JTM 45 build 400v plate voltage, how to calculate OT primary impedance?

[J M Fahey]

2 details:
1) that way of winding transformers is **annoying** to say the least: you set up the machine for, say, 0.22mm wire (primary) .. wind 1 layer ... bring wires out ... set machine for 0.8mm wire (secondary) ... wire 1 layer ... rinse and repeat, switching wire, **adjusting advance screw and transfer wheel every time**
Maddening job.
It *might* be more practical nowadays if you have a CNC controlled winder, where advance screw is driven by a computer controlled stepper motor and adequate software, but I am certain he did that "by hand".
Even with conventional winders itīs more reasonable if you use a "stick winder", where you wind, say, 8 coils at the same time on a single full sheet of Nomex or Mylar , do the same job as I mentioned above but at least you end up with 8 identical coils and not just 1.
Then the 3 or 4 ft long "stick" is cut into 8 individual coils with a special fine toothed band saw.
But this presumes high volume Commercial production, not practical for one-offs as Mr Turner did in his day.

2) Heīs an old guy, retired since 2012 or so, heīs just selling remaining stock (which may take years), last page update was in 2016 or 2017, just *hope* heīs still among us
Donīt think he currently worries too much about market share or competition.

3) in a way, that winding method can be better hand-made, no need for expensive complicated machines when you can just hand guide wire, and let it self space, easier done than explained.
All you need is one of these:


and a little patience

In fact, I would not be surprised at all if thatīs how he wound those.

As a side note, a friend of mine, successful Tube amp maker, has a similar machine, but driven by a plain cordless screwdriver.
Heīs wound hundreds of Transformers so far.
Low RPM lets him hand guide wire with little effort.

[mikepukmel]

2 details:
1) that way of winding transformers is **annoying** to say the least: you set up the machine for, say, 0.22mm wire (primary) .. wind 1 layer ... bring wires out ... set machine for 0.8mm wire (secondary) ... wire 1 layer ... rinse and repeat, switching wire, **adjusting advance screw and transfer wheel every time**
Maddening job.
It *might* be more practical nowadays if you have a CNC controlled winder, where advance screw is driven by a computer controlled stepper motor and adequate software, but I am certain he did that "by hand".
Even with conventional winders itīs more reasonable if you use a "stick winder", where you wind, say, 8 coils at the same time on a single full sheet of Nomex or Mylar , do the same job as I mentioned above but at least you end up with 8 identical coils and not just 1.
Then the 3 or 4 ft long "stick" is cut into 8 individual coils with a special fine toothed band saw.
But this presumes high volume Commercial production, not practical for one-offs as Mr Turner did in his day.

2) Heīs an old guy, retired since 2012 or so, heīs just selling remaining stock (which may take years), last page update was in 2016 or 2017, just *hope* heīs still among us
Donīt think he currently worries too much about market share or competition.

3) in a way, that winding method can be better hand-made, no need for expensive complicated machines when you can just hand guide wire, and let it self space, easier done than explained.
All you need is one of these:


and a little patience

In fact, I would not be surprised at all if thatīs how he wound those.

As a side note, a friend of mine, successful Tube amp maker, has a similar machine, but driven by a plain cordless screwdriver.
Heīs wound hundreds of Transformers so far.
Low RPM lets him hand guide wire with little effort.

Thanks Juan, great post. Im a bit too much of a project glutton, get too many things going at once. Character flaw. But I would absolutely love to learn more about how to do this, and actually build a few output transformers. Do you ever see these hand winding machines onsale online used? Must be a fairly rare tool.

Re output transformers you mentioned winding one layer. Wow, ONE layer at a time? for a typical double ended amp (2 x 6v6 or 6L6, 8k range pri) how many layers do you wind? So, you wind each layer, one pass, then solder the ends together between the same side layers? Where do you do all the soldering? I tore apart a few dead transformers back in the day (dead ones) and don't recall seeing very many soldered connections, looked like a couple of big loops. Maybe they were crappy transformers, though.

