Resonant Royer Oscillator questions

[GadgetBoy]

So, I'm experimenting with an old project that I'd given up on a while back.

On paper, this makes sense, but I want to take a 12v doorbell transformer. Rectify and filter it's output, feed that into a resonant royer oscillator, then the output into my Cockroft-Walton adder. (See attached schematic)

I tried this previously with a 12v switch-mode supply, and it worked for about 5 minutes, then the supply just died with no fanfare. I figured that there most have been some sort of feedback that the switch-mode supply disagreed with?

Questions:

1. There are no fancy active components here, so I don't have to worry about the supply dying, right?

2. Do I need to common the neutral from the CW adder with the mains neutral?

3. Am I baking up the wrong tree?

[T3sl4co1l]

You may want a filter capacitor.

There are many reasons why an SMPS might die, and not all of them are eliminated here.  You can always short out a transformer and make it catch fire.

FYI, Royer is a saturation-commutated class D oscillator, and not resonant.  The resonant version (with series supply inductor) is properly called a Baxandall oscillator.

The high voltage community has internalized this error so this seems a particularly ineffectual bit of pedantism I admit.

Tim

[GadgetBoy]

Filter capacitor on the 12V  dc side? Got one, although I suspect it's not big enough (100uf), so I'm rummaging for the bigger caps I know I should have before testing.

So that's a no on the neutrals, Roger.

Baxandall sounds more evil-lair sci....I meant garage hobbiest, anyway. Thanks for leading me towards some interesting reading.

[Zero999]

The circuit should be isolated from the mains, for safety reasons. Now that might sound a bit perverse, given the output is probably a few kV, but the current will be limited by the impedance of the converter and the capacitors, so is likely to be less dangerous than the mains.

Have you considered using an isolation transformer and a voltage multiplier? That would eliminate the need for an inverter, but would require larger capacitors.

[chris_leyson]

The resonant version (with series supply inductor) is properly called a Baxandall oscillator.

Thanks Tim. The current fed Baxandall converter is a nice simple topology and really has very little to do with the Royer converter as Tim pointed out. Most people just repeat what they read on the internet without looking into the history and development of push pull self oscillating converters.

A little history. One of the first applications of the Royer oscillator was a linear voltage to frequency converter, this cut and paste from Wikipedia.

Royer, G. H. (1955). "A switching transistor D-C to A-C converter having an output frequency proportional to the D-C input voltage". Transactions of the American Institute of Electrical Engineers, Part I: Communication and Electronics. 74 (3): 322. doi:10.1109/TCE.1955.6372293.. A 1955 paper by Royer on his circuit.

The Royer was a popular topology in DC to DC converters back in the 60's to convert 12V to a higher voltage for tube based mobile transceivers. PYE used it in their VHF/UHF mobile tranceivers back in the day.

One improvement on the Royer was the Jensen converter which moved the saturating transformer from the power stage to the transistor base drive stage, it offered better efficiency at the expense of an additional drive transformer. It still gives you propertional base drive current which optimises transistor switching times. See Swithed mode power supplies by Billings et al.

A similar topology to the Jensen converter was used quite a lot in early PC/XT power supplies but only to make it self starting. Contol loop didn't rely on saturation and used TL494 or similar driver.

Jim Williams wrote some excellent application notes on current fed resonant self oscillating converters used as CCFL power supplies. Rather than reinvent the wheel you could use a small CCFL converter to generate your HV output, just add rectifier diodes.

[GadgetBoy]

The circuit should be isolated from the mains, for safety reasons. Now that might sound a bit perverse, given the output is probably a few kV, but the current will be limited by the impedance of the converter and the capacitors, so is likely to be less dangerous than the mains.

Have you considered using an isolation transformer and a voltage multiplier? That would eliminate the need for an inverter, but would require larger capacitors.

I've been continuously running into an issue getting this CW adder to work on 120v (otherwise I'd just run it directly off the mains). My research on the subject has indicated that the adder works better with a higher starting voltage, and/or higher frequency. 120v - 240v (or higher) step up transformers are too expensive and way too huge to make my project worth it, so I've been experimenting with stuff I have on hand.

Using the 12v SMPS with the baxandall (image attached to make sure we're talking about the same time) worked REALLY well. I had an audible hissing coming from the carbon fibre emitter, and a light corona effect (only visible in compete darkness). Output on the baxandall was about 400v @ 40kHz. I don't have any equipment capable of reading the voltage at the end of the 11 stage CW adder, but my NCVs all started freaking out about 12" away from the emitter, so there was definitely an emission of electrons.

The issue was that the SMPS started buzzing, then the baxandall started buzzing, then the SMPS died, and that was that. I have a second cheap 12v SMPS, but after what happened to the first, I'm a bit wary of risking a second one.

Thus, this frankensteinian mess.

[vk6zgo]

The circuit should be isolated from the mains, for safety reasons. Now that might sound a bit perverse, given the output is probably a few kV, but the current will be limited by the impedance of the converter and the capacitors, so is likely to be less dangerous than the mains.

Have you considered using an isolation transformer and a voltage multiplier? That would eliminate the need for an inverter, but would require larger capacitors.

I've been continuously running into an issue getting this CW adder to work on 120v (otherwise I'd just run it directly off the mains). My research on the subject has indicated that the adder works better with a higher starting voltage, and/or higher frequency. 120v - 240v (or higher) step up transformers are too expensive and way too huge to make my project worth it, so I've been experimenting with stuff I have on hand.

Using the 12v SMPS with the baxandall (image attached to make sure we're talking about the same time) worked REALLY well. I had an audible hissing coming from the carbon fibre emitter, and a light corona effect (only visible in compete darkness). Output on the baxandall was about 400v @ 40kHz. I don't have any equipment capable of reading the voltage at the end of the 11 stage CW adder, but my NCVs all started freaking out about 12" away from the emitter, so there was definitely an emission of electrons.

The issue was that the SMPS started buzzing, then the baxandall started buzzing, then the SMPS died, and that was that. I have a second cheap 12v SMPS, but after what happened to the first, I'm a bit wary of risking a second one.

Thus, this frankensteinian mess.

If you can find an old centre tapped  tube type power transformer, with what we used to call "200 volts (or more) a side", you will have access to 400v or more across the  complete secondary.
These are still quite common in junkboxes in Oz, but understandably may be rarer in NA, due to the dominance of the abominable transformerless "All American Five" in that  market.

Tube stuff used in industry was still quite likely to have power transformers, however, so you might find one.
Mains transformers are a mature technology, so you are most unlikely to find one that is unsafe, but obviously, testing is still a good idea.

They aren't near as large as you might imagine, those used in Australian "mantel" radios of the 1950s, fit easily into the palm of your hand.