Output transformer interleaving?

[Steve Conner]

Hi all

I'm doing a case study for a 300W bass O.T. based around a huge C core that I salvaged from a 1kVA transformer. I'm just trying to figure out how interleaving works. I heard the leakage inductance varies as the inverse square of the number of interleaves, but how the heck do I count them?

If I have the primary in one lump sandwiched between two secondaries, how many interleaves does that count as? 2?

What about two identical bobbins on a C core, each with a secondary sandwiched between two primaries? (this is the simplest arrangement that will equalize primary DCR) Is that 3, 5 or 6?

What about the next step up: two identical bobbins each with 5 layers arranged like so:

Secondary-Primary-Secondary-Primary-Secondary

Is that 3, 5, 10, or what

In both of the above cases, the secondaries would all be identical and parallelled, and the primaries would all be identical and seriesed.

[kg]

steve, i don't know off hand and i don't have a copy of the rdh4 here at work but i will forward your message to someone i think might know the answer. if they don't come through i can look it up later today.

ken

[Gregg]

My experience with guitar OT interleaving is you don't need more than 5 layers. This is enough even more than enough. I prefer the configuration I-II-I-II-I. The opposite also works OK.

[Wakculloch]

If you have acess to the uni library, the April 1955, and Jan '56 copies of Wireless World had good articles about OT design; much of it is the same as the BB "Coil design and construction manual", although this leaves out a lot of the maths.
Serial interleaving of the primary can lead to problems, in that the outer coils are longer and have more resistance, upsetting the balance between the two halves. This led to the standard arrangement of:-

Half primary-anode 1
Half secondary
Primary-anode 2
Half secondary
Half primary-anode 1

Unfortunately have no idea about how many interleaves any of this is!

Alan.

[Gregg]

Serial interleaving of the primary can lead to problems, in that the outer coils are longer and have more resistance, upsetting the balance between the two halves.

The key is how you connect the primaries. For more info check out this page:

http://www.mif.pg.gda.pl/homepages/tom/transfor.htm

Pictures 8 and 9 show how to do it properly. This way your primary halves will be within +/- couple of Ohms range.

[Steve Conner]

Are you sure? I can see how DCR gets equalized to a first order in figure 7a. (The mean turn length of sections 1 and 5, taken together, will average out around the same as that of section 3.) But I don't see how it works in 8 and 9.

For true balance at AC and DC, I think you need something like figure 12, which is what I was thinking of doing, since I have two bobbins to fill.

Wakculloch: Strathclyde uni library have Wireless World back to 1914

[Gregg]

Are you sure?

I'm sure because I'm using it in my transformers and I have several of them wound this way. The DCR difference between two primaries halves is usually 1-3 Ohms. It works both for 50W OT (3x600 primary) and 100W OT (3x400 primary). Leakage inductance measurements show a bandwidth of up to 30kHz.
Winding an OT as shown on pic 12 is best (for Hi-Fi) but IMHO is an overkill for a guitar amp.
The best way to find out which one is OK for you is to make some experiments and listen to the results.

[Wakculloch]

The balancing of twin bobbins is beginning to sink in. Dimwittery due to evening posts and beer.
Here's another subjective musing for you.
Just done a messy boogie with a dodgy OT, so stuck in a three section, same ratio, but physically much larger part as a tester. Test passed, ordered and installed the diddy mesa five section OT and amp was returned to that slightly constricted and uptight MB sound.
LOts of iron is good.

[gmoon]

There's also some good info in the Radio Designer's Handbook, Chap 5 (PDF) on winding capacitance, mixing windings, interleaving, etc.

Probably not as well illustrated as the Polish link, but easer to follow...

[kg]

Originally Posted by gmoon

There's also some good info in the Radio Designer's Handbook, Chap 5 (PDF) on winding capacitance, mixing windings, interleaving, etc.

Probably not as well illustrated as the Polish link, but easer to follow...

holy shit, someone scanned it in!

edit: steve, check page 20 of the pdf (p218 of the book), as it shows the breakdown of interleaved sections.

[Steve Conner]

Thanks guys! I actually have a real copy of RDH4 here, but it's such a colossal book I can never remember what's in it. I looked up that table on p218 and it answers my question, sort of: it doesn't say what happens when you make two identical bobbins of a certain winding arrangement and parallel all the sections.

I guess since the two bobbins are electrically connected, but only coupled magnetically by the core, the result will be to halve the leakage inductance, hence double the value of N^2 as they show it in the table. Not quadruple it, like if you made a single bobbin with twice as many interleaves.

So I suppose two identical bobbins of what Wakculloch called the "standard arrangement" would have a N-squared factor of 32. Does this seem right?