Experimenting with waveguides using the LiteVNA

[joeqsmith]

Discussion of Errors in Gain Measurements of Standard Electromagnetic Horns
R.W.Beatty

1967 National Bureau of Standards

https://www.govinfo.gov/content/pkg/GOVPUB-C13-61d10f6b178c42c8794fe546fc9a7700/pdf/GOVPUB-C13-61d10f6b178c42c8794fe546fc9a7700.pdf

[joeqsmith]

Two new sets off a different printer arrived.  Standard PLA and ESD safe.  The ESD plastic has a very high resistance and I doubt could be directly plated based on my previous experiments with the conductive plastic.   

Front left black is the wet sanded, primed, 2 coats of conductive spray.  To the left of it is the orange is raw printed, no primer and 2 coats of conductive coating.  Once these dry, we can get started. 

[joeqsmith]

Plating one of the test horns.  Different donors for different areas.  Plan to plate the flange, then the WG tube and finally the horn.   

[Kosmic]

Apparently if you hookup your part to a motor and have it rotate in the solution, the coating is supposed to be more evenly distributed.

Ex: https://youtu.be/vsrlrH3omZc?t=287

[xrunner]

Be interesting to experiment with conductive PLA for your project. I have no idea if it would work for waveguide but might make for an interesting investigation -

https://all3dp.com/2/conductive-filament-brands-compared/

[coppercone2]

figured out what the hell was going on, I am printing a WR42 horn now. I had the wrong programs but the abandoned workstation was setup properly so I got that done. Cleaned up the rubble around my work area after completing a few repairs, it looked like a shelf fell over lol

[coppercone2]

I put the setting 3 for height of the horn section and 4mm walls

What I don't like is the 90 degree corners on the flange. Paint on a 90 degree corner sucks. I think it would be good to add a copper shim under the flange with the flange dimensions that has bends that go into the hole, and glue it in, then paint over that. If its thin enough it should not matter. I figure I can print just a few sections of the bottom piece to have a object to form thin copper over before transferring it Or gently rounding the corners with a very fine file.

I mean print out a waveguide slice, then put foil over it wrapped inside to mold it correctly (cut an X into the foil with the dimensions of the rectangle and bend inwards), then glue it to the real horn, then paint that over. There will be a bit of uglyness at the corners of course, but it might be a more reliable transition then just having paint going around a 90 degree bend into the flange. With very thin material it might be comparable to the paint layer thickness. I imagine if you work carefully 95% of the flange bend will be covered by foil, so your paint-foil interface will be triangles on the interior of the horn that are about 0.5cm high, and the flange to flange connection will be 100% metal, so you don't need to worry about screws cracking the paint. And having a subsurface copper shim that has paint on top of it, is IMO much better of a transition, then clamping a wave guide to metal paint. At least there is some kind of chemical bond going on there between the conductor and the paint, not a metal to paint bond mechanical bond. So then there will be a metal to metal bond at the flange interface and then a paint to metal chemical bond on the interior of the horn.


Past that you would need butt joints on foil to make it better I think..... guess you need a shrink ray to do that one.

[joeqsmith]

Be interesting to experiment with conductive PLA for your project. I have no idea if it would work for waveguide but might make for an interesting investigation -

https://all3dp.com/2/conductive-filament-brands-compared/

I have been.

Paper discussing using it for waveguides with one example being a horn
https://www.eevblog.com/forum/rf-microwave/experimenting-with-waveguides-using-the-litevna/msg4642243/#msg4642243

My first attempts to directly plate it (also following two posts)
https://www.eevblog.com/forum/rf-microwave/experimenting-with-waveguides-using-the-litevna/msg4648288/#msg4648288

[joeqsmith]

Apparently if you hookup your part to a motor and have it rotate in the solution, the coating is supposed to be more evenly distributed.

Their two plates may help as well, plus the larger tank. 

He talks about using constant current which I think for the conductive PLA, this is how it will need to be done.   

[joeqsmith]

...
I mean print out a waveguide slice, then put foil over it wrapped inside to mold it correctly (cut an X into the foil with the dimensions of the rectangle and bend inwards), then glue it to the real horn, then paint that over.
...

One way would be to use foil tape over the entire surface.  I think for simple shapes, like horns, this would be fine.  That or using PCB material or bending metal.

I would like to find a 3D plastic print process we could use with more complex shapes.   

[coppercone2]

oops, I put it so the horn is supported by the flange base, so that part is ugly because of all the supports and requires flat sanding etc. I think it will need to remain in the printer over night till I buy some more denatured alcohol to spray it down with.I was thinking how to prevent aperture distortion, not surface finish on the faces. If you put it side ways you get possibly a different problem with warping, which is much harder to deal with, so maybe its the right call. If you print it large apature down, then you either need internal supports or no supports, which I don't like either.

