Created
11/04/06

Ever since I took apart my first microwave and got my hands on that little four inch cube that is directly responsible for the magic of the instant noodle cup, I've always wondered - what does a magnetron look like inside? Eventually free time and curiosity coincided and I hacked one apart on my workbench. This was my chosen victim: a Matsushita 2M204-M2. This is the gizmo in your microwave oven that emits ridiculous quantities of radio frequency energy into the cooking cavity to heat your food.

At one end is a sheet-metal box with a crimped cover. I uncrimped it with pliers and removed it to expose the internal terminal connections. Both terminals connect through coil chokes to connections on the inner magnetron tube. If memory serves, one of those connections sits at a potential of around -2 kV relative to the chassis; the other is at only a few volts relative to the first terminal. A large current flows in one terminal and out the other to heat the cathode. Perhaps the chokes are meant to limit the inrush current when the device first turns on? Unless my LC meter is completely worthless (the jury is still out on that one) each coil has an inductance of 3 nH.

The next step was to somehow crack open the tough steel frame. The frame tightly holds together the magnetron tube, two magnets, and several aluminum cooling plates. It is quite strong; the tabs that hold it together could not be budged by my pliers. However, I have a cheapo drill press!

Obviously, the drill won the fight. The silver knob-thingy is the antenna shroud that, when mounted in an oven, protrudes into the waveguide and sends craploads of microwave energy on its way to the cooking cavity. The gold ring is an RF gasket that is made from several layers of fine woven metallic thread, possible brass or something similar. Its purpose is to make an RF-impermeable seal between the magnetron assembly and the waveguide, so that no stray microwave emissions can escape into the guts of the oven and fry the electronics (or escape through vent holes and fry nearby humans.) Beneath that is a thin steel spacer; beneath that, a thick cermet ring magnet. Under the magnet you can see the aluminum cooling fins - these keep the magnetron from melting down and setting your kitchen on fire.

Further disassembly yields this, the true beating heart of the microwave oven. This is the magnetron tube. Of course, it isn't really a magnetron until you put the two magnets back onto it. The super-simplified explanation of how it works is that the tube uses high voltage to accelerate electrons into the vacuum inside; as the electrons travel through the magnetic field produced by the two permanent ring magnets, their direction of travel is bent into a quasi-circular path. As they zip along this orbit they pass by small resonant cavities formed by an arrangement of regularly spaced radial metal vanes, and this produces rapid electrical oscillations in the vanes. This oscillation is tapped and radiated away via an antenna as microwave energy. You'll see the vanes here in a bit when I chop the thing open.

One thing I've always been curious about is the internal structure of the antenna probe. On every model of magnetron I've seen, this probe has a hole in the end, and you can sort of see some blade-like thing down in there, but can't really tell what it is. I decided that it was time I got to see what it is. First, I carefully made a shallow cut all around the probe wall with my dremel. I cracked the probe body off to reveal that the blade-like thing is, in fact, the tube's pinch-off. This is where the tube was sealed to maintain internal vacuum. Glass vacuum tubes sometimes have a pinch-off, or sometimes they have a little 'nipple.' This tube happens to be made of copper, so the pinch-off is conductively coupled to anything touching it on the inside of the tube.

It is apparent, following a deeper cut, that it is touching the microwave antenna. This is how the microwave energy gets from inside the tube to the waveguide probe.

I make a more ambitious cut with the dremel around the circumference of the main cylinder. Being made of copper, the tube gets to be quite hot in the process. I could barely hold onto it with a triple-folded shop rag.

The antenna is rooted on one of the cavity vanes, and exits through a hole in the end cap. The end cap shows some interesting burn patterns. I think that this is probably the visual representation of whatever probability density function describes the impact location of an electron forced to run this crazy gauntlet of electric and magnetic fields.

These pictures simply do not do justice to the immaculate beauty of the insides of this thing. The copper is pristine. I think that the gray disc in the center is the top of the cathode, and it looks like it is probably coated in some material to aid in thermal emission of electrons. The two rings connect alternating vanes together, which supports a resonance condition by forcing successive vanes around the circle to be opposite in phase. That is, unless I completely misunderstand how a magnetron works. Note that this is more than likely.

Well, that's that. I'm out of pictures. It's a lot of work getting inside of one of these things, but it's worth it. The intricate collection of little copper vanes inside a magnetron really is one of the prettiest damn things I've ever seen that wasn't named Katie.