3

I would like to try and build a basic microwave plasma torch using an oven magnetron. I've had experience with S-Band RF back in my physical chemistry days but am not liking the prices for waveguides. Is it possible/practical to couple a microwave magnetron to a 50 Ω coax to terminate in a basic plasma torch head?

ocrdu
  • 8,705
  • 21
  • 30
  • 42
nitrous dude
  • 31
  • 1

2 Answers2

7

it's not possible. You need to couple out the wave from the cavity, through an appropriately shaped waveguide. You can then convert that waveguide to whatever shape you want as long as you maintain impedance, and the waveguide withstands the necessary powers. For example, a coax cable is a feasible waveguide for most microwave oven frequencies, but not at all for microwave oven powers.

Now, whether or not that's going to be the cheap coax that – I'm very afraid – you been thinking about, I'd have my doubts: "plasma torch" says "high power", to be specific, high enough power to ionize gases. At these power densities, anything will melt, burn, disintegrate: anything but vacuum and large-crossection air-filled waveguide / cavity with a single, stable, well-defined field maximum in the place you blow gases through.

So, you cannot just take your TV cable and hope for the best, unless you're in for hurting yourself and starting fires. The size of your waveguide is dictated by the need of the maximum field strength not being high enough within the wave guide to ionize whatever is used as dielectric in there.

Marcus Müller
  • 88,280
  • 5
  • 131
  • 237
  • Some very large coax might be able to handle the power. Something like LMR-1200, perhaps. But yes, definitely not going to work with RG-58 or -174 or whatever cheap coax the asker is thinking of. – Hearth Mar 27 '23 at 21:29
  • 1
    This cable https://www.rf-microwave.com/en/times-microwave-systems/rg115-51513/coaxial-cable-rg115-50ohm-ptfe-10-3mm-high-power/rg115/ supports about 500 W at 2.45 GHz and costs £15 per metre. – D Duck Mar 28 '23 at 08:06
  • @DDuck wow, that's actually pretty nice cabling! Bookmarked. – Marcus Müller Mar 28 '23 at 08:13
  • 1
    Whenever I hear about people fooling with microwave oven components, whether transformers or other, I always hope they have paid up their life insurance policy, and if it covers accidental electrocution. – Michael Harvey Mar 28 '23 at 11:49
  • yes Michael, my insurance is well paid up. I've found a fairly detailed paper outlining a torch that will function as low as 20w using a commercial oven magnetron. As for the risk of electrocution, as I said, I worked in a lab for a number of years during grad school. My familiarity with the subject is somewhat distant, but I'm quite risk adverse. Again guys, thanks for the feedback :) – nitrous dude Mar 28 '23 at 22:17
  • Just for completeness, I'm thinking about 3d printing some suitable waveguides. Plating the inside surface is quite easy. Again, thanks for the feedback. – nitrous dude Mar 28 '23 at 22:23
  • Make sure to make the conductivity of that surface good and the surface flat - displacement currents heating up cracks in waveguides is not a new phenomenon – Marcus Müller Mar 29 '23 at 20:01
6

Is it possible? Yes. Can you afford that coax? Probably not, if you can't afford the waveguides. Power RF folks please correct me, but I'd imagine S-band waveguides are a low denominator, if not the lowest, and per foot are some of the cheapest in terms of manufacturing overhead on top of raw material price.

Coax that can push 1-2kW in S band is not usually stocked at non-specialist distributors, AFAIK. So buying might involve minimal quantities. And the coax is not necessarily any easier to deal with than the waveguides - the connections are still key to safe performance, so the tooling used in terminating the coax, as well as the connectors, have to be professional grade. That means expensive. Eye-wateringly expensive. And you need to practice a bit before you get the connections right. And everything has to be tested before any significant power is applied. So you need T&M that can deal with that - a VNA with a cal kit suitable for the purpose, and high-quality interconnects.

Whereas with waveguides, a small test transmitter and a homebrew RF leakage detector will at least ensure the RF will stay inside until needed, if the waveguides are otherwise not damaged, and are assembled with good workmanship.

Kuba hasn't forgotten Monica
  • 32,734
  • 1
  • 38
  • 103
  • Yeah, with a half-kilowatt-plus RF transmitter (the magnetron), even 10% of insertion loss at the coax connection is going to melt that connection. And if that half-kilowatt ends up leaking out into the open while testing, the whole neighbourhood is going to lose their WiFi (and the relevant regulatory authorities might come for a chat). – TooTea Mar 28 '23 at 07:21
  • Bigger problem is the RF being reflected off the mismatch and back into the magnetron – D Duck Mar 28 '23 at 08:09
  • 1
    @DDuck hm, for a consumer microwave oven, I would've expected that to *not* be a fatal problem – after all, if you put nothing into your oven and turn it one, the thing mustn't explode, either (so, probably some thermal feedback? Don't know, actually.) – Marcus Müller Mar 28 '23 at 12:01
  • 1
    @MarcusMüller I'm not sure, but I think the oven cavity is designed as a purely reactive load for the magnetron when it is empty. So there is no actual power delivered to the cavity (bar losses) when there is no food in it. It's like having an ideal capacitor being put in parallel to an AC power supply: it doesn't draw any power. The inserted food becomes a lossy/resistive "dielectric" in the cavity, and that turns the cavity in a lossy cavity. – LorenzoDonati4Ukraine-OnStrike Mar 28 '23 at 12:20
  • @LorenzoDonatisupportUkraine exactly, but since it doesn't draw any power, *all* power emitted by the magnetron gets ends up being reflected – where else would it go? So, the magnetron system needs to be somewhat robust against that – Marcus Müller Mar 28 '23 at 12:23
  • @MarcusMüller Interesting problem. I suspect there is no electronic feedback system at play for such a basic operation mode (wouldn't it be too dangerous? microwaves ovens were invented a long time ago). I have a somewhat lingering idea that the magnetron+waveguide+oven cavity is a system that draws no power when it is empty. It is initially "charged" with a sationary wave pattern which resonates and doesn't require any more power to sustain. I guess that whole systems looks reactive to the power supply it is connected to. I'm just mixing a wild guessing and some ancient memories, though. – LorenzoDonati4Ukraine-OnStrike Mar 28 '23 at 12:29
  • @MarcusMüller It would be interesting if some power-microwave guy chimed in and cleared the doubt. – LorenzoDonati4Ukraine-OnStrike Mar 28 '23 at 12:31
  • Power does go back into the magnetron, usually there is a thermal cutout to keep things from melting. Other methods exist: https://patents.google.com/patent/US6008483A/en?oq=6008483%2c In a radiotherapy linac there is a circulator or an isolator to protect the magnetron but here you put a 100 A, at 45 kV on the device which produces about 2 MW at 3 GHz – D Duck Mar 28 '23 at 19:02