Also, do you keep winding, then measure along the way, or only go by dc resistance of the total length of wire (i.e. theoretical resistance vs measured)?

Do you put paper between the layers, so the layers sit close together? (I can think of way too many questions that would make this post way too long).

****
Back to the original question, after I started reading various blogs, some people said the original original marshalls were 'too close' to edge of design, and sometimes burned up the output transformers if you used the wrong output impedance. Not sure if it was hearsay or fact or a mix. Also, didn't know that you could swap in, say 6.6k vs 8k for the same tube set, without something terrible happening.

[J M Fahey]

Thanks Juan, great post. Im a bit too much of a project glutton, get too many things going at once. Character flaw. But I would absolutely love to learn more about how to do this, and actually build a few output transformers. Do you ever see these hand winding machines onsale online used? Must be a fairly rare tool.

Actually they are quite common, and quite inexpensive, the simplest being a piece of iron pipe holding a shaft, with a hand wheel at one end and a simple transformer bobbin holder at the other.
Not even a counter, you count in your mind and , say, every 100 turns add a tick mark on a piece of paper, or you mod a pocket calculator as counter, search the proper yt Video.

I suggest straight driving the shaft with a cordless drill.
I have one for prototypes, basically like this one:

Re output transformers you mentioned winding one layer. Wow, ONE layer at a time?

you aalways wind a layer at a time.
Usual way is you fill one layer, apply a sheet of insulating paper (pressphan) or Mylar/kapton/nomex, then wind next layer, and so on.
You only bring out fullb section ends.
What is special in Turner is that he brings each layer ends separate, he cuts the wire leaving pigtails, he then winds another with different diameter wire, again brings pigtails out, until he finishes it.
He states it: 10 primary layers (each one may have, say, 50/80 turns, whatever fits) , interleaved with 9 secondary layers, again all brought out separated.
Allows VERY fine adjustment of impedance, both plate and speaker, and of course is EXPENSIVE.

"Normal" transformers, say Marshall, are not that detailed (unnecessary for Guitar), typically wind half primary (which may be 4 or 5 layers) , 0 to 4 ohm secondary (say, 2 or 3 layers of thicker wire), 1/4 primary, 4 to 16 ohm secondary (with an 8 ohm tap), remaining 1/4 primary.

In USA a good transformer can cost some U$60, about 1 hour bench rate, and Mail is relatively inexpensive and fast, so 99% of people just buy one.

In Argentina, Brazil, and many Countries of the World, same transformer climbs to U$120 because of freight and Customs, easily takes 30 days to arrive and be liberated, and bench rates are much lower (in U$) so it pays to have a local winder make one.
Lots of people in Europe are discovering excellent Polish winders.

for a typical double ended amp (2 x 6v6 or 6L6, 8k range pri) how many layers do you wind? So, you wind each layer, one pass, then solder the ends together between the same side layers? Where do you do all the soldering? I tore apart a few dead transformers back in the day (dead ones) and don't recall seeing very many soldered connections, looked like a couple of big loops. Maybe they were crappy transformers, though.

Not that detailed, see above.
4 or 5 (or 15) layers are wound with one continuous wire, no soldering needed if you do not cut it to begin with.

Also, do you keep winding, then measure along the way, or only go by dc resistance of the total length of wire (i.e. theoretical resistance vs measured)?

You count turns, which is the important parameter.

Do you put paper between the layers, so the layers sit close together?

Yes, always.
Only transformers which are "scramble wound" , meaning not layer by layer, nor even wire smoothly side by side simply because itīs not practical, is smallest of the lot, those in wall warts, maybe a Reverb driver , which use very thin wire, impossible to control with precision.
Similar to winding pickups and for the same reason.

Mind you, I often see relay coils, microphone transformers, etc. , smoothly wound with very fine wire, but I guess those are VERY expensive machines.

