How do you define a transformer that runs "too hot"? Any rule of thumb such as being able to touch it for x number of seconds before it hurts too much?

- Geoff

Excellent question Geoff. I look forward to an answer from one of the resident experts. Do you think we might use a meat thermometer to put a number on it?

Barry

If you don't have a thermometer to check,
My rule is,
If you can't cup it your hand and keep your hand there forever, It is too hot.
I usually just check Pin voltages and if they are high, I got to lower incoming voltage one way or another,
I shoot for Filament voltages in the 6.0 range on 6V tubes. Not 6.5, Not 6.7, But all the way down to 6.0.
Radio will be fine,
Trans temp I look for is 115 to 125,
Peter

Thanks for the tip Peter. I've always heard about a transformer being "too hot" but until now I've never heard anyone actually define what that menat. "Too hot" is too subjective. It's good to be able to get a definition for people that are new to the hobby and people like myself that have been involved for awhile but still don't know what "too hot" means.

- Geoff

There is no such thing as a temperature standard for "too hot" using the touchy feely method. Some insulation ratings will run too hot to touch.

Here is a basic breakdown and it has combined many of the older letter classifications:
http://www.wbdg.org/ccb/DOD/STC/twewg_tp5.pdf

When a restored radio is run at its era voltage that becomes the proper temperature. This is not the place to go into transformer design but suffice it to say that it kept improving as knowledge and better materials came along.

Many of my sets run so that I can hold on to the transformer laminations for 30 seconds but its not comfortable but it is correct. This is at their designed for voltage, typically 115 or 117VAC

When run at a lower AC voltage and a lower temperature I feel their life has been substantially extended.

Carl

I absolutely agree, Carl. Heat causes aging of insulating materials. The more heat the faster the aging. Better insulating materials can take more heat longer, but they all age. Stress accellerates aging, just like with us.

Transformers have a finite life because of this aging problem. Heating will eventually causes insulation failure. The goal is to stave this off as longs as possible. So, if you can keep the heating down without sacrificing adequate performance by reducing input voltage you can extend that finite life a lot.

The other thing that is sometimes forgotten is the affect of heat on the termination wires exiting the bell housing of transformers. That wire often exits through a small hole in the metal with its only protection being the insulation of the wire. The older wiring doesn't hold up well in this regard. So it is not unusual to develop shorts right at the exit points which can also lead to transformer destruction.

Curtis Eickerman

Here's a method of determining the inner core temperature of a power transformer using your ohm meter. You might find it interesting to try with and without the bucking transformer to compare temperatures.
viewtopic.php?f=10&t=126380

If you have a good idea of what the set is supposed to draw in current, you can measure it with a good old fashioned analog ammeter. If excessive current on B+, there is trouble in voltage divider capacitors (especially filters) or resistors. Nest the coupling capacitors or gassy tubes. Follow the volts. Sizzling of filament or pilot light wiring can similarly be detected. If the transformer is designed for 110 vlts, and you have 125 volts at your mains, you will have to install a bucking transformer or huge resistor to tame it as you are doing.

Measure AC secondary voltages at plates of rectifier with rectifier removed. Should be virtually the same with reference to center tap.

Wow. That is cool... or would that be hot?

Thanks, Terry. I had never seen that before.

Curtis Eickerman

Put it in the bucking transformer 120 VAC supply line. It will then protect against both radio shorts and very unlikely transformer shorts.
I would never use the series choke notion suggested above since, as a reactance, it drops voltage proportional to current - you don't want that. You want a lower voltage that is more or less constant with load up to, say, a couple of amperes. The most useful set-up is to use a filament transformer, 120 AC to 12.6 VAC, centre tapped (6.3 + 6.3) rated at 3A on the secondary. The posted sketch picture shows this.
Cheers,
Roger

The thing is Roger, when wired as a bucking transformer, the secondary winding is also in series with the load and is acting as a reactive ballast, providing a voltage reduction all by itself, regardless of whether the primary winding is energized or not. Most of the voltage reduction seems to come from the reactance of the secondary winding, not from the bucking action of the primary as you might expect.

