Hot Transformers

[mcai7et2]

I finally got round to fitting my Audiocom Superclock into my CD player yesterday, and decided to see what I could do about a power supply at the same time.

I picked up one of those little PCB mount transformers (16VA, 2x24V outputs, encapsulated in a blue plastic case) and used half wave rectification to get the voltage down to about 18V. This is smoothed by a 4700uf cap, bypassed with a 100nf cap which feeds a pretty standard LM317T circuit (the one from the datasheet with the bypass caps and safety diodes) to give a regulated 12 volts.

I am currently testing this PSU but have noticed that the transformer gets very hot after a while. I have only used great big toroids in the past but this never happened, is it normal?

 

[martin clark]

To an extent, yes; small transformers are fairly inefficient in the sense that the magnetising current (ie losses) are fairly high compared with the rated current they can deliver. 2 or 3 watts in such a small volume makes things hot quickly.

However, I'd look again at the half-wave rectification, because this is adding to your problem. Since the caps are charging only 50 Hz (instead of at 100Hz with a bridge/full wave rectification) the charging current peaks are twice as great, and therefore the copper losses four times what full wave would incur. Result, one hotter transformer.

I'd suggest you try connecting the two windings in series and using two diodes to give fullwave rectification. The output would be roughly 1.41 x 24v, plus up to 20% offload (because small transformers have poor regulation) - 34 to 40v in other words.

The LM317 can usually stand at least 35v dropped across the regulator, but it'll need a small heatsink at least in order to dissipate 2 or 3W. You might want to consider using a transformer with lower output voltage to reduce waste heat...

M.

 

[mcai7et2]

The idea behind using half wave rectification was simpler implementation and lower noise of a single Schottky diode compared to a silicon bridge. Currently the second secondary is sitting on an open circuit, would it help running this in parallel with the 1st?

Also, are there any circuit diagrams around which show what you describe (fullwave rectification using 2 diodes).

Thanks,

Ed

 

[martin clark]

Small fullwave schematic attached*. You'd still only have one diode in series with the output. Connecting the two secondaries in series would provide the necessary centre tap - this is how Naim rectify power for all their gear.

You should certainly try connecting the two secondaries in parallel with your existing setup - I'd expect the temperature to drop slightly becaus eth transformer will now appear to have a slightly lower DC winding resistance.

M.

*bad edit of an image nicked from somewhere else on the wibblywobblyweb...

 

[mcai7et2]

Hmm, clever stuff. This will work very easily with the layout I have on the Veroboard. Just to confirm, running this from a 2x24v transformer will give (roughly) 1.4 x 24 volts over the cap?

Thanks,

Ed