Toroid dual 240V primary

[glarsson]

Quite an ingenious way to handle the task

A friend acquired a very expensive amplifier (cheaply) with the transformer completely toasted. Someone had twisted all secondary wires together and applied power.

[BurningTantalum]

The secondary winding is/was about 45 metres of dual 2.0mm copper.
Strawberry- Unfortunately the varnish is burned or blackened all the way through the windings. Had it just been on the surface I think your suggestion would have worked. There is a disturbance in the layers at this point, as if there was a kink in the wire when the transformer was wound, which may have been the case as it was only about 50mm from the lead-out, so maybe it has been a hot spot for all its life and failed one day when it was being thrashed in high ambient temp.
Bud- Nothing will surprise me anymore about electricity and 'the great unwashed'. The number of times I have seen a failure where the owner has started swapping wires around to 'get it working again...' This is the same process as the chainsaw/strimmer/mower that stops and won't start so the owner screws both carb mixture screws around wildly.
BT

[ArthurDent]

If you counted all the turns on all three windings so you know the amount and size of wire to rewind and if you don't have friends to run laps, you could use a method that was used in the mills in the northeast U.S. in the 1800s. To weave cloth they had 2 sets of threads that alternated up and down and a shuttle with a bobbin of thread that would be passed from side to side to make the cloth. If you can make a crude version of the shuttle that will be small enough to pass through the window of the toroid when it has all the wire on it, you might be able to wind it by yourself. Here is a photo of one of these shuttles to give you an idea. You would probably have to snug each turn tight or if you wind it too loosely you could run out of room to pass the shuttle through.

If there was some way you could test the rest of the amplifier by temporarily using external transformers before you spend a lot of time and effort to rewind the toroid, only to find that isn't the only major problem, and it really isn't worth fixing, that could save you a lot of grief.

[BurningTantalum]

Arthur- This is what I will end up doing when I have exhausted all other avenues, and it has returned to the top of my inbox.
Actually I lived for a large part of my life in a World Heritage area where these looms were invented. The designs were a guarded secret until an enterprising New Englander got a guided tour and memorised the designs until he got back to Pawtucket, or somewhere.

IanB- I chopped off the primary today and counted 460 turns (actually 920, but they are in parallel and can be ignored for voltage purposes).
The schematic showing the connections labelled 'AC27' cannot be correct as it would give 38VDC, and the reservoir capacitors for this rectified output are 35V working.
37 turns/460 turns x 240 volts =19 volts, which is 27VDC by coincidence, and more likely to have 35V capacitors, so maybe the Yamaha diagram has been mis-written.
This doesn't help with the main output as 280/460 x 240 = 146V which is 206VDC. The res capacitors are listed as 160 working voltage.
Or have I missed something? I ran a Dremel once round the toroid and each short piece of copper would be 1 turn. I divided the number by 2 as there are 2 windings, in parallel and in phase (unlike the secondaries.)
Regards, BT

[ArthurDent]

First I'd like to apologize for my ancestors copying your weaving methods. They probably thought of it as 'innovation'. 

Yes, the turns count confused me as well. I assumed the main secondary should be 75-100VAC for each winding given the 160 volt caps and the '27' winding could be 17-19 volts given the 35 volt caps.

If the '27' secondary is about 17 volts and has 37 turns that is about 2 turns per volt so the primary for 240 volts should be 480 turns, which is close to what you counted. That would make the 280 turns you counted on the main secondary 140 volts which makes sense using math but seems very high from the capacitor rating. With low audio output the load on the power supply should also be low allowing the main supply voltage rise to near peak value or around 200 volts. If the main secondary voltage was 70 volts then the number of turns on that winding should be exactly half of the 280 you counted or 140. Is it possible you counted the entire length of the windings and not just one half?

