reason I ask is that it would be neat to show up the handful of transformer guys, like Bud, that have had us at their mercy all these years.

Best Regards,
TerryO

[moonbird]

Originally Posted by TerryO

Do you have a spread sheet that will enable us to to plug in the values needed when the design has progressed to the point where one usually has "certain parameters provided, and just have to fit them into the empty spaces?"

The reason I ask is that it would be neat to show up the transformer guys that have had us at their mercy all these years.

Thank you Terry!! Thas wha I'm talking bout!!!

[moonbird]

Originally Posted by dsavitsk

Some background reading http://www.jensen-transformers.com/a...rs Chapter.pdf

Thank you dsavitsk - a great resource.

[Johan Potgieter]

Apologise for the silence, folks. Like Schwartzenegger, I will be back.

[Tony]

hello there Bud,

my new year's resolution for 2010 is to wind myself an OPT based on your guidelines, i have the z11 cores, i have bobbins, what i do not have at this time is time.....but will get there....as always i have the power transformers made....that one is eassy...

again many thanks for your tips...

[Johan Potgieter]

To continue

Originally Posted by TerryO

.... which seem to defy the normal rules and ends up depending on the experience and skill of the winder.

Again, with respect, I hope after this that would prove not to be quite the case. Skill of the winder: Yes, but only inasmuch as proper winding is concerned, not slapdash windings-pouring-over-the-side sort of thing.

Do you have a spread sheet that will enable us to to plug in the values needed when the design has progressed to the point where one usually has "certain parameters provided, and just have to fit them into the empty spaces?

Yes, www.dissident-audio.com/OPT_da/OPT_da-322.zip

A friend of mine has used this with success. I have not tried it, having used RDH (Radio Designer's Handbook) since before there were internet and pcs. That is then also what I will refer to below, finding it quite convenient. (To me the RDH procedure allows me a kind of 'overview', where if necessary I can more easily do back-engineering if required!)

Further I have been given by others:
www.turneraudio.com.au/output-trans-pp-calc being a complete calculated procedure.

The reason I ask is that it would be neat to show up the handful of transformer guys, like Bud, that have had us at their mercy all these years.

... says the man with exemplary diplomacy, that I must put my typing finger where my mouth is!

But no - I am not here to 'show up' BudP or anybody else. Bud seems to know his stuff well, also being in the business professionally (see his profile). Only, Bud and I had two different approaches. I only noticed near the end of his first post that he was actually describing single-ended transformers (normally with air gap), which are quite specialised! With respect, I judged that it would be best to start with principles of the normal push-pull (p.p.) output transformer before advancing to specialities. It would show that reading wider about the principles of transformers would help the uninitiated to better understand, and that is included in the RDH chapter kindly supplied to the thread earlier. I would then kindly ask that those interested down-load the RDH reference first. It would save reams of narrative to be able to refer to the graphs there, etc. (One of my difficulties is that I do not know Moonbird's background, thus apologies if I seem to under-estimate.)

Just going back to calculating from the previously determined primary inductance, the common formula here is

Lp = 3,2*A*µ*N²/l*10(power 8)

with L in Henrys, A = core cross sectiona area (inch²); l = length of magnetic path (inches); N=number of primary turns (to be determined from this), etc. (RDH p.211)

In fact, everything else appear in RDH p.215 - 228. (Reading section 5 about power transformers gives good magnetic basics.)

Thus what remains after my initial contribution is the matter of leakage reactance (Lr) which determines the upper transformer limit, barring C-effects.

Most important and easy to use are the graphs on p. 218 - 219. (For myself I have made enlarged copies of these to use for each design, thus saving the book. These have thin lines; I hope the reference supplied here would print clearly. Otherwise someone with the original could enlarge the figures only and post here.)

It should appear logical that in order to have tight coupling between P and S, these should be sectionalised and interleaved. But that should occur in a certain way often ignored, even by reputable manufacturers. In fig 5.13E is a table showing that sectionalising in the right way offers substantial advantage over wrong ways. To summarise, if one 'cuts' primary and secondary windings into equal sections, and interleave in such a way that the beginning and end sections are ½Ps each, it gives a 4x improvement over the wrong way.

Example: One wants three secondaries (explained later). The above way appears under N² = 36 (top, right hand column). [And darn it, they would use the symbol 'N' here. It has nothing to do with the number of turns itself - rather it should have been F² or S² or whatever - this 'N²' is only valid for figs. 5.13E and 5.13H. Also note that we shall see that the higher N² the lower the leakage.] But to continue, it will also be noticed that for the same factor, immediately under the top 3 Ss, there is a sectionalising with all of 6 Ss - 12 P and S sections altogether, but giving the same factor because of unwise proportioning. Alternately, doing our 3 S-sectioning the wrong way, lands one under N² = 9 only.

One can now go determine Lr from the graph in fig. 5.13H. The keys are shown in figs. 5.13F and 5.13H, which should be clear. One starts with the previously determined N² at the bottom, and follows the route shown in fig. 5.13G.

And that is it!

OK, not quite; there is still the question of wire gauge and the dimension 'c' depending on the intersection isolation material, and a few practical things. But this is getting lengthy and the hour late; perhaps tomorrow.

Thanks for patience.


PS: I have often used these procedures over many years; how accurate are they? In my latest design of a 100W p.p. output transformer with 3 secondaries etc., the calculated Lr figure as above was 6,5 mHy. Two transformers measured 6,2 and 6,7 mHy respectively. I would submit that that is more than acceptably accurate.

[moonbird]

Originally Posted by Johan Potgieter

To continue
Further I have been given by others:
www.turneraudio.com.au/output-trans-pp-calc being a complete calculated procedure.

Thanks for all the time and effort kind sir - however this link seems busted?!?!

Just looking around on this site ... have not found the calculator yet but check this out ...

output-trans-theory