Mounting Magnetron

Every fusor and fusion system seems to need a vacuum. This area is for detailed discussion of vacuum systems, materials, gauging, etc. related to fusor or fusion research.
David Kunkle
Posts: 284
Joined: Thu Apr 17, 2014 12:43 pm
Real name: David Kunkle

Mounting Magnetron

Post by David Kunkle »

I am trying to mount a microwave oven magnetron directly to my chamber. I left a spare 2.75" CF port on my chamber just for this. But so far, I haven't come up with any easy ways to get it mounted and sealed. Looked thru the archives but no good pics or descriptions for what I'm doing.
If your experiment needs statistics, you ought to have done a better experiment.

Ernest Rutherford
Dan Tibbets
Posts: 578
Joined: Thu Apr 17, 2008 1:29 am
Real name:

Re: Mounting Magnetron

Post by Dan Tibbets »

First, a magnatron can be dangerous. Outside of the appropriate mounting in a shielded microwave oven, microwaves can be dangerous. I understand that the lens of the eye (or is that the cornea?) is very sensitive to the microwave frequencies. So you probably wouldn't explode like in the movie "Kick Ass" or fry your brain like in an industrial accident in Great Britten, but you might damage your eyes. Be very careful and informed about the necessary shielding.

Having said that, a glass port may be reasonable for mounting. The magnatron would not need to be exposed to vacuum. There is a reason a fine metal grid is placed over the glass window of microwave ovens. Keep in mind that any glass ports, including the glass port the magnatron sits on could be a route for the microwaves to exit the metal Fusor and again expose you or your equipment.

Dan Tibbets
David Kunkle
Posts: 284
Joined: Thu Apr 17, 2014 12:43 pm
Real name: David Kunkle

Re: Mounting Magnetron

Post by David Kunkle »

I have a microwave leak detector on the way- I have no intentions of cooking myself. At least I'll know when and where it's not sealed.

I thought it would be easier to install and seal by installing the magnetron antenna inside the chamber to cut down on leaks. i.e. epoxy alumina or metal tube over the antenna and attach that tube thru a hole cut in a 2.75"CF blank. Now I'm not sure if microwaves only exit the hole in the antenna or if the entire metal cap radiates the microwaves- In which case I probably can't have metal in contact with it? A youtube video showed a magnetron close to a evacuated bell jar with sparks coming off the antenna and striking the glass- eventually putting a hole thru it. Looks like the antenna is wired directly to the anode in the magnetron?

I was going to put a piece of microwave door screen over the viewport, but your post got me thinking if microwaves could come thru the ion gauge. That could be tricky to seal up. I suppose I could blank off the ion port when using microwaves. I have a good idea where the vacuum is on my system depending on what's going on inside.
If your experiment needs statistics, you ought to have done a better experiment.

Ernest Rutherford
John Futter
Posts: 1848
Joined: Wed Apr 21, 2004 10:29 pm
Real name: John Futter
Contact:

Re: Mounting Magnetron

Post by John Futter »

David
Yes you can put the microwaves into the chamber through a view port the size of the 23/4 view port would be considered a microwave iris ie it is not ideal and there would be a lot of reflected energy back into the magnetron shortening its life.
It is unlikely that the energy will escape through your gauge port as this can be thought of a a wave guide beyond cutoff.

You will only need a couple of hundred watts maximum
see what Doug Coulter has done.
I have designed a few microwave assisted PECVD systems the biggest using 25 kilowatt of RF energy.

A simple detector is easy to make, all you do is get a small 50-100uA analog meter, one of those edgewise tuning meters out of cheap radios is ideal and put a diode across the meter terminals leaving the excess lead hanging out both sides of the meter terminals a 1N914 1N4148 is ok but a hot carrier diode such as HP8200 series is much better. The last time I did this a couple of years ago I used a BAT84 schottkey
User avatar
Rich Feldman
Posts: 1470
Joined: Mon Dec 21, 2009 6:59 pm
Real name: Rich Feldman
Location: Santa Clara County, CA, USA

Re: Mounting Magnetron

Post by Rich Feldman »

