Experimenting with waveguides using the LiteVNA

[joeqsmith]

Some time back I had read George C. Southworth's  Forty Years of Radio Research, published in 1962.   This book covers the history of waveguides.  I've often thought about experimenting with rectangular waveguides at home but the lack of a need and high costs of even used parts deterred me.   Outside of putting together a circular waveguide antenna from pork & bean cans, I haven't progressed.   

With the LiteVNA now having supporting a harmonic mode above 9GHz, I am once again thinking about it.   Many common sizes available far exceed the frequency range of the LiteVNA.   This company offers various surplus parts.
https://www.surplussales.com/index/RF-Waveguides.html

Link of waveguide dimensions.
https://www.everythingrf.com/tech-resources/waveguides-sizes

With ham bands providing  5.65–5.925 GHz, 10.0–10.5 GHz & 24.0–24.25 GHz, I tried some searches to see if anyone was playing around with it but didn't really find much out there.   

http://www.ham-radio.com/sbms/presentations/Walt_Clark/DROplexer.pdf

I'm curious if any members here are experimenting with them?

[Mechatrommer]

good luck digging the wormhole, we will follow where it suits... waveguide practicality only in tens of GHz domain and up, below that you can do in pcb (from what i read) but there is a DIY SA using sort of wave guide/chamber/tank at 1GHz region so you may want to look it up its Scotty's SA i think https://scottyspectrumanalyzer.us/ edit: yup it is cavity filter... https://scottyspectrumanalyzer.us/cavity.html

[rf-messkopf]

Tobias Plüss from Switzerland has built some pretty impressive waveguide devices, such as a bandpass filter, a directional coupler, and a termination. Also a cavity filter. For such projects you not only need the requisite test equipment, but also a well equipped machine shop, plus the required machining skills.

[joeqsmith]

http://www.w1ghz.org/QEX/Rectangular_Waveguide_to_Coax_Transition_Design.pdf

http://www.bobatkins.com/radio/10GHz_diode_detector_mixer.html

[coppercone2]

I saw that waveguide mixer on some old webpage to link a shed to a house with a 10GHz RF link IIRC

[joeqsmith]

Waveguide primer

https://www.microwaves101.com/encyclopedias/waveguide-primer#introduction

Contains lots of Rules of thumb

Very rarely does something in microwaves work as it was designed. Tuning waveguide structures requires some tricks. One such trick is to use a steel ball bearing inside the structure, that is moved around using a permanent magnet from outside the waveguide, while you monitor the part's performance using test equipment with signals applied. Once you find a spot that improves performance, mark it with an "X", then you can either drill and tap it and insert a tuning screw, or it's "hammer time" and you can use the concept of "dent tuning."

[coppercone2]

btw I had utterly horrible luck buying parts on ebay like adapters between waveguides the terminology is basically butchered to hell to the point that for small parts I really wanted to make castings out of silver

when you get to the gasketed parts its a total mess, buyer beware. I think relatively few people know the proper nomenclature of these things so you are going to have a serious struggle.IMO its worse then stuff related to pipes.

[joeqsmith]

I've been looking at photos of used parts.  Much of it seems like scrap metal.  I'm thinking small signal experiments with the VNA without any plans to gas fill. Flat flanges only. 

Looks like kirkby offers some waveguide cal kits.
https://www.kirkbymicrowave.co.uk/Support/FAQ/How-do-I-perform-waveguide-calibration/

[coppercone2]

yeah but alot of the good deals you can get is old airplane stuff from like cargo planes etc that have gaskets even if you don't need them so it will end up being tempting for alot of people. Or navy stuff that is also gasketed because of the ocean

[joeqsmith]

I saw what looked to be a cal standard last week.  No idea what size.  Seller had no information.  It was sitting at under $20 with three bids.    I just looked for it again but looks like it was sold.  

[coppercone2]

there is alot of weird educational things for waveguides for teaching RF physics (still used today). The idea was to teach the intro to electromagnetics class with demos. If you see very odd measurement aparatus it might be stuff meant to teach first principles of RF, not for industry+ use, e.g. the measurement is made in a round about way so the low order equations could be used to perform a lab (often very mechanical). Also hard to destroy and you might actually catch the person running with it since its heavy, the waveguide is used to teach reflection, I think I saw a catalog of those things some where (so the professor can do a sanity check with 'standards'). Easy to chain it down too. Back when the 141T was a hot commodity the students in the state schools/community colleges (when electronics vocational training was a thing) had some seriously strange apparatus (which could probobly be repaired/manufactured/supplemented by the university machine shop without needing IC etc). Kill two birds with one stone (advanced physics and advanced machining students make the stuff).

