Radio Antenna Design Discussions

[warren52nz]

Hi. I haven't been able to find a lot of technical discussions on the design of aircraft antennas ("antennae" is plural for an insect antenna only) so I thought I'd start one as I'm a Radio Engineer with a degree in Physics and I've been working in radio and TV broadcasting for 25 years.

I've seen a few designs for "miracle" antennas but, at least some of them, are ignoring some basic laws of physics. My intention in this forum is to gather knowledge from those of you that have some understanding of radio physics and from those who have direct experience in trying specific ideas.

After all, radio physics is akin to "Black Magic" unless you delve in really deeply and there's nothing like trying the theory to determine whether you've got it right or not.

That's why I want to say right now, that a field test using trial and error is always going to be the ultimate way to know you've got it right.

You don't have to understand or be interested in this thread to take advantage of the findings. Hopefully we will come up with some answers that are useful to everyone. I don't have those answers yet by the way.

Please keep in mind that I haven't studied this stuff for decades so I might make a mistake from time to time. I don't mind if anyone notices something I've got wrong and points it out. I'm not perfect. And I'm not up to speed with the latest designs so I might step on some toes during this discussion. If I do and I'm worng, I apologise in advance.

Oh, and I work in metric like everyone else in the world (except the US and, perhaps, the UK?). I'm sure you can convert if you need to.

So here goes.... Radio Physics 101 :rolleyes:

First of all, there is no substitute for length when you're trying to capture a radio signal whose wavelength is much longer than your antenna. For 72 MHz the wavelength is about 4 metres and you don't get really good results until you get to a length of 1/4 of a wavelength or 1 metre, in this case. You can calculate wavelength in metres by dividing 300 by your frequency. So 30 MHz is 10 metres.

But we don't have 1 metre to play around with on most models, especially the small ones. My 3D Pro would struggle to allow 30cm. So we're all stuck with antennas that are too short. This has 2 major penalties:

1. the "radiation resistance" will be too low and
2. the impedance of the antenna will naturally be a poor match for the receiver

Thankfully, it's possible to tune the antenna to make it match the receiver. That's what coiling the antenna around a tube is all about. More on this later.

But a low radiation resistance is not fixable, you have to live with it. And it's radiation resistance that gathers the signal. The matching process is a way of preventing the signal from "bouncing" of the receiver input and going back out into space.

So there are two things to deal with, getting as much signal onto the antenna wire as possible and making sure that it all gets into the receiver.

I don't want to write a book in this first post so let me cut to the chase and someone else can hopefully add something (as I will when I get time).

A short antenna can be matched by a coil of wire (at the receiver end of the wire) but the number of turns is critical ! The point at which the match happens may be only 1 or 2 turns away from a point that is a crappy match. i don't know about you, but my Hummingbird 3D Pro has about 30 turns of wire on it, which was just the number of turns it took to use up all of the wire that's attached to the receiver and allow a bit to run out to the tail (to increase my radiation resistance).

Now I'd be astonished if I've got anything even APPROACHING a match with such an arbitrary procedure so I want to develop a way of testing it so that I can get it exactly right.

The first thing that occurs to me is to use a Spectrum Analyser or a Network Analyser both of which I've got here at work but they're worth over $50,000 each so most of you would be stuck at this point! I haven't thought of a way to do this that doesn't require such equipment yet but there might be a way. Trial and error springs to mind.

I'm hoping I can come up with some results that you can use without you having to test it. The results are purely based on the geometry of the antenna.

It's lunch time here in New Zealand so I'll leave it at this for the first post.

Anyone interested?

[vintage1]

(i) Antennae is the correct plural for antenna irrespective of usage

(ii) you are right in that size matters

(iii) tuning on antennae is far from critical. If the average front end/antenna combo has a Q of more than 5 I would be VERY surprised.

(iv) omnidirectionality is far more important than e.g. tuning a vertical whip for best performance - at least with fixed wings. I concede that helicopters do not spend much time other than horizontal.

