magnetron microwave launcher design 2

[ProfDrK]

I plan to direct the output from an oven magnetron from a launcher/horn combination to dry wet soil as efficiently as possible. I have played with a couple of launchers, but despite an auto tuner showing no reflections, the (aircooled) magnetron overheats (over 100 C in a couple of minutes). When mounted back into the oven, it levels at 90 C. I think airflow is not an issue in my configurations. I noted two launchers have a grid of small holes and a dimple opposite the emitter on the sidewall of the waveguide..for what benefit? I will make another launcher and want to know besides the 1/4 wavelength back standoff, is there a trick I am missing that prevents overheating? Should the exit flange be placed at 1/2 wave from the emitter or is 1 lambda better? or is some other fraction better? All hardware is W284 size and I wonder if it shouldn't be bigger?

 

[biff44]

WR284 is only recommended down to 2.6 GHz, but you are still well above cutoff frequency of 2.09 GHz, so you should be ok there. I assume you just have a big reflection coefficient and the magnetron is receiving that reflected power.

How are you trying to launch the energy into the soil? Water has a dielectric constant of 80, dry soil has an Er of around 5, air is 1. So if you are trying to go from the waveguide to air above the ground with some sort of antenna, when it hits the wet earth you will have nearly total reflection from the Er of 80. If you eliminated the air layer, and went from waveguide into some sort of dielectric (maybe an Er=9.9 alumina sheet 0.4 inches thick--a quarter wavelength) you could get better penetration into wet soil.

 

[Comcokid]

I posted a magazine reference to a microwave magnetron soil heater in thread239-145266 . Maybe the article has some helpful pointers in your application.

 

[ProfDrK]

Thank you biff44 and Comcokid. Regarding interface to soil, I was aware some reflection might occur, but the soil has only of the order of a percent of water. Importantly, my overheating problem occurs even when there is no load at the exit of the waveguide/horn (whether there is a horn or not) with 700W forward, 1W reflected. Regarding the weed killer article, I saw it in Microwave Journal when it was published but my project started way before that. BTW, others have used microwaves to sterilize soil for replanting.

 

[biff44]

Lets start off with some basics here. You have a magnetron that puts out power. By "horn" I assume you have an antenna horn that is professionally designed to send microwave energy from the WR284 guide into free space. SO, have you tried just sending the energy into the air (point the horn up at the sky, get the heck away to not fry your eyeballs) and see if the magnetron still overheats? This will tell you if the waveguide is working with your magnetron.

One question you have to ask is "does the waveguide horn have a low reflection coeeficient at 2.46 GHz"? If it is a standard WR284 horn, it is probably tuned for the best radiation efficiency (ie lowest reflection coefficient) in the band center, which is 3.275 GHz. It may not work well at 2.45 GHz, causing a big reflection of power to the magie.

Tell us if the magie works fine when you are pointing the horn into the sky.

 

[biff44]

BTW, have you tried blowing a big-mamma fan on the magie?

 

[ProfDrK]

Sorry biff44, I addressed those points as part of the introduction to the problem. The problem is baffling because 1. the launcher/automatic tuner/15 db horn is operating in a large sheilded room with idle equipment and furniture in it to absorb reflections (that is equivalent of open to the sky in my opinion, but another way to think of it is a 10x10x10 foot oven cavity with food) and 2. there is a dedicated fan on the fins of the magie with a paper tunnel to direct it through the fins where I have a thermocouple pressed against the magie body between the fins. As I said in the first and fourth post, the tuner power readings say 700W forward and 0 back and airflow for cooling was not an issue. I wouldn't ask if it wasn't tricky. This is why I suspect the commercial launcher and want to make one that won't overheat the magnetron - which is the point of the original question.

 

[logbook]

Well at the risk of sounding stupid, if there is "no reflected power" the launcher can’t be overheating the magnetron. Either the directional power meter is telling lies or the magnetron is overheating itself.

I suggest that the situation is logically "baffling" because one of the presented "facts" is false.

 

[Higgler]

Most overheating problems that I've seen are center conductor heat buildup.

If you leave the magnetron in the oven, cheat the 3 door switches and connect waveguide over the rf output, then to your horn (or wg-coax. transition), life may be better.

I've done similar before without problems.

kch

 

[ProfDrK]

logbook, it's tricky. I blame the launcher. Nobody wants to give me that I might be right. I don't have the equipment to prove the launcher SWR is correct. The backspace from the dipole is 1.75 cm, and as far as I know, it should be more like 3.05 cm. This launcher has 3 arrays of small holes in the 3 narrow sidewalls around the dipole. The maggie end has slightly larger dimensions than the flange end (opening) forming a box which tapers down to the flange for 3/4 the waveguide length (overall 15.5 cm). I am concerned this geometry is incorrect and wanted to try making a better launcher.
Higgler, I was considering your idea but thought a 'proper' launcher might do better. I still need advice (?) for that. Can you list the causes for the 'center conductor heat build-up' if other than reflected energy? Am I correct to assume an auto-tuner can't fix the short backspace of the launcher?

 

[biff44]

And, you know that the "weird" launcher you just dexcribed is not arcing over by what means?

 

[Higgler]

Center conductor heating is usually IR losses due to thin wire or poor plating.

