Any RF antenna design boffins around?

[icon]

Hi

I have made, and am making some more, 5.8GHz antennas of a circularly polarised, 'clover leaf' design. These are used to transmit video from... well anything really.

The first thing that surprised me (I don't think in GHz) was that these are a 'short' at DC - the centre conductor of a length of RG316 is simply soldered via 3 short wires arranged at 120deg, to the shielding braid.

My first go was with 1.6mm silver solder wire, which I happened to have lying around. Surprise, surprise it actually worked! I haven't tested the range yet.

I'm going to have another go, this time with much thinner copper covered steel (actually MIG welding wire). at 5.8GHz, I presume the 'skin' is thin enough that the signal is carried in the copper.

My questions are these: what actual effect does the thickness of the material have? 1.6mm vs 0.6mm? Thicker wire is obviously less prone to physical distortion, but would it be better (better how?) to accept the fragility of thinner wire? Second question; the silver will tarnish - the copper is very thin so eventually the steel underneath will rust. Wouldn't it be great to dip it in Plastidip, or spray it with lacquer? Nay! This will disrupt the pixie dust which is essential to its function. What effect does this kind of thing have? Does it affect the frequency the antenna is 'tuned' to (if that even means anything)? If so, is it possible to guess how much, and make changes accordingly, or do you need fabulously expensive, multi-GHz-capable equipment?

I realise that this probably requires a three-year course to explain, but I'd be grateful if someone had a go.

Cheers
John

[DimitriP]

Regarding wire thickness: skin effect causes the wire to appear thinner as the frequency is increased
Fun reading here https://en.wikipedia.org/wiki/Skin_effect

[vk6zgo]

There is copper & copper----the copper layer on the welding wire may have impurities due to being optimised for other than conductivity.
The layer is also quite thin & subject to corrosion.

A sprayed layer of lacquer shouldn't detune the antenna much----a heavy layer of Plastidip probably will.
Basically,you have to work out which way the resonant frequency changes,& correct for it.

No "pixie dust",but you probably need some method of  checking resonance,if you want to optimise the things.

[retrolefty]

Many decades ago when more active in ham radio I build several wire and metal antennas. One thing I can say is that none flew right off the design/calculation board. A SWR meter was always used/needed to 'trim' for best performance.

 Testing and adjustments is more important then what wire gauge you use in my opinion.

[Howardlong]

It will have an effect on the effective length of the antenna elements and have an effect on the impedance.

You might want to invest some time in a modelling program like EZNEC. I haven't done much antenna work for a while, but I did use modelling programs like EZNEC quite extensively some years ago, there may be more modern alternatives nowadays. You can see the effect of changing things like length and thickness.

Personally speaking, I'm a bit old in the tooth and like to do final impedance tweaking using a VNA. There are some wizards these days who can model the whole thing and be confident it'll be pretty much there from their simulation alone. I found once I'd used a VNA and could see interactively the effects of changing things in real time I just didn't want to go back to random and somewhat aimless tweaking. Luckily these days there are some pretty reasonably priced PC based VNAs, but I'm not sure they go up to 5 or 6 gigs.

One final thing is to be very wary of mechanical supporting structures of any kind at microwaves. Insulators have dielectric and loss properties not the same as free space and can be catastrophic on radiation patterns for example, not to mention matching. Note that your antenna might, of you're lucky, present a reasonable match... but then so will a 50 ohm resistor!

[T3sl4co1l]

Yeah, if your only gauge of performance is "hey, it worked, I guess" -- just put up a 1/4 wave whip or something like that, and leave it be.  It's not that you won't do better -- you certainly can -- but you won't know, and without measurement, any conclusion is meaningless.

Checking signal strength is helpful, but not ideal.  It has to be done carefully.  Example: an antenna which radiates modestly in the forward direction, but strongly in the sideways direction, might still measure better than a merely "good" unidirectional antenna.  Or worse.  How?  That strong side lobe could take a long path back to the receiver, thanks to reflections in the test area.  The test site is critical.  An open air test site is preferable, and an anechoic test chamber even better still!

Signal strength indicators are also not consistent.  RSSI is an arbitrary scale.  Bandwidth is adaptive, and often lazy (i.e., it might not renegotiate for a faster channel, unless it has to).  Power level might be adaptive too (i.e., it only uses what it needs), which would *really* throw off your measurement!

So it's very tricky to do, and really needs to have good equipment to get a sound answer.  Without that equipment, be very weary of, and careful about, any possible sources of interference: side lobes, reflections, mismatched loads, and dubious measurements.  Doubt everything -- yourself foremost, as you are the easiest person to fool -- and also remember that, if it's not falsifiable, it's not science.

Tim

[hendorog]

Have you seen this forum?
http://fpvlab.com/forums/forumdisplay.php?16-DIY

According to ibcrazy, who popularised the cloverleaf for FPV, thicker wire on a cloverleaf increases the bandwidth of the antenna - i.e. essentially the antenna becomes a bit more tolerant of manufacturing errors.

