Mystery Output transformer

[R.G.]

I didn't go down the damage path because the OP said it was loud. That indicated to me (quite possibly wrongly! ) that there was an issue identifying the windings.

The burned windings do look suspicious, all right.

Measuring inductance with a meter or measuring magentizing current (leakage??) will do it too, but you either need an L-meter or the ability to limit the current into the transformer if you count on high magnetizing current. The magnetizing current can be very high on shorted-turn mains transformers (which this is not).

I personally can't imagine testing transformers for shorts without the neon bulb tester.

[martin manning]

I'm a big fan of the neon bulb test too, for OT's anyway. I find I can short the filament winding on a known-good PT and still get a flash, whereas shorting part of the secondary on a good OT will result in no flash.

With that caveat, a very easy way to do this is with a hardware store neon line tester and a 9V battery. Clip-lead the neon tester across the OT's primary or secondary, and then tap the clips on the battery's terminals.

[R.G.]

I'm a big fan of the neon bulb test too, for OT's anyway. I find I can short the filament winding on a known-good PT and still get a flash, whereas shorting part of the secondary on a good OT will result in no flash.

Good to know. There's always some point where the volts per turn on a high voltage winding will still get a flash. The nature of transformers is that the volts-per-turn is proportional to the rate of change of the magnetic field. Every turn has the same voltage across it when the field changes, and the way you get more volts is to get more turns.

For a single shorted turn, the voltage goes to something as the M-field collapses, even if only from wire resistance. In a coupled high voltage winding, N times the voltage on one turn may be enough to flip over the neon, which is a fixed voltage.

I suspect this could be completely circumvented by always connecting the neon across the lowest voltage (or a low-voltage) winding. That prevents the N-times-a-small voltage from getting you a flash on high turns ratio transformers. A short on a single turn or even a layer in the high voltage winding will keep the low voltage winding from exceeding the trip voltage of the neon.

... and as I typed this, I came up with an other scenario for getting flashes from shorted low voltage windings when measuring on a high voltage winding. Leakage inductance, especially in a power transformer.

OTs will be specifically designed for low leakage inductance, PTs mostly ignore leakage, and may be deliberately designed for high leakage, as in the side-by-side bobbin winds.

A leakage inductance is by definition an inductance that does not couple through the core to other windings. So loading up current into a high voltage winding in a trannie with significant leakage also loads up the leakage inductance. A shorted low-voltage winding can only inhibit flux change in the iron, but can have no effect on the leakage flux, which can be enough to cause a flash on its own.

The neon bulb flash brightness does have imbedded in it the amount of energy delivered to the bulb, but the simple version of the test only looks at did it flash, not how bright and how long was the flash.

The more I think about this, the more I think that leakage may be the culprit. The answer for the simple tester is to test the lowest voltage winding that's practical. That ought to do it, I think. Leakage inductance, like primary inductance, goes up as the square of the number of turns, so a lower voltage winding dramatically reduces it.

[martin manning]

Well, if I connect the bulb across a low voltage winding (5V or 6.3V filament) on a PT I have here, I can't get a flash at all. Putting it across the primary will produce a strong flash, and shorting a filament winding reduces it to a weak flash. From this I could suggest using the primary winding when testing PT's, and looking for a strong flash to indicate a good transformer.

[R.G.]

What kind of battery are you using?

[martin manning]

9V 6LR61 "PP3" type.

[JMFahey]

RG, I'm afraid my brain can't keep up.

Let me get something straight. **If** the secondaries were wired separately, we wouldn't see continuity through all four wires, right?
I'm getting readings through all four, suggesting to me that this is a tapped OT. Am I getting the lingo right?

Did you mean, stick some AC on the primary, and check phase & volts on secondary? Did that, but didn't check for phase. (I do have a scope.)

BEFORE THAT THOUGH...

This is getting complicated; my experience tells me when that happens, to go back to principles.

I measured DCR on the secondaries again. I used shorter leads and got some results that at least showed a difference between wires. I even used a second meter, not quite as accurate, but the readings from that supported the Fluke's.

Both meters measured ohms to one decimal place, so when they danced between .1 and .2, I used an arbitrary .18.

And along with all the data I collected yesterday, I've drawn a plausible conclusion.

The big '30' turns ratio support the idea that green & brown are close to each other, and that orange & black are close to each other.

Primary impedance: I'm still scratching my head. With green tap @ 16 ohms, and brown tap @ 8 ohms, you get results close to each other. (around 1,110) Orange @ 4 ohms is way off. If it was for a 1-ohm speaker... Now I'm talking crazy.

I see it in another way.

5600 ohms looks quite right for a pair EL36/6L6 if you are not going for *maximum* power (around 50W) but are happy with, say, 30W ... which seems to be the case, so that load value seems most reasonable.

Now 1000 ohms PP looks not reasonable at all, so look at it other way: if we consider 5600 primary fine, secondaries would now be :

Brown: 1414/137*4 ohms=41 ohms

Green: 1414/74*4 ohms=76 ohms

Huh? .... who makes 40 and 75 ohms speakers?

Nobody, but who said those are straight speaker outs?

I *guess* (may be wrong of course) that that transformer was recycled/repurposed/pulled from some old PA amp, where the main outputs were *voltage* outputs (think 25/50/70/100V) to feed long lines (each slave speaker would have its own little step down transformer) ; many such amps had also a "lowly" 4 or 8 ohms output to straight feed speakers placed a few yards away.

Of course in a guitar amp priorities are reversed, the "local" speaker is the important one.

In a nutshell, and based on your measurements which I trust, I'd use just the Black+Orange taps (I guess it was already wired that way) and safely tape and ignore Brown and Green.