Switching power supplies are familiar to Hackaday readers, whether they have a fairly conventional transformer, are a buck, a boost, or a flyback design. There’s nearly always an inductor involved, whose rapid change in magnetic flux is harnessed to do voltage magic. [Craig D] has made a switching voltage booster that doesn’t use an inductor, instead it’s using a length of conductor, and no, it’s not using the inductance of that conductor as a store of magnetic flux.
Instead it’s making clever use of reflected short pulses in a transmission line for its operation. Electronics students learn all about this in an experiment in which they fire pulses down a length of coax cable and observe their reflections on an oscilloscope, and his circuit is very similar but with careful selection of pulse timing. The idea is that instead of reflected pulses canceling out, they arrive back at the start of the conductor just in time to meet a pulse transition. This causes them to add rather than subtract, and the resulting higher voltage pulse sets off down the conductor again to repeat the process. We can understand the description, but this is evidently one to sit down at the bench and experiment with to fully get to grips with.
[Craig]’s conductor is an alternative to a long coil of coax, a home made delay line of the type once found in the luminance circuit of some color TVs. It’s a coaxial cable in which the outer is formed of a tightly wound coil rather than a solid tube. With it and a high-speed gate driver he can light a couple of neon bulbs, a significant step-up, we think. We’re trying to work out which component is being abused here (other than the gate driver chip he blows) as the conductor is simply performing its natural function. Either way it’s a clever and unexpected circuit, and if it works, we like it.
This project is part of the Hackaday Component Abuse Challenge, in which competitors take humble parts and push them into applications they were never intended for. You still have time to submit your own work, so give it a go!
Um, how do you think those reflections happen? Transmission line effects are the result of the distributed inductance and capacitance of the conductors that make up the line, and (while there are many equally valid perspectives to view this from) the voltage spike where a pulse reaches the open-circuited end of a transmission line can be directly attributed to that inductance.
The point is that the voltage conversion principle used can not be properly explained well with lumped circuit theory. The wire is used as a proper transmission line.
It is not the simple lumped inductance idea, although you can arguably say that it can be approximated with many lumped elements a-la telegraph equation.
I find it a brilliant hack, I have yet to see transmission line reflections used to boost dc power deliberately, and it’s very cool.
Thanks! Very good points regarding emphasis on transmission line reflections. I have always conceptualized the idea in terms of wave reflections. My thought process was as follows: a wave wave travels from the source, reflects off the end of the line, and the reflects back to the source, and on an on. I thought…hmm, is it possible to add a bit of energy to the reflection each time it arrives back at the source, and then worked through the concept using bounce diagram sort of logic. I then realized it’s a a quarter-wave transformer like principle. I haven’t seen it used for power conversion either, though it is similar to how a dipole antenna provides a voltage boost, and I’ve seen some references to long power lines causing a “nuisance” voltage boost when they approach ¼ wavelength in length.
The principle work probably work even for a medium that doesn’t fit the lumped capacitor/inductor model that well, such as a waveguide, or antenna wire in free space.
Nope, I use open circuit line reflection to get up to 100kv or greater pulses. I didn’t come up with the technique either, a lot of clever people work at US national labs.
Or did, till jackass in chief got in.
Cool! Know of any papers or other publicly available resources on it?
That’s no way to talk about Obama.
elegant explanation imo
This is essentially a tesla coil.
Yup, the homemade delay line cable is admittedly purpose-built for boosting voltage. I more recently substituted three RG-58 test cables connected end to end as a “voltage boosting” element, and dropped some photos and data in the latest project log. Not sure if that qualifies as abuse, but it’s at least outside of their intended usage for connecting up test equipment. I was able to push 1W through it at 72VDC (at the output of the voltage doubler rectifier) through the RG-58 cable. Unfortunately, I couldn’t get more power out of it with the measly gate driver that I’m (ab)using to drive the cable. Its output resistance is way too high and limiting what I can do for now.
Anyone desiring a more thorough (mathematical) explanation of voltage-stacking along a transmission line can simply query http://gemini.google.com with the following:
Consider a narrow width voltage pulse of constant frequency driving the Source end of a section of coax that is open ended at the destination; What explanation can be given to account for a open voltage many multiples of 2 times the input amplitude?
If you’re interested, check the project files. I have a document there with a fairly thorough derivation for output voltage as a function of characteristic impedance, source resistance, and load resistance when applying a square wave source to a transmission line. It shows how I arrived at the design equations in the Details section. I found Gemini helpful for simplifying some of the equations and solving an infinite geometric series in that derivation.
Same principle as a LASER ? But replace photons with electrons
This technique dates back to the early 1940s and was used in British AI radar to generate narrow pulses with sharp rise and fall times. It was AFAIK Blumlein who invented this. In the modern era it’s been used in microwave sampling gates found in frequency counters and microwave synthesisers.