Audio-grade rectifying diode?


Are there any out there? Thanks!

clustrocasual

Don't bother with caps across diodes. Add a 0.22ųF cap in series with a 10R resistor across the power transformer secondary. 

Post removed 

When transformer's sinewave voltage becomes smaller (at the peak) than capacitor's voltage diode becomes reverse polarized. In spite of reverse polarization it will conduct for a moment in reverse direction (limited response), to snap back to zero.  Fast diodes like Schottkys respond faster but also snap back faster creating very short current spikes.  Amplitude of the spike in not as important as time.  Extremely short spikes or fast changes (dv/dt) can couple to any, even smallest inductance in the circuit.  Soft diodes, like HEXFREDs snap back slower.  Is it audible?  I suspect it depends on the circuit.  It will likely do more harm in DAC circuit (containing clock) than in power amp, but I've never compared different diodes.

Here is article about it with drawings that will help:

https://www.allaboutcircuits.com/technical-articles/fast-ultrafast-soft-standard-schottky-selecting-the-right-rectifier/

Bypass any rectifier diode with a 0.047uF polypropylene cap and it becomes an audio grade rectifier.

This statement is false. The old saw about bypassing rectifiers really doesn't work.

The problem is that the rectifier has a capacitance in its junction and the power transformer has an inductance. Together they form an electrical resonance usually at a pretty high frequency owing to the tiny capacitance involved.

Added more capacitance does not solve the problem- it merely moves the resonance down to a lower frequency.

This might not be an issue because the peak might not go into excitation. But if it does you get 'diode noise' (which is really coming from the transformer) that leaks about readily.

The solution is to bypass the input of the diode from the power transformer with a resistor in series with a capacitance to snub the resonant peak.

If this is done correctly then the choice of diode won't have much effect on the circuit as proposed in the quote.