There is also a cheap and cheerful SWR meter you can build for 5.8GHz which will allow you to test that theory.
Using that device you could at least adjust for the frequency shift effect of the effect of plastic dip/lacquer.

You can get somewhere with very cheap DIY gear. There are other options as well before you get to an Agilent VNA for $$$

[icon]

There is also a cheap and cheerful SWR meter you can build for 5.8GHz which will allow you to test that theory.

You couldn't point me to the one you're thinking of?


Thanks
J

[PA0PBZ]

You couldn't point me to the one you're thinking of?

http://fpvlab.com/forums/showthread.php?2768-SWR-meter-Do-it-yourself-project

[icon]

Ah, right. The sticking point there is the directional coupler, which might spoil the notion of 'cheap'. Also opens the possibility of getting the wrong sort? Is there any particular thing I should be looking for (or avoiding)? What's a 'negative detector'? ('Directional coupler with negative detector')

Cheers
J

[hendorog]

Ah, right. The sticking point there is the directional coupler, which might spoil the notion of 'cheap'. Also opens the possibility of getting the wrong sort? Is there any particular thing I should be looking for (or avoiding)? What's a 'negative detector'? ('Directional coupler with negative detector')

Yeah there aren't many on ebay at the moment. Just a case of waiting and jumping on one when it comes up cheap.
It's quite possible to build one as they are very simple, but brace yourself for a lot of trial and erroring. There are a couple of examples in that thread.

You want something which covers the frequency range of interest and has the right coupling to get the best out of the detector.
4-8GHz is the octave which covers 5.8GHz. Wider frequency ones are available but the directivity is generally worse. They tend to be more expensive on ebay as well.
For a 200mW Tx and the standard diode detector you want about 10dB coupling.
You probably want SMA connectors, although I did notice some reasonably cheap N type DC's if you already have or can get hold of adapters.
e.g.
http://www.ebay.com/itm/Narda-3044-10-4-to-8-GHz-10-dB-Type-N-F-Coaxial-Directional-Coupler-/221906405127?hash=item33aaa70707:g:lkEAAOSwNNxWFB4j

The Directional coupler with negative detector you mentioned sounds to me like a coupler which already has a detector built into it.
I don't know for sure but it might/should be possible to just use that without needing the detector circuit from fpvlab.
I'm sure someone here will know exactly how it works

[Howardlong]

If it's any help, I started my microwave career on directional couplers and crystal detectors, with AC millivoltmeters on the detector and an AM modulated signal source. While it's definitely possible to make measurements with this technique, in these days of instant gratification I'm not sure how many people will have the patience to want to follow it through. Certainly when I was using that technique, the very same people who'd suggested and explained it to me were using VNAs and RF power meters with calibrated heads themselves!

I now understand why. It's fine for the odd rare measurement, but for more than that, especially for frequency sweeps, I'd not want to go back.

[LukeW]

Regarding the OP antenna being a DC ground short, that's not unusual for many antennas.

[icon]

Regarding the OP antenna being a DC ground short, that's not unusual for many antennas.

I'm sure you're right, but as a non-RF bod it caused me a "Wait...what?" moment.

J

[TheSteve]

Having had a go at building some cloverleaf antennas myself I can say that the odds of it actually being tuned exactly where you want it is slim to nil. I haven't even found a commercially built one that was on frequency yet(note that I have not tried one from ibcrazy). They were all several hundred MHz from the ideal spot. Of course they still work, but not quite as well as they actually can. I doubt very few of the antennas sold commercially are ever tested on a proper machine, they just measure, cut, bend, solder and ship - hoping to get them in the ballpark.  I did build a 2.4 GHz one that was nearly bang on - they are much more forgiving at lower frequencies.
My test device is a calibrated Agilent N1996A 6 GHz spectrum analyzer with built in return loss bridge and SWR measurement suite.

[hendorog]

Having had a go at building some cloverleaf antennas myself I can say that the odds of it actually being tuned exactly where you want it is slim to nil. I haven't even found a commercially built one that was on frequency yet(note that I have not tried one from ibcrazy). They were all several hundred MHz from the ideal spot. Of course they still work, but not quite as well as they actually can. I doubt very few of the antennas sold commercially are ever tested on a proper machine, they just measure, cut, bend, solder and ship - hoping to get them in the ballpark.  I did build a 2.4 GHz one that was nearly bang on - they are much more forgiving at lower frequencies.
My test device is a calibrated Agilent N1996A 6 GHz spectrum analyzer with built in return loss bridge and SWR measurement suite.

They can be tuned by increasing/reducing the feedpoint gap. The idea is to not cut the centre conductor too short until you've finished tuning.

[TheSteve]

Having had a go at building some cloverleaf antennas myself I can say that the odds of it actually being tuned exactly where you want it is slim to nil. I haven't even found a commercially built one that was on frequency yet(note that I have not tried one from ibcrazy). They were all several hundred MHz from the ideal spot. Of course they still work, but not quite as well as they actually can. I doubt very few of the antennas sold commercially are ever tested on a proper machine, they just measure, cut, bend, solder and ship - hoping to get them in the ballpark.  I did build a 2.4 GHz one that was nearly bang on - they are much more forgiving at lower frequencies.
My test device is a calibrated Agilent N1996A 6 GHz spectrum analyzer with built in return loss bridge and SWR measurement suite.