John Futter wrote: A simple detector is easy to make ...
and calibrated survey meters to measure 2450 MHz leakage are plentiful. I think the Holaday brand is respected in the trade, and their products are up on ebay today for between $20 and $75.
holaday.PNG
holaday.PNG (181.75 KiB) Viewed 8245 times
Suppose a magnetron out on the table were radiating 1 kW with an isotropic beam pattern. (not physically possible, but it simplifies the figuring.) At a distance of 50 inches, the power density would be less than 5 mW/cm^2. That's the officially tolerable leakage field strength for microwave ovens. It's also the FCC's Maximum Permissible Exposure for 2450 MHz in occupational settings, according to: http://sp.ehs.cornell.edu/lab-research- ... _Guide.pdf
I can attest that MWO leakage can jam a household WiFi network while a Holaday probe in near field hardly budges the pointer.

"Do not look into waveguide with remaining eye."
All models are wrong; some models are useful. -- George Box
David Kunkle
Posts: 284
Joined: Thu Apr 17, 2014 12:43 pm
Real name: David Kunkle

Re: Mounting Magnetron

Post by David Kunkle »

003.JPG
002z.jpg
001z.jpg
Here's my best idea for mounting the magnetron directly to the chamber. 2.75" CF X NW40 reducer. The CF part will bolt right to an extra nipple I originally built into the chamber. The original steel gasket (around the magnetron insulator and sits on top of the magnet) and the brass woven gasket both sit nicely in the gasket groove on the NW40 flange. I plan on JB Weld around the steel gasket and insulator; and under the steel gasket and the magnet to seal the vacuum. Very little epoxy should be exposed to vacuum, and the brass woven gasket should keep microwaves from reaching the epoxy at all. The tip of the magnetron is still further from the reducer tube than the distance of the insulator.

Thought I'd run it by the group to see if anyone sees any catastrophe problems I've overlooked.
If your experiment needs statistics, you ought to have done a better experiment.

Ernest Rutherford
User avatar
Rich Feldman
Posts: 1470
Joined: Mon Dec 21, 2009 6:59 pm
Real name: Rich Feldman
Location: Santa Clara County, CA, USA

Re: Mounting Magnetron

Post by Rich Feldman »

Very pretty, David. Isn't "serendipity of fit" a delight when it appears?

For the magnetron to oscillate, you need a flux return path from top surface of upper magnet to bottom surface of lower magnet. It must be wide and thick enough. Stainless steel is probably not permeable enough. How about mild steel, which could be plated on vacuum side with e.g. nickel to prevent rusting.

Those ferrite ring magnets are surely magnetized after assembly, as are the ones in loudspeakers.
I don't know if taking the assemblies apart partly unmagnetizes the material, as in some Alnico applications.
And don't know how sensitive the magnetron is to B field strength. Experts?
All models are wrong; some models are useful. -- George Box
User avatar
Mike Veldman
Posts: 155
Joined: Mon Aug 13, 2001 4:38 pm
Real name: Mike Veldman
Location: Stillwater, Oklahoma

Re: Mounting Magnetron

Post by Mike Veldman »

In the majority of microwave applications I've worked with, the probe from the RF source is perpendicular to the waveguide not parallel. You will also need a waveguide sized to the frequency of the RF source. I suspect the arrangement pictured above will cause magnetron heating, possibly to the point of destruction.
I tried to contain myself, but I escaped.
User avatar
Richard Hull
Moderator
Posts: 14991
Joined: Fri Jun 15, 2001 9:44 am
Real name: Richard Hull

Re: Mounting Magnetron

Post by Richard Hull »

Possibly heating to destruction, possibly death from high power port escapes of ideally focused death rays. If one is not a microwave engineer, playing with maggys outside of tuned, shielded cavities with controlled VSWR, etc., you are playing with a loaded gun. The ideal situation............Your maggie burns up or your supply catches fire and you live to tell the tale over a camp fire.

Richard Hull
Progress may have been a good thing once, but it just went on too long. - Yogi Berra
Fusion is the energy of the future....and it always will be
The more complex the idea put forward by the poor amateur, the more likely it will never see embodiment
User avatar
Bob Reite
Posts: 576
Joined: Sun Aug 25, 2013 9:03 pm
Real name: Bob Reite
Location: Wilkes Barre/Scranton area

Re: Mounting Magnetron

Post by Bob Reite »

JB weld, as I recall is non-conductive. You need to have an intimate connection between the magnetron body and the flange you are using, or it will leak microwave energy. Finger stock might be better if you can find some fine enough for 2.4 Ghz. Maybe stainless steel bellows.
The more reactive the materials, the more spectacular the failures.
The testing isn't over until the prototype is destroyed.
prestonbarrows
Posts: 211
Joined: Sun Jun 24, 2012 1:27 am
Real name:

Re: Mounting Magnetron

Post by prestonbarrows »

This will not work well at all.

