I’ve just picked up an old NAD 6325 tape deck. The belts are stuffed, the capstan is dried out, and it’s 110V.
Belts and capstan are terribly straightforward - I’ve ordered new replacements.
The voltage I initially thought would simply be a case of changing some links on the primary side of the transformer. However that’s not the case - the transformer has just one primary winding.
Not wanting to give up, I popped the transformer off the PCB and had a look - definitely one winding. However it’s got its own bobbin, not under the secondary. Plus it looked like there was just a few hundred turns.
So I disassembled the laminations, which was relatively straightforward once I’d coaxed the first one out, then unwound the primary, counting some 837 turns of 0.3mm (0.07mm^2) enamelled wire, giving me 131mV per turn.
My thought is simply to rewind it with 1750 turn of 0.21mm (0.035mm^2) wire, giving me 230V nominal. I very well may go mad counting out 1750 turns, but I guess I can do it over the course of a few evenings.
My rationale is simply to halve the cross sectional area, then double (actually a bit more than double as 230 is more than twice 110) the number of turns.
Is that reasonable? I have a vacuum pump, so plan on putting it in a jar with some lacquer and pulling a vacuum once reassembled to ensure it doesn’t rattle.
Alas it isn’t simply a matter of replacing it with an off the shelf one, as it has several secondaries.
Belts and capstan are terribly straightforward - I’ve ordered new replacements.
The voltage I initially thought would simply be a case of changing some links on the primary side of the transformer. However that’s not the case - the transformer has just one primary winding.
Not wanting to give up, I popped the transformer off the PCB and had a look - definitely one winding. However it’s got its own bobbin, not under the secondary. Plus it looked like there was just a few hundred turns.
So I disassembled the laminations, which was relatively straightforward once I’d coaxed the first one out, then unwound the primary, counting some 837 turns of 0.3mm (0.07mm^2) enamelled wire, giving me 131mV per turn.
My thought is simply to rewind it with 1750 turn of 0.21mm (0.035mm^2) wire, giving me 230V nominal. I very well may go mad counting out 1750 turns, but I guess I can do it over the course of a few evenings.
My rationale is simply to halve the cross sectional area, then double (actually a bit more than double as 230 is more than twice 110) the number of turns.
Is that reasonable? I have a vacuum pump, so plan on putting it in a jar with some lacquer and pulling a vacuum once reassembled to ensure it doesn’t rattle.
Alas it isn’t simply a matter of replacing it with an off the shelf one, as it has several secondaries.
It is, at least from a theoretical POV, however you might encounter lowly practical problems: transformers are almost always machine-wound, and manually matching the packing factor achieved by these machines can be challenging.
It all depends on the copper window available: if there is some room left, manual winding will pose no major problem, but if it's full, you will have to be thrifty, and if the winding is "old-fashioned", ie. orderly layers, you will need to follow the same pattern, otherwise there is no chance your turn number will fit in.
Note that even if the primary was "scramble-wound", you should take some margin for your manual winding (unless you have a real winding machine available - hand-drills or even lathes are no valid substitutes)
If the original winding well utilized the winding window I'd drop at least another notch on the diameter (to 0.19mm) as for thinner wire the proportion of the available area used by insulation is higher. Taking on board Elvee's comments about hand winding vs machine winding, I'd go down another one (to 0.17mm). You halved the copper area but more than doubled the number of turns, so if the original was a tight fit you'll have no hope at all of fitting it all in.
I’m cleaning the lacquer off the laminations - I soaked them overnight in acrylic lacquer thinner, which softens the lacquer nicely. I can then remove most of the goo with a chux cloth. I tried a green scotchbrite pad for a couple of stubborn buts but it’s too aggressive and damages the oxide layer, so I’m then able to measure a low resistance across the lamination with my multimeter.
Best approach seems to be using wooden tongue depressor as a spatula to remove the stubborn bits without damaging the oxide. It’s a fair bit of work. When I reassemble it I’m going to have to be careful not to put the scotchbrite damaged laminations next to one another.
Best approach seems to be using wooden tongue depressor as a spatula to remove the stubborn bits without damaging the oxide. It’s a fair bit of work. When I reassemble it I’m going to have to be careful not to put the scotchbrite damaged laminations next to one another.