[joeqsmith]

All of mine were printed with the flange on the bed, horn pointed up. 

[coppercone2]

yeah but I put supports. I also had a small crack happen on the flange when I was removing supports. I think I can do this without supports, but I think I Can also sand and patch this one with some epoxy putty.

If I print another without supports I can compare them to see how much they do for all the problems they cause.

[joeqsmith]

Shown is a closeup of a raw PLA print.  All parts were printed without supports.   

One of the first tests will wet sand and prime one horn and leave the other raw.  You can see it's not quite finished in photo.

***
Looks like the drag and drop photo attachments is once again not working.   

[msat]

Thanks for taking one for the team and trying the copper gasket spray, Joe. I never used it on anything other than a head gasket, and that was long ago, but I recall it applying a very copper-looking coat, and none of that aqua-colored slime from your pics. Strange. Do you recall which product you used? I also wouldn't have expected it to provide a particularly robust coating, but hopefully good enough for a bit of testing as long as you don't get too touchy-feely with the stuff. Electroplating is definitely a better option, if a little more involved.

edit:
I only skimmed through this thread after you referred to it in your yt vid, so I haven't looked at the reference you used for electroplating, but in the past I looked up some blogs and videos about plating 3D printed parts in general. Some of the results looked quite good.

Thought about using a friend's SLA printer and even came across some low resistance resin in the neighborhood of milliohms per square, but it was like $500 a bottle 
Even if that borderlines too high a resistance, I bet it would take electroplating pretty well. but that price..............

[joeqsmith]

That was some spray-on Permatex.  I checked the other stuff I have and it wasn't copper based.   If you find something that can be applied to plastics without damaging them,  is mechanically stable, highly conductive and cheap, let us know.   

There is a conductive plastic that has MUCH lower resistance that the stuff we are playing with.  Cost is about $200 for a small spool.  Enough to print one horn.   MG Chemicals also has what appears to be a better coating at about $200 a can.  Both could be a bust.   

[joeqsmith]

Measuring the small PLA ESD conductive horn.   I don't think we would see much gain from it, if any....

[msat]

I thought I came across gasket spray that looked like it might have had some promise, but I must have been mistaking. 

I guess this ultimately comes down to what skills/abilities you have, what tools you have access to, and how much cash you have to burn. Obviously It doesn't really make sense to spend a lot of money on fancy filament/paint/resin when it approaches the price of pre-made hardware. Yeah, the Pasternack stuff tends to be expensive even on the used market, but various parts from things like microwave links can be had for pretty cheap (for less than $100, I got TX and RX amps, two filter waveguides, circulator, and two coax to waveguide adapters). For homemade parts, you probably can't get much cheaper than your PCB builds - even if you had them come pre-cut from a Chinese board house, or your bent sheet metal parts.

Since we can't [economically] 3D print a part and call it a day, I guess it's a matter of cheapish and effective ways of electroplating. Looks like the Caswell copper conductive paint (https://caswellplating.com/copper-conductive-paint-4oz.html?sku=CCP) is at least a good base coat for electroplating (which is what they basically market it for) at a reasonable price - $50 for ~20sq/ft. It's recommended to be sprayed on using an airbrush or or preval (though that might end up being wasteful). Meaning an additional $50-100 for an airbrush kit.

[coppercone2]

well working with the mg chemicals conductive paint can is a nightmare to say the least. I put a nasty coat of conductive paint on it. probobly should pour it out of the can and apply with a brush, or just try a dip coat. I think you can use a trash bag twister tool and a drill press to make a slowly rotating dip coating apparatus also. dodgy is the word.

[joeqsmith]

I thought I came across gasket spray that looked like it might have had some promise, but I must have been mistaking. 

You may have.  I imagine there are many products available but manufactures don't typically provide a spec on how conductive they are.  So you would need to buy and test them.  May cost you a fair bit of money to try them and you may not find anything.

I guess this ultimately comes down to what skills/abilities you have, what tools you have access to, and how much cash you have to burn. Obviously It doesn't really make sense to spend a lot of money on fancy filament/paint/resin when it approaches the price of pre-made hardware.

And may not perform as well, assuming that's a goal.

Yeah, the Pasternack stuff tends to be expensive even on the used market, but various parts from things like microwave links can be had for pretty cheap (for less than $100, I got TX and RX amps, two filter waveguides, circulator, and two coax to waveguide adapters). For homemade parts, you probably can't get much cheaper than your PCB builds - even if you had them come pre-cut from a Chinese board house, or your bent sheet metal parts.

Making complex parts from PCB will have some drawbacks, like soldering.  The board house would need to be able to plate the edges.    Having shapes inside the tube would require soldering on the inside.  Seems like a bad idea.   For a simple horn, certainly doable. 