Ultimately the problems resolves itself into a collection of wire or wires that cover the most area possible in a plane parallel to the transmitter antenna. Its probably fair to say that a collection of wires in 'star' mode going to every possible extremity of the airframe is as good as it gets, and then some mild form of tuning to get a little bit of signal boost.

Better of course if you can get motor generated interference down, is a more sensitive receiver.

[hoppy]

Here's some field test work performed by RC-CAM -
http://www.rc-cam.com/ant_exp.htm

I'd really like to see your take on the subject.

[vintage1]

I said it before elsewhere. Everything in that article is consistent with my relatively crude gut feel about how things behave.

By the ay, his worst case gain reductions - 12dB or so, only amount to a 4 times loss in volage senitivity.

The difference between a park fly cheapo reciever and a top of the range reciever is probably more like 100 times voltage senitivity (40dB)

If you don't have a weight problem, fly the real McCoy receivers on shortened antennae. Not the parkflies on whatever short antennae they come with.

[warren52nz]

(i) Antennae is the correct plural for antenna irrespective of usage

Well I can't be bothered arguing about something so trivial but:

http://lists.contesting.com/pipermail/towertalk/2002-August/049948.html

ANTENNA pl. -tennas (for 1), tennae (for 2) 1. a conductor by which
electromagnetic waves are sent out or received; aerial. 2. Zoological--
one of the jointed, moveable, sensory appendages occurring in pairs
on the heads of insects and most other arthropods.

I've been a broadcast engineer for 25 years. I manage all of the telcommunciations facilities on Sky Tower in Auckland which is the tallest free standing structure n the Southern Hemisphere. It has 10 floors for telecommunications and a 93 metre mast. You can search me out in Google with ""Sky Tower" and "Warren Harding" if you're interested.

Anyway, I agree with most of your other comments. I'm working in a vacuum of information about the front end of these receivers so it's good to hear your comments. I'm going to measure mine on our $50,000 Network Analyser here at work. I'll let you know what I find.

As for helicopters spending most of their time horizontal, that's hardly true of modern 3D models which can fly in any orientation, including backwards. The problem with helicopters and antennas is that they have this big "Mix Master" whirling around upstairs so you need make the antenna rigid otherwise when you go inverted the antenna will wrap around the mast.

If you've never seen a helicopter being put through its paces, I recommend this video about 2/3 of the way down the page described as:

"A beautifully flown Hornet II. This flyer has unbelievable skill Flown superbly"

http://www.helihobby.com/html/rc_helicopters_videos.html

But I'm getting off topic.....

I'm getting a growing feeling that you're right about the tuning not being critical. You don't need a huge range especially for a chopper whose orientation you have to ALWAYS be certain of. With a plane you can pretty much see what it's doing because it's always going forward (I have a plane too).

I started this investigation because I was getting a lot of glitching and I wanted to improve my signal to noise ratio. I finally eliminated it by putting the ESC control wires through a toroid (last night, actually). But I think it's still worth sorting it out once and for all.

I'll have a look at RC-CAM's link (thanks). he seems up on the subject.

Cheers

[vintage1]

I love trivia. Actually one source lists either form of plural for both.

I suspect the true state of affairs is that the term was orginally coined by zoologists, who knew Latin, and stolen by US radio engineers, who could barely speak English...

In the UK we say 'Aerial' of course;)

[warren52nz]

Haha, good one. 8^)

Aerial it is then, So is the plural aerii? LOL Now where were we..... oh yeah...

I've been thinking about it a lot and don't have much to work with by way of specs but it seems likely that a receiver manufacturer would have to allow for a huge range of receive antennas.... I mean aerials, because of the wide range of applications. On a plane you can dangle the aerial out the back because it always (ALMOST always, I suppose) goes forward but in a chopper you have to have a rigid aerial because it can fly in any orientation, so you can either shorten the aerial or wind it into a coil.

So I suppose they'd have to keep the Q low to allow for whatever the end user does. And this will sacrifice gain and selectivity.