Back to the original question, after I started reading various blogs, some people said the original original marshalls were 'too close' to edge of design, and sometimes burned up the output transformers if you used the wrong output impedance. Not sure if it was hearsay or fact or a mix.

Not sure, such early amps are Collectorīs stuff and never reach these shores, so never saw one on my bench.

Also, didn't know that you could swap in, say 6.6k vs 8k for the same tube set, without something terrible happening.

Nothing terrible.
In fact most tube amps can tolerate up to 2:1 mismatch, meaning double or half impedance, certainly wonīt care about meager 20% (the difference between 8000 and 6600 )
In fact my generic OT for a couple 6L6 or EL34 is 5500 ohm to 4 ohm (sometimes 8 ohm)
WHY?
Because my experience is that practically *any* tube, *any* brand is happy with that load, and gives me honest 40W clean which is more than enough for Club play.
Imagine not having to *agonize* over tube brand, date, NOS vs modern, ultra careful biasing, etc.
All work unless very worn or dead.
4000 ohm plate to plate which is more of a classic load *can* give me 50W ... but amp becomes more choosy, is not happy with all, requires more precise matching, after 1 or 2 years tubes start showing their age, etc.
All for 10W extra.

A Forum Member, kmz6x , lives in Russia for over 10 or 15 years now.
he used to have the largest Service Shop in the San Francisco area, was *the* Factory service for lots of brands , but got fed up with a lot of things and moved to St Petersburg, swearing he would never plug in a soldering iron again.
But you know, this is a virus we all carry inside, and after some time he started again, as a hobby (his main Business now is Tourism, big time) .

He was *amazed* at getting at his new tiny Lab-in-an-appartment-room amps for servicing, most Tube type of course, from Artists such as "Uncle Ivan and his amazing accordion" , the King of Weddings, Birthday parties, Baptisms and any other social or Family party, working 6 nights a week, non stop, for the last 30 or 40 years ... most with the *original* tubes !!!!!!!
And still working, although at reduced output.
He couldnīt believe it, until he found all those amps (made in the 60īs and 70īs) used the tubes "by the book", exactly under those conditions "nobody uses", at least not Leo, Jim or Dick.
Hereīs what the old RCA datasheet says:

[mikepukmel]

I don't quite have a mental picture of how you'd wind a OT with one continuous piece of wire for each winding, without cutting and soldering. Can you finish the current loop, then mount the spool someplace and let that spin around while winding the next loop of the other winding? It looks like the winders are all made to run off one spool.

[J M Fahey]

Not sure about your doubt, but in any case: suppose Iīm winding a *simple* 50W Fender OT, 4000 ohm plate to plate primary, 4 ohm secondary.
This is the most basic OT design which "works".

Primary is wound with about 0.25mm diameter wire, you set wire guide advance screw for that diameter for even winding.

You put the plastic bobbin on the winder holder, start winding the first layer and end to end you will fit about 140/150 turns.
Wire beginning exits through a bobbin side hole for later connection.

You wrap around that first layer a 0.1mm sheet of Mylar or treated paper, itīs about the thickness of regular printer paper, you hold it with a piece of tape.

Then you wind the second layer, which again is wrapped in insulator paper, and so on and on, layer by layer, until you wind half primary.
We are talking, say, almost 1000 turns so you have about 6 layers of primary wire.

Until now, wire simply "continued on the top next layer", without cutting or soldering anything, itīs a continuous wire.

When you finish, you bring out the end and now cut it.
You end with 2 pigtails which are the ends of half primary winding.

You wrap two insulation layers for safety, set machine advance screw for 1mm to 1.25mm (secondary) wire , obviously you have less turns per layer, say 30 to 40, so usually 2 layers are enough.

Again double insulation thickness and you wind remaining half primary.

So you end with 6 pigtails: for 2 half primaries and a single secondary.
You burn/scrap insulation, twist and solder whatever you need and solder colour wire leads for identification.

