Protecting a PCB wideband antenna and making it portable

[DaJMasta]

I've got a hold of a spectrum analyzer with a decent bandwidth and I'd like to do some passive signal hunting - pointing an antenna at things/places and seeing if I can see what is out there.  Since it's purely for my enjoyment and curiosity, the budget isn't big, and to get something that will pick up as much as possible (of the bandwidth, I don't really need high signal strength), it seems like something like a PCB log periodic antenna seems ideal, especially considering how cheap some of them are.

So if I get my antenna and attach an SMA cable on a bit of rigid or semirigid coax, how do I house said antenna so that I can point it at whatever I think looks interesting?  The analyzer is a battery operated one, so ideally I'd like to use something that will give me enough structural support to give me some sort of grip to point it at things and so that I don't damage the coax connection or the board when it shakes around while I carry it.

I've been looking around for materials that are fairly RF transparent up into the few GHz range (the analyzer goes to 7GHz), but I don't see all that much available data on what would be suitable.  From what I can gather, standard plastic enclosures wouldn't be suitable, but perhaps plexiglass or fiberglass ones would?  It's been recommended that styrofoam or low to mid density polyurethane foam are great choices, but I'm not sure they're really give me the protection or in-hand grip that I think I'm looking for.  I also know that commercial versions of this exist, and at least from the pictures, they seem to be using some kind of plastic as the housing.  While something like the Aaronia HyperLOG series looks like what I'm picturing, I can't justify $400-600 for my own playing around (some of that budget is already in the analyzer, after all).

So anyone have DIY experience covering your wideband RX antennas or making something like this?  With my limited understanding, my best simple design is probably a piece of plexi bent in an L shape, with the antenna attached to one face and the SMA connector poking through the other.  Then cutting down the SMA face or bending it to make something of a grip.  Alternatively, a sheet of plexi with the antenna on it and a piece of PVC with a slit cut in the end to slide the sheet of plexi into to use as a grip.

[T3sl4co1l]

Most plastics are transparent to RF, but avoid those with high losses (e.g., PVC?) I suppose.

PCB material isn't the best, but the nice thing is you can make via connections wherever you need, the dimensional tolerances are excellent, and you have complete control over element widths as well as lengths and positions.  I would recommend milling out most of the space between elements, leaving a few ribs for rigidity: this reduces the dielectric load between elements.

The antenna needs to be at least as wide as the lowest frequency of interest, and at least as long as the directivity you desire (you can make a very long, shallow-angled antenna that has directivity like a Yagi, but the directivity is much poorer than a Yagi of the same size; that's the price of bandwidth).

You might also consider an exponential horn shape, which has better directivity (I think..?), and also uniform phase response (if you should ever get a scope that fast, you can do fun radar stuff!).  The ridged horn isn't very easy to construct (you need the ridges, plus housing), but a Vivaldi antenna is basically the same thing sans horn.

Tim

[KD0CAC John]

Put a piece of what ever you think make work , plastic in the micro wave , start out with low power and short time , working up to high power for maybe a min. - if it doesn't heat up , your good to go .

[DaJMasta]

I've heard the microwave thing mentioned before, is a microwave's radiation wide enough band to be a good measure of it overall?

Basically, it seems like a great simple test, but it would miss out on checking whatever frequencies weren't being tested, and I was under the impression that they were tuned somewhat to work on water, so microwaving something wouldn't actually be an indicator for good wideband performance.  I suppose it would still be worth testing, at least.


I looked at some of the commercial similar antennas and they seem to be preferring a fiberglass cover, and that there are omnis that go well beyond the range I'm looking at that still use fiberglass housings.  Wonder if I can find something that will fit right, as I don't think I want to be fabricating my own fiberglass parts in a custom mold... would be neat but seems like a lot of trouble.

[T3sl4co1l]

Water has an absorption peak ca. 30GHz, actually, so it's just ordinary dielectric and induction heating at ~2GHz.  Also helps that the average sample of water is moderately conductive.

For simple materials, dielectric loss varies slowly with frequency, roughly as sqrt(f).  If it's low loss at 1MHz and medium loss at 1GHz, it's not going to be much worse at 10GHz (I suppose FR-4 would be a good representation?).  If it's medium loss at 1MHz and high loss at 1GHz, it's going to be pretty bad anywhere up there (e.g., PVC?).

There are tables of dielectric loss for various materials, at various frequencies.  Most ordinary polymers and composites are "simple", i.e., they don't have weird peaks and stuff.  If molecules can move around within the material, it's probably not going to be low loss (e.g., anything conductive, of course; plastics with plasticizers, or absorbed moisture; liquid crystal displays; etc.), and will have loss peaks in some frequency ranges (the frequency corresponds to the relaxation or diffusion time constant related to the impurity).

Tim

[DaJMasta]

Interesting, thanks!

Just zapped a little 2"square or so piece of 1/8" plexi for a minute and while it's warm to the touch, it's under 100F, so while it will probably do the trick, I'll see if I can get a hold of a few other sample materials and try them out.


I will consider messing with the PCBs, but other than mounting holes I may not do much - I'm getting ones which are already designed and characterized by a guy in Texas and sold pretty cheaply ( http://www.wa5vjb.com/ if you're curious).  Of course, especially at the price point, it could be interesting to dremel out some of the dead area of the PCB and seeing how it changes performance.