Transformers; identifying which turn is shoted

[Mick Bailey]

I've no trouble in identifying a transformer with a shorted turn, but the difficulty is once this is achieved, how can the affected winding be identified?

A major short between several layers of a winding can sometimes cause a notable difference in resistance, assuming that the 'good' resistance is known beforehand, or an identical transformer is to hand for comparison. I have a Victoria Double Deluxe in right now where I suspect the 5v heater winding is shorted. The problem with low-voltage windings is they read low resistance anyhow, so even a substantial short can be difficult to identify.

[nosaj]

I've no trouble in identifying a transformer with a shorted turn, but the difficulty is once this is achieved, how can the affected winding be identified?

A major short between several layers of a winding can sometimes cause a notable difference in resistance, assuming that the 'good' resistance is known beforehand, or an identical transformer is to hand for comparison. I have a Victoria Double Deluxe in right now where I suspect the 5v heater winding is shorted. The problem with low-voltage windings is they read low resistance anyhow, so even a substantial short can be difficult to identify.

Does it put out 5 volts?

[Mick Bailey]

No 5v. A fuse-blower. 100% on there being a shorted turn, though even if another winding was shorted it still wouldn't put out 5v. Just want to find out if there's a way of positively locating which winding.

Actually with this amp I'm now pretty sure that the 5v winding is affected as Mark at Victoria measured the resistance of a good tranny. He reads 0.5 Ohm, I get .01 Ohm. I'd say that's probably fried internally.

But it would be nice to know if there was a sure way of knowing which winding had gone where the resistance readings all look OK.

[J M Fahey]

Maybe, but trhat's not the point.

Suppose it's the 5 volt winding ... so what?

It will be probably close to the core, buried below 500 turns primary, 1500/2000 turns secondary.
FWIW it m ight even be below the 6.3V winding, although that should be the least of your problems.

Anyway you see it, it's rewinding meat, all you'll salvage is the core.

Not forgetting that the wires are (should be) impregnated in very tough oven curing varnish .

[Mick Bailey]

Suppose it's the 5 volt winding ... so what?

If it's a failed HT winding, well that points to either a PSU or output tube fault, or the design/material spec of the transformer. I see a lot of these in general. Nothing the owner can do to control that. In this particular case there's no other external fault.

The early DD amps have a 5v/3A winding. Some owners fit 5U4GB rectifiers, which is right on the transformer spec. This amp has been fitted with one when the owner bought it (used) - it should have a 5AR4 or 5Y3 to remain within spec. So, did 'user error' cause the fault? Maybe, maybe not.

So this prompted me to think how you would go about identifying which winding has failed, without having to do a post-mortem. Not necessarily this particular amp, but in general.

[nosaj]

If it's a failed HT winding, well that points to either a PSU or output tube fault, or the design/material spec of the transformer. I see a lot of these in general. Nothing the owner can do to control that. In this particular case there's no other external fault.

The early DD amps have a 5v/3A winding. Some owners fit 5U4GB rectifiers, which is right on the transformer spec. This amp has been fitted with one when the owner bought it (used) - it should have a 5AR4 or 5Y3 to remain within spec. So, did 'user error' cause the fault? Maybe, maybe not.

So this prompted me to think how you would go about identifying which winding has failed, without having to do a post-mortem. Not necessarily this particular amp, but in general.

So when you toe the line with transformer specs isn't that asking for trouble since there is no leeway?

[Mick Bailey]

Some transformers are conservatively rated - Hammonds for instance. I've bought lots and they'll run hard all day and every day. But running anything right on the spec then means that in a momentary overload situation (such as with a shorted rectifier or cap) then the safety margin is lessened - especially where the only fuse is on the primary side.

In addition there's a shift towards solder-through insulated wire and the greatest number of transformer failures I see are where this has been used.

In guitar-amp land we run lots of stuff beyond spec. EL34s with 500v screens, EL84s with 480v plate voltage and Matchless amps running 110% dissipation at idle. We can take a look at the history and see what can and can't be done, what's reliable and what isn't. Maybe running a different rectifier can work out and there could be dozens of people out there doing it. I know that Victoria has upped the PT spec since this amp was made. Taking a chance with a tube is one thing, but a PT is another matter.

In the end I want a repair to go out of the door and not come back, so I like to know what's failed and why, and to advise the customer accordingly. Sometimes we never know why and have to settle with just getting the amp fixed.

[R.G.]

So this prompted me to think how you would go about identifying which winding has failed, without having to do a post-mortem. Not necessarily this particular amp, but in general.

Good thought, good motivation. But it's going to be very, very difficult to do.

