Transformer heating in linear power supply

[oliverbm]

Hello,

I built for myself a small linear power supply for a project (230V input 2x22V 0.1A output). During testing, the circuit worked fine and all the voltages were what I expected.
However, the transformer is heating to high temperature, hot enough that I cannot touch it, and even after power off, stays hot for ten minutes easy.

I have disconnected the secondaries of the transformer and measured the following :
220 Ohm primary resistance
25.5 ohm for the 1st secondary resistance
23.3 ohm for the 2nd secondary resistance

When loaded with a 1Meg resistor on each secondary I noticed the same heating effect, and measured 32.6V and 32.7V for each secondary.

The transformer is a Noratel FM50/22. I was not able to find a datasheet but I sent an email to the manufacturer.

Do you have any idea what is going on or what other tests I can do ?

Thank you for your help and I wish everybody all the best for 2020.

[IanB]

If the transformer gets hot with no loads on the secondary windings it means the transformer is not designed for the primary supply it is being given (voltage or frequency).

It is not usual for a mains transformer to be described with a part number like FM50/22. It is more typical to have a primary voltage and frequency rating (like 230 V, 50 Hz), a list of secondary voltages (like 24 V - 0 - 24 V), and a total VA rating (like 48 VA) or a secondary current rating (like 1 A).

Since you have listed none of these things, and since your transformer is getting hot, I wonder if it is really a mains step-down transformer?

[Kleinstein]

Seeing some 32 V from a 22 V transformer is about normal for a low grade small transformer. The 22 V are under nominal load.
Small transformers tend to run relatively hot, even without a load. 4.4 VA is still not that small, but if a type pushed for lowest costs it could still run close to saturation and thus have high loss even under no load. Some are designed to run up to 120 C internal temperature under nominal load.

One could check the approximate internal temperature, from the rise in DC resistance with temperature. So measure the resistance cold and hot when just removed from the power. Copper resistance is about proportional to Kelvin temperature.

For the load keep in mind that a rectifier and buffer cap load draws more AC current than DC current taken out. For a small transformer the AC current may be some 20-35% higher than DC current. So it would be some 70 mA DC maximum.

[oliverbm]

@IanB : My mistake, I forgot to put that information : the diagram on the sticker specifies 230/115/0 V 50Hz 5.0 VA on the primary and 0/22V 0.114A on each secondary. So I am pretty confident that it should work, and all the wires have a different color so I feel safe saying that I did not put 230V on the 115V wire.
@Kleinstein : I hope to get the datasheet from the manufacturer and at least have some idea of expected temperatures, but it is running hot enough to damage the label...
I will try to measure the resistance when it's hot and cold  to compare. And I can also make some current measurements, maybe something will pop up.

Thank you for your help 

[tunk]

Those 22V most likely is AC, i.e. around 30Vdc fully loaded after rectification and smoothing.
Unloaded or lightly loaded it will be a bit higher.

[andy3055]

Try connecting an ammeter in series with the load to check on the current and make sure it is less than the rated 0.114 Amps.

[Mechatrommer]

my bet is you throw that transformer away and get a new one. the transformer reaches saturation even in idle, so either its not designed as intended (as IanB said) or something went wrong inside, for example, few windings are shorted inside due to burnt enamel, is it new? its probably 300ohm primary when new who knows? anyway, if you continue to use it, you'll waste power guaranteed. or make it worst until it melt by itself or trips your ECB..

[floobydust]

OP, if your transformer is a toroid make sure the mounting hardware is not making a shorted turn. It should not run that hot unless it has a shorted turn. Your regulation is 50% (22V 0.1A or 33V no load) which is terrible even for a 4VA part.

Cheap power transformers run the flux density quite high, so they will run warm especially with high line.
I measured this primary current on a 166JA12 12V 1A E-I core to see what happens. Note magnetization current is out of phase with voltage and not a sine wave. It's not all wasted as heat. Above 140VAC the transformer growls and vibrates quite bad but not saturated.

[IanB]

@IanB : My mistake, I forgot to put that information : the diagram on the sticker specifies 230/115/0 V 50Hz 5.0 VA on the primary and 0/22V 0.114A on each secondary. So I am pretty confident that it should work, and all the wires have a different color so I feel safe saying that I did not put 230V on the 115V wire.

One thing is slightly odd: if a transformer is designed for either 230 V or 115 V operation, then it would typically have two parallel primary windings. You put the primaries in series for 230 V and in parallel for 115 V (taking care of phasing to avoid fireworks).

