Experimenting with waveguides using the LiteVNA

[coppercone2]

released from cold war agreements.......

I think someone better call the dog catchers on them with what I have been reading in the papers for the last two years. I keep seeing rehashes of the Adams family, Russian aggression and China paranoia. Thought it was alternate reality 1950. And all the... cures.. that got rid of microwave backbone are going bad too, banned compromised networking/comm equipment, extreme ASAT weapons proliferation, rapidly accelerating third world nuclear weapons development (Iran/NK), cyber attacks on infrastructure (and these are the drunken uncoordinated hit with a 'foam pool noodle' kinds) etc.

Heartwarming tale about setting up microwave links to broadcast freedom
https://tech.slashdot.org/story/23/03/08/2224256/taiwan-suspects-chinese-ships-cut-islands-internet-cables#

[joeqsmith]

My attempt make a 1/4 wave offset short for a calibration standard.   Also shown is my short.  Bought a decent load from Narda.

Showing VSWR measurements of the Narda sliding load (1.353) and two of the low end loads from LRL.   

I tried to measure the two LRL loads directly with the LiteVNA.  Shown compared with the data from the old Agilent.   

[0culus]

I have been following along with some interest. It's pretty neat what you've achieved here.  I have not attempted to roll my own yet. However, I recently acquired some more surplus WR-90 waveguide pieces, so I've finally got enough to do some experiments. I also recently acquired a partial HP WR-28 cal kit for a song because some !@$#%@! robbed the precision load and offset shorts (and the floppy disk) out of it. I got three pieces of WR-28 line and the tools though!

[coppercone2]

I removed like 4 garbage bags of rubbish from the basement and cleared most of the tables, ready for another print and to make a little shelf to hang in front of the VNA off the rack. Amazing how much trash repairing that equipment generated. Screws all over the floor, component leads, baggies, q-tips, wipes, solder, sand paper, blocks, sponges, jigs, panels, cut components etc. I just let it accumulate when I was trouble shooting some of the stuff, mixed with furnace parts, plumbing parts and turnings and wow that got out o hand.

Step 1: ability to place cal kit on flat surface instead of broke equipment stonehenge.

[coppercone2]

are you doing experiments with dielectric fill or something btw? looks like a gas syringe.. sf6?


I also wanna make some waveguide stands now when I saw your setup. I wonder how hard it would be to get useful dimensions that interface well with many parts. I am just thinking to glue wood blocks together to make loose fit ones. I hate those long setups with coaxial parts, stuff is ends up floating at weird angles. A good one is like a welding table almost.

Did anyone that worked for commercial endeavors on wave guide have such a setup ? Like a grid/optics table dedicated to microwave parts setup? Once you accumulate enough waveguide parts it seems reasonable to think about development of such a thing.

[joeqsmith]

There was a couple of complete sets of Anritsu calibration standards on eBay for cheap.  One set was for WR90.  The old Agilent uses a 1/4 offset.  The Anritsu set came with the 1/8 and 3/8. 

The stands that came from LRL are nice and heavy.  They have three rubber feet.   Downside, they can't be adjusted, can't fit next to one another with the lip at the base, and of course support WR90 only.   Makes sense as that's what the training kits all used.  ARRA made a much nicer university kit.   You can buy stands, or I've seen some 3D models you could print your own. 

First Google hit:
http://www.lteq-microwave.com/waveguide-stands.html

[joeqsmith]

The Keysight PN# 11664-20021 1/4 shim is 9.78 +/- 0.02mm thick.   

[coppercone2]

What is your bath chemistry? I have gotten copper sulfate

Did you build a hull cell for testing the plating ? (basically a tapered chamber that allows for a field gradient on a single plane to extrapolate plating current)

What I came up with is to make plastic welding easier, is to use a thicker glue to put the cell together out of acrylic, and then place it at an angle then flow liquid silicone in there, let it cure to tacky, then rotate the cell to the next edge and repeat till there is a bead of filler on the inside, because gluing those things tight is hard. For calibration purposes it should hold IMO, only a few hours probobly. And to use thick materials for the base. 4 hour dwell time should seal any gaps with the liquid silicone. I did this for a rather heavy built photo development tray, since its heavy you can sway the liquid with little risk of spilling, if its light you are liable to get splatter if you make a mistake, but it weighs several pounds. Dampens your body movement. For those without a PhD in aquarium building.

[coppercone2]

started the second horn print, hopefully the new spray can of silver paint will work nicely. This time I will paint the apature with a paint brush before I spray paint it, to try to maximize surface quality

[joeqsmith]

Used copper acetate.   I have not made anything beyond what I have shown. 

I thought about running a few samples to get an idea of the thickness, for the skin depth.  Instead, tried a test sample and saw within a few seconds @ 500mA it had taken.  Left the part in the bath for about a half hour and called it good. 

[coppercone2]

[coppercone2]

Have you looked for leakage at those plastic wave guide joints? I plan to test that also, I wonder if there is a need for gasketing or something.

