High speed diodes question

I've got a question regarding high speed diodes for power supplies. Should you always stick with the same specs (in this particular case stick with a UF4004 to replace a 1N4004) or is there an advantage to increasing the specs to say a UF540X diode?

Also, what about Schottky diodes? What is the advantage to using them? What would be a suitable Schottky replacement for a 1N4004?

Thanks.

No advantage on overspec'ing diodes that I can see, assuming they were spec'd correctly to begin with. One advantage to Schottkys is low forward voltage drop, so you can use them in low voltage power supplies where there is not a lot of extra voltage available to begin with to do the ripple filtering.

No real advantage, but it shouldn't hurt anything.

I don't think I would want a situation where the lead sizes of replacement rectifiers (or any diodes) were larger diameter than the originals because the solder pads on the PCB may not like it.

Curiously Some have learned (realised) that retrofitting a Hi Zoot Ultra fast Soft recovery Diode gives some benefit, possibly even an audible one
BUT only One is needed in a bridge to achieve said claimed advantage.
Not all 4.

It is often reported by people who have made the change to 'fast' or 'soft recovery' rectifiers, that the amplifier in question sounds better in some way. However, I have not seen any test results which might explain this phenomenon.

However, this little nugget intrigues me...

bare said:

Curiously Some have learned (realised) that retrofitting a Hi Zoot Ultra fast Soft recovery Diode gives some benefit, possibly even an audible one
BUT only One is needed in a bridge to achieve said claimed advantage.
Not all 4.

Click to expand...

I would like to see a technical explanation of why this might be so... :scratch2:

What kind of power supply are you building? Is it a linear supply, or a switcher?
Fast recovery rectifiers and Schottky rectifiers reduce losses in switching power supplies. In linear supplies, their fast turn on / turn off characteristics can actually introduce a bunch of high frequency noise, especially if you are using a beefy transformer with a low resistance secondary, like a toroidal transformer.

Some people believe that these fast diodes sound better in linear power supplies. I haven't experimented with them to be able to say one way or the other. I have seen the noise they make on a scope when used in linear power supplies, though.

Tom

Schottky diodes have no reverse recovery at all, being a majority carrier device, like a vacuum tube. In this aspect they are superior to standard diodes. The low voltage drop is useful for applications like making a 6.3V DC heater supply from a 6.3V AC supply.

Silicon Schottky diodes are limited to low voltages; SiC Schottkys can handle high voltages, but they have a higher voltage drop than Si Schottkys. I use Schottkys exclusively in my amp builds, even the 2.3kV supply for my 833 amps (20 diodes in a full wave bridge) . For preamps I prefer tube rectification.

I've always believed that RF in audio circuits is bad. Any suitably rated diode will "work" in terms of building a power supply, but the question is what diodes give the cleanest output, free of unwanted RF when they switch. I usually put a small bypass of about 0.1 uF across cheap slow diodes, but you may do better with faster diodes- probably only using a bypass or snubber to kill the higher frequency noise mentioned above. An old trick (and they may not exist anymore) is to sweep the antenna of an AM transistor radio over the circuit in question. If you hear a lot of noise, check the power supply for RF. If you hear any kind of tone, check the amp for oscillation.