The way you are talking about the transformer you do not mention four wires for the primary? I would be looking for two separate primary windings each with the same resistance.

If you put two such primaries in series you absolutely have to connect them the right way round or there will be a short circuit and damage will occur. Check for the correct phasing using the data sheet for the transformer or using a low voltage AC supply.

[jdragoset]

Normal convention for winding is to label the start as one and the ending at two, the second winding will be three and four, and so forth.
Since they will all be wound in the same direction, the "dot convention" will be all the odd numbers dotted.
This implies that parallel connection would be 1,3 and 2,4 connected (low voltage primary) and series would be 1,4 input and 2,3 connected.

[George Edmonds]

Hi

The problem is called magnetising current, most manufacturers run the iron core into saturation, particularly if on a 50Hz supply and the core gets hot.

George G6HIG

[IanB]

Unless it is somehow a really low quality transformer--I have never seen a mains transformer get even slightly warm when running unloaded, even little ones. Normally I expect them to sit there at about room temperature unless loaded up. Have they cut corners so much on modern transformers that people think getting hot is normal?

[Kleinstein]

For the really small transformers, like 1 VA it is pretty normal that they run hot. Some even get hotter when idle than under nominal load - but this is a specialty for the really small ones.  A 5 VA transformer normally should not get that hot. The usual no load loss should be at some 1 W. There are special types with lower loss and chances are really cheap one could be worse, especially when primarily made for 115 V 60 Hz and not tested for 50 Hz.

[oliverbm]

I have made some DC resistance measurements under no load. Hot was after roughly 1 minute of operation (under no load).
Terminal   Cold    Hot
230-0       169     288
115-0       115     261
0-22 1st    20.7    35.7
0-22 2nd    22.1   40.5
Primary AC current : 410mA
I have not found any shorts between primary and secondary (measured the DC resistance between every possible combination of wires). As you can see from the picture, every wire is in a different color and clearly specified, so I feel pretty confident that my wiring is not an issue. You can also notice that the label is damaged due to heat...

At that point I see these possibilities :
a) This is normal behaviour : it is just a little bit crap
b) This particular one has some kind of failure
c) This item is mislabelled and is not supposed to do what the label says it would
d) There is something really dumb that I did, and it is so obvious that I cannot see it
I hope to get an answer from the manufacturer for some documentation but I don't think it will be before next year. If anyone has an idea or something I could test I am still curious to know what is going on, but with only a multimeter at hand I don't think I can do much more.

[IanB]

Something seems wrong with the resistance of the 230 V winding. If the 115 V connection measures 115 ohms, then I would have expected the 230 V connection to measure something like 350 ohms. 169 ohms doesn't seem right.

The 230 V has to have twice as many turns as the 115 V, which for the same wire thickness would make it 230 ohms. But since the magnetizing current is less with more turns the wire can be thinner, so it could be 300 ohms or more. 169 ohms seems wrong, like it is shorted out or something.

[StillTrying]

Measuring the AC voltages 0-115 and 115-230 might reveal something, such as the 115 tap voltage not being in the middle of the 230V.
I think it's .

[aqibi2000]

During your experiments at some point you may have overloaded the transformer which resulted in damage to the winding enamel - thus your overheating and finding this issue now.

If you had a Q Ring Tester and the another duplicate transform you’ll find the damaged winding

[oliverbm]

Ah Ah !
I measured the AC voltage on the primary side :
230-0       232V
115-0       192V
230-115   40V
There is definitely something wrong with the winding on the 230-115 side of the primary.
I feel confident declaring this transformer dead !

Thank you all for your help, I hope I can return the favour and I wish everybody a merry Christmas and a happy new year

[Mechatrommer]

if you declared it dead and got a new replacement... why not tear it apart? but it could be a daunting task unwinding primary on that kind of transformer, but it could be something to learn. even tough the outer core looks ok, the inner core usually heated the most first and thats where usually burning started. i managed to fix a shorted audio toroidal transformer by doing this, unwind and carefully rewind and avoid short connection by adding insulation tape inside, but the outcome looks a bit ugly, but it works again. but audio (car 12V input amp) transformer is easy since not many turn in windings. the enamelled wires can also be salvaged for another purpose i've done that on CRT winding i really dont know what it was but its a square/rectangle wire winding around the CRT tube some sort of magnetizing the tube i dont know, not a CRT expert. the ferrite core is also a good part to keep ymmv.