Baffled by extra transformer wire

[WaveyDipole]

Before I started testing I made sure that I had only a 3amp fuse in the plug otherwise it might have been a different story! My LCR did show all the windings as inductances as well as giving a resistance measurement so at least I know that it has not blown open, but are there any other tests I can do to determine whether any other damage has been done, e.g shorted windings?

PS Just noticed a pinned thread on the subject. Will perhaps try out one or two of the ideas here. Hope I'm not overly cautious about this!

[julie_m]

If you make a tuned circuit with the longest section of your primary winding and a capacitor, and excite this with a much lower-frequency square wave, you should be able to observe with an oscilloscope several cycles of ringing at each transition. Short-circuited turns will tend to damp the resonance within only a few cycles. If deliberately shorting another winding (WATCH OUT FOR HIGH VOLTAGES!) makes a big difference, the transformer probably is O.K. If it makes no or only minimal difference, the transformer may have shorted turns.

Applying too high a voltage to too short a winding will have caused the transformer core to saturate (imagine there are a finite number of those tiny little imaginary compass needles; they are spring-loaded, and potential energy is stored by altering their alignment; and once they are all aligned the same way, any surplus energy has got to go somewhere else) and suddenly begin to draw a very high current, causing it to overheat. If you were lucky, the plug fuse will have blown before the transformer got hot enough to do any permanent damage. Otherwise, the transformer will need to be rewound.

[WaveyDipole]

I will have time to have a go at this at the weekend or sometime next week. The only square wave source I have is the 1khz calibration signal from the oscilloscope. Just been working it out. Given the approximate 20H inductance of the winding when configured for 240v, a 470pf capacitor in parallel should give a resonant frequency of about 1.6khz. Is 1khz sufficiently 'much lower' to make this work?

[WaveyDipole]

Well I had a go at this, but I couldn't get the trace to 'ring' at all. I also tried capacitors of different values. I'm not sure I'm doing this correctly, but when Julie said the logest winding, does this mean when its set to 240v and all windings are in series, or does this mean the longest single winding segment?

In the meantime, I checked over the power board. Firstly, one of the fuse carriers was removed from the board and broken and I effected a repair. Secondly I checked all components including electrolytics with the LCR and all seemed fine. I worked out where all of the connections go and drew a diagram of the board for future reference. I also connected a LED with suitable dropper resistor to the 6v secondary winding to give me an indication of when the power is on. Finally, when I was sure that everything was ready to go, I connected the transformer to the PSU board and checked the input and output voltages. I got around 25.5v AC in and +-35.5 DC out which is pretty close to the expected. I maybe need to run this on 230v, but ran the transformer for 5 minutes without any load on the DC out and it remained cold.

If I had a neon I would probably try thaat test as mentioned in the pinned thread. In the meantime, any help in setting up this 'ringing' test properly would be appreciated.

[julie_m]

I meant the longest contiguous section of the winding; that would be the 120 volt section. But it would not matter if you used more than one section in series. If there were shorted turns anywhere, they would show up on test. Or maybe not; it's conceivable that if the ringing frequency was too high, it could have been swallowed up by eddy current losses.

Anyway, if the output voltages are correct when the transformer is powered and it does not become hot after running for an hour on no load, it probably is O.K. If you say the fuseholder was broken, that indicates that it blew violently, which is actually a good sign: it means the overload was mercifully brief. Creating shorted turns requires not only for the temperature to be raised to the melting point of the insulation, but also the latent heat of melting, and is only likely to happen under sustained overload. It sounds as though your fuse failed -- or succeeded, if you prefer to look at it that way -- quickly enough to prevent that sort of damage.

[WaveyDipole]

Julie, sorry if my comment caused some confusion. The fuseholder on the AC input (i.e. in the plug) is OK. It was just the fuse that blew. The broken fuse holder that I was referring to was on the seperate power recitfier board and is on the DC output. One of the legs was broken and holder had been removed from the PCB by the previous owner. It was quite straightforward to repair.

Anyway, back to the transformer, I had some time today to try the neon bulb method described here:

https://www.vintage-radio.net/forum/...ad.php?t=14317

I set up as per diagram and tried it on a known good transformer first and found the test worked worked Ok on this unit. As expected, the neon would flash when the 6v DC was disconnected. I then tried the same on the suspect transformer, testing every combination of winding starting from the 110v setting and working up to 240v. The neon flashed on each setting so assuming the test itself is consistent, this would seem to indicate that everything is OK with the transformer.

Phew!