Toroidal power transformer shielding

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
My3gger said:
Does anyone really hear difference between SS and tube rectification with preamps runing in class a? Could be difference because of less stiff psu from old schematics.

The load current stays the same in Class A, so a less stiff power supply shouldn't make a difference. With a choice, you may want to use tube rectification because it won't have the following problem.

One thing that does make a difference is the "ordinary" solid-state rectifiers such as 1N4004-1N4007 have a turn-off time after the voltage is reversed, so they conduct for a short time. Once they stop conducting, the voltage is at a strong negative, and the sudden drop in current causes ringing (oscillations around the 100kHz range) due to stray inductance of the transformer resonating with stray capacitance. This can be picked up capacitively/electromagnetically by the audio circuits which rectify them like the radio signal they are, and cause a slight 60Hz buzz (which may make you think you need more shielding).

There are two common fixes. One is to put a snubber across each diode and/or each secondary winding. This consists of a small capacitor (about 0.1uF) in series with a small resistor (maybe 100 ohms). These damp the oscillations. The other is to use fast-recovery rectifiers, which don't have the turn-off problem that regular rectifiers have. If you're a belt-and-suspenders type person, you could do both.
 
Through testing of a few combinations, i found faster soft recovery diodes don't induce spikes into circuit.
Schottky diodes consistently seem ok for heaters, their low Vf along LDO reg also allow getting ~6,3Vdc from 6,3V trafo.
Unbypassed ultrafast diodes never got me spikes or other diode problems into phantom supplies. Grounding is bigger issue in phantom, i see lot of manufacturers get it wrong.
So a little thought and not much expense compared to 1n400X give very clean supplies. I scoped supplies and complete circuits (tube mostly) to find this out.
 
Back
Top