The cutoff frequency for circular waveguide is f_c = (1.8412 * c) / (2 * pi * r)

For a 2.75" CF nipple, this gives about 2.8 GHz. A kitchen magnetron is going to put out 2.45 GHZ ideally, a frequency lower than the cutoff of the tube meaning that you will not get propagating waves. The wavelength is essentially larger than the tube and can not fit through it.

And that is in the ideal case where you have the magnetron properly coupled to your circular waveguide, which is absolutely not the case in your pictures. RF is very sensitive to structures on the order of its wavelength, in this case a few centimeters. You can't just arbitrarily strap RF components together and have power be transmitted properly; very specific geometries and positioning are necessary. You would probably get some radiation past the nipple into the chamber by virtue of it being so short, but the large majority of the output power will just bounce off the cylinder right back into the magnetron and burn it out.

As others have mentioned, the best case you can expect from what you have shown is to quickly fry the magnetron without burning your house down or cooking your eyeballs. What are you trying to accomplish here anyways?
David Kunkle
Posts: 284
Joined: Thu Apr 17, 2014 12:43 pm
Real name: David Kunkle

Re: Mounting Magnetron

Post by David Kunkle »

After some experimenting, pretty much everything everyone posted is correct. It will not work without the "top plate" that I removed. It does work with the SS fitting in place, but the output is significantly reduced. I only need to run it for seconds at a time, but I would rather have more breathing room before it burns itself up. I had seen a post about using a 2" copper pipe, but no longer seem to be able to find it. Is 2" large enough to transmit microwaves? I'm certain that I can seal the whole thing against vacuum after my little experiments with the magnetron, but if 2" will work, I'll have to tear down the chamber again to weld in a nipple with a 2" diameter tube. 2 NW50 nipples would do the trick.

I was using Preston's equation, but for some reason I can't get what looks like a correct answer using c as speed of light in m/s, and r as radius of pipe in meters.

No signs of it turning into a death ray yet- although there's always hope.....

"If only that stupid death ray had worked, there wouldn't have been any ass whuppins."
Sheldon Cooper, Ph.D.
If your experiment needs statistics, you ought to have done a better experiment.

Ernest Rutherford
prestonbarrows
Posts: 211
Joined: Sun Jun 24, 2012 1:27 am
Real name:

Re: Mounting Magnetron

Post by prestonbarrows »

David, we can check the units on the equation in the last post to make sure everything works out. speed_of_light / radius => [m/s] / [m] => [1/s] otherwise known as Hertz, the unit of frequency.

You can further simplify by lumping the constants together with c approximated as 3*10^8 [m/s]. So the lowest cutoff frequency for circular waveguide to just f_c = (8.79*10^7) / (r) where r is the interior radius of the cylinder in meters.

Plugging in the dimensions for 2.75" conflat tube gives 8.79*10^7/0.031242 = 2.81352*10^9 [1/s] or roughly 2.81 GHz.

A proper mathematical treatment for waveguides gets complex fast. Basically where these numbers come from is solving Maxwell's equations with the boundary conditions of zero electric fields along the conducting surfaces of the waveguide. The constants pop out of Bessel functions which describe the fields.

Basically, the cutoff frequency is where the wavelength would go to infinity which can't happen in the real world so you can't have a real wave under those conditions.

Studying up on 'waveguide modes' will help you understand a bit more about what is actually going on and why you need certain special geometries to couple a magnetron to a waveguide or any other RF device.
David Kunkle
Posts: 284
Joined: Thu Apr 17, 2014 12:43 pm
Real name: David Kunkle

Re: Mounting Magnetron

Post by David Kunkle »

Thanks for setting me straight. Simple enough equation- I just had a brain cramp involving a few zeros. Duh.

