Rectifier Diode Question

[Clintstone]

I'm in the process of buiding another amp and I am all out of 1N4007 diodes for my rectifier cct. What I do have is a bag of 1N5408's which are huge compared to the 1N4007's. Is there any need to use 4 of these in a full wave rectifier or is 2 going to be plenty stout? From what I understand using 2 or 3 diodes in series was just somthing they did at the time because the smaller diodes were considerably cheaper.

 

[Ten Over]

I'm in the process of buiding another amp and I am all out of 1N4007 diodes for my rectifier cct. What I do have is a bag of 1N5408's which are huge compared to the 1N4007's. Is there any need to use 4 of these in a full wave rectifier or is 2 going to be plenty stout? From what I understand using 2 or 3 diodes in series was just somthing they did at the time because the smaller diodes were considerably cheaper.

1N5408 and 1N4007 have the same voltage rating and it is the voltage that matters. For a non-bridge rectifier circuit, two 1N5408's in series would be prudent.

They used series diodes because the individual diode didn't have a sufficient voltage rating. The voltage rating adds when the diodes are in series. The labor involved with installing six diodes would have exceeded the cost of large diodes, so it is unlikely that diode cost was the determining factor.

 

[2L man]

Instead of 1N4007 I have bought UF4007 lately because they do not cost too much more. This after reading HiFi forums where builders have found ultra fast diodes produce lower voltage spikes which benefit especially SE amp builds.

There is another common ultra fast type which I have used but don't remember its type number. Its voltage and current values are higher I recall 1kV and 3A and each cost perhaps about $1. It might be UF5408 which is designed to replace 1N5408

 

[sds1]

This after reading HiFi forums where builders have found ultra fast diodes produce lower voltage spikes which benefit especially SE amp builds.

I was looking into this recently, I came away with conclusion that faster recovery was not necessary for rectification at mains frequency. I was especially interested in whether rectifier diode recovery speed was consequential on power supplies to an MCU. What did you find out?

 

[Phrygian77]

@Clintstone

Diodes have a peak inverse voltage (PIV) rating that cannot be exceed or they will fail, and dump AC into your circuit, which will destroy your power supply caps.

That PIV rating must be higher than the peak to peak voltage for a full wave (grounded CT) rectifier. You need to take the no load RMS voltage for each half of the transformer secondary and multiply it by 2.8284 to get the peak to peak voltage.

Vpp = Vrms * 2 * sqrt(2)

Since the PIV absolutely cannot be exceeded, you need a buffer to account for fluctuations in mains voltage. This is the reason why 2 to 3 diodes are used in series. The third is there as a failsafe.

 

[FenderLover]

I was looking into this recently, I came away with conclusion that faster recovery was not necessary for rectification at mains frequency. I was especially interested in whether rectifier diode recovery speed was consequential on power supplies to an MCU. What did you find out?

Merlin has scope pictures in his power supply book. The switching glitch is considerably less. Regardless what diode is used, a parallel disc cap also suppresses the glitch. The reason for suppressing the glitch is that the switching noise can get coupled through the transformer and find its way to the audio, particularly on cheaper transformers. Since it is measurable and diodes are so cheap, why not use a good one? I bought 100 UF4007's so I'm set for life.