Primay turns for SE 5W OT and other winding questions

[Phil_S]

I had a SE bad transformer that I took apart. I realize this may be a fools errand, but I think I might have some fun and it's a low cost project. I've been doing some reading and I think I understand the math for determining the number of primary turns and wire size. However, most of what I'm reading must be meant for the hi-fi market and I'm getting numbers that seem rather high. I basing this on published specs for an 18W PP OT with 3000 turns on the primary and other information that seems directed at guitar amps, but I have several different results. Is there anyone out there who can tell me if I'm on the right track and the number of primary turns you'd use if you were winding a typical SE 5W 6V6 guitar amp output transformer?

As an aside, I realize the PP 18W is really 1500 turns per half, that one EL84 is about the same as one 6V6, and that output in watts = 9. I figure by this logic about 750 turns on the primary is plenty, but I don't like it that I've skipped several steps, etc. But it is the only bit of consensual validation I've been able to find.

With apologies for how long this makes the post, here's some of the math I'm working with.

The core area is 1.8mm x 1.9mm = 3.42 sq mm = .52 sq in.

Formula for calculating primary turns per volt:
10^8 / 4.44 * A * f * BM

where:
A is core area, above
f is the frequency 82Hz for guitar low E
BM is flux denisity in gauss/sq in.; I see suggestions of 40,000 to 90,000, but don't know how to choose a number.

Q1: What to use for f?
Q2: what to use for BM?
Q3:I figure a single 6V6 is usually a 4.5W amp at 45mA on the plate. Let's round to 5W for some padding. So, 5W @ 45mA is 111V. Once I have turns per volt, multiply by 111?

Running this through the equation, I get turns per volt:
BM = 40000, turns/volt = 13.2, 13.2 * 111 = 1485 primary turns
BM = 70000, turns/volt = 7.55, 7.55 * 111 = 838 primray turns
BM = 90000, tunrs/volt = 5.87, 5.87 * 111 = 651 primary turns

For those using the metric system, BM in gauss per sq mm and core area 3.42 sq mm:
BM = 6000, 1485 primary turns
BM = 10650, 837 primray turns
BM = 13700, 650 primary turns

I don't know enough to judge what's reasonable. I'm inclined to say I need about 900 turns, but I get to 650 a lot quicker.

I've come to the conclusion that the primary needs to be 34 AWG, but I should probably use 28AWG if it fits because the small size will be hard to handle without professional equipment. I get this from a lookup chart on wire size and current capacity. I've got .045 mA, which is 35 AWG, and then allow for some padding.

I've settled on 19 AWG for the secondary; the secondary current is the primary current * turns ratio. For a 5K to 4z OT, the turns ratio is about 35. So .045*35 = 1.6A. Using 700cm/A, that is what the lookup chart says is right. How much of a misjudgment is it to use 20 AWG; maybe I wind it for 8z and 16z only, or I just don't worry too much about making a 4Z tap and pushing it a bit? Of course some of the becomes easier to decide once I know the primary turns.

Am I on the right track? Any conversation about this would be appreciated. Yes, I think I'm aware, the challenge will be to get it wound neatly and make it fit in the window. I'm willing to try. What I really want at this point, is to know how much wire to buy and I'm not sure!

Thanks.

Phil

[Structo]

I too am interested in this.

You know, in the old days it wasn't uncommon for a HAM guy to wind his own transformers for his radio gear.

There definitely seems to be some magic about winding transformers as you can take two similar OT's that will sound a lot different from one another.

So I guess you start with the core material and what it's made from.
The laminations and how they are stacked.

I'd like to see how a machine wraps wire around an enclosed core.

The Englishman that owns Champ Electronics in Nottingham winds his own trannys. He has some info about it there.
http://www.chambonino.com/

[Structo]

Check out these miniature scissor jacks he uses to elevate the chassis.

Pretty cool although I like a cradle better.

[img:750:405]http://www.chambonino.com/work/selmer/selm7h.jpg[/img]

[Phil_S]

Thanks, Tom, I'll take a look.

This story is a bit embarrassing. It seems this OT that I thought had the smoke let out may have been good. Good for what, I don't know. It has 100 or 101 secondary turns. What is 10:1 good for? We need ratios of about 17, 25 and 35 for typical 5K to 4-8-16 ohm secondary.

To continue the story, there was a spark when I was testing it with low voltage and I may have reached the wrong conclusion. It seems the meter that was in use now likes to start up showing 256K ohms on the display. So much for that meter. The OT, itself was a salvage, so I am not losing any sleep over having separated some of its parts. I unwound the secondary and removed a few turns on the primary.

I've got one bit of primary wire that is 63" long and has 3.6 ohms end to end. The remaining primary on the bobbin is 642 ohms. If we assume the primary turns were 1000, I've lost 63 turns. I've got a twin to this OT and its primary is 673 ohms. I measured the bits and pieces of the primary I removed (damaged when removing the paper). They add up to about 60 turns. All this adds up pretty tightly. I'm thinking I've got 933 turns. If that is OK for a primary turns count, I just need secondary wire. At about 4" per turn, 100 turns is about 33 feet and ratios suggest I need less than that, more like 20'.

