Powersupply problems

[mcs]

I have just built two small amp. boards using the LM3875. One of them is working fine with a 2*15V transformer. Then I got a 2*22V transformer, so I decided to test the other board with that.

I connected an oscillator to the input, an 8 Ohm resistor, a scope and a voltmeter to the output. I turned everything on, and a nice 1kHz sine wave appeared on the scope (low voltage - about 5Vpp). I then turned up the volume and everything seemed normal. But suddenly I heard a strange noise, and I think I saw some fireworks. Before I could do anything the fuse blew (which may have been what I saw). The 4 rectifier diodes were very hot, and the two 4700µF electrolytics felt warm also. When I tested the diodes one of them was shorted. The diodes in the working amp don't even get warm after the amp has been running for hours.

But what happened? The fuse was a 1.6A slow-blow type (1.6A is a lot in 230V territory), so a lot of current must have passed through the transformer. The transformer is a 160VA toroid BTW. The diodes I used are 10A 45V schottky diodes (labelled B1045) - are they too "weak"?

You can see the schematic (and the still-working board) here: http://stiftsbogtrykkeriet.dk/~mcs/LM3875/

Best regards,

Mikkel C. Simonsen

[Mr Evil]

The diodes have insufficient voltage rating. For a normal bridge rectifier, the diodes see a maximum of 2 * the peak output of the transformer, which in the case of the 22V one will be 2 * 1.41 * 22 = 62V. The diodes need to be rated for more than that.

[mcs]

Thanks for the help - I thought twice the RMS voltage was enough.

I guess the PBYR10100 should be a better choice then...

Best regards,

Mikkel C. Simonsen