R.G. Keen ne-2 bulb transformer test

[Charlie Wilson]

So I have been playing around with this test for shorted transformer windings using a ne-2 bulb and a low voltage power supply ( 6v lantern battery) and although it seemed pretty black and white, there may be some grey areas. I am using a Mercury Magnetics Pro Reverb power transformer(known to be good) and a 1972 Twin Reverb power transformer(unknown if it is good). Both transformers will light the bulb if it is connected across the B+ secondaries and the power supply is touched across the primaries. However, if the bulb is connected across the filaments of the Twin transformer and the power supply is touched between the primaries the bulb will not flash. It will with the Pro Reverb transformer. Not sure how to interpret this. Anyone else play around with this test? RG you out there? https://www.premierguitar.com/the-super ... mer-tester
CW

[wpaulvogel]

I just hook the transformer up to 120 volts ac through a variac and test secondary voltage output after doing simple resistance/continuity tests for verifying no shorted windings and confirming wire/winding groups. I’ve heard something about the bulb test but you’re kinda making it act like an ignition coil using dc. Using a light bulb limiter and a variac to slowly bring up the voltage is my preferred method for beginning the transformer testing.

[martin manning]

DC resistance checks are not likely to reveal shorted turns, and that's the reason for doing this test.

I've fooled around with the neon bulb test a couple of times, with "good" transformers only, and I would say that if you can get a flash in any way, there are no shorted turns anywhere, at least not that will occur at relatively low voltage. The lack of flash with the bulb connected to the heater winding in one case but not the other could be due to a difference in leakage inductance. I tried it just now on a new Trainwreck PT, and I can not see a flash with a hardware store neon tester on the 6.3V winding using a 9V battery as the DC source on the primary. I did get a strong flash with the neon tester on the HV secondary.

Another test you can do is apply power to the primary and measure no-load power consumption with a watt meter. My Trainwreck PT is showing 6.7W unloaded on a Kill-a-Watt. You'd need another Twin PT to get a really good comparison.

[xtian]

Similar experience here. I've had a neon-bulb-test rig in the toolbox for a few years. With a known good OT, I can reliably get a flash one way (primary to secondary, or opposite...I forget which) but not the other. With known BAD OT, no flash at all. Note that this is while using a low current DC source--a 9v battery!

[Charlie Wilson]

Thanks guys. Yeah, not so black and white a test. I was playing around with the Twin and Pro transformers to get a feel for what to expect. However, the main reason was to test a power transformer in a blackface Bandmaster that I have been working on for a bit and would like to use regularly. The transformer has elevated filament voltages,6.6vac at 118vac mains, and gets very hot after sitting idle for over an hour and I mean burn your hand hot. I assumed it probably had some shorted turns in the primary but it passes the bulb test. I plan on sending it to Bryan Sour to have a look at and possibly rewind.
CW

[martin manning]

...the main reason was to test a power transformer in a blackface Bandmaster that I have been working on for a bit and would like to use regularly. The transformer has elevated filament voltages,6.6vac at 118vac mains, and gets very hot after sitting idle for over an hour and I mean burn your hand hot. I assumed it probably had some shorted turns in the primary but it passes the bulb test. I plan on sending it to Bryan Sour to have a look at and possibly rewind.

I'm not sure what would make it run hot with no load if there are no shorted turns, and that will probably be an issue if it is put into service. It would be interesting to hear what your guy has to say about that.

[Guy77]

Thanks guys. Yeah, not so black and white a test. I was playing around with the Twin and Pro transformers to get a feel for what to expect. However, the main reason was to test a power transformer in a blackface Bandmaster that I have been working on for a bit and would like to use regularly. The transformer has elevated filament voltages,6.6vac at 118vac mains, and gets very hot after sitting idle for over an hour and I mean burn your hand hot. I assumed it probably had some shorted turns in the primary but it passes the bulb test. I plan on sending it to Bryan Sour to have a look at and possibly rewind.
CW

Hi Charlie, regarding your transformer running hot at idle. I experienced the same thing with the Hammond replacement Twin Reverb transformers. They run very hot to the touch while idling but I have had them running for years in several builds with no issues at all. I remember reading that the early Fender Twin Reverb transformers also ran very hot to the touch and they have stayed the test of time as well. So maybe all is fine here as well but I understand the need to be certain.

Cheers
Guy

[martin manning]

CW are you talking about a PT running hot with power applied and no load connected, or idling in an amp with heater and HT current flowing? If it heats up with no load and no shorted turns, that would have to be eddy current losses.

[R.G.]

Wow. That article was fifteen years ago. It was heavily edited for available space, as I remember.

If I remember correctly, the origin was an email from ... Wayne Horner (? iirc ?) noting the trick to me. It worked on all the transformers I had to hand and seemed theoretically sound. More importantly, to me, it worked on a single-turn short that I rigged up by snaking a thin wire through the cracks around the center leg of transformers.

It gets chancy where you load the M-field into the core through a winding of few turns and then look for flashes from high-turns windings - like loading a filament winding and looking for flashes from a high voltage winding. The physics of the situation favor high inductance windings and high currents, of course. This isn't perfect.

As you note, it's not a black-and-white test, but a useful indicator. It's pretty sure that if you run it on the high turns winding with both the battery and neon on the high turns winding, it'll be a good test.

