Behavior of Microwaves over Hemisphere

Mark, I would expect it to increase that tendency. Look at the behavior of ground waves in MF radio transmission.

(I don't _know_ this -- I just expect it. Due diligence is left as an exercise to the reader.)

--
Tim Wescott 
Control system and signal processing consulting 
www.wescottdesign.com

Here's a start;

I would guess it makes little difference. Obviously, the space inside the hemisphere will be quieter than near the back face of the plate. I would guess the intensity at the same point (i.e., some distance behind the flat plate vs. the same distance behind the perimeter of the hemisphere) will be very similar.

Tim

-- Deep Friar: a very philosophical monk. Website:

Hemisphere with convex side to transmitter would INCREASE leakage to the opposite side compared to flat metal disk of the same radius. There would be noticeable difference if the wavelength is comparable with radius.

Vladimir Vassilevsky DSP and Mixed Signal Designs

Easy. I do something like that with antenna models.

Here's a flat plate reflector model: What you're interested in is the leakage off the back of the reflector, which in this case is the 3 lumps on the left: However, the spacing between the dipole driven element and the flat plate reflector will produce dramatic changes in the front to back ratio, and therefore the leakage around the back side of the reflector. Rather than have me try to guess the parameters, it would be helpful if you would disclose:

1. Frequency of operation.
2. Distance between radiation source and hemispherical reflector.
3. Type of metal (yes, it makes a difference).
4. A better guess as to the diameter of the hemisphere

Note that one optimize the model for best front to back ratio but that's probably cheating.

Grounded to what? At microwave frequency, any wire that claims to be a ground wire is actually an inductor.

I don't know. That's why I do models.

That also begs the question where are you measuring the RF field? On the surface of the reflector or some distance away. If you get far enough away, and into the Fresnel Zone(s), edge diffraction becomes the prime determination of the field strength, which I suspect will be much worse for the sphere. If you're fairly close to the reflector, then surface conduction (as you describe) will be the main contributor. However, if you just want minimum overall RF behind the reflector, then just minimize the area inside the reverse part of the antenna pattern.

Incidentally, you can obtain a bad guess experimentally. Use a stainless salad bowl as a hemispherical reflector and a similar size piece of aluminum foil on cardboard for the flat plate reflector. Hide a cell phone operating on 1.9GHz behind the reflectors and compare signal strengths. If that doesn't play, try a smartphone on wi-fi frequencies.

Do we get a clue as to what you are trying to accomplish?

--
Jeff Liebermann     jeffl@cruzio.com 
150 Felker St #D    http://www.LearnByDestroying.com 
Santa Cruz CA 95060 http://802.11junk.com 
Skype: JeffLiebermann     AE6KS    831-336-2558

Thank you for offering your advice. What software do you use?

1.8 GHz

100M

Aluminium

1M

Yes, no ground.

Let's say 10cm off 1) the back of the flat plate, and 2) the same from the back edge of the hemisphere. At the axial mid line.

Is there any way to limit this effect, perhaps by modifyng (flaring) the outer edge of the hemisphere?

I have tried using an EMF meter, as you suggest, and it appears to show the hemisphere is more effective when you count the extent of shielding within the radius of the enclosed space.

As you say, however, the edge defraction may be worse. I am looking at which topology shields the greatest net area behind it, and where the break-even point is in terms of ditance behind the reflective surface.

My reason for asking here is simply to understand the factors involved.

4NEC2. See: Faster multithreaded NEC2 engine: Both are free.

Oh-oh. Cellular.

The 100 meters may be a problem. I've never tried a feed that far away. NEC2 was intended for modeling wire antennas and this may be a stretch. We shall see.

Ok. I'll throw together some models in the next day or three. It's late and I'm going to have a busy day tomorrow. Bug me if I forget.

ok. Fairly close to the reflector.

I don't think so. It really depends on what you are trying to accomplish. If you are just experimenting with the isolation produced by various shapes, I don't have an answer. However, if you're trying to minimize the RF behind the reflector, I suggest you look into RF absorbers, such as carbon foam. I think (not sure) that a carbon foam absorber around the edge of the aluminum disk may also help prevent edge diffraction. However, NEC2 won't model dielectrics and absorbers so I can't simulate what might happen.

What type of EMF meter? If it's one of those sold on eBay for detecting ghosts, they quit at about 2KHz. There are a few that work well into the microwave region, but without any frequency selectivity, they will also detect everything from power line radiation to broadcast stations. For example:

I think you might mean "greatest net volume".

I can't help you much. I'm fairly good at making things work, not so good with the theory, and really sloppy with the math.

--
Jeff Liebermann     jeffl@cruzio.com 
150 Felker St #D    http://www.LearnByDestroying.com 
Santa Cruz CA 95060 http://802.11junk.com 
Skype: JeffLiebermann     AE6KS    831-336-2558

For a solid disk you could count the knife-edge diffraction attenuation.

You could also look if scattering theory might help you.

What is behind your test device ?

• Absolutely flat metallic horizontal surfaces

• Average sandy or grassy ground

• Trees, poles big stones (at what distance?)

• Buildings at what distance ?

Note that the environment behind your test setup will often reflect back quite a lot of signal power.

Also note that if you point the convex side towards the signal source, the reflections will enter the concave side and there is a focal point somewhere behind the shield, making situation worse than with a simple disk :-).

Can you put your device within a complete metallic sphere ?

Thanks for the offer. I will be most interested.

My EMF meter is lab grade and directional.

Is carbon foam a high tech item, or could it be found in some form at an industrial supplier? Or maybe I could try shielding paint on stiff foam. I once made my own from charcoal and graphite.