Now, if anyone can answer the not so simple question of how many primary turns, I think I can finish this off PDQ. I can wind 53 or so turns of #19 or 20 wire by hand fairly easily.

Thanks.

Phil

[VacuumVoodoo]

3.42 sq.mm? That's about as large as area of the hole in a typical eyelet on a component board.

[Phil_S]

USA resident who is metric challenged. What's a decimal place among friends? 18 mm * 19 mm = 342 sq mm or 3.42 sq cm or roughly the size of a thumb print. Thank you Aleksander!

From what I have gathered, the core area is adequate to support the OT requirements. I'd be happy to have a second (or 3rd) opinion. Of course, I have no idea what sort of material the lams are made from.

I downloaded a calculator
http://www.dissident-audio.com/OPT_da/Page.html
I am fairly certain it is for hi-fi, and though it take many of the inputs I expected, it keeps telling me I need 2363 tuns of 31 or 32 AWG for the primary and that I need 19 or 20 AWG for the secondary if I want a 4 ohm winding; 21AWG for 8 ohm; 22AWG for 16 ohm. At least I'm getting some confirmation on the secondary wire. Secondary turns, of course depend on primary turns.

I've been fooling with various information to indirectly estimate the gauge of the existing primary. I have come up with almost anything between 31AWG to 38AWG. I think, based on the number of turns in a layer, I stick with my earlier estimate. The best match I see in the paper I dowloaded (auth: Robert G. Wolpert) suggests I've got EI-75 lams, given mean length of one turn is 4.2", turns per layer 34 AWG = 132, 33AWG = 118. I think I took off half a layer, estimated about 60 turns. I don't have any sort of instrument to properly measure the wire gauge, so I've got to guess.

I have not been doing all the homework I can. Sitting here typing, I realized that I've got a twin to the OT. I just pushed 6.5vac into the secondary, got 222.5vac out the primary. If there are 100 turns on the secondary, then there are 3423 turns on the primary. Confirming in the other direction with 6.7vac on the secondary, 0.212 comes out the primary and that's 3160 turns. Given the very low voltage out, we can assume the second measure lacks precision but reasonably confirms the primary number of turns. I guess, then, if I only lost 60 or so turns, the questions change a bit. I've got enough turns; what is the wire gauge?

Probably the table I want to use comes from the Wolpert paper. I measure 3.6z for 63". Wolpert gives a table of ohms/1000". Math says 1000" should give me 57 ohms. Wolpert says 1000' of 38AWG is 56 ohms. OK, that's very close. Let's say I've got 38AWG. The twin primary meters at about 673 ohms. 673/56 = 12; 12,000' / 3 = 4000 turns. (Divide by 3; 1 turn is about 4".) This suggests further that the electrical measure of 3400 turns is reasonable. Do you agree? And if so, then should we conclude I've to 38AWG on the primary? Can I go that small for a 5W primary? Lacking any practical experience here, what I'm reading says no, but I have this other feeling that says maybe I'm OK with it.

The lams, including two screws to keep it together, weighs 6.7 oz, so, let's call it 6.6 oz. or about 187g and it looks to be consistent with a 5W OT.

Anyway, still looking for help here. Sorry, this is getting long. I hope it is coherent. Thanks.

[VacuumVoodoo]

What about the air gap? It's a SE amp, right? Too short and the core will saturate, too long and you loose low end. It's a big difference between 0.05mm and 0.15mm. Just to make a point about scale of things.

[Phil_S]

I was saving the air gap for later, but, as long as it's been raised, we might as well go there. I am inclined to leave what was there. I'm attaching a photo. It is hard to capture. Look at the I lams -- there are two bits of cellophane like tape across them, considerably thinner than a piece of paper. I have read that typical copier paper is 0.1mm thick. http://hypertextbook.com/facts/2001/JuliaSherlis.shtml I have also seen the upsurd suggestion to use a bit of cardboard from a shoe box; that has got to be way too thick.

I've got a bit more information about the primary wire. The bobbin (primary wire + cardboard) weighs 1.1 oz. (Sorry for mixed measures.) Wolpert's table says that 38 AWG has 12887 ohms per pound and 37 AWG has 8077 ohms per pound. Let's assume 1.0 oz for the wire at 642 ohms. x16 for a pound = 10272 ohms. I don't have a precise measure for the weight of the wire alone.

Previously I said the wire matches closely to 38 AWG on the basis of ohms per 1000 inches. I found another confirmation, 38 AWG is 2.164 ohms/m. My sample of unwound bits calculates to 2.25 ohms/m. If anything, that suggests something smaller than 38AWG, but I think ohms plus weight, above, is close enough to finally confirm the gauge of the wire.

These two together seem to be the best evidence that it is wound with 38AWG. The metric equivalent is ~0.10 mm diameter or ~.008 sq mm.

I'm thinking the charts say the current capacity for 38 AWG is .13A for chassis wiring http://www.powerstream.com/Wire_Size.htm . This is well above .045A we expect to see with a 6V6 or even .054A we might see with a 6L6. Am I applying this chart information correctly?

What do you think? Get on with winding a secondary over the existing primary? (I'm not sorry to skip winding 3400 turns for the primary.)

And what about the air gap?

Thanks.

Phil