The actual inner diameter of a 2.75"CF tube is only 1.375", giving 5 GHz as the cutoff. Working backwards for 2.45 GHz, I get a r of 1.4". So if that's correct, I would need a pipe of about 3" to conduct those waves? In an actual oven, is it the complex shape including 2 "dimples" in the sheet metal that allow the energy to be conducted into the oven chamber?
If your experiment needs statistics, you ought to have done a better experiment.

Ernest Rutherford
John Futter
Posts: 1848
Joined: Wed Apr 21, 2004 10:29 pm
Real name: John Futter
Contact:

Re: Mounting Magnetron

Post by John Futter »

David
No the microwave is fed with a classical probe fed waveguide with the probe the thing sticking out the base of the magnetron being a quarter wave from the shorted end.
There is also an impedance matching section as it enters the microwave via that mica sheet.
You can fool your tube section into looking larger in diameter by dielectrically loading it with something other than air.
I have used teflon sheet rolled up and stuck in an undersized waveguide to do this.
Teflons dielectric is 2 so something better like aluminium oxide which is ten would be better but flat sheets of this are expensive and brittle
David Kunkle
Posts: 284
Joined: Thu Apr 17, 2014 12:43 pm
Real name: David Kunkle

Re: Mounting Magnetron

Post by David Kunkle »

When you say Teflon sheet rolled up and stuck in a waveguide, do you mean just lining the inner surface of the tube with the Teflon, or are you rolling it up into more of a solid cylinder of Teflon and inserting that into the tube?

I see the dielectric of Pyrex is about 5. Titanium dioxide is 100 and cheap enough, but what to do with all that powder?

Thanks.
If your experiment needs statistics, you ought to have done a better experiment.

Ernest Rutherford
prestonbarrows
Posts: 211
Joined: Sun Jun 24, 2012 1:27 am
Real name:

Re: Mounting Magnetron

Post by prestonbarrows »

David Kunkle wrote:The actual inner diameter of a 2.75"CF tube is only 1.375"
Good catch. Dumb mistake on my part, I read the wrong number off Lesker's tables. A smaller waveguide means higher frequencies.

A microwave oven generally is set up with a makeshift rectangular waveguide of cheap sheet metal with the stub on the magnetron mounted through the wide side 1/4 wavelength from the closed (shorted) end. This is a standard transition from coax to rectangular guide.

Rectangular waveguides have a whole different set of equations but the principle of a cutoff frequency and smaller waveguides needing higher frequencies still hold true. When using dielectrics in the waveguide, you can get significant losses depending on the material. This shows up as heat in the dielectric and can lead to melting if not well cooled (easy to do in a vacuum).
David Kunkle
Posts: 284
Joined: Thu Apr 17, 2014 12:43 pm
Real name: David Kunkle

Re: Mounting Magnetron

Post by David Kunkle »

I'm still interested in using dielectrics in a waveguide. Losses probably won't matter. I'm already using a 600W magnetron which will probably be overkill even if I only get a fraction of the energy into the chamber. Is the dielectric used as a liner in the waveguide, or is it loosely packed in there, or something else? Thanks.
If your experiment needs statistics, you ought to have done a better experiment.

Ernest Rutherford
John Futter
Posts: 1848
Joined: Wed Apr 21, 2004 10:29 pm
Real name: John Futter
Contact:

Re: Mounting Magnetron

Post by John Futter »

David
The dielectric is worked on by the e field which is present so it is a volume percentage thing nothing to do with the internal surfaces.
Preston
Teflon has the lowest loss tangent of any dielectric almost immeasureable at 2.4GHz

David if you really want to get the microwaves into the chamber the way you suggest, all you have to do is extend the probe on the magnetron. Using any number of half wave extensions using your 1.33" tube as the outer you have just made some coax. If you knew the impedance of your chamber you could taper the extension so you have a transmission line transformer to match the magnetron impedance
David Kunkle
Posts: 284
Joined: Thu Apr 17, 2014 12:43 pm
Real name: David Kunkle

Re: Mounting Magnetron

Post by David Kunkle »

Finished mounting the magnetron to the chamber. Covered the big viewport with microwave oven screen. Both are on the far side of the chamber. Seems to hold vacuum just fine. When I turn it on, my leak detector stays well below 5 mW/cm2- at least where I stand. When I dangle the detector into the chamber and switch on the magnetron, the alarm goes off immediately and it shuts itself off. So I know I'm getting plenty of microwaves into the chamber. Only downside is the camera goes blank which is right up against the screen covered viewport.

