QUestion about diode mod for rectifier

[Gingertube]

Guys,
I had a bit of a think about this from 1st principles.
So where does that semiconductor switching noise come from?
You have a junction with an electric field across it - hence a junction capacitance which has a charge on it when conducting. To get the SS diode to turn off, as the voltage reverses, you get a "splat" of current in the reverse direction to discharge that junction capacitance or if you prefer to think in other terms - to sweep out the junction charge.
The tube rectifier will not allow that reverse current spike/splat and so contrary to what I said above it probably doesn't matter what sort of SS diode you use, AND that anode to anode capacitor has no purpose and should be left out.

Does this make sense?

For your possible interest only-
The difference in normal SS diodes, the ultrafast soft recovery types and schottky is the magnitude of that junction charge and hence the magnitude of the reverse current spike required to "sweep it out".
From memory ONLY (from an investigation I did 2 years ago) the junction charge values were someting like:
normal diode - 500 nano Coulombs
UF types - 100 nano Coulombs
Schottky - 20 nano Coulombs

Newbies - a Coulomb is the unit of charge, 1 Coulomb per second flow of charge = 1 Amp of current.

Cheers,
Ian

[trobbins]

The severity of the snap effect is related to the dI/dt of the current being passed by the ss diode at the time it is asked to withstand a reverse voltage. dI/dt is also the main contributor to the transient voltage developed across the secondary winding's leakage inductance (via v=L di/dt).

As such, the dI/dt can be simplistically calculated from:
dI = load current of power supply
dt = time taken for the mains AC voltage waveform to traverse from where the tube rectifier was handling exactly the load current (ie. not charging the capacitor any more), to where it is equal to the filter capacitor voltage. This time is a lot shorter for an ss diode, as the AC voltage only has to change by about 1V, whereas for a tube diode the AC voltage have to change many tens of volts (depending on tube, and lengthened somewhat by the filter capacitor voltage falling a bit during that time).

The AC waveform voltage is not linearly decreasing at that time, but could be assumed to be for simplicity, so in comparison, the dI/dt of the valve rectifier could be many tens of times lower than that experienced with a ss diode.