Gathered in 2012 off line from Herbert Krause Micro-Mechanik web site. He is now a silent key and the web page is gone from the internet.
Its a Hi Q Filter for milimeter bands designed by OE9PMJ, this sketch is dated June 1989.
Hohlleiter-Anschluss means wavegude aperture. The dimensions are all in millimeters.
Translated by babelfish:
Watch the location of waveguide joints!
At the turn the screwdriver vote fo recovered!
Kcomm about f0/2 1 eliminiest's to the about talk by the E0A resonance, this was however by the waveguide cutoff .
According to KØCQ:
Watch the location and orientation of the waveguide flanges.
The screws set the resonant frequency.
There can be leakage at ½ the resonant freuquency at a fundamental resonance of the cavities, but the waveguide cutoff frequency prevents any leak through from being a problem.
Prepared November 10, 2015 by KØCQ
Extended November 11, 2015 by KØCQ
In DUBUS 2/2009 DL2AM reported building a scaled OE9PMJ dual cavity for 76GHz. He found he had to enlarge the aperture between the two cavities to get the insertion loss under control. He also found for use and for measuring he had to use isolators before and after the filter. He reported insertion loss when adjusted was 3.4 dB on 76.032 GHz. Part of that was probably the pair of isolators.
Some notes on fabrication:
I'm sure the thickness of the web between the two cavities affects the coupling, so if that dimension isn't held tightly, there may be a need to adjust the diameter of that aperture. Same thing for the webs coupling to the waveguides. To keep those under control, drilling the cavities in the center (I'll call it the “cavity plate”) plate the positioning is critical. If drilling with common 118º drill bits, a pilot hole is crucial because that drill bit grind doesn't cut at the center so tends to run away unless there is a divot wider than the drill bit web thickness. A center drill is very stiff and one good way to mark holes. In England its called a Slocombe bit. A spotting bit is a better solution but harder to find. A 135 degree split point bit cuts to the center and doesn't wander. Its worth the bother of finding because it doesn't need a pilot bit hole,
Plan to drill the cavity holes undersize. The loss in a cavity resonator depends on the surface resistivity and that is much higher for the rough surface a drill bit leaves behind. It may be pracitcal to use boring bar in the milling machine to get the holes closer to the design diameter.
After drilling, finish to the final size with an adjustable straight reamer or by sanding to size. To sand to size turn or acquire a dowel, metal rod, or metal tube that just fits in the drilled hole, Preferably with enough clearance for two layers of wet or dry carborundum paper or emery cloth. Cut a slot in the tool lengthwise with a Zona saw or jeweller's saw to take a strip of the carborundum paper or emery cloth. Make the slot longer and the strip wider than the thickness of the cavity plate. Start with about 100 grit, spin with a drill to polish and enlarge the brass to the desired diameter. It may require lengthening the strip to add thickness to reach the desired diameter. Finish with about 400 grit to polish. Much finer is possible these days..
I'm not sure about the cavity mode used here, but typically in a right circular cylinder the current in the wall is greatest at the end plate. In this design that's not a solid connectioin, but based on pressure between the three plates. If the contact is imperfect the cavity loss will be increased. Silver plating may help, but making sure all four surfaces are truly flat should help. In the extreme one would want the rim of the cavity hole to be a few thousandths of an inch higher than the rest of the cavity plate so that flat cover plates will concentrate the contact pressure at that rim. That's not easy to accomplish. Flat is easier to accomplish by using the 400 grit carborundum paper supported face up by a flat surface, like a surface plate or plate glass. Wet is handy and it will make the surfaces flatter than the milling machine or factory roll will do. Be careful about dragging fingers on the wet carborundum paper, there's no pain as it takes away flesh. I learned that using that technique to grind quarts crystals eons ago. Valve grinding compound on the flat surface also works but is messier though it won't grind the finger tips.
The plates could be soldered together, but any solder inside the cavity has higher resistance than the brass or copper and its hard to clean out excess solder after the second plate is soldered in place.