If you hook it up to a low power level (maybe through a attenuator) and place it on a copper sheet, it should have bad reflections but also allow you to measure leakage with another horn. just need to make sure equipment wont be damaged. Would be interesting to know how much is leaking out, rather then just being dissipated on the non ideal surface. If the 'end cap' leaks too much, I guess you need to make a conductive opposite piece to plug the horn with... if it has enough over lap that should work even if not ideal I believe..

I also think I can sand the launcher face from the nickel painted one and give it a silver paint over it. I don't think the paint should be too thick at the interface, but I don't think it will cause too much a problem on the rest of the horn. Definitely want to keep track of undesired exposure...

This might be a good opportunity to test high silver content silicone gaskets. Easy to make a little mold with the bare horn. I think a proper woven one would need to be laser cut or something (I think that material exists, kind of like very fine braid)... maybe they sell specific parts too.

Also had some success cutting the sanding sticks with a exacto (takes a while to cut through the wood strip) to make narrow sanding sticks to try to sand the interior of the narrow portion of the waveguide.They don't sell ones that have a fine grit, so you can use a cutting guide and keep cutting with a fresh blade.

[joeqsmith]

Outside of running spherical sweeps in the far field, no.   As part of that testing, I did compare the printed horns with a production part.   I ran the sweeps using the LiteVNA with my poor man's extender at roughly 10GHz.   

I could certainly take the best horn of the group, tape it off with copper foil to seal it as you suggest, and repeat the sweep using my old Agilent VNA.   It's limited to 9GHz but has a much lower noise.   

I wet sanded the flanges on each horn.   Consider the skin depth and the manufactures thickness for the coating (datasheet) along with how many layers I applied.   I never used any sort of gasket.   Flanges were bolted together.  I did go over the choked flange design. 

Once you have yours built up and tested, let me know your results and I can try to replicate them with my setup.  Keeping in mind, everything I have shown is in the X-band.   I would expect things to get worse at shorter wavelengths. 

[coppercone2]

well the usual happened with MG chemicals spray products, they work like shit. I damaged the horn while cleaning it after a failed paint job because I dropped it, so I would have to make a new one. The can was heated, shook for 2 minutes, and then clogged 1 seconds into the work. Also unlike the nickel paint, this one does not transfer at all with a manual brush, the nickel one you can kind of mis use as normal paint if you are quick.

Now 3/3 MG chemicals spray product (nickel paint, silver paint and ESD dissipative coating) all had severe problems during application, even when purchased from digikey. And leaky cans

Not going to try again until I find some kind of paint product that I like.

There is some kind of problem with surface tension and sharp edges too, but I will not even experiment when the aerosol flow is so dodgy

Only MG chemicals spray product I recommend is alcohol. Maybe someone will figure out a 'consumer grade low skill' spray on product in the next 10000 years. Paint with out metal particles is slightly easier, but still.... they uh... over do the easy portion of this... too much art involved, and the rate of cash flow out of my pocket is way too high when I am misting fucking silver.

And this can only works if you shake it while you are spraying, you have about.. 1/3 second after you stop shaking that it messes with the flow. Maybe if you add a vibra motor to the can it would work better IDK,or you get a musician that treats it like a maraca or whatever, and you get random clumps of silver

The usual shit on friday when you think you can finish something off eh, feel the same way about this as I do about aluminum brazing  :'(

difficulty level > pizza dough

I almost want to try again, but this time rough up the surfaces to 60 grit, and do enough coats to smooth it out, but probobly that will end badly too.

[joeqsmith]

I painted all those test horns with several coats each and have maybe a half can left.   

[coppercone2]

well I dunno, I managed to do the last one with nickel paint, the silver paint is just extra evil. The second I saw what the can did I knew its over

Some people are not good at math or throwing things, but I am not good at dealing with pastes and paints.

At this point a kiln is gonna be cheaper.

[coppercone2]

I think the best solution might be hybrid.

Cast the end launcher in a open mold with a minimum of silver, then glue the 3d printed horn to it with epoxy. Then you can do a sand casting, with a regular 3d print, so long you can machine the top flat. No aperture to paint. The whole thing is hard to cast probobly, and expensive.

Infact I think I can cut the end of basically any wave guide and glue a pyramid to it..

I will see how many ml of silver i need, maybe its small enough that I can melt it with a torch in a small crucible and pour it open into a sand mold made with a 3d print.

[joeqsmith]

My only goal was to get a feel for if 3D printed plastics was a viable option for making prototype waveguides.  The horns were are a simple enough shape to try out.  I have no use for them beyond the experiments shown.   

From my previous posts, attempting to use the LiteVNA to directly measure VSWR, even with the offset load standard wasn't very good when compared with my old Agilent.  When using harmonics above 6GHz, the LiteVNA's performance is just too poor.   When I measured the horns during the video, I mentioned I had used an old Narda coupler and used port 2 for the reflected signal.   I mentioned the problem with this was both the coupler and amplifier I used was only rated to 8GHz and the lower end of the extender is about 8GHz as well.   

Still, shown looking at the two Narda loads with the extender+coupler, compared with using the LiteVNA to directly measure the load.  Using the extender to measure S11 at higher frequencies could be solved with $$$. 