This is a somewhat crude way of doing it, but gives you the visual image:

[mikepukmel]

Thanks for the video Juan, sorry for the stupid question, ok I think Ive got it. Start with thin ga wire. Wind a whole bunch of turns, one piece. This is wind 1. When you're done with wind1, stop, bring the ends out, tape. Start with wind2 (thicker wire). Wind a bunch of turns. When you're done with wind2, stop, bring the ends out, cut. And so on. So, each wind, i.e. a wind for a particular gauge, may have many layers, all in one piece, but when its done, you bring the ends out and cut them so you can switch gauge.

That was the part I didn't "get". I thought they had some way to interleave 2 different gauge wire, and EACH gauge was one continuous piece throughout the whole transformer, so you only had, say 2 ends for each gauge at the end. Can't do that. Will have 2 ends for each wind, each interleave.

It must, then be fairly critical or at least important, how you join the ends. I mean, the charge travels through the wire in a particular direction. If you connect the ends wrong, then instead of cooperating, you might make two wind's battle each other. (right hand rule for B field around the wire, they all need to rotate so they cooperate between winds)

Also, the wire has to be stripped a bit, and what do you do put heat shrink or some kind of shellac over the joints?

An aside, unrelated, the Australian guy who is now retired, wrote on his pages that he had originally bought metal for the cores from a company that aligned the iron molecules a certain way during manufacture, and that they don't make those plates any longer. Wondering if some high end company,like McIntosh buys stuff like that.

***
I think I found a new hobby!

[mikepukmel]

Aside, I don't like the new "like" and "dislike" buttons so much. I used to just like "like" not necessarily, because I like the post (most often I really do) but more of a "Thanks for posting" kind of thing. I have to check 3 times now that I didn't click "unlike" which, I don't think id ever do.

[mikepukmel]

So that's how they wind toroids!

https://www.youtube.com/watch?v=tT44CnqH15I

[J M Fahey]

Start with thin ga wire. Wind a whole bunch of turns, one piece. This is wind 1. When you're done with wind1, stop, bring the ends out, tape. Start with wind2 (thicker wire). Wind a bunch of turns. When you're done with wind2, stop, bring the ends out, cut. And so on. So, each wind, i.e. a wind for a particular gauge, may have many layers, all in one piece, but when its done, you bring the ends out and cut them so you can switch gauge.

That was the part I didn't "get". I thought they had some way to interleave 2 different gauge wire, and EACH gauge was one continuous piece throughout the whole transformer, so you only had, say 2 ends for each gauge at the end. Can't do that. Will have 2 ends for each wind, each interleave.

It must, then be fairly critical or at least important, how you join the ends. I mean, the charge travels through the wire in a particular direction. If you connect the ends wrong, then instead of cooperating, you might make two wind's battle each other. (right hand rule for B field around the wire, they all need to rotate so they cooperate between winds)

Yes, thatīs why you keep careful track on hich end is start, which end is finish, so they are connected in phase (exactly like speakers in series in a cabinet).

Also, the wire has to be stripped a bit, and what do you do put heat shrink or some kind of shellac over the joints?

Canīt trust shellac, I (and most) normally use paper tape, but heatshrink is a deluxe option.

An aside, unrelated, the Australian guy who is now retired, wrote on his pages that he had originally bought metal for the cores from a company that aligned the iron molecules a certain way during manufacture, and that they don't make those plates any longer. Wondering if some high end company,like McIntosh buys stuff like that.

Fancy way to say he used grain oriented silicon steel strips.
I forgive HIM, heīs not actually lying

You can get such strips any day of the week (hey, *I* can even being in Argentina, and within 30 minutes driving from my home ), and I guess some grain oriented C cores are still made somewhere; Audiophile Market is very small but still exists.

Hey!!! look what the cat brought!!!:
http://www.bridgeportmagnetics.com/c...2_C-Cores.html