Could a variable capacitor divider replace a Variac?

[Omegatron]

Hmmm... You can definitely drop down the voltage, and ideal capacitors don't dissipate any power. So it seems, at first glance, that you could use a capacitor divider as a lossless voltage step-down device for AC.

So you could use a cheaper variable cap divider as a replacement for a Variac? And it would be continuously variable, too, whereas the Variac is only variable in discrete windings.

There has to be something wrong with this. What am I not thinking of?

 

[GENIERE]

Hmmm... You can definitely drop down the voltage, and ideal capacitors don't dissipate any power. So it seems, at first glance, that you could use a capacitor divider as a lossless voltage step-down device for AC.

So you could use a cheaper variable cap divider as a replacement for a Variac? And it would be continuously variable, too, whereas the Variac is only variable in discrete windings.

There has to be something wrong with this. What am I not thinking of?

With a constant frequency (50-60hz) a variable capacitor can be used to vary the impedance of a circuit so it need not be part of a divider circuit. There is nothing that is loss less and it might be difficult to make one to handle large currents. Offhand, I don’t think it would be too difficult to power an LED directly from the mains and alter its intensity with only a variable capacitor in the circuit.

...

 

[SGT]

Capacitors can be and are used as voltage dividers not only with AC but with DC voltages. The divider probe of an oscilloscope uses a capacitor voltage divider.

 

[Omegatron]

Capacitors can be and are used as voltage dividers not only with AC but with DC voltages. The divider probe of an oscilloscope uses a capacitor voltage divider.

Yeah but at DC the caps are high impedance, so you can't get any useful power transfer. At low freq AC I guess there would be some impedance seen, too. But caps don't dissipate power; they just slosh it around, so it seems like it would be lossless step-down for ideal components.

 

[SGT]

Yeah but at DC the caps are high impedance, so you can't get any useful power transfer. At low freq AC I guess there would be some impedance seen, too. But caps don't dissipate power; they just slosh it around, so it seems like it would be lossless step-down for ideal components.

Only if the load is high impedance, like a scope input.
Since the AC impedance of a capacitor is frequency dependent, if you use a capacitor voltage divider with anything but a sinusoidal signal, each frequency of a composite signal will see a different impedance and the signal would arrive distorted to the load.

 

[Antiphon]

The full answer is this.

Yes you can do exactly what you want to do.

There is a difference between a variac and voltage divider. A variac
is a transformer and (if idealized) can supply unlimited current at a fixed
voltage.

The variable capacitor will be an adjustable series impedance and as such,
the voltage will fluctuate will the load.

As a practical matter, to deliver lots of current the caps would have to
have large values. But for dividers powering low current circuits this is a cost
effective alternative to a transformer.

 

[Omegatron]

Only if the load is high impedance, like a scope input.
Since the AC impedance of a capacitor is frequency dependent, if you use a capacitor voltage divider with anything but a sinusoidal signal, each frequency of a composite signal will see a different impedance and the signal would arrive distorted to the load.

But the impedance of both capacitors would change at the same time, so the level of each frequency component wouldn't change. Do you mean phase distortion? Power in power lines is by far the fundamental, though, and I doubt phase lagging the higher frequency components would have very much effect on the power you could deliver at all.

The variable capacitor will be an adjustable series impedance and as such,
the voltage will fluctuate will the load.

But if the caps are small compared to the load, a change in the load impedance wouldn't affect it much.

As a practical matter, to deliver lots of current the caps would have to
have large values. But for dividers powering low current circuits this is a cost
effective alternative to a transformer.

Yeah.

Hmm... I know I could figure this all out quite easily just by doing that math, but I don't feel like it. :-)

 

[SGT]

But the impedance of both capacitors would change at the same time, so the level of each frequency component wouldn't change. Do you mean phase distortion? Power in power lines is by far the fundamental, though, and I doubt phase lagging the higher frequency components would have very much effect on the power you could deliver at all.

Don't forget that your load will be in parallel with one of the capacitors, so you will have phase lag. If you have a sinusoid, this phase lag is not important, but if you have any other waveform, each sinusoidal component will have a different phase lag and the signal will be distorted.
Also, remember that your load does not change, so the voltage divider will be between a variable capacitor and the parallel of another variable capacitor and a fixed load.

 

[berkeman]

One other handy property of a variac... Most variacs are built as autotransformers, with the input AC voltage brought in with one connection at the bottom end of the winding, and the other connection brought in about 75% of the way to the other end of the winding. The variable tap wiper can sweep all the way up to the top of the winding, so Vout can be higher than Vin if you run the wiper all the way to the top of the windings. This is very helpful for things like switching power supply development and testing, since you can vary the AC Mains input to your supply under test all the way from high line to low line.

Also keep in mind that since variacs are generally autotransformers, they are not isolated. You should use a series isolation transformer with the variac if you are going to be working with exposed AC Mains voltages.

 

[Omegatron]

Don't forget that your load will be in parallel with one of the capacitors, so you will have phase lag. If you have a sinusoid, this phase lag is not important, but if you have any other waveform, each sinusoidal component will have a different phase lag and the signal will be distorted.

Yeah, but not a problem for power supplies.

Also, remember that your load does not change, so the voltage divider will be between a variable capacitor and the parallel of another variable capacitor and a fixed load.

The load can change, so this is an important consideration. <s>Somehow there would have to be a balance between low-impedance caps for output voltage steadiness and large-value caps for high current capability.</s>

Silly me. Large value caps are low impedance. So the bigger the caps, the better.

This is very helpful for things like switching power supply development and testing, since you can vary the AC Mains input to your supply under test all the way from high line to low line.

Yeah, I use them at work. I was looking at it and thinking "man, those are expensive, I wonder if there's a cheaper way to get the same effect."

Also keep in mind that since variacs are generally autotransformers, they are not isolated. You should use a series isolation transformer with the variac if you are going to be working with exposed AC Mains voltages.

So just use a fixed step-up transformer before the cap divider and get the increased voltage and the isolation at the same time.

 

[berkeman]

>So just use a fixed step-up transformer before the cap divider and get
>the increased voltage and the isolation at the same time.

Interesting idea, Omega. I think that the main problem with using caps instead of magnetics will be the volumetric energy density. With effective load capacitances of hundreds of uF for most power supply inputs, your top divider capacitor element will need to be on that order. But most variable caps that I've seen are in the pF-nF range, and I can't imagine how you'd make one in the many-uF range. I guess if you could make the mechanicals tight and reliable, and immersed the moving plates in water or something with a high dielectric constant, you might be able to get there. Have you calculated how big physically the variable cap portion would have to be?

 

[Omegatron]

Interesting idea, Omega. I think that the main problem with using caps instead of magnetics will be the volumetric energy density. With effective load capacitances of hundreds of uF for most power supply inputs, your top divider capacitor element will need to be on that order. But most variable caps that I've seen are in the pF-nF range, and I can't imagine how you'd make one in the many-uF range. I guess if you could make the mechanicals tight and reliable, and immersed the moving plates in water or something with a high dielectric constant, you might be able to get there. Have you calculated how big physically the variable cap portion would have to be?

Of course I haven't. :-) It was just an idea. I guess it would be just as expensive as a Variac, or more. You could use a string of caps and have only discrete output values...

You mean oil, right? Water would conduct.

 

[berkeman]

You mean oil, right? Water would conduct.

I was thinking DI water, so it shouldn't conduct. Oil would probably be okay too, and I guess there'd be less of a worry about corrosion...