I know there are very real serious waveguide related apparatus, but there is also the low accuracy learning stuff, I am almost sure I saw a big catalog full of it where it was marketed as such for much lower prices... and like technican grade stuff for getting things 'in order' before the wizards showed up to get it tuned just right, like during airport construction etc... just so the fitters don't get it totally wrong.

Waveguide VNA calibration is interesting, that is getting into TRL type stuff, state university will not give you that.

[TWMIV]

I did my Masters work in synthetic aperture radar using waveguides. There are a lot of interesting things you can do with them. We did microwave NDT and materials analysis, almost entirely with waveguides as they make great antennas in the nearfield.

Have an object of unknown dielectric properties that can fill the aperture of the waveguide? You can inverse the properties of that material out. See doi:10.1109/19.816128 for the basis of the technique. 
You can also extract the permittivity of some planar materials though a SAR scan, but I have not had much luck extracting the loss factor using SAR. I believe the phase difference between measurements was causing issues getting the loss factor, and I had to finish my thesis on a different topic before I explored that any further.

[joeqsmith]

It appeared to have been a few stepped shorts.  No slit and I assumed EDM.   Not sure what else it would have been used for beside calibration.

I would say the educational level is several steps up over where I am with my bean cans.  I had put an offer down on this kit some time ago but no bites.  No real goals in mind outside of playing around.    

https://www.ebay.com/itm/143596987580?_trkparms=amclksrc%3DITM%26aid%3D111001%26algo%3DREC.SEED%26ao%3D1%26asc%3D20180816085401%26meid%3D2cc6f70776bd4fb2a9cbd4e222bf8e46%26pid%3D100970%26rk%3D1%26rkt%3D1%26sd%3D143596987580%26itm%3D143596987580%26pmt%3D0%26noa%3D1%26pg%3D2380057&_trksid=p2380057.c100970.m5481&_trkparms=pageci%3Ae8ec71f6-75ba-11ed-8daf-42aeea92eaef%7Cparentrq%3Ae9b2b9041840a768e9b39eeeffffefc2%7Ciid%3A1 

[coppercone2]

thats nice for a physics lab class thats not garbage

I wonder what that ends up looking like when you get broke ass waveguides instead of a broken breadboard and out of spec transistors with a power supply that is stuck at 6.2V for a microwave class. All the bolts are cross threaded so use these pencils and paper clamps, please take the folded gum wrapper that the TA found in the trash to seal the waveguide structural cracks after the janitor reamed the waveguide that kids were playing stick ball with with a chair leg (9:30 pm friday physics final 30% grade lab)

[TWMIV]

If you just want to get started what you need at a minimum would be a coax adapter, a short (flat plate), a spacer, and a load.

That would let you make calibrated reflection measurements. Add a through section to help kill off any evanescent modes and with a decent vna your in business. With a second coax adapter you can do trl cals in addition to ssol and can play around with cross polarization or transmission stuff.

Horn antennas, and even more complicated structures can be effectively 3d printed and coated in conductive paint at and below x-band. Surface finish becomes more important with increased frequency as you would expect.

We were mounting a tv to use as a monitor in a lab one day and needed to find the studs. Threw an x-band waveguide on a Field-fox (keysight portable vna) and just rubbed it along the wall to find the largest s11 return. That might actually be a fun project to build up from scratch...

[joeqsmith]

Getting started, I wonder if the VNA would even need calibration.     I'm thinking just a couple of transitions to start with.  Saw a matched set yesterday that may be a good fit.

The WR90 is fairly common but I wonder how stable it even is.   Cal the VNA at the ends of the cables with SOLT,  bolt the two transitions back to back and insert them in the path.  Would we expect S21 to be < +/-1dB over the range of the NanoVNA?   

The cutoff of the lowest order mode for WR90 is above  what the Nano can produce without harmonics.  At least with mine, once you get above 8G, it gets pretty bad.  At 9 after cal, S21 is maybe +/-5dB?  Even when using the leakage term, the floor is maybe 15dB down is all at the highest frequencies.   So there's a pretty limited range that it may be able to be used.   I did buy an old PNA and am working on increasing it's frequency range but still, it's limited to 9G.  It's certainly much more stable and lower noise than the Nano though.   

If the WR90 components are typically stable down to their cutoff,  it may be fine.  From what I have been reading, I suspect that some parts may have a very narrow range it can be used and it may be well outside the range of the Nano.  With many parts being old surplus, finding data sheets seems to be rare.   

***
I have attached some data I recently collected that compares the low cost VNA with my used Agilent.  One is looking at a high pass.  The other looking at the floor.