I've searched the forum on ways to limit the power output of a magnetron, and it seemed not doable or very complex. I happened to run across this site today: http://www.cyclotrons.net/showthread.ph ... ton-source

Interesting part for me was this:
"First, find an old 'working' microwave oven. This must be the older type with a big, heavy MOT (microwave oven transformer) not the newer 'inverter' type. (I'm working on a method utilizing the inverter type, and I'll post details later, but it is much more complicated)
Remove the MOT, MOC (microwave oven capacitor), HV diode and magnetron. Take care not to damage the gasket seal around the waveguide on the magnetron.
Discard the MOC, but replace it with one about a 50th of the capacitance. The voltage rating must be the same or better than the original.
Using a smaller capacitor reduces the power available to the magnetron (reduces the current, to be precise, and it is this that controls the magnetron output. Fig.3 in the paper linked to here: http://accelconf.web.cern.ch/accelco...ERS/TH4055.PDF demonstrates this).
This achieves three things,
1.It allows you to run the magnetron in CW mode (constant waveform), rather than the intermittent mode it runs in in the oven.
2.It reduces the power to ~20 Watts, which is plenty for this purpose.
3.It makes it much, much safer.
Connect these items together as they were connected originally, but using the smaller capacitor. (details of these 'doubler' circuits are plentiful, use google)"

My magnetron is out of an el cheapo, but new $50 700W oven. He talks about new ones using inverters. Not that I'm an electrical engineer, but mine certainly looks and feels to me like a big, heavy MOT with the usual wiring, diode and cap I've seen in all the diagrams. Is it still possible I have an inverter?

The original cap on mine is 2100V and .75uF. I've found a cap on ebay that is 2000V and .015 uF. (next best I could find was 3000V) The 100V difference won't matter much, right?
If I follow the above correctly (assuming they're talking about a 1000W oven), this would give me 14W output instead of 700W which I figure is enough to do the job.
From what I've read, it seems too simple. Is this post full of crap? Looks like I can find out for sure for $6.50.
If your experiment needs statistics, you ought to have done a better experiment.

Ernest Rutherford
User avatar
Bob Reite
Posts: 576
Joined: Sun Aug 25, 2013 9:03 pm
Real name: Bob Reite
Location: Wilkes Barre/Scranton area

Re: Mounting Magnetron

Post by Bob Reite »

So you got it to work without killing the Magnetron due to a mis match? The best way to do it if cost is no object is to hook the Magnetron to a three port circulator, with the third port connected to a dummy load. That way, any reflected power due to a mismatch goes to the dummy load and not back down the magnetron. If you want to get fancy, after the circulator output, put a directional coupler then a tuning box. You can adjust the tuning box for minimum reflected power. If you want to reduce power, you can deliberately mis tune it; the magnetron will be protected by the circulator dumping the reflected power into the dummy load.
The more reactive the materials, the more spectacular the failures.
The testing isn't over until the prototype is destroyed.
David Kunkle
Posts: 284
Joined: Thu Apr 17, 2014 12:43 pm
Real name: David Kunkle

Re: Mounting Magnetron

Post by David Kunkle »

Yeah, seems to work OK so far without blowing up. Though I haven't run it for an extended period, but I plan on not having to. I follow you on the circulator, but cost is at least partly an object. I have that smaller cap on the way. I'll try that first and report back on how that works out. If it doesn't, I'll probably try to build the circulator next. Thanks.
If your experiment needs statistics, you ought to have done a better experiment.