This suggests that, for a specific application, you could raise the Q to get a better result. Tuning it up like this would have the dual benefit of increasing the signal and rejecting the noise.

Specifically, what I had in mind is getting as long an antenna as I can reasonably manage, then winding turns at the base and measuring the return loss at the Rx input on the Network Analyser until it dips at 72 MHz.

Alternatively, wind too many turns at the base and then put a variable capacitor at the receiver end and tune out the excess inductance. That should raise the Q.

What do you think?

[vintage1]

I try not to think about RF. It always gave me a headache. You know its possible to do the maths on a single tuned circuit, but even something as simple as a double tuned loosely coupled one is a complete nightmare.

Then when I got into doing a bit of it professionally, I threw away the maths books because just waving a hand near a coil could put a receiver off tune by miles.

With a computer, you can just about calculate the response of ceramic and SAW filters...but anyway you leave that to the designers and accept the funny shaped graphs they send you in the spec sheets.

I used to read 'Wireless World' - funny name - Do you remember Toytown, with Larry the Lamb and Dennis the Dachsund - the evil sausage dog with the thick German accent - going to see The Inventor, who was busy Inventing Wireless 'The trouble with Wireless, is that it involves an awful lot of WIRES'... - anyway Wireless World, which did indeed involve huge quantities of Wire, used to have every 6 months or so articles with titles like 'A Novel Antenna design For SideBand Suppression In the 32 Cm Band' or somesuch.

Featuring pages of impressive graphs, mathematical formulae, and a blurred photograph of things that looked like spinning wheels attached to large glowing valves. Then the next 6 months would be full of letters by earnest members of Academic and Wireless Establishments explaining the Flaws in the Mathematics, and demonstrating that it could all be done better with a baked bean tin, some old candle wax, and a piece of silver plated water pipe etc etc.

Then when I joined briefly an establishment actually making advanced radio and radar equipment for aircraft and miltary use, it was all a bit of a shock to be confidentially advised that 'what the boffins say is usually utter bunk: We knock up things that basically have half a chance of working and we use that lot' - pointing to about £50,000 quids worth of VSWR's spectrum analysers, wobbulators, sweep generators, cables and what looked like silver plated sink waste pipes and all the wonderfully expensive parahernalia that is used as an excuse to deprive taxpayers of their immoral earnings ('all in the name of National Security') 'and just eff around with it till it's good enough, and hope to **** that someone knows how to mass produce it'..

Which is, essentially, how you go about it. Assemble £50,000 worth of expensive test gear and eff around till it works better than when you started - by literally fiddling with one bit at a time and using what theoretical knowlege you may or may not have, to eff around with the right bits....

RF engineering, as the digital boys used to say is a Black Art, and many a tunnel diode was sacrificied on many a moonlit night in the Evocation of the Perfect Mixer.

Just knock up some arrangements of Wire, feed em into the reciever, stick the tranny down the other end of the bench, and start doing minor adjustments to assorted inductors and capacitors between the reveiver and The Arrangemnt Of Wires. Sacrifice a few chickens, go and see your confessor, pray hard, and maybe Divine Inspiration will show you that a Pentgrammatical arrangement or even a Cross, will in fact give you 0.07dB more gain, coupled to some arbitrary bit of gubbins that magically refuses to work off the bench, till you discover the loose coil of wire under the insulating mat from last weeks failed experiment that made it all work so well in the first place...;)

And the funniest thing of all, I ain't joking.:D

I had to use a tunerhead once. FM broadcast. Came from Philips/Siemens. IT HUMMED. anywhere in a magnetic field this damned thing HUMMED and it was NOISY.

Turned out in their efforts to get the most stable oscillator, they had used a zener diode - one of the noisiest things known to electronics, to stabilise a voltage controlled oscillator. Naturally the noise on the stable voltage cause it to jitter...the hum? Oh they used a ferrite cored induictor as part of it as well. Did you know ferrite changes its reluctance in a magnetic field? No? Nor did I. We had to substitute about 1000 of those stupid coils with brass cored ones that didn't...and stuck a fat electrolytic across the zener to cool it down. It still rumbled a bit, but no one important noticed.

