Someone Told Me This Wouldn't Work?

[funk1980]

You got it.
Concerning the components; the speed of the diode has to do with it's recovery time: the time it takes to go from forward voltage to reverse. When the polarity is reversed, a diode does not stop conducting right away, introducing noise and such. A fast diode will recover much faster after this switch. In the case of half wave vs. full wave, you won't really benefit from this.
It's always good to use proper low ESR caps. Very complex PSU's depend on specific values though, so 'lower is better' doesn't always go.

 

[Regenpak]

True, but if you also draw current from the positive half of the secundary voltage you introduce extra half wave pulses. And thus noise. The power is almost nothing but not zero. Maybe max one watt. In the end I don't think the power transformer gives one iota whether its bias load is more or less balanced or not. The one in my Joepie China Dynaclone doesn't complain.

 

[DF96]

I understand now. Using a halfwave means you have a series of pulses, in this case below zero, punctuated by a series of gaps where the diode doesn't conduct. In a full wave you have a continuous series of pulses without gaps. So the the halfwave depends more heavily on the smoothing caps. So to optimize a halfwave you would want a fast diode and fast, read low ESR, caps. Correct?

No, you don't understand. Half-wave and full-wave rectification can both involve big gaps if a capacitor is used. Without a cap full-wave gives very tiny gaps when the inpout voltage is below the diode threshold voltage (slight oversimplification) while half-wave gives gaps about half of the time.

Half-wave has half the ripple frequency so you need twice the size of cap to get the same ripple voltage. In this context I'm not quite sure what a 'fast diode' is - we are only talking about 50/60Hz! Cap ESR has much the same effect in either case.

You seem to be trying to gild a lily before understanding what a flower actually is. You can't balance off a cap input circuit against a resistive load on the other side, and there is no reason why you need to anyway.

 

[kevinkr]

Fast diodes are used to reduce EMI, otherwise the cap idea above is fine.

Note if you really want to balance out the current you need to build a positive version of the bias supply complete with caps and the same overall load current. (no pot) I suppose you could use it to light an LED. I would tend to doubt there would be significant benefit as even your tube rectifier is unlikely to have identical forward conduction angles and internal resistance when conducting so some imbalance is going to remain from that source.