[AMPS] ? Voltage Doublers as Anode Supplies

[Scott Townley]

Heath and I'm sure many others made a bit of coin using full-wave voltage
doublers for their anode supplies.
I've been trying to "design" one myself.  I've been using the output
voltage curves in Langford-Smith (p. 1174, 4th ed.) (published in slightly
different format in any ARRL handbook; recent ones are chapter 6) to find
Vout given C, Rload, and Rsec.  Rsec is generally given as the sum of the
forward resistance of the rectifiers and the transformer itself, and in the
curves the parameter for finding Vout/Vac is Rsec/Rload.  And of course
Iload=Vout/Rload.
(I use Langford-Smith because many more values of Rsec/Rload are avaiable,
especially for the doubler case).
Findings with some questions:
Vout/Vac is *very* sensitive to Rsec/Rload in a doubler configuration,
compared to a standard full-wave bridge.  I suppose this is illustrated by
the oft-published statement "doublers have crappy static regulation".  It
almost seems unusable...under load you might be 2x Vac and with no load
2.8x Vac...yuk
	a. When a tube is spec'd "Maximum anode voltage 3kv", I assume that means
3kV under ANY conditions, right?  So building a doubler that floats no-load
up to 3.8kV or so is a no-no.
	b. If memory serves the Heath SB-200 doubler supply was 2750V no-load and
2400V under load.  How did they get that good of regulation?  Practically
as good as a full-wave bridge.
	c. Say I have a high-quality transformer (maybe only 50 ohms secondary
resistance) but I want 2x Vac,rms, not 2.8x...the curves would imply that
the path to get there is adding some resistance in the secondary line (my
case says 130-140 ohms)...is this a real and accepted practice?  Seems
silly to blow 130W with a 2.5kVa supply...?  But how else can you "dial in"
the load voltage of a doubler?  Or perhaps the clue is to intentionally
design as near to 2.8x Vac under load?
	d. As an exercise, I wanted to see the variation of load voltage with load
resistance.  So I started at a design point and found the resultant
voltages for load resistances doubling and then halving from the design
point, to make a curve of Vout vs. Rload.  It's interesting in that the
quantity Vout^2/Rload (or, kVa) moves around a LOT...increasing
significantly with low load resistances (high currents).  Obviously one
cannot draw infinite current at zero load resistance; what physical effect
am I missing?  Do these design graphs assume some effect to be zero, which
becomes dominant at low load resistances?  I don't have the original
Schrade (sp?) article in 1930-something Trans. IRE.

I'm inclined to think that Heath did it by using a transformer that many
would consider sub-standard; that is, having a fairly high secondary
resistance.  Again, though, all that excess heat to dissipate seems like a
poor design choice.  Does anyone have any particular insight along this line?

Would love to hear anyone's experiences, etc.  And,
THANKS IN ADVANCE!!!
------------
Scott Townley		
nx7u@primenet.com
------------
Collector of:
	Stoddard Aircraft EMI/RFI receivers and accessories
	Big Parts for that Big Linear Amp 
	70's era RF test equipment HP/GR/Tek
	Radio-related technical reference material 1940+
	...anything else that will keep me off the streets at night

--
FAQ on WWW:               http://www.contesting.com/ampfaq.html
Submissions:              amps@contesting.com
Administrative requests:  amps-REQUEST@contesting.com
Problems:                 owner-amps@contesting.com
Search:                   http://www.contesting.com/km9p/search.htm