I spent nearly three years of my life getting that lot of stuff working, and after that I got a little jaded, and went on to Extremely Loud Musical Amplification, which had a lot less wires than Wireless, and did at least have the virtue of annoying pensioners and little old ladies, as well as being Definitely Not What Professional Engineers Did.

Mind you, even some of those were capable of acting under unfortunate conditions as both recievers for and transmitters of, extremely high levels of VHF frequencies if you failed to get the output stage stable or stick the right magic blob on the input... As transmitters, they usually Got Hot And Made A Nasty Smell, accompanied by a burst of static in the radio you usually had on in the corner of the lab 'just checking the RF output, Sir, honest'..before a glowing component resulted in a complete rebuild...


Which probably doesn't answer your question at all, but I feel much better for sharing it with you all.















L

[warren52nz]

Well THAT's about the funniest thing I've read in who knows how long!!! :)

And boy, can I identify with it.

So I'm off to buy a Berg Rx and an Azzra antenna! F it. :D

Seriously, that's what I always say at the end of my RF advice.... "after you've done all that, take it out in the field and try it!"

Years ago I also decided to design a high power 12 Volt, inverter driven MOSFET amp for my car and did so. Still working after bouncing around in my car for almost 20 years!

And I used to be a professional musician too (guitar and vocals). Still do occasional gigs with my brother who also moved to New Zealand.

In fact I combined my musical and engineering know how into a simple way to determine your headspeed. The note that comes off the tip of the blades can be directly converted into headspeed by using a piano or other musical reference to determine the note. Then it's just a look up table. I posted it on RunRyder thinking it was a stroke of GENIUS only to get replies saying they "couldn't tell an F# from a fart" and comments like that. Oh well... *sigh*. At least I know MY headspeed now.

As for the antenna thing, I think I'll just measure my Rx input impedance and play around with wire length and base loading coils to get as good a match as practical and leave it at that. After all, if my chopper gets more than about 50 metres away it'll be toast through visual orientation errors anyway.

It's just that I'm battling with a mate who doesn't buy the physics argument and even has his antenna doubled back on itself making his effective length about 4" and he won't listen to my advice. He just says things like "Well it works for me and yours is glitching". Smiling smugly. So I'd like to show him what a bit of engineering can achieve even if it's Black Magic. I'm fortunate in that I have access to some VERY nice RF test gear.

I'm wondering whether I'd be better off measuring he return loss on my antenna and dipping it at 72 MHz or creating a field with my Tx and measuring what's coming off the end of my antenna into 50 ohms. Then there's the issue of my Rx input impedance if it's not 50 ohms. Someone was saying that typical front end has a matching transformer, presumably to transform the low radiation resistance up to 50 ohms or 75 ohms, whatever it is. So the input might be 10 ohms or something else unfortunate!

[vintage1]

Most aerials either go into a single turn of wire around the input tuning coil, or vai a couple of pf into it somewhere.

In order to keep the tuning coil more or less of some sort of Q its a good idea to decouple the aerial as much as possible. I.e. the last thing you want is an arbitrary bit of wire being closesly coupled to a tuned circuit, because the tuning will change as you move objects around in its field.

This makes most of the stuff about 'tuning the antenna for 72Mhz' complete tosh.

For a chopper you definitely want a vertical whip if you can get it. You are severely limited on height by the rotor...its a question of getting the best out of it you can and the azarr is probably about as good as it gets.

In my old 27Mhz supergegen days, I had a caopacitr come unsoldered around the antenna circuit. I still had 20 meters of range with no antenna whatsoever..

4-6 inches of piano wire whip with 2-3uH in the base to sort of get a few dB more out of it by some crude approxiamation iof what a 40" piece of wire looks like at 72Mhz will get you close enough I dare say. That's basically what an azar is, behind all the techno hype.

