Reactance: Switching capacitance for an inductive power source

[ACharnley]

Hi,

This is a further consolidation of my previous posts which hopefully draws my questions to a final solution.

I have a highly inductive AC source, a dynamo which varies in frequency and voltage. System efficiency is imperative and so is power extraction. To that end, maximum extraction can be had by switching in capacitance at various points in the frequency.

By implementing active rectification I had hoped to be able to switch in the capacitance at the DC side via an N-fet low side. The bridge FET's (unlike an ideal diode) don't prevent the reverse flow so I theorised it would be cutting down components and complexity. A win win. This assumed capacitance could be added in parallel to have the same effect as in series.

LTspice seemed to agree. Frequency sweeps over various capacitance show the change in power.

So, I prototyped it and ran some real-life tests. The "none" shows the base-line. "230uF series" shows the ideal extraction (it would fall pretty rapidly if the frequency were any higher but then a smaller cap would switch in).

Clearly whether the capacitance is in series or parallel has a much greater difference than I expected, and while the active rectification based switch in does work it's no where near as performant.

Can anyone explain why LTSpice has got it wrong and/or why the results are as they are?

Cheers,

Andrew

 

[alec_t]

This assumed capacitance could be added in parallel to have the same effect as in series.

Eh? You've lost me. The only way I can see the effect being the same is if the effect is zero.
In parallel/series with what? I don't see any capacitor switch-in going on in that circuit.
It would help us to check circuit operation if you post your asc file.

 

[ACharnley]

Here's the asc (you'll probably have to swap the N and P fets out to something else).

The capacitor is in series here, you can move it in series with AC1/AC2.

In real-life the results are the opposite - minimal effect in series and quite a bit in parallel.

 

[alec_t]

Folk here are leery of opening zip files. You should be able to upload the asc file unzipped.

 

[ronsimpson]

L1 = 0.1 H. That seem very high.

 

[ACharnley]

It's correct Ron, actually 0.12H.

 

[ACharnley]

Eh? You've lost me. The only way I can see the effect being the same is if the effect is zero.
In parallel/series with what? I don't see any capacitor switch-in going on in that circuit.
It would help us to check circuit operation if you post your asc file.

So C1 is across the power source at the moment, move in series with the inductor and AC2. In real-life test the capacitor has a much greater effect in series, but LTSpice says otherwise.

 

[ronsimpson]

but LTSpice says otherwise.

When SPICE does not agree with real life then you have not told SPICE something. or you don't understand something.

Most people do not include series resistance in a inductor. Capacitors have internal resistance. Most inductors have a saturation point.
I am questing using a 0.12H simple inductor. It could be much more complicated.

I don't have answers. Just questions and comments.

 

[alec_t]

So C1 is across the power source at the moment, move in series with the inductor and AC2.

If in series with AC2 you will be applying AC to what is presumably an electrolytic capacitor. Make sure the cap is an AC type or you will have fireworks!

In real-life test the capacitor has a much greater effect in series

Effect on what? What sort of effect?

 

[ACharnley]

The capacitor changes the phase and should be applicable in parallel or in series with the inductor. The capacitance has to be matched as it becomes a frequency blocker and can also reduce output.

Real-life - in series the greatest output is possible "the hump". In parallel some additional output is possible but not quite as much.

LTSpice - in parallel the output matches my real-life series test. In series it doesn't work at all.

As Ron has stated I've included capacitor resistances.

 

[alec_t]

The capacitor changes the phase and should be applicable in parallel or in series with the inductor.

Your post #1 sim doesn't have a cap either in series or in parallel with the inductor.
If B1 is supposed to simulate a generator, shouldn't both amplitde and frequency be swept?

 

[ronsimpson]

highly inductive AC source, a dynamo

What dynamo? Part number or data sheet please.

 

[ACharnley]

The companies don't give them out (I tried). The 0.1H came from the manufacturer. A German source measured one at 0.12H and my multimeter has another at 0.11H.

 

[ACharnley]

Your post #1 sim doesn't have a cap either in series or in parallel with the inductor.
If B1 is supposed to simulate a generator, shouldn't both amplitde and frequency be swept?

For series you will need to move the cap, I don't know if there's a way to run two simulations alongside each other (which would be fantastic). Or do I just duplicate the circuitry?

For parallel the fets allow reverse flow so it has the same effect (ignoring fet resistance).

If I can learn how to sweep both then great!

 

[alec_t]

If I can learn how to sweep both then great!

Here's one way, using the 'Modulate' model in the Special Functions folder :-