Another way to double check this is to measure the primary wire size in circular mils and the same for the main secondary which would give you a rough ratio of the current rating for the windings. If the primary was AWG 24 with a cir mil area of about 400 and the main secondary was AWG 18 with a cir mil of about 1600 then the ratio of 4 would mean the secondary might have 4X the current but would have to have 1/4 the voltage so each winding would have about the same VA rating. I've ignored the '27' winding which is low current and won't influence the calculations that much.  Other than that, I'm at a loss to explain this

[BurningTantalum]

Hi Arthur, and thanks for persisting with this!
I counted 860 of the single turns pictured below. I had already pulled off 30 turns of double wire before I got fed up and chopped the rest off with the Dremel, so I halved the 860 to 430 to give the actual turns, and added the 30 = 460. That seems valid to me. It is a mystery. A new 500 or 650 VA transformer is $200 plus shipping so I will leave this for a rainy day, and sweet talk my mechanical friend about the wonders of toroidal coil winders, and mention Joseph Arkwright and the water powered cotton loom.
Incidentally, the primary is 0.9mm diameter over the varnish; the secondary is 2.1mm over the varnish. This suggests 1/2 the voltage (very roughly) =120VAC which is more in the ballpark. My guess is that it should be 100 or 110VAC.
Thanks again for the help.
Regards, BT

[tautech]

I'll drop this here in case it's of assistance:
https://ludens.cl/Electron/trafos/trafos.html

Good luck.

[ArthurDent]

BurningTantalum - "Incidentally, the primary is 0.9mm diameter over the varnish; the secondary is 2.1mm over the varnish. This suggests 1/2 the voltage (very roughly) =120VAC which is more in the ballpark. My guess is that it should be 100 or 110VAC."

The diameters of the wire seem reasonable and close to stock wire sizes but I said "measure the primary wire size in circular mils', not the diameter. Using cir mils would give a ratio of about 5:1 which is probably a somewhat high ratio. All these calculations are wild approximations but a voltage in the 50-80 volts range would make more sense than 100-120 volts.

[SilverSolder]

You know the total redneck fix for this amp...  is to wire the remaining toroid so it powers both left and right!

This is assuming you are OK with 200W per channel rather than 450W, of course! 

[ArthurDent]

SilverSolder, I believe you'll find the open space is for the input board and the filter cap board, not for a second transformer. There is only one transformer.

[BurningTantalum]

Thanks for that Tautech, I will peruse that extensive info when I have a moment.

I have not come across 'circular mils' (When I went to school in 1961 we learned the cgs metric system, despite being in the UK,  as it was a primary school that was intended to feed into an elite grammar school.) but it appears to be an area measurement so I will have a think about that when I have a moment.
BT

[SilverSolder]

There is only one transformer.

Must be a massive transformer - 1000VA or even more?

[ArthurDent]

SilverSolder - "Must be a massive transformer - 1000VA or even more?"

Check the specs, the power consumption is 500W or 650VA. Output ratings often are wildly different from power input depending on the output power rating method used. Max output is rated from 1200 to 1400 watts.

[ArthurDent]

The cir mil listings of the two wire sizes used (.9 and 2.1) are 1288 and 6529 or about a 5:1 ratio for current capacity. The link below is to a wire table with the wire’s mm diameter in the next to last column and the cir mil numbers in the 4th column from the left.

If the 240 volt input is 650VA then the current is about 650VA/240V=2.7 amps. That would mean the main secondary current could be up to 5 times that or as high as 13.5 amps. Assuming power out is about equal to power in then 650VA/13.5A=48V. There are also transformer losses and another 17 volt secondary winding that subtract current and would make the 13.5A figure lower. The actual main secondary current rating might be closer to 10A which would make the voltage closer to 65V which would make more sense. The power supply load varies with the audio output levels as well so using a larger secondary wire size than needed would lower the coil resistance and help with the power supply stability. 
 
Another way of looking at it is that the 460 turn primary is for 240 volts input or 1.917 turns per volt. If the main secondary is 280 turns, 280turns/1.917turns per volt=146 volts which gives almost 200VDC peak and this would mean the VA rating of the secondary would be much greater than the power in.  Half of that 200VDC (which has to be wrong), is about 100VDC or 73 volt RMS which is within the ratings for the filter caps.