[warren52nz]

I hear the term "piano wire" a lot. I presume they mean a piano string. Except there are heaps of different diameters. I'm going to try a guitar string. Just not sure how to connect a wire to it. Soldering certainly won't work on hardened steel. Your suggestion about doing that with a bit of inductance on the base is exactly what I have in mind.

I'll have a go on the weekend. It's Friday here in New Zealand.... we're right next to the date line.

Thanks for your advice. Nice to talk tech with someone who knows what they're talking about.

PS Is that your photo in your signature? You sound much older than that photo looks.

Cheers

[hoppy]

Small diameter hardened wire is called "piano wire". It comes in sizes as large or larger than 1/8" and is usually in 36" lengths. You can heat the end to "soften" it.

[warren52nz]

Oh, thanks Hoppy. Having heard the term many times over the years, I've never had it explained to me before. Is it a hobby shop thing? Or where do you get it? I take it you can solder to it?

A hollow tube would work better electrically as it would widen the bandwidth or range frequencies it would work at. Maybe a small brass tube. The signal only travels along the outer few microns of the metal so having a solid wire is unnecessary weight.

Cheers

[hoppy]

Here's one source - http://www.hobby-lobby.com/cgi-bin/search.pl/hobbysearch/

I also get it at hardware stores and local hobby shops. It's commonly used for pushrods and landing gears.

As an antenna - Beats me.:)

[vintage1]

Its got good properties - bebnds but doesn't break or stay bent.

Another useful possibility is a bit of GF rod (NOT carbon) with copper wire wound down it in a helix.

[cmhshh]

As transmitters, they usually Got Hot And Made A Nasty Smell....
Ah that reminds me of a book by CF Rawnsley called Nightfighter. We was one of John Cunninghams radar navigators in WWII.

While talking about the tempermental nature of the early airborne radar sets he mentions 'The awful smell which usually resulted in volumes of smoke and anxious enquiries from the sharp end of the aircraft if the set was not switched off quickly'. :)

SH

[vintage1]

I think thats where I - and the guys I worked for - cribbed it from.

My first boss was an ex RAF technician so its probably down in the book of useful pharses that every radar tech learns.

I've got that book too. Excellent.

[warren52nz]

Don't recognise "GF" rod. But what you're describing sounds like guitar strings and I've got HEAPs of those! The helical winding on the larger guitar strings doesn't seem to make them any more rigid, just thicker so it's probably not a good idea, weight vs strength.

I quite like the idea of thin, hollow copper or brass tube. Good bandwidth, weight and rigidity. And "solderability". And if it kinks you could just straighten it. Not as critical as drive shafts, etc. which need to be perfect in order to work.

[warren52nz]

Actually, just remembered that, in broadcast coax, the ridges on the outer conductor, which are there to make it flexible, contribute to losses so smooth is better.

Just don't know how I'd make a good connection to a guitar string... hardened steel.

[flyingwings]

Attached is a photo of an antennae wire that is embedded in a Skyray wing. Total length of wire is 39.5 inch. Half length of Skyray wing is 8.25 inch.

From your comments cross coupling will probably kill the antennae length. Will adding inductance at the receiver help at all?

Paul

[warren52nz]

I think you'll find it hard to find an RF engineer who would speculate on the effective impedance of that layout. The curves will add inductance and the proximity of the wire where the curves are close together will add capacitance making it some sort of transmission line (as coax is).

Measurement would be best. Others who have played around with designs like this might like to add something from their memoires.

No matter what you do, field testing is the ultimate test.

[flyingwings]

No matter what you do, field testing is the ultimate test.


Thank you for the response. I was anticipating that field testing would be the only final answer.

Paul

[vintage1]

Don't recognise "GF" rod. But what you're describing sounds like guitar strings and I've got HEAPs of those! The helical winding on the larger guitar strings doesn't seem to make them any more rigid, just thicker so it's probably not a good idea, weight vs strength.

I quite like the idea of thin, hollow copper or brass tube. Good bandwidth, weight and rigidity. And "solderability". And if it kinks you could just straighten it. Not as critical as drive shafts, etc. which need to be perfect in order to work.