Be sure transformer core is rated for 50Hz mains.
See -
Voltage and Frequency
Section 4 - frequency.
See -
Voltage and Frequency
Section 4 - frequency.
Not quite that straight forward Suzyj. Doubling the number of winding and decreasing wire size increases both resistance and inductance, hence would affect copper losses, magnetizing current, regulation and VA.
Since you have already dismantled the transformer you have not many options available than what you suggested and hope that the transformer specs was not too tight to begin with, then there may be no problem except fitting the windings onto the bobbin.
Have fun....
Since you have already dismantled the transformer you have not many options available than what you suggested and hope that the transformer specs was not too tight to begin with, then there may be no problem except fitting the windings onto the bobbin.
Have fun....
In principle, yes, and the voltage ratings of modern enamels also allow for it, but I wouldn't choose such a solution: for all the winding length, the potential difference between the two adjacent wires will be 115V (for a 230V connection), and I don't feel comfortable with this: a temporary voltage spike of several µs could trigger arcing and lead to an unpleasant failure, ruining all of the previous work
Even with a single winding there’s still the full 230V between the first winding as it comes out from the bottom of the bobbin and the windings at the top. I was surprised to see there was no attempt to add any insulation at this point. I’ll probably put a bit of kapton over the wire down to the first winding.
I could do a dual primary by winding 875 turns, then winding another 875 turns on. I’m pretty unlikely to ever need that though.
I could do a dual primary by winding 875 turns, then winding another 875 turns on. I’m pretty unlikely to ever need that though.
Well that didn’t work so well.
I started winding by hand. I did 350 turns then put the over-temp cutout in, then wound over that. I broke the wire at about turn 550.
I didn’t buy enough wire to rip it off and start over, so I joined the wire, put some lacquer on it to insulate it, and went back to winding. Feeling a bit dispirited I opted to put it in my pedestal drill and finish it off more quickly.
Winding with the pedestal drill is very fast but very messy. I gave up counting and just filled the bobbin, reasoning that I’d test it and remove wire to get to the right secondary voltage.
Reassembly was straightforward - I used a bench vice to squish it for the last lamination.
So then I took a deep breath and applied power. I got outputs of +/-12 and 19V, which was higher than it’s supposed to be (around +/-9 and 15). I figured I should really load it up a tad before continuing to account for droop under load.
So after running it on the cassette deck for a few minutes, the primary over-temp popped and that was it. I presume my dodgy broken wire resulted in a shorted turn, which overheated the primary.
Anyway it’s way too fiddly for my tastes. I think I’ll bin it and replace it with a 9V dual 15VA toroidal, and a 15V 15VA toroidal. They should fit nicely.
I started winding by hand. I did 350 turns then put the over-temp cutout in, then wound over that. I broke the wire at about turn 550.
I didn’t buy enough wire to rip it off and start over, so I joined the wire, put some lacquer on it to insulate it, and went back to winding. Feeling a bit dispirited I opted to put it in my pedestal drill and finish it off more quickly.
Winding with the pedestal drill is very fast but very messy. I gave up counting and just filled the bobbin, reasoning that I’d test it and remove wire to get to the right secondary voltage.
Reassembly was straightforward - I used a bench vice to squish it for the last lamination.
So then I took a deep breath and applied power. I got outputs of +/-12 and 19V, which was higher than it’s supposed to be (around +/-9 and 15). I figured I should really load it up a tad before continuing to account for droop under load.
So after running it on the cassette deck for a few minutes, the primary over-temp popped and that was it. I presume my dodgy broken wire resulted in a shorted turn, which overheated the primary.
Anyway it’s way too fiddly for my tastes. I think I’ll bin it and replace it with a 9V dual 15VA toroidal, and a 15V 15VA toroidal. They should fit nicely.
Or it could have been you had too few turns and so were running into core saturation. A quick check of the magnetizing current on first switch on might have been indicative, though of course hindsight is always 20/20
Just for the fun of it I have two microwave-oven transformers for change of windings. I will only work on the secondary windings because it is less work and much less delicate than for the high voltage carrying primary. And, I do not run into core saturation with the original primary.
The cores are huge and used to supply some 800VA.
The cores are huge and used to supply some 800VA.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.