[AMPS] Voltage Doublers

[Ian White, G3SEK]

Rich Measures wrote:
>
>>
>>Glenn McNeil wrote:
>>>
>>>A few questions about voltage doublers. I'm building a single tube 4CX250R 
>for 
>>
>>>70cm using a voltage doubler 
>>>supply. This is my first attempt at FWD supplies.
>>>
>>>I have a transformer with about 750vac secondary, 12ohms dc resistance. I'm 
>>>planning on using IN5408 diodes and 
>>>a string of electro's. 
>>>
>>In one word, DON'T. The secondary resistance of the transformer...
>
>€  He stated that sec. R was 12-ohms.  The secondary-R in  a SB-220's HV 
>transformer is 11.7-ohms and it delivers good regulation on SSB with a 
>2300w-PEP load.   
>
That's for a significantly higher-voltage transformer (1175V?), and I
seem to remember more than 10% droop at full load. 

For this application Glenn needs every last volt he can get, because
even the off-load voltage is only just enough to operate a 4CX250
efficiently at 432MHz. On-load with a poorly regulated supply, he's
likely to be in trouble.

>>combined with the effects of primary resistance and diode on-resistance,
>>which you also need to factor in - will ruin the voltage regulation. You
>>will get disappointingly low voltage at zero-signal anode current, and
>>it will droop severely down from there. 
>
>€  Taurine feculence.
>
I checked the design charts before expressing my opinion. Did you?

>>The result will be quite hard
>>limiting of RF output on speech peaks. Been there, done exactly that,
>>and sorely wished I hadn't built a complete amp around that useless
>>transformer! 
>>
>€ What was the secondary resistance of the schlock transformer?
>
That was before I learned to measure such things first, and it's long
gone.

>>It was only many years later that I really thought about the standard
>>FWD design charts, which were there in the ARRL Handbook all the time...
>>and they could have told me exactly what was going to happen.
>
>€ Amen, Ian.
>>
>>The best practical advice is "don't ever use FWD unless you KNOW the
>>transformer was specifically designed for FWD." 
>
>€  Tranformer winding resistance tells one all that one needs to know.  
>
You need both the secondary resistance and the primary resistance. The
primary resistance has to be multiplied by the turns ratio squared, and
then added to the secondary resistance. This will often double the total
effective resistance of the transformer. Then you must also add the
effective diode resistance (see below).

And even when you have all the numbers, it's unwise to give a throwaway
opinion without looking at the design charts to see what they mean.

>>Otherwise it is not
>>going to have a very low secondary and primary resistance, because they
>>aren't needed for bridge or biphase operation. 
>>
>€   Methinks Ian would do well to look at the family of curves in the 
>ARRL Handbook for full-wave bridge (FWB) rectification.  Secondary R  for 
>a FWB is no less important than it is for FWDs.  

I looked carefully this morning, and just looked again. What you say is
not true. Secondary R for a FWB is *much* less important than for a FWD.

Let's take a specific example, for a supply giving 3330V off-load, and
0.75A at full load. Use 30uF for FWB, or 2 x 60uF for FWD. What values
of total effective series resistance RS do you need for 3000V on-load?

R_load (RL) = 3000/0.75 = 4000 ohms

Full-load / no-load voltage ratio = 3000/3330 = 0.9 = 90%

FWB: omega*C*RL = 45 (@ 60Hz)
>From the chart, RS/RL for 90% regulation = 2.2%, so RS = 88 ohms

FWD: omega*C*RL = 90 (@ 60Hz) - remember, twice as much C.
>From the chart, RS/RL for 90% regulation = 0.4%, so RS = 16 ohms - 
yes, only *sixteen* ohms! 

Also remember that those figures include the total contributions of the
secondary, primary and diodes. RMS rectifier current is about 1.5x the
RMS DC output current, so with say 6 diodes in series that's 6 x 0.7 /
(0.75/1.5) = about 8 ohms. So there goes half of your 16-ohm maximum,
before you even get to thinking about the transformer.

>Also, a transformer for 
>FWB rectification has 2x the number of secondary turns of Thinner wire.  
>This increases secondary R. 

But not in the ratio of 88 to 16! That's my point: you won't find a
transformer with winding resistances as low as they need to be for FWD
unless it was specifically intended for FWD. In a transformer originally
designed for FWB you're most unlikely to find such low winding
resistances, because FWB simply doesn't need it.

> Anoher problem is that FWB transformers use 
>way more paper than an equivalent FWD transformer.  
>
I'd still guess that truly equivalent transformers - ones that gave
exactly the same voltage regulation - would come out very similar in
size, and the FWB one would be significantly lighter because it would
have to use much more copper (half the turns but much more than twice
the cross-section). Usually FWD transformers are smaller and cheaper
only because the voltage regulation is poorer too.


>>>(3) Is it good practice to place a diode across each cap in the stack, 
>cathode 
>>>towards the + side, for reverse current 
>>>protection.
>>>
>>AFAIK, the only time that reverse current could occur in a capacitor is
>>when a charged stack is being completely discharged through an external
>>load (this is not what happens at routine switch-off, when each cap
>>discharges through its own voltage-equalising resistor). Even in a 'big
>>bang', a single cap could only reverse polarity when all the others were
>>almost completely discharged too, so there's not much risk of severe
>>reverse current. And finally, reversal could only occur if one or more
>>caps is severely mismatched from the others in the stack. Unless there
>>is some other scenario I haven't thought of, those diodes don't seem to
>>be necessary.
>
>€  In one of my amplifiers, the rectifiers shorted - after c. 10-years - 
>sending AC to the electrolytic filter caps.  The caps blew their pressure 
>vents and went kaput.  

That's a good point, if no fuses blow and you don't hear the transformer
groaning. Diodes directly across the capacitors would certainly take
care of blowing the fuses, though there are probably better ways to
tackle this problem. 

73 from Ian G3SEK          Editor, 'The VHF/UHF DX Book'
                          'In Practice' columnist for RadCom (RSGB)
                           http://www.ifwtech.com/g3sek

--
FAQ on WWW:               http://www.contesting.com/FAQ/amps
Submissions:              amps@contesting.com
Administrative requests:  amps-REQUEST@contesting.com
Problems:                 owner-amps@contesting.com