Oh,. Glass fibre = GF. Like the end of a fishing rod.

Its stiff, but still flexible, and you could simply wind a helix of wire up it and coat it with glass cloth and epoxy to make ain interesting sort of helical whip.

I can't remember exaxtly how this affects perfomance oher han to say that it resonates at a lower frequency than the euivalent length of wire straight up it would.

[warren52nz]

I did some reading up on helical antennas about a month ago in the "Antenna Engineering Handbook" by Jasik (thicker than the Bible!) and, from a once over lightly, it appears that there are 2 modes you can design into helical antennas:

1. one which radiates perpendicular to the helix and
2. one that radiates out the end

Didn't go too far into it as my detailed physics knowledge is rather (!) rusty but you sure wouldn't want an antenna that fired straight out the end... too directional!

I think it's a transmission line of some sort.

I'm patently hit on the idea of finding an RSSI (signal strength) pin on my Rx IC if I can find the number, then plug that into an LM3914 LED bar graph driver mounted right on the chopper. That will give me a continuous readout of the received signal strength. Someone else has done it. Then I can just play around with the antenna until the signal peaks. Now THAT's what I call field testing!

8^)

[vintage1]

What reciever chip are you using?

There is genrally an AGC line somewhere in them.

Its a good idea I have to say.

[warren52nz]

Yeah, not MY idea, unfortunately. 8^)

I've used the LM3914 for O2 sensors on cars before and they have a sufficiently high input impedance to connect directly to the oxygen sensor (over 1 Meg). So that won't be a problem.

The trouble is, the main IC on mine is buried under the IF can by the look of it and it's a surface mount IC (of course) which means microscopic surgery for soldering. But I'm pretty brave, I'll have another, more aggressive look. I'm really chuffed on the idea of having a signal strength meter on the model. Perhaps receiver manufacturers should consider this as an option. It would be a lot easier to implement in a factory.

[warren52nz]

Chickened out of the lobotomy.... ultra miniature soldering in there! I might never get it going again. Only other option I can think of is to randomly scope each pin on the bottom and see if one of them has a DC voltage that tracks the signal strength. Might try that.

[warren52nz]

I've bought some thin (about 3mm) brass tubing which I intend to stick out the front or back of my landing gear as a proper, rigid antenna. Then I'll solder the antenna wire to that and put a few turns on it at the receiver to give it some base loading. Hopefully that will give me a good result.

Two of our club members have bought and installed Azaar antennas which have an active stage of some sort in them (because there's a battery connection on them). The capture area is about 2" long and 1/2" wide (flat). I thought the effective length looked rather tiny but wondered how well they work so we compared one of them to mine which at the time was just the stock antenna wire wrapped around the plastic "straw" that came with my Trex for the antenna with about 6" hanging off the end, straight (well, straight like spaghetti more like it).

Using the same type of transmitter (Optic 6) both on 72 MHz with the Tx antennas down and one at a time, my range was approximately the same as the Azaar, maybe a couple of extra meters for mine before glitching set in.

Now there are a number of other things that could effect the range including the quantity of glitching signals in one model vs the other but it was a worthwhile experiment I think for the effort. I was impressed that such a little antenna could equal a much longer (but much messier) antenna. I guess it has an amp in it which makes me wonder how immune it would be to intermods in high RF fields.

But I'm skeptical of claims that it will "outperform" a longer, passive antenna set up properly although it wasn't a very controlled experiment. It must have virtually no radiation resistance.

[vintage1]

Most recievers are noweher near optimum sensitivity: If the azarr has a booster it could easily perform as well by simply boosting the smaller signal.

[Damske]

Directly from Matthias Schulze:
But it is better to replace the cutted part by a laquer insulated copper wire (use varnished wire - do not use piano wire -> this is not conductive for HF).


I'd think that the man knows what he is talking about
This goes for airplanes, the antenna is not shortened, we just like less drag