I'm pretty sure something could be done with research-lab grade equipment somehow, but with what's in most repair shops, you need to get out the scalpel.

The only thing I can think of is to look for differences in leakage inductance. I speculate that this might be able to figure it out.

First some background. Windings couple to the core imperfectly. Not all the magnetic field they generate goes into the core material. With more than one winding, not all the magnetic flux that couples to the core from one winding couples out of the core to the second - or third, fourth, fifth, etc. Leakage inductance is the measure of this flux not coupled from winding to winding.

For a two-winding transformer, you can measure two inductances on each winding, the inductance with all windings open, and the inductance with the other winding shorted. Shorting a winding makes it very difficult for any other winding to generate an M-field inside the core. But a winding CAN generate a field inside the leakage area, the flux that would not be inside the cor and/or coupling to the shorted winding. This is by definition the "leakage inductance" from that winding to the shorted winding.

In a multi-winding transformer, there are different leakages from one winding to other windings, as you would expect; there are different areas for the flux to leak through to not couple the other winding.

I speculate that if you measured the leakage inductance on a winding in a faulted transformer, and then selectively shorted the other windings, you should see some change in measured leakage as you shorted good windings. But the shorted winding is already internally shorted, so there should be no difference shorting and un-shorting it.

Whether the differences are big enough to be reliably measured probably depends on the nature of the short and how accurately you can measure the changes and non-changes.

Er..., I speculate...

[Enzo]

If you just want a project, or if your forensic itch just cannot be scratched any other way, go for it. But if I spend an hour tearing down a transformer to see why it failed, and that assumes I will find out, how do I charge my customer? "Here is your amp, all fixed $200. Oh plus another $100 to satisfy my curiosity."

[Mick Bailey]

More of a side-project to satisfy curiosity. The customer only gets billed for time taken to replace the transformer. And that's my trap with this game - billable time is actually quite low due to distractions and the wider interest that each job has the potential to generate.

Plus I talk too much.

[Alex R]

More of a side-project to satisfy curiosity. The customer only gets billed for time taken to replace the transformer. And that's my trap with this game - billable time is actually quite low due to distractions and the wider interest that each job has the potential to generate.

Plus I talk too much.

oh god some of my customers talk WAY too much bless them. That's what comes of putting a couch and a coffee machine in the shop I'm thinking of charging admission, or offering a range of artisan breads or something.

[J M Fahey]

I suggest you put a couch like this :

[Silvertone Jockey]

Here is an idea that might work. One I would pursue if I were in your position of needing to determine the cause of failure on a regular basis. All transformers including power types will pass audio frequencies as well. So in theory you should be able to hook up a signal generator through a suitable capacitor for isolation and scope all the windings to see which one has drastically diminished output. You could start with the suspected winding acting as the primary in which case all of the other windings would have diminished output, or use that hook up to verify your findings

I have never tried this mind you but I would give it a shot if the need ever arose YMMV

[R.G.]

Unfortunately, that won't work all that well. The problem is that windings are not really signal generators, the flux changes in the core is. The windings are only "sensors" of what's happening in the core, viewed through the distance of the leakage inductances. The windings will all have drastically diminished output.

A shorted turn anywhere in the winding window acts like an almost infinite load on the flux changes in the core. The way my motors professor would have put it is that, ignoring leakage for a moment, the volts per turn on all turns must be the same. So a shorted turn will have some milliohms of resistance, and the driving signal will force that turn to carry enough current to be limited by the driving source's impedance.

For shorted turns that are not heated to smoking and breakdown by the current, the volts per turn will be very low, probably best measured in millivolts. And for all the other turns coupled to that turn through the core, the volts per turn will be the same - millivolts.

That's only true to the extent that the other turns are coupled to the shorted turn by the core, and leakage inductance is literally a measure of how UN-coupled they are. So if a transformer winding has a large number of turns and only one or a few of them are shorted, the other turns will be carrying the current from the shorted turns, and will generate a leakage field outside the core material. Other secondaries will also share some amount of this leakage field and will show some signal, but vastly smaller than they would if the shared flux were in the main core, not the surrounding air.

This is actually a bit of expanding on the differences in measured leakage inductance I mentioned earlier. Directly measuring the leakage to all the other windings is probably simpler, although being able to measure the low signal voltages is another way of measuring the same thing.

[Mick Bailey]

oh god some of my customers talk WAY too much bless them. That's what comes of putting a couch and a coffee machine in the shop I'm thinking of charging admission, or offering a range of artisan breads or something.

Artisan breads. I like the idea. Just putting on a sourdough right now.