Rating transformers for power and temperature.

[soldar]

I have quite a few transformers in my junk pile and I am planning on using some in some projects but I am trying to figure out if there is a better way of gauging their power handling capacity.

I have some 12V transformers which I am going to use to make an unregulated PSU for a car 12V air compressor. Since the compressor will only be used for a few minutes at a time the transformer does not even have time to get too hot.

On the other hand I have some projects (correctly pronounced with short o, proh-jects) where I want the transformer to work for hours or indefinitely. I do some tests and a transformer feels cool to the touch after 10 minutes, warm after 20 minutes and hot after 30 minutes. Maybe this temperature is OK, maybe if it continues to rise it would shorten the life of the transformer.

The easy solution is to oversize the transformer so it does not get hot but that means an unnecessarily heavy PSU.

Maybe use a resettable temperature switch which would switch off in case of dangerously high temperature.

Any advice on how to best rate a transformer for indefinitely long time use?

[soldar]

I am doing some crude tests with a small transformer of which I only want to use the primary as autotransformer. It has dual 220/125 V taps and the "upper" part of the winding is thinner wire because the current would be lower in that part when connected to 220.

But when used as auto transformer that part of the winding generates most of the heat. Interesting.

[Thunderer]

Yeah, Mr. Joule is having a joke on us. This is why the "upper" part of the winding is getting hotter.

Now, first thing first, you seem to be a beginner. No problem, but play safely. You have no isolation with auto-transformers, and transformers do not care what is the neutral, nor the hot wire. So, you may transfer the hot wire directly to the output.

1. Measure with a caliper (like this one :https://en.wikipedia.org/wiki/Calipers] [url]https://en.wikipedia.org/wiki/Calipers[/url]) the diameter of the enameled copper wire.
2. Look here on the maximum ampacity of the wire https://www.farnell.com/datasheets/1942907.pdf (I took the first Google link for "enameled copper wire ampacity").

Now, you have a first idea of the rating of the winding. Going above the maximum ampacity will translate in high Joule loss = heat, and slowly in winding burn-out.
Well, even if the good wire is used, you may not have enough iron core, so it goes hotter again.

Note: some newer or cheap transformer are made to run hot, they are economically built (less turns). So, it is normal to get hot. But not safe, if not properly monitored.

[johansen]

you can reverse engineer your transformers by plugging them into my spreadsheet.

https://www.eevblog.com/forum/projects/softwarespreadsheet-for-transformer-calculation/

it includes time per degree Celsius temperature rise.

you will need to measure the volts per turn, then plug the appropriate numbers into the transformer spreadsheet, and play with the wire gauge so that the resistance calculated is the same as the resistance measured. if you do it right, you won't need to measure the wire gauge. (which can be difficult sometimes without cutting open the transformer)

[soldar]

I have loaded the transformer with about 30 W and measured input and output volts and amps , not with precision, but roughly 32 W input and 29 W output so about 3 W is wasted as heat in the transformer. About 10%. Sounds reasonable.

I measured the windings resistance at 20C and then connected the transformed with load for an hour.  During the hour the temperature rose slowly. After 15 minutes it felt very slightly warm. After 30 minutes it felt quite warm. After 45 minutes it felt hot but not too hot. After 60 minutes it felt hot to where I could hold it for a few seconds but not more.

While it was still hot I measured the resistance of the windings and it had increased by about 16%. The temperature coefficient of Cu is 0.00393 so an increase in 16% indicates an increase of 40C. 

40C + 20C = 60 C which seems to be quite safe and is compatible with my touch temperature test.

I am quite satisfied the transformer is apt for the job and I think this system of testing it is quite good and reliable.

[srb1954]

I have loaded the transformer with about 30 W and measured input and output volts and amps , not with precision, but roughly 32 W input and 29 W output so about 3 W is wasted as heat in the transformer. About 10%. Sounds reasonable.

I measured the windings resistance at 20C and then connected the transformed with load for an hour.  During the hour the temperature rose slowly. After 15 minutes it felt very slightly warm. After 30 minutes it felt quite warm. After 45 minutes it felt hot but not too hot. After 60 minutes it felt hot to where I could hold it for a few seconds but not more.

While it was still hot I measured the resistance of the windings and it had increased by about 16%. The temperature coefficient of Cu is 0.00393 so an increase in 16% indicates an increase of 40C. 

40C + 20C = 60 C which seems to be quite safe and is compatible with my touch temperature test.

Remember that calculating the temperature rise based on the increased resistance will give the AVERAGE temperature over the whole winding. It may pay to add a safety margin of, say, 10℃ to allow for hot spots within the winding.

[soldar]

Remember that calculating the temperature rise based on the increased resistance will give the AVERAGE temperature over the whole winding. It may pay to add a safety margin of, say, 10℃ to allow for hot spots within the winding.

That's a good point which I shall keep in mind but in this instance I think there is still plenty of margin.

Another thing is that when transformers get too hot they start to give off that characteristic toasty smell well before they suffer any permanent damage.

I am more concerned with the fact that I did the test with the trafo in open air at 20C but in the end it will go in an enclosure and may be at higher ambient temperature.  For anything which generates heat I prefer a metal enclosure but they are much work to make. In my opinion enclosures of all types should be cheaper to buy. Maybe I will just mount it all on a wooden base and fit a wire mesh around it. Or something. That might be another topic in another thread.

[gbaddeley]

For long term reliability at close to rated load, anything over about 50 deg C will significantly heat up the surroundings, components, case, connectors, and reduce their life. Transformers shouldn’t feel more than moderately warm to touch.

[soldar]

For long term reliability at close to rated load, anything over about 50 deg C will significantly heat up the surroundings, components, case, connectors, and reduce their life. Transformers shouldn’t feel more than moderately warm to touch.

I agree in general but each case needs to be assessed on its own. I have a crude PSU to power my 12V tiny air compressor for bicycle wheels and it only gets a few minutes of use every few weeks. The transformer does not even have time to get up to working temperature and the only electronic components are a rectifier bridge and a capacitor.

Then, like in this case, it is for a battery charger and is for intermittent use. The transformer will heat up slowly and will reach its stable working temperature where it will remain for probably short time because the battery will have charged. I will need to do some testing but it seems to me heat will not be an issue.

Still, I agree that heat kills electronics and it is best to keep any heat away from electronic circuits.