They can be tuned by increasing/reducing the feedpoint gap. The idea is to not cut the centre conductor too short until you've finished tuning.

I might try making more 5.8 GHz models when I get some time, see how long it takes to get one bang on frequency.

[T3sl4co1l]

I would caution that, if mere 5% errors in frequency are substantial in performance/characteristics, it's a dangerously precise and high Q antenna, and probably won't serve you well anyway for those reasons.  A 1/4 wave whip will cover a much wider bandwidth (particularly with a relatively wide, or flared, element), and has only one element to tune.

Tim

[hendorog]

I would caution that, if mere 5% errors in frequency are substantial in performance/characteristics, it's a dangerously precise and high Q antenna, and probably won't serve you well anyway for those reasons.  A 1/4 wave whip will cover a much wider bandwidth (particularly with a relatively wide, or flared, element), and has only one element to tune.

Tim

The point of using these for fpv video is the circular polarisation. The multi-pathing is pretty terrible on 5.8GHz without that.
Its not that hard to make and tune for good performance as long as there is some way of measuring it.
The 1/4 wave antennas that come with the tx's need tuning with an SWR meter as well.

The other common option is a helical. It is forgiving in bandwidth and easy to build.

[icon]

Having had a go at building some cloverleaf antennas myself I can say that the odds of it actually being tuned exactly where you want it is slim to nil. I haven't even found a commercially built one that was on frequency yet(note that I have not tried one from ibcrazy). They were all several hundred MHz from the ideal spot. Of course they still work, but not quite as well as they actually can.

And then the 'band' I'm working with runs from 5.645GHz to 5.945GHz, which is 300MHz anyway. I suppose that allows for some selective assembly!

J

[Zero999]

Second question; the silver will tarnish

I doubt that will be a problem. If the silver on the surface tarnishes, then it will become an insulator and the current will continue to flow underneath it. The only time there would be a problem is if the dielectric loss of the silver oxide layer had a high loss at the frequency of interest or if all the silver oxidised.

[thewyliestcoyote]

http://www.urc.cc/pubs/URC-2002a.pdf

While silver oxide does seem to be kind of lossy, the thickness of the layout would be very thin. This is no different than a dielectric loaded wire antenna. First component as one would expect in this relationship is the thickness of the coating.

No good engineer in this feed does it with out the aid of software to simulate designs. It is just like building things but cheaper, faster to do, and ... . Look up a software called NEC (Numerical Electromagnetic Code). Versions 1 and 2 are open sourced and free.

https://en.wikipedia.org/wiki/Numerical_Electromagnetics_Code

There are many ways to make an antenna wider bandwidth. If that is the only problem my I suggest the conical dipole. That said antennas for these frequencies are so cheap now days. Unless than you are wanting to learn then grab a good book on the subject, good simulation software, and some free time and go to town.

Testing antenna performance is much much more than just the input match. A 50 ohm termination has a great VSWR, but a real crappy far field pattern for some applications  .

[TheSteve]

The clover leaf is already modeled and verified to be a solid antenna, especially for FPV video uses at 5.8 GHz. The only real challenge is to build one that is properly tuned as some pretty detailed precision is needed.

[thewyliestcoyote]

The clover leaf is already modeled and verified to be a solid antenna, especially for FPV video uses at 5.8 GHz. The only real challenge is to build one that is properly tuned as some pretty detailed precision is needed.

I have no dough there is a close form solution to the ideal clover leaf antenna. I am a little confused if you have it already modeled and simulated why is this not just a monte carlo simulation problem.

If you have a physical antenna the thing to do is mount it on the air craft and use a signal generator and a spectrum analyzer or a vector network analyzer. Use these to measure 2 things the S11 or the VSWR and far field pattern(This would be S21 on the VNA). With antennas where circular polarization is desired this very important know what polarization is as a function of the angle. If is not just a constant.

You have to include the air craft as it is most certainly in the reactive near field of the antenna and will act as part of the antenna system.

If it is a tuning problem, what previsions are there for tuning? Matching networks, trimmable parts of the antenna, and so on.

[T3sl4co1l]

The clover leaf is already modeled and verified to be a solid antenna, especially for FPV video uses at 5.8 GHz. The only real challenge is to build one that is properly tuned as some pretty detailed precision is needed.

That sounds rather contradictory though: if it's "solid", then why are people apparently having trouble building it to the exact frequency?  Why is the frequency so particular?

Like I said earlier: if the bandwidth is so narrow that it's hard to get it right, it's not a good design.  Good design means good everything: not just "oh well, you didn't build it right" -- a good designer must accommodate the tolerances of the manufacturing process!

Offhand, I don't see actual measurements and discussion of this, so I have no idea how troublesome this is.

Tim