[joeqsmith]

All the bolts are cross threaded so use these pencils and paper clamps

Watched a few videos where they use the paper clamps to hold the parts.  I guess in a lab environment your not expecting a lot of movement and get away with it.  I couldn't find it but here is one where they use some wood clamps:

https://youtu.be/l_hlrQDcJAo?t=152

[coppercone2]

Lol. It might actually be a good project to make special spring clips to hold the waveguide pieces together. Clico clips also come to mind. Good project for wire that’s not make another diy hose clamp. You just need to be sure it does not scratch.

[TWMIV]

Strong clamps work well if everything is aligned. Well fitting shoulder bolts and some clamps will get you most of the way there if it is a temporary setup that isn't going to move.

I used shoulder bolts for anything that didn't have gauge pin holes for alignment.

[joeqsmith]

For everyone's amusement, I attempted to create the worlds worse WR90 waveguide with back-to-back transitions using some FR4.   The seam is in the worse possible location and it's held together with O-rings.  The dimensions are not precise.  There was ZERO attempt to tune it.  I'm not sure what would make it worse.

https://www.rfcafe.com/references/electrical/waveguide-chart.htm

I ran a SOLT cal on the LiteVNA and measured S21 with the cover off, then with it on sweeping from 100k to 9.3G.  It appears to have a 3dB cutoff around 6.9G and remains somewhat flat up to the limit of the VNA (again, the VNA is VERY poor in this region.) 

[joeqsmith]

Running a SOLT from 7 to 9GHz then looking at S21.   It hard to believe any used production made transitions would be this poor.    Gives me some sanity check.  Then again, the Lite certainly wasn't designed for this.  Too bad that PNA is limited to 9GHz.  I wonder where it starts to fall off on the high side. 

[joeqsmith]

My first waveguide experiment.  Took out my hacksaw and proceeded to chop my new WR90 waveguide in half.  The humanity of it all!!  Well, not half.  The base slipped in the vise and it's anything but square.  So much for precision machining....

Two new transitions were placed onto some high tech mounts.   I then measured S21 with a wire polarizer with the wires vertically and horizontally mounted.  Up till about 8GHz, there's a nice 10dB drop.    Like my homemade polarizer? 

[0culus]

Interesting experiments!

I have a growing collection of waveguide components, mostly WR-90 with some WR-62 thrown in for good measure. My most recent "build" is a simple doppler radar using surplus Raytheon circulator and a Scientific Atlanta WR-90 crystal mixer with adjustable stub. The microwave source can be either my 8672A fed through a TWT amplifier to get a little less than 2 watts, or a X band gunnplexer. It's quite sensitive. I've been using my Tektronix 547 with a Type 1A7A high gain diff amp to view the IF of the xtal mixer. That allows me to cancel out 60 Hz noise and use very small volts/div settings.

I always am on the lookout for deals, plus I've gotten some hand me downs from elmers (that circulator mentioned above was a particularly good one to get as a hand me down!) Coaxial transitions, especially for WR-90, are stupidly expensive. I only have one. I have three WR-62 transitions as well as a few for proprietary Litton ribbed waveguide that is similar in size to WR-90 but has a lower cut-off.

[Mechatrommer]

The bare shorting wire in ceramics filled fuse tech that is in waveguide tech i suggest to get a pro/free em solver sw to back it before wasting wr90 materials.. why dont build for lower freq where litevna is most sensitive? Even if you can make a successful tunnel at 9GHz, the other components will be eye watering expensive, ymmv.

[joeqsmith]

The bare shorting wire in ceramics filled fuse tech that is in waveguide tech i suggest to get a pro/free em solver sw to back it before wasting wr90 materials..

While jumping a fuse because we never learn how not to pop them in the first place is one work around, certainly starting out with some scrap materials to run some simple experiments does reduce the costs.   

I did play a bit with some free simulators a while back in response to someone's question on layouts.   The free simulators are very limited and I spent more time trying to work around those limitations than reaching my goal.   So for now, solder and PCB materials are my tools of choice for trying out the basic concepts.  No doubt this will become a limiting factor later on but fine for the basics.       

why dont build for lower freq where litevna is most sensitive? Even if you can make a successful tunnel at 9GHz, the other components will be eye watering expensive, ymmv.

I guess I wasn't clear on that.   I don't see a problem showing the limitations of the LiteVNA.  At $120, I don't think too many people are expecting a lot out of it.  The developers included harmonics to extend the frequency range so customers could play with it.  That's what I intend to do.   WR90 is fairly common in the used/surplus market and falls into that extended range. 

In this video the teacher is running a similar experiment on polarisation.  Beyond talking about the experiment itself, he is also covering  the tools, their age, their costs...   Personally I would get more out of a class using the low cost VNA and some basic materials where I work the math and build the structures, than watching a teacher perform them with purchased parts.