Mark Newman

The common sheets of conductive anti-static black foam will not work thanks to too high resistivity (typically 10^5 ohms/square). Metalized conductive foam for RF shielding and grounding won't work because of too low resistivity (typically 0.5 ohms/square). What you want is something in between which when installed, looks like the impedance of free space (377 ohms) so that nothing is reflected: Google:

I ended up with a surplus roll of the right stuff that was used to patch holes in RF anechoic chambers. I'll see if I can find it.

In the distant past, I got tired of plugging RF leaks with expensive die cut shapes and wanted something cheaper. I decided that mixing flake graphite with RTV would produce something that could be applied as needed. In order to get the graphite flakes to overlap and conduct, very little RTV could be used. The result was a brittle and crumbly mess with no strength. However, I suspect someone else has produced something either conductive or loaded with ferrite that can be dispensed with a caulking gun. I'll later (later).

Meanwhile, try using a wet towel. That's what I use when I want to block RF reflections or cover part of dish antenna in order to estimate fade margin or dish illumination. There are some obvious disadvantages to using a wet towel, but the convenience and low cost are so tempting. Try cutting up a sponge and paste it around the edge.

--
Jeff Liebermann     jeffl@cruzio.com 
150 Felker St #D    http://www.LearnByDestroying.com 
Santa Cruz CA 95060 http://802.11junk.com 
Skype: JeffLiebermann     AE6KS    831-336-2558

I think they arer adding carbon fibre now to bridge the gaps. But all this is expensive, unless a chunk of stealth technology falls into your yard.

Thanks, that's a great tip. Can the liquid be something that doesn't dry out? Would it help to be conductive, as in salt water or silicone electrode gel?

Mark Newman

Not sure if NEC2 will handle this, but a field solver type application is the right choice.

With these assumptions i do not see any of the curved surfaces beating the flat plate. The advantage of all the curved surfaces occurs closer the surface than the cut off plane.

Area or volume? Big difference.

Strictly behind or simply behind the "shield" surface?

Clarity is necessary.

Finally, the "shield" will also have a dramatic effect on the available RF pattern of any transmitter "behind" the "shield"

It appears that NEC2 is probably not a good choice for this. I've been struggling with the problem for the last 3 evenings. I'm getting wortheless plots, strange numbers, and sanity check failures (i.e. average gain test). I'll keep trying for a few more days.

--
Jeff Liebermann     jeffl@cruzio.com 
150 Felker St #D    http://www.LearnByDestroying.com 
Santa Cruz CA 95060 http://802.11junk.com 
Skype: JeffLiebermann     AE6KS    831-336-2558

Thank you for your most generous assistance.

Would it help to reduce the distance form the emitter to the shield surface until some valid result is achieved?

Information on any effects of the hemisphere would be useful.

Mark Newman

IOW the advantage of the hemipshere is only that the volume it encloses is better shielded that an equivalent _unenclosed_ volume behind a flat flat?

Excuse me for invoking a water-like analogy, but does the lesser angle of incidence between the microwaves and the edge of the hemisphere's rim, compared to a flat plate, reduce what might be called "turbulence"?

By "behind" do you mean within the hemisphere, or trailing its outer edge? In general, how might the effect be described?

Mark Newman

Pretty much IIUC.

I do not think that is a reasonable analogy, diffraction is a bit stranger than that.

Most definitely within the partial sphere, paraboloid, ellipsoid, hyperbaloid or other shape. A tenth size behind any of the cutoff plane of curved surfaces would not perform much different than the flat plate IIUC. Unless that was near the focus point then it would be very directional. Shape dimensions matter.

No. I'm getting useless garbage like this: That's a 1 meter flat plate, with an 1800 Mhz dipole feed, 10 meters to the right, at 5 deg resolution. I'm ignoring overlap and segmentation errors in the flat plate for now. If you want to play with it, the NEC2 deck is at: Wire 1 is the dipole. Everything is in meters. The "10" on the wire

1 line is the distance between the flat plat and dipole feed in meters. Starting with Wire 2, everything else was generated with the 4NEC2 "Build" program.

The problem is that 4NEC2 is looking at the entire antenna and not just the flat plate. As the dipole source moves farther and farther away from the flat plate, the effects of the flat plate reflector become less and less. The result is that the plots look like just the dipole radiation, without the dish.

In order to make use of a combined plot, it would be necessary to produce a plot of just the dipole, and subtract it from the combined plot, leaving only the radiation from the reflector. Besides not being certain that this is a valid method, I'm a lousy programmer and don't have time to do that. So, after a trivial dry run, I'll see if I can convince one of my acquaintances to do it. Bug me again in a week.

I can build a reverse hemispherical reflector, but I'm sure it will just produce more garbage.

--
Jeff Liebermann     jeffl@cruzio.com 
150 Felker St #D    http://www.LearnByDestroying.com 
Santa Cruz CA 95060 http://802.11junk.com 
Skype: JeffLiebermann     AE6KS    831-336-2558

Thanks for taking interest and trying. Very much appreciated.

Mark Newman

That is nice. But i thought that he wanted the antenna 0.1 m from the flat plate. You might want to run that situation as well.

?-)

Naw, it's ugly. I'm resisting the temptation to ask in the NEC mailing list because I don't have the time to do the inevitable reading, experimenting, programming, and following the experts suggestions.

Nope. He wanted the source to be 100 meters away. The 0.1 meters was the distance behind the reflector where he wanted numbers for the field strength. Quoting

--
Jeff Liebermann     jeffl@cruzio.com 
150 Felker St #D    http://www.LearnByDestroying.com 
Santa Cruz CA 95060 http://802.11junk.com 
Skype: JeffLiebermann     AE6KS    831-336-2558