Making power converters with transmission lines

[craigMWS]

I've been experimenting with using transmission lines for power conversion.  Leaving a note here in case anyone’s interested.  It's documented on Hackaday: https://hackaday.io/project/204242-boosting-voltage-with-a-cable, and I submitted it for their component abuse challenge.

The basic idea is to drive a transmission with with a low impedance square-wave.  If the transmission line’s time delay is one quarter the source’s period, the line will boost the voltage.  It’s essentially a quarter wave transformer, using a pulse source.   The voltage gain for a lossless line with zero source resistance is:

Vo/Vi = R_L / Zo

and with a non-zero source resistance:

Vo/Vi =  (R_L * Zo) / (R_L*R_s + Zo^2)

Where, R_L is the load resistance, Zo is the line’s characteristic impedance, and  R_s is the source resistance.

Is it practical?  Not sure yet. But it is fun to experiment with.

Cheers,

Craig

[makebit]

Sounds good,  can the boosted voltage be "collected" into a capacitor and used to supply a load with?

[UnijunctionTransistor]

You mentioned that you have been experimenting with it. This is very good. What have your results been so far?

[craigMWS]

@makebit: Sure, the output can be rectified (it's generally AC bipolar) and filtered or stored with a capacitor.  It's pretty easy to generate high voltages.  I've been using a load resistance across the capacitors, along with a DMM for monitoring to make they're discharged when powering down to work on the system.

So far, I've been experimenting with a voltage doubler rectifier.  Since the frequencies for this type of converter tend to be high, the diodes must be pretty fast. The diode capacitance can also degrade performance.  I used some air-wired 1N914's in series (to reduce capacitance) in the attached example.  This system is running at about 10MHz, and is generating 236V across a 100 ohm load (0.56W).  I've used a custom "delay line" sort of coax to allow running at "only" 10MHz with a 12in cable.

[Benta]

Certainly it's possible. A transmission line has an L and a C (and a lossy) component.
But what's the point? A normal  LC (+ R) combination would do the same.

[David Hess]

A quarter wave section can be made with lumped elements to do the same thing, but allowing lower frequency operation.  This seems like an alternate way to describe some resonate power conversion schemes.

Galvanic isolation is usually desirable, leading to some kind of transformer.

Another place I have seen unusual high voltage conversion is with the use of a ceramic transformer with piezoelectric elements at either end.