Voltage doubler questions

[db!]

Do the capacitors on a voltage doubler need to be non-polar? Is there a type of capacitor recommended for use with a voltage doubler circuit?

I have also noticed that there are different "topologies" of doublers (delon/bridge, CW, villard); do they all perform the same?

I'm trying to pull 270V@120mA from a 120V 0.5A transformer; is this doable?

thanks

[metalsculptor]

Doubler capacitors can be standard electrolytics, for your application 470 uF 200V electrolytics in a full wave doubler configuration will give you a little over 300V with more than enough current.

Most older computer power supplies use this method when operating on 120V changing to a bridge rectifier with series capacitors when switched for 240V operation. So you can strip a computer power supply for parts if need be.

[Redshift187]

You will not get quite double the voltage. Do you mean 270VDC from 120VAC? 120mA from 500mA sounds reasonable. Capacitors do not need to be non-polar, but you need to be a little more careful about polarity.

I use a Delon half-wave voltage doubler for a microphone pre-amp, to get +/- 10VDC from a 9VAC transformer.

[db!]

Thanks for the replies!

I plan to use the doubler to raise the voltage and an RC filter circuit for the final output. I have simulated it on PSUD2 and it theoretically should work. Also, how should an electrolytic capacitor be connected in terms of polarity? Most VD circuits don't indicate polarities...

metalsculptor - Thanks for the tip, I just garbage picked a few PSUs last week.

[metalsculptor]

Polarity, look at the diodes, and think which way the current flows in them. If in doubt in the simulation put a volt meter in each side of the capacitor of interest.

[Elvee]

Originally Posted by db!

Do the capacitors on a voltage doubler need to be non-polar? Is there a type of capacitor recommended for use with a voltage doubler circuit?

I have also noticed that there are different "topologies" of doublers (delon/bridge, CW, villard); do they all perform the same?

I'm trying to pull 270V@120mA from a 120V 0.5A transformer; is this doable?

thanks

For the Greinacher/Schenkel topology, the input capacitor should be non-polar, as it can see a reverse in polarity when powered on.
In addition, the ripple frequency of these doublers is the line frequency, which is why the Delon/Latour topology is preferable where possible: but the input and and output have no common connection, which can be a problem for some application.
With properly sized capacitors, you should be able to get your 270V out.

[AndrewT]

Originally Posted by db!

..................I'm trying to pull 270V@120mA from a 120V 0.5A transformer; is this doable?

I don't think so.
The doubler still effectively loads the transformer with a capacitor input filter.
The manufacturer of the transformer will have a de-rating factor for this type of load.
It is commonly around 0.7, but does vary from type to type and from manufacturer.

120V @ 0.5A is 60VA. Applying the de-rating factor results in a maximum continuous wattage of 42W.
At 270Vdc, 42W allows a maximum continuous current of 155mA.
At this output level the transformer is running at 100% capacity.

But the voltage of the output has been pulled down from ~340Vdc to 270Vdc to attenuate the output ripple. This, I believe, means that 340Vdc should be used to determine the transformers maximum rated DC current from the doubler.

340Vdc @ 120mA is equivalent to 40.8W or about 97% of rated capacity.

I think the transformer will run hot.

[Elvee]

Using realistic parameters for the transformer, this looks possible (to be confirmed by measurements).
A series bleeding resistor R2 has to be added; its actual value depends on the copper resistance of the windings.

[metalsculptor]

Originally Posted by AndrewT

I don't think so.


But the voltage of the output has been pulled down from ~340Vdc to 270Vdc to attenuate the output ripple. This, I believe, means that 340Vdc should be used to determine the transformers maximum rated DC current from the doubler.

You are assuming that a transformer is a perfect AC source with the the winding resistance in series. It is not, there is a a reactive component as well which is quite high in small transformers especially split bobbin types, this makes the transformer behave like it has a small choke in series with the filter capacitor. Effectively this reduces the output voltage without generating additional heat. A toroidal transformer will be closer to your worst case assumption.

Microwave ovens also use a voltage doubler and deliberately introduce additional leakage inductance to better utilise the transformer (split bobbin with low permeability spacer between windings)

[db!]

Thanks for the replies!

Sorry for not mentioning earlier, 120mA is my "conservative estimate" for the maximum current draw. The quiescent current will be about 32mA; if it get anywhere near 120mA, the transformer will be the least of my worries. The max plate current and dissipation is 31mA(+/-9mA) and 8W respectively, I doubt it will draw any more current. The transformer I was planning to use is a hammond 167G120.