On transformers running hot: Transformers can take more heat than you think they can. What fails is the insulation on the wires. The copper and iron will still transform properly up to the curie temperature of iron, about 770C. Most quality-manufactured transformers these days use Class E/120 or Class B/130 insulating materials, so they are still OK at temps that you would NOT keep a fingertip touching.

I used to worry about the fingertip test, but it is, like the flash test, an indicator more than a black-and-white test. Ever since I got a US$25 non-touching temperature measuring gun, I very much prefer to actually measure temps. If you ever wonder how hot something is, you owe it to yourself to get one of these. Fingertips are hard to calibrate; Gordon Liddy famously would let a cigarette (... uh, what was that??) smoulder in his palm, so people do differ in how much temperature they can stand. The real thing to do is to measure how hot the trannie is running. Manufacturing variation, especially in potting the trannie can cause variations in the outside temperature.

And it takes a long time for a transformer to heat up. The thermal time constant (time til 63.2% of final temp rise) for a 300-500W E-I transformer can be in the range of an hour, so under constant loading, it may not reach final temp for a few hours. See the neat graph here: https://www.mmc.co.jp/adv/dev/english/d ... tor08.html. Just like with Rs and Cs, you have thermal resistance and thermal capacitance.

The gold standard for measuring transformer temps is the change-of-resistance method. Copper's resistance changes by +0.393% per degree C. If you have the right equipment, you can literally use the copper as a temperature indicator. For low-resistance windings, this may need a good four-contact/Kelvin milliohmmeter, but for high resistance windings, maybe 100+ ohms, an ordinary resistance meter might work well enough. For a dead-room-temperature transformer (which might need it to be off for 8+ hours to get cold), you measure the resistance of a primary as accurately as you can. Worry about really good connections to the wires! Write that down. Then run the trannie at full load for while, maybe thirty minutes or so, and re-measure the resistance. Write that down. Over time, you'll develop numbers that can be placed on a temperature-versus-time chart and figure out not only what the thermal time constant is, but what the FINAL temperature will be at some time later. Your meter has to be able to resolve down to about 1% of the initial temp for this to work well, though.

[Charlie Wilson]

Thanks again for the responses. Martin, the transformer gets slightly warm with no loads, warm approaching hot with the filaments connected, and hot sitting idle with everything connected. I have a Pro Reverb of the same year and even with a tube rectifier and both tubes sitting right behind the transformer and it does not get nearly as hot.
The other thing is the filament voltages, 6.6vac at the original mains voltage seems too high and I can't help but think that both the issues are related. What I usually do with old Fender amps that I buy and are going to see a lot of playing use is get a Mercury power transformer, put the original in a box, and not worry about it. However, after reinstalling the original power transformer in my Pro reverb after 15 years, I realize how much I prefer those original transformers. I read about Bryan Sours and he seems like he really is into reproducing the early Schumachers the way they were originally wound and felt when the amp is played. I originally contacted him to rewind this transformer but now I think I should put it in a box and have him make me a reproduction.
CW

[R.G.]

No load temperature rise in run-of-the-mill power transformers is highly dependent on the magnetizing current, which is in turn highly dependent on how far into saturation the designer lets the iron go. If the original design had fewer primary turns, or not as much iron in the stack, the iron gets worked further into its magnetic field limits. This both makes for more heat in the iron in all loading conditions, but also causes more resistive heating in the primary copper at no load.

You mentioned original Fender transformers. They were designed for a lower line voltage than more modern ones, as the nominal line voltage used to be 110, then 112, then 115, then 120. The typical voltage at my house is 124-126. This has the same effect as pinching pennies too hard on the original design: high magnetizing losses, and higher no-load magnetizing currents. This is the basis for AC voltage bucking transformers, what I called a "Vintage Voltage Adapter".

An original or accurate reproduction of a transformer designed for a lower input voltage cannot sound like it originally did (within the range of how much, if any, a PT affects sound) unless it's fed the original voltage. The sound of an original transformer on higher mains voltage may be nice, but it's not going to be "original", subject to that "if any" thing.

[Charlie Wilson]

Thanks RG. Yeah I guess I probably should have stated that the mains voltage were I live is usually at 118vac so I’m pretty close to the 117vac that most blackface Fenders back plate say they run at. Although that brings up a question, where the mains back in the 60s ever actually 117vac or did Fender just choose a mid point between 115v and 120v as a suggested voltage. Or in other words, is there a chance my transformer was actually wound with a 115vac primary? Finally, I prefer the sound of vintage amps at their original voltage especially the tweed Fenders which to my ears sound really awful at elevated mains voltages.
CW

[martin manning]

...were the mains back in the 60s ever actually 117vac or did Fender just choose a mid point between 115v and 120v as a suggested voltage?

"After World War II the standard voltage in the U.S. became 117 V, but many areas lagged behind even into the 1960s. In 1954, the American National Standards Institute (ANSI) published C84.1 “American National Standard for Electric Power Systems and Equipment – Voltage Ratings (60 Hertz)”. This standard established 120 volt nominal system and two ranges for service voltage and utilization voltage variations." https://en.m.wikipedia.org/wiki/Mains_electricity

How 117 became standard, it doesn’t say. Only that it had been creeping up from the original 110VDC used by Edison. I didn't know that the 120V standard dates all the way back to 1954, the wide-panel Tweed era. The nominal range is given as +/-5%, so 114 to 126.