Voltage Multiplier Questions

[Raven Luni]

Greetings,

I'm considering using voltage multipliers for a number of things, small things initially like getting useful voltages out of my 2.7V ultracapacitors but maybe some high voltage stuff further down the line.

Questions:

- If a simple oscillator is used to create the initial AC, Does it have to be sinusoidal or would a square wave work as well?

- Would the frequency have to be matched to the capacitor values chosen for the ladder?

- Is a higher frequency more efficient?

- Is the voltage difference at each stage the same or does it go up exponentially? (this is important for choosing properly rated components).

Thats all I can think of for now. Feel free to point out any other important bits as well

 

[Raven Luni]

Bumpity bump, chumps

 

[duke37]

For efficiency, a square wave would be best.
An initial transformer to get the voltage up to several volts should be used to reduce the losses in the multiplier diodes.
The frequency will be a compromise, depending on the transformer type, capacitor sizes and diode speed. Probably 100kHz would be a good start.
If the capacitors are in series, then adding stages will give less and less advantage when the capacitor reactance comes into play.

 

[Raven Luni]

Ah - thanks
I think I mostly get it. So there would be both inductive and capacitive reactance (minus the diodes) so a tunable circuit might be a good way to go?. I've seen people saying that stepping voltages with transformers is more efficient at higher frequencies - but isnt inductive reactance bigger at high frequencies?

 

[BobK]

Yes, which allows you to use a transformer with many fewer turns.

bob

 

[duke37]

One problem will be the very low supply voltage of 2.7V. You may need a diddy transistor blocking oscillator to give, say, 10V to drive CMOS which would be necessary to get very low voltage drop switching devices.

You may be able to boostrap the main inverter so that it is only inefficient over the first few cycles. Some design work would be needed.

 

[Raven Luni]

Hmm - lots to think about - most interesting
A blocking oscillator on its own might be enough for lower voltages as long as theres enough current to keep a reservoir capacitor charged