[coppercone2]

alright well I gave it another shot because I had to wait for something to finish nearby,it looks like I got a nasty coat of paint on the interior. If I am able to do a full cure, then very very carefully sand it down a little with sanding sticks, and then maybe touch it up, it can be finished. I think this paint cures pretty quick and its rather tough, so it might actually sand OK. The flange has a chip on it, but that will fix easily.

I am kind of wondering if its possible to texture the horn somehow before printing, and then use a gap filling primer, sand that down, and then paint that.

[coppercone2]

after a 1 hour bake at 72C, it seems to be sanding good with the sanding stick (240 grit). Looks like I need to do another pass at 320 and do a very minor touch up on a pin hole that formed. But on the final coat, I am either going to have to stand it, at least spot sanding, because every time the can leaves a few lumps.

Need to get that plastic parts primer, when I ordered it online before they were out of stock and I keep forgetting to do it.

The nickel spray paint seems like its.. rough after application, it reminds me of galvanzing paint. The silver paint seems smooth on finish, it reminds me more of a clear coat type finish.

I might recommend that if its feasible, to just go with nickel paint for electroplating, the silver paint is harder to apply.

I believe I can try to plate both of the horns I made.


Also, for clear prints, do not paint the outside till you are sure of the inside. You can use a powerful flashlight to see if there is any uncovered areas.

[coppercone2]

Well ok, I put a third coat on it pain stakingly. The wave guide portion is sealed, but there was some pin holes on the flare that I could not seem to seal, even when I was heating with a hot air gun to spot cure the paint and re seal. I dunno why its so hard to seal up, it looks fully coated to the naked eye, yet light escapes some how. Defiantly need primer of some kind for future production.

but as a precaution, I will be putting copper tape on the outside of the horns. It seems like a good idea anyway, incase there is flaking or something, so it does not radiate in the wrong direction. Might be electrically annoying. Or I will just plate them.

It looks similar to the degrading coating put on those black paint covered light bulbs with a apature to direct light in a particular direction. When the paint flakes off because of age, it looks like that.


I don't know if they have translucent materials for your 3d printer joe, but if they do I recommend making a translucent one and seeing how well that paint and plating adheres, if there is any pin hole leaks. 

Kinda wonder too, if its all done, would putting gold leaf on it hurt it any?

[coppercone2]

lol ok, I was just doing comparative tests between the two antennas to see how they effect the reflection. The silver one has slightly deeper nulls and the same general shape. Gremlin attack the VNA though, seems to need to reset the source sometimes when there is a big disturbance on screen (like swapping DUT)

Based on the 18-20GHz range, which is not even in spec of the test fixture, the silver paint one got ~0.3 swr better and the ~ same shape when it was pressed into the coupler without screws, and this seemed repeatable

and it turns out the second adapter I got is for a slightly smaller wave guide, it has no response within the range of the instrument, the seller screwed me. there is some kind of force preventing me from getting this work done, i even managed to crack the silver antenna because I tightened a screw too much

so, their kind of similar and the silver one seems to have a generally very slightly lower SWR response when doing comparative testing, and the SWR is much better then with only the coupler (adding either antenna... levels out some of the hills) . Need another coupler to try transmission though. I can do this frequency better with a external directional coupler and SA/SG,maybe I will set that up later, so I am within the spec of the equipment  and with the new incorrect coupler I can make a even smaller harder to measure antenna (TY ebay).

but hey, at least nothing exploded, pretty happy. Gonna get conductive rubber gaskets though, so I don't risk cracking shit again.

[coppercone2]

BTW its alot to scan through your posts and videos, did you have a before and after comparison of some electrical plot before and after plating ?

[joeqsmith]

Take aways from the last video showing the 3D printed plastic horns.   

1) Only the first horn was left not primed.  Learned that lesson on the second horn.   
2) I allowed the paint to fully dry between coats.   
3a) I saw a measurable increase in the horn's forward gain when going from 2 coats to 5 coats.  Adding a 6th coat I saw no measurable increase.
3b) Lowest DC resistance did not equate to better gain.
4) I saw very little difference when comparing a wet sanded surface with a one just lightly sanded.
5) If you want to measure VSWR, make sure you have some way to validate your setup. 
6) MOST IMPORTANT,  wear a wrist strap!!!  ESD is fierce competitor.   

BTW its alot to scan through your posts and videos, did you have a before and after comparison of some electrical plot before and after plating ?

I don't believe so.  Only the effects on forward gain by adding more coats and some of the ESD and conductive plastics without any coating.   In a few cases, the signal actually attenuated as I assume the uncoated plastic was absorbing some of the signal.   

...  and it turns out the second adapter I got is for a slightly smaller wave guide, it has no response within the range of the instrument, the seller screwed me. ....

   
 
I've searched various sources for used components and it's common that the seller is clueless.   In many cases, the OEM is gone along with the data for the device.   It's up to you to ask detailed questions.  At least with WGs you could ask the seller make some quick measurements to at least let you know if you're in the ballpark.   

[joeqsmith]

Using my homemade calibration standards with the LiteVNA.   The twist, no nut, bolts or clamps were used during the calibration.