Are you certain your voltage is the same inside the oven as outside. Have you checked the current draw on the unit in both conditions, maybe it's suppose to be higher heat due to higher current draw.

kch

 

[ProfDrK]

Higgler; Power from the wall socket was measured and is nearly identical whether in the waveguide or the oven. With a working maggie, current draw is unchanged. I can eliminate conductor heating failure since the operation in the oven is repeatable and the temperature levels-out at 90C at power setting of 10. In case it matters, this oven setup has an inverter powersupply (and an ammeter added to maggie return line) not transformer. Tests involve moving the maggie only and everything from the HV leads back is in the oven structure. A maggie was killed from overheating already. Symptoms of failure were small output and lower than normal current draw while always getting very hot (150 C and more) in 4-5 minutes. biff44; I have no commercial arc detector but I have heard arc noises from the dead maggie. I see no tracks inside the launcher nor on the purple ceramic (antenna) of the failed unit.

 

[Higgler]

In the normal microwave oven, isn't the probe that sends out the rf energy directly connected to the magnetron, or is there a cable and connector setup (I think it's directly connected). I picture you removed this section and surrounded the probe with your own launcher. I wonder if you have a really good rf ground connection in the launcher you created compared to the rf ground in the oven? Current flows equally on the center conductor or the launcher and the surrounding groundplane lets call it. The transition or connection for the current to flow onto the ground may be better when attached in the oven than in your launcher.

If I'm not picturing the setup, please excuse the error.

Maybe describe what's removed from the oven, is it the magnetron with integral launch probe and without ground cup shaped launcher??

kch

 

[ProfDrK]

I unbolt the magnetron from the waveguide fabricated on the oven wall and rebolt it into a comercial aluminum launcher which is wr340 size at one end and wr284 at the other. Its about 6 inches long terminated with a circular flange. The 'probe' passes through a 1.25 inch aperature centered on the broadside of the waveguide that seals to the magnetron mesh gasket in both cases. Of the many microwave ovens I have dissasembled, the magnetron mounts to an elbow waveguide feeding the cavity of the oven (launcher). In all cases I describe, there is no coax or cable. I understand the term launcher can refer to a coax to waveguide converter, but I never intended that interpretation here. Maximum power transfer by beaming to the target without use of an enclosed cavity is my goal.
I am still looking for guidelines to make a proper launcher (even though I am making a couple of them right now as I wait for ideas from the experts). I don't have the funds right now to buy another commercial one.

 

[logbook]

My posting here is a comment not advice because I am both legally and technically incompetent in this area.

I am looking at my 1947 edition of Very High Frequency Techniques. Magnetrons were new and novel at the time; flushed with success during the war. Several of the drawings show "coupling loops" inside the magnetron cavity. The output is then on a coaxial line. This makes sense now that I think about it; the magnetron is a vacuum cavity device and that doesn’t sit well with a waveguide output.

I just pulled the cover off the magnetron output on my microwave oven. It has what looks like an E-field launcher in the waveguide, albeit with jumbo proportions. It is very simple. I think you should stick with the standard terminology of "launcher" for the coaxial to waveguide transition and "horn antenna" for the waveguide to free space transition.

I think you will find the E-field launcher is around one quarter lambda from the short-circuited end of the waveguide. The "back-short" is used to tune the launcher. If you dismantle the magnetron and use your own waveguide then the position of the back-short is going to be crucial. How you can adjust it at full power and not blow the f*ck out of it I don’t know. I would think the original designers set it up with directional couplers and a tuneable back-short and only applied full power after it was tuned.

Some pictures of the before and after assemblies would be very nice to see. If your microwave oven is like mine you would have to cut up the original waveguide to see how the launcher is positioned in the guide relative to the back-short and indeed if it is an E-field or an H-field launcher.

 

[logbook]

This unit would seem useful for your application:

http://www.coaxialpower.com/specifications/mg6k.htm

 

[ProfDrK]

Thank you logbook. I will add this to similar devices I have seen for future purchase. My problem is coupling the output of such a beast (or its little brother) to almost dry soil in situ in order to further dry it. I have several other stipulations, such as compactness, but that is for the future. As for the launcher, all I know is the backspace must be quarter-wave and I have reasoned that the launcher waveguide should be 1 lambda long. My construction methods are no better than the quality of the oven structure. For a 6kW unit, I suspect details like arc detection would be important in a launcher design. Once I can demonstrate feasibility using oven parts as much as possible, then possibly further funding that would pay for a 6kW unit will follow.

 

[logbook]

There are some interesting considerations in the launcher design depending on the characteristic impedance of the magnetron output. From what I can understand, the height of the waveguide is adjusted to match the real part of the magnetron output impedance. A standard waveguide has an impedance of around 300 ohms, a half-height guide being around 150 ohms and the impedance being reduced by reducing the height further. The backshort is adjusted to match the reactive portion of E-field probe’s reactance.

Ramo, Whinnery and van Duzer (Fields and Waves in Communication Electronics, 3rd ed) fig 8.11g shows an example of the matching but refers to Collin (Field Theory of Guided Waves) which I don’t have, but which I have now ordered!

 

[Higgler]

I have heard of termite killing using microwave heating, and I am guessing that they use microwave oven sources. I've seen the setup using four horn antennas to create a 12"x48" heating zone. I googled termites and microwave heating and found alot of sources of info but initially no commercial company although I know they are available. I checked again and found a website which shows a picture of an operator with a single magnetron and horn heating termites in the wall of a home. Maybe contact them about their hardware?

http://www.hi-techtermitecontrol.com/

kch