Ernest Rutherford
User avatar
Rich Feldman
Posts: 1470
Joined: Mon Dec 21, 2009 6:59 pm
Real name: Rich Feldman
Location: Santa Clara County, CA, USA

Re: Mounting Magnetron

Post by Rich Feldman »

What part of a magnetron is at risk from excessive reflected power?
Perhaps David just needs an anode temperature monitor.
Attachments
thermom2.jpg
thermom2.jpg (27.67 KiB) Viewed 7320 times
All models are wrong; some models are useful. -- George Box
prestonbarrows
Posts: 211
Joined: Sun Jun 24, 2012 1:27 am
Real name:

Re: Mounting Magnetron

Post by prestonbarrows »

A magnetron is basically an arc discharge between coaxial electrodes inside a magnetic field that causes the electrons to orbit in a circle on their way from the cathode to anode. A set of specially tuned cavities interact with the electrons as they spiral around and push them into bunches. Each time a bunch passes, the slow ones get sped up and the fast ones get slowed down. The shape resonates at one specific frequency like blowing air over a half empty bottle. One of these cavities is tied to the output and as the bunches pass, you get EM output at that frequency 2.45Ghz for all kitchen models.

A standard kitchen style one is something like 50% efficient; so about an equal power gets delivered to your load (food) as to the anode. This is why that section is buried under heat sinks. You typically can not change the power output of this style, only change the duty factor while still at full output.

With high reflected power, you get two main things happening. First, the reflected wave is generally out of phase and at some reduced amplitude. Remember this is directly connected to one of the cavities; the returning wave also yanks around on the electron bunches but is not matched up with the other cavities and tends to scatter the nice electron bunches. Generally, this will cause the frequency to shift and hurt efficiency. Neither of these points are very tightly regulated for cheap kitchen models anyways. Secondly, you will get standing waves of voltage/current in the microwave structure when waves are traveling in opposite directions. This causes localized spots of extra resistive losses which need to be dealt with. Again, kitchen models are basically made as cheap as possible, so you probably don't have overhead on the built in cooling system.

This is basically what happens if you run with no load (food) in the microwave. The hardware will get hotter than normal, probably to the point of catastrophic failure if unchecked. But, there is nothing fundamentally that would cause the magnetron to instantly explode.
GE Appliances wrote:Do not operate the oven while empty to avoid damage to the oven or the possibility of a fire.

If by accident the oven should run empty for a minimal amount of time (approx 5 min), no harm is done. Do try to avoid operating the oven empty at all times--it saves energy and prolongs the life of the oven.

However, if running the oven empty should occur, it may overheat and shut down. Allow the oven to cool down and reset itself. Try using the oven again normally to make sure it is heating.
David Kunkle
Posts: 284
Joined: Thu Apr 17, 2014 12:43 pm
Real name: David Kunkle

Re: Mounting Magnetron

Post by David Kunkle »

It does work! Swapped in the 2kV, .015 uF cap instead of the factory 2.1kV .75 uF. Started it up and got nothing with the leak detector where I'm normally standing at the controls. Well, I guess this doesn't work at all. Crap! Then decided I should try around the back where the viewport is. Shoved it into the screened viewport and started getting a reading! Obviously the magnetron is putting out a much lower amount if I have to stick it next to the viewport as opposed to getting the same reading on the other side and 6 ft. away.

I also discovered there is a little radiation that seems to be coming directly off the magnetron, but the majority of leakage is coming out the ion gauge! The mounting starts as a 2.75" CF nipple on the chamber, then a 90 degree CF elbow, and the ion gauge mounts to that. I may try wrapping the magnetron and gauge in something that will absorb microwaves (any suggestions?).

According to the link, this should be reducing the output by a factor of 50- meaning I've reduced output to 14W. Seems plausible given the reduced readings I'm seeing. I had to buy 5 of these caps. Now I figure if I want to increase the power, to say 28W, all I have to do is wire another one in parallel with the first. The page also claims this converts it to CW mode. I was thinking this is the same as continuous output?- which would be another bonus for me. I've studied another page showing how the doubler circuit works in a microwave, but not sure how a conversion to continuous output would occur. But, seeing how they were right about swapping a smaller cap to reduce power.........

Thanks to Preston's last post, if a microwave can run at full power for 5 minutes without damage, this magnetron reduced to 1/50th power should be able to run indefinitely- far longer than I need. I may even scrap putting a cooling fan on it. I'll have to measure it with my infrared over a period of time.

I have just confirmed this solves another problem for me. Before with the magnetron on, the camera would black out. Just tried it with the reduced power magnetron, and the camera works! AWESOME! There's just a little interference, and the picture is almost the same as without the magnetron running.
If your experiment needs statistics, you ought to have done a better experiment.

Ernest Rutherford
Post Reply

Return to “Vacuum Technology (& FAQs)”