Since we can't [economically] 3D print a part and call it a day, I guess it's a matter of cheapish and effective ways of electroplating. Looks like the Caswell copper conductive paint (https://caswellplating.com/copper-conductive-paint-4oz.html?sku=CCP) is at least a good base coat for electroplating (which is what they basically market it for) at a reasonable price - $50 for ~20sq/ft. It's recommended to be sprayed on using an airbrush or or preval (though that might end up being wasteful). Meaning an additional $50-100 for an airbrush kit.

The verdict still out on that one. 

[joeqsmith]

well working with the mg chemicals conductive paint can is a nightmare to say the least. I put a nasty coat of conductive paint on it. probobly should pour it out of the can and apply with a brush, or just try a dip coat. I think you can use a trash bag twister tool and a drill press to make a slowly rotating dip coating apparatus also. dodgy is the word.

Without seeing your process and results, it's hard to say.   If the 3D print or the prep work is poor, the coating is not going to correct that.  There could be a chemical compatibility problem as well but I assume you started out with a test sample.   

The biggest problem I see is you are working with WR42.  There are a few reasons I am starting out with WR90 to get a feel for some of the basics.  Construction is certainly part of it.   

Last would be your level of expectations.  Because I have no idea what yours looks like,  I have attached a photo showing a closeup view of the coated surface for one of my non wet sanded test horns.   This should give you some idea of what to expect.       

[coppercone2]

the problem is the can. I renember it was dodgy when it came in, when I looked at it last night it had crust formed on the sprayer, so I  tried to clean it out with a brush and some paint thinner, and put a new sprayer on it, but it was like working bad and soaked my gloves. Maybe its defective from distributor. You also need to shake it very frequently compared to normal paint, and the pressure is pitiful. But for a $40 can of spray paint thats really bad IMO. I remember last year when I tried to paint a sample piece it was also basically bad. I think what happened is when I first used the can, after using it, it got stuck on, and I took the sprayer out and cleaned it out with paint thinner and left it there, then loosely reinserted the sprayer,  but over a year it got crusted over... (but I have paint cans (rustoleum) next to it that are 3-5 years old, while rusty, they have 0 problems). The ESD dissipation paint spray can I got from MG chemicals a while back too also had poor performance (for resistor lead bender project). They have something going on with their spray cans.

I just would have thought for such a expensive product they would make it a little better, because 5$ cans beat it that are 4x older).

I bought some paint brushes so I can rework the feed to the horn because a bit of paint accumulated in there and when I was cutting it off to make the apature rectangular again, I left some bald spots, so I will sand it with a stick and repaint that area by hand.


But the resistance from MG chemicals nickel paint seems low, from the front of the horn to the rear flange (fully coated on both sides), the resistance is <1.5 ohms, but I know that having non conductive spots on the rear feed channel of the horn will be bad, but repainting a square centimeter by hand is not the end of the world, I got a set of tiny brushes that have a head the size of a match head. I was initially going to try with a Q-tip but I thought to just get some brushes lol


I maybe recommend if you get MG chemicals products, just buy it from digikey, not the cheapest distributor.

tldr: bad experience in general with MG chemcials spray products but fine experience with other products.

I need to make a tiny sanding stick, which I am not looking folward to.

[coppercone2]

also for the next horn I will try to see if I can make the copper foil feeder/flange, so the paint cover near the feed is not as critical.

[joeqsmith]

I've been shaking the can a few minutes before using it each time and a few shakes along the way.  End by flushing the port.  Thin coats.  I wouldn't expect these metal coatings to act like paint but so far, that is how I treat it.  There has been no signs of it clogging.   Shelf life for the one I wanted to try was one year.   Not sure what the MGC is.  They don't seem to publish it. 

Resistance will depend on coating used and thickness.  Attached showing one of the large test horns.   You can see my first measurement was 2.6 ohms.  I added a few more thin layers to get it down this far.   Dots mark where I measured it.  Note it's fairly consistent.   I never tried the copper foil tape.  It will be interesting to see what you come up with. 

Also shown is a test print made from the conductive plastic I had attempted to directly copper plate earlier. 

Slowly getting everything together.   

[coppercone2]

thats looking good. I plugged mine into a 8510C VNA for SWR plot.

It looks like under 15GHz the SWR is very high, and it has a sine wave pattern, and of course I did not calibrate it. But I see that between, 15-17 its wavey but some areas near 1, 17-20GHz (limit of my test set) the SWR is under 1.8. Won't get good data till I get a SMA cal kit, I will get ome ictures later

But I guess it does work, if I wave my hand infront of the antenna it makes the SWR plot act up like the real horn antenna do. I think you could transmit with it. And I got nickel because it was supposed to shield lower frequenices, so copper is doing much better I see.

It does fufil the criteron of SWR<2 for 18-20GHz (well under 2).


basement is cold right now, dont wanna be down there, and its an obstacle course near the racks