Using waveguide section as cavity

[ChristofferB]

Hi! Sorry for sorta-double posting, but I made a thread about this previously that didn't really convey my question.

I'm trying to make a small 10-12 GHz system that's basically just kinda like a grid dip meter: microwave oscillator feeding a lenght of WR-90/WG16 waveguide, and a directional coupler to couple out the energy from that cavity. See attached.

I'm going to put a sample in the cavity and a large swept magnetic field around it, but that's not important for now.

Question is - If I want this as a resonant cavity, is the lenght of the waveguide of consequence? I've stocked a lot of strange WG components over the years but never actually put any energy in them, so I just want to double check so I don't kill my gunnplexer.

thanks for looking!

[Kire Pûdsje]

Not sure about the setup you are showing.

I assume you want to measure material properties of some sample at the bottom.
If these effects are too small in a normal waveguide run (travelling wave), in general the sample is placed in a cavity. The idea is that the Q of the cavity essentially amplifies the effect of the sample in a normal waveguide by the same factor Q. (Think of the wave bounding around approximately Q times, experiencing therefore Q times the effect of a normal terminated waveguide)

The cavity needs to be closed at both ends. The bottom seems shorted, however the top needs to be closed by either an iris or a post.
The length of the cavity is definitely of importance on the resonance. If you are using say WR90 waveguid, for this frequency range, it will support te TE10 mode. The numbers being the number of half sines along each dimension. For a cavity another dimension is added (the length). Also eg. a TE101 mode. Just use one of the online calculators. The length should be chosen so that the resonance is within your range.

For determining material properties, with a transmission measurement it is in general much easier to determine these, than from a reflection measurement.

Finally if you are worried about the gunnplexer experiencing full reflection, for hobby purpose, just glue a piece of blackened (by fire) wood on the side in the waveguide between the gunnplexer and the cavity, acting as a small attenuator.

BTW, lookup "microwave wavemeter" for frequency measurement.

[ChristofferB]

Thanks for the reply!
Yeah, I'm trying to cobble together an EPR spectrometer. I see, putting in an iris plate sounds like the way to go! - Putting in a small absorber in there might be a good safety feature - maybe a bit of antistatic foam?

I simply cannot find a TE101 calculator, could you perhaps point me to one? The lower section of waveguide is (if memory serves) 12 cm long - that should in my mind be about 4 wavelenghts or 8 ½-wavelenghts at 10 GHz.

For determining material properties, with a transmission measurement it is in general much easier to determine these, than from a reflection measurement.

-Could this be achieved simply by flipping my directional coupler? Another alternative is to put in a 'magic tee' as seen here:




The big drawback is that I have no idea how strong my magnetic field will be (aside from estimating with online calculators) so the resonance frequency is hard to know in advance.
My field will probably be about 6000 gauss, but it's hard to tell.

http://www.pci.tu-bs.de/aggericke/Lehre/Spektroskopie/ESR_NMR/ESR.htm

this is the article I'm loosely following, though stepping up the waveguide game a bit - in that article, the entire assembly is the cavity.

thanks for the interest!

[Kire Pûdsje]

The linked ESR story uses a single cavity for both the diode and the sample.
your image shows another setup using reflection measurement.
In the attached image I roughly tried to show what happens in reflection and transmission measurements. In reflection it will be much harder to determine the Q. This can realistically only be performed on well calibrated setup.
In transmission, only the shape is important. No need to have well calibrated losses. just a matter of measuring the 3dB points.

Since total Q of a setup is all separate Q's in parallel. Therefore the loading due to coupling should be kept very small.
in a transmission you need to couple energy to the cavity and extract it through another exit. As shown in the lumped, the ring resonator setup (microstrip) and the waveguide image. The idea stays the same.

The linked ESR story talks about maximizing coupling to the E-field. However there the sample is placed in a location where the E-field is minimum.

[CD4007UB]

To add to Kire's advice, the setup you're considering is similar to one of our student experiments on high-field ESR at 10GHz. Our resonant cavity is made from a piece of waveguide with an adjustable plunger at one end (to tune the resonant frequency). The sample is inserted through a small hole in the E-plane (thin side) at the centre of the cavity. The cavity input consists of a sheet of copper (sandwiched into the waveguide joint) with a small hole in it to couple the cavity to the main waveguide. The size of the hole is adjusted empirically for minimum reflection from the cavity ('critical coupling', which is similar to impedance matching for a radio antenna). We originally used a magic tee, but swapped it later on for a 3dB directional coupler. We also used to use a Gunn diode oscillator (which replaced an old klystron, as in Kire's diagram) and then moved on to a tunable YIG oscillator (sometimes see them cheap on eBay).

For ESR at 10GHz, you need quite a strong (about 0.35 Tesla) and homogeneous magnetic field. Our electromagnet weighs >100kG - not the sort of thing you'd find outside a physics lab. For homogeneity, you need large-diameter pole pieces (ours are about 6" wide) and a small gap. We use a test sample of DPPH, which gives a very sharp ESR line with a known g-factor (to calibrate the field). A YIG oscillator allows you to wobble the microwave frequency and you can then see the cavity resonance (using a didoe detector and an oscilloscope) and estimate its Q. With a Gunn diode (which has its own resonant cavity and is suceptible to back reflections) you can modulate the amplitude but not the frequency - they are also easily damaged.

We also have a low-field ESR setup. It uses a marginal oscillator at 60MHz (again, with DPPH). It's much, much simpler to make and only needs a weak field (~2 mT), which can be produced with a homemade solenoid (~500 turns in 6"). It's worth considering as an alternative to working at 10GHz. (Our marginal oscillator is a very old design, and if we were building it again we'd probably use a more modern Robinson oscillator.)