The bucking action from the primary causes a little further reduction and, I grant you, provides somewhat more consistent output voltages relative to load. But, at least with ordinary transformers, the difference does not seem all that significant, especially over the range of currents encountered in vintage radios.

I disagree. The bucking transformer provides a low impedance voltage source that is out of phase with main AC input, so it subtracts. The very low impedance of this source, and it's minimal leakage inductance (reactance), drops next to no extra volts. Thus, once connected to the low impedance supply side (120 VAC line), there is no inductance in series with the line, so no choke effect. A low loss transformer connected to a close-to-zero impedance supply has a nominally zero output impedance. If you don't believe this draw out the equivalent circuit, add the transformer "dot" (phase) notation and compute the output.
Cheers,
Roger

These are the schnizzle, nontechnically speaking;) Made two more after I discovered how much cooler my radios run with them:)-Gearhead

It is to be noted that the results you obtain depend to a certain extent, on the specific transformer you are working with.

Small, low power transformers made for current-limited applications have higher impedance secondaries, and can behave like ballasts if their primaries are open-circuited. Larger transformers have lower impedance secondaries, and consequently less ballast effect. It is necessary to energize the primary if you want a real buck or boost action to take place. This means that the output voltage remains lower or higher by the amount of the secondary voltage, over the load range of the transformer.

One other thing I've noted in experimenting along these lines is that the inter-winding capacitance plays a much bigger role in determining the overall power factor (and hence the efficiency) of the arrangement, than one might suspect. When a two-winding transformer is hooked up with the windings opposing (subtracting), you get the maximum capacitance and a lower power factor. The power factor is better if the two windings are connected in series across the line as an autotransformer. However, when a two-winding transformer is connected as an autotransformer, the voltage regulation is not as good as a bucking transformer.

I think you might be on to something there, Chris. In the circuit below with the switch open, it's a simple reactive ballast. With the switch closed, it's a "bucking" (or boost) transformer. It's plain to see that the secondary is in the supply ciruit either way.

Roger may be speaking theoretically with ideal tranformers, or at least transformers designed for the purpose. With the inexpensive tranformers we use, the reactive component of the secondary must be considerably greater than ideal and will always be a large factor. Unless throwing the switch somehow makes the reactance of the secondary disappear. I can't see that though.

This thought experiment might help you see it: does the output of a filament transformer act as if it has a large reactive (inductive) source impedance?

Instead of this continuous squabbling over such a minor point why not just build and test? And start off with a transformer that is rated for 2X or more of the radios rated wattage.
Start off by measuring the magnetizing current.

With a radio using a single ended class A audio there will be very little current variation. OTOH when running AB or B there will be a wide variation.

Carl

Can't argue with that, Carl. It's only one connection different so it's real easy to see which way gives the kind of perfromance that might be preferred. If he puts that switch in there, he could even flip back and forth between the options.

Curtis Eickerman

If you want to see how reactive things are connect a small resistance in series with the secondary and see how the voltages add up.

RRM

You don't need the bucking transformer to be 2X the load rating, quite the reverse. A 3A, 12.6 VAC (centre tapped) filament transformer (say 38 VA) will buck (or boost) a nominal 117 VAC load up to over 300 VA (3A x 117V = 350 VA.) I use a 15 volt, 10 amp buck/boost transformer (industrial type) on my bench supply (with primary run from a variac so I can vary the buck voltage) and it will regulate to over a kilowatt load (except that my added Hammond 1:1 isolating transformer limits me to 200 VA.) I just tried it with a 300 watt hair-dryer (for a few seconds!)... there was barely a 5 volt drop from no load. Taking out the Hammond, the drop was barely 2.5 V up to 1250 watts! Most of the time we are bucking for 38 watt AA5's or 55 watt AC superhets so, at about 0.3A and 0.5A respectively, any little filament transformer will do the trick.
Another point... in the excellent diagram above, if the switch has a third position shorting the bucking transformer primary (in isolation!), the transformer secondary reactance that started all this drops to nearly zero... no buck, no loss, full supply volts to load.
Cheers.
Roger