Both of the above are S.W.A.G. (scientific wild-ass guesses) as I have no idea what the design parameters are and I probably couldn’t follow them if given. However the main secondary voltage can’t be as high as 146 RMS or 200 Peak and 50V from the first guess to the 100V from the second guess looks somewhat believable with closer to 100V looking like the best guess. If the turns count on the main secondary were half of what was mentioned as the counted turns then you’d get reasonable numbers.

http://www.ihiconnectors.com/AWG%20wire%20sizes.htm

[BurningTantalum]

Thanks so much Arthur, you are tenacious ! The chart is useful and I have saved it to file.
I am certain that I counted the secondary turns correctly. The first 30 were bifilar, so 30 is valid. The chopped off windings I counted in 10s into a waste bin and marked off on a pad in the traditional '4 strokes and a slash' ie 50 turns. Then divided the total by 2 and added the 30.
I have a couple of 50-0-50 toroids so may lash them up as 100-0-100, plus a small transformer or PSU and check the amp works which I'm sure it does. The output transistors are OK.
What I really need is some kind soul with a working amp to measure the AC from the transformer, but I still don't know what I have done wrong.
I am surprised that the quoted output rating seems to revert back to the days of Japanese 'music power' etc ('both channels instantaneous switch-on thump peak power' as we used to call it...)
Regards, BT

[SilverSolder]

One way to get an approximate verification on the winding count might be to weigh the copper leftovers on a good postal or kitchen scale?  -  you can weigh a known number of turns as a reference, or go with an "average diameter" of the winding, to do the math on how many windings would have made sense.  This kind of thing should get you in the right ballpark, at least.

[BurningTantalum]

The secondary winding was unwound in one length. I have it wrapped on a cardboard tube and there are two ends at each end. I have paced out the rough length of it, as I had tied it to a verandah post to enable it to be wound onto the tube, and it went round and through the house. It was roughly 45 metres of kinked wire, unstretched laid on the ground. Each turn was roughly 200mm (the inner primary turns are about 175mm). This gives at least 240 turns which is not far from what I counted.
I am going to check the amp works and leave this until I find more info and can source a reasonably priced alternative (maybe a pair of smaller toroids to make x-0-x)
Thanks all.
BT

[SilverSolder]

That definitely sounds thorough!   

[fenugrec]

Whoa, someone else rewinding a P4500 toroid ?
I just recently found the service manual too, and decided to verify idle current properly and run some high power tests. Decided to check here just for fun; my unit is wired for Canadian / US 120V mains so my information may or may not be useful.

Around 2013 I rewound both secondary layers to repair mechanical damage - the toroid clamp bolt worked itself loose and the thing banged around in the case for who knows how long, severing both secondaries. Primary was luckily intact. I vaguely remember getting a quote of about 4-500$ for a new transformer, so I instead went the manual rewind route. It took me a long time to work up the motivation and the supplies to do so, but it turned out to be roughly a 2-day job IIRC.

Back then I was a bit careless when taking notes, but I wrote "33 turns sec, double wire" (implying centertap of course, not parallel) for the outer layer, measured as 0.43mm dia (with / without varnish ? no idea). I replaced that with 25ga to be safe.
Inner secondary is 159 turns, that "go around" 3 times around the toroid, so to speak (i.e. "36h around a clock"...) Dia  ~ 1.25mm, so I went with 16ga (1.29mm).
I wrote "pri, 2x 200t ??" but that is of course only a guess since I didn't touch that layer. Also I don't recall seeing a thermal fuse, which was possibly hidden under the paper wrapping of the primary.

I used a poor-man's version of a shuttle : I took a bike rim, split in one place, measured + rolled the wire onto it, removed the coil by bending inwards both ends of the rim, then folded the coiled wire into a thin oval that just barely fits through the toroid (using a pvc pipe to form a radius at both ends to prevent kinking). Rubber bands at both ends of the "shuttles" to keep them disciplined.

When powered on, at idle, I measure 135VDC on the B+ / B- bus. Filter caps are rated 160V so "it works, mate". I'm fairly confident this is within a few % of specs.

Let me know if you want other measurements or info. If I were to do it again I would get better adhesive tape for the end instead of black e-tape as I did (the clear mylar insulating tape I found was non-adhesive). Other than that, I think it's a viable method for expensive / hard to replace toroids like this with a reasonable number of turns.