Dual antenna for FM

[gordguide]

I am trying to construct an FM antenna for automotive use. I'm trying to eliminate the vertical dipole antenna on the fender of the vehicle; I realize that this is an excellent design for FM, but it is subject to occasional vandalism when parked in the usual public spaces. Put another way, once is enough, twice is too many. We're at "too many".

The vehicle is a (somewhat collectible) 1990 convertible Mazda Miata. I'm rebuilding the car with performance mods. There is no need to be strictly stock in appearance, although the overall "look & feel" will be maintained.

The 1990 and 1991 models did not come from the factory with a power antenna, although it's an option for 1992~1996, which are cosmetically identical. The problem being that a fender brace is missing from the earlier sheet metal, and fender damage from the weight of the power antenna is a (very real) potential problem. So, I've decided to explore other options.

There is a fairly large area forward of the front radiator braces where (on stock models) a plastic composite nose is fitted OEM; on my car eventually this will be replaced by a fibreglass part, and in that case there will be at least the same amount of room there; probably more.

The nose can possibly accommodate 50" horizontally, although that is subject to the ultimate nose modification planned. It is relatively free of metal and could be thought of as mostly empty space.

Because any potential hidden solution will mean horizontal installation, there is an issue of directional reception in a moving vehicle. I have not tested any installations to check exactly how big a problem that may be. A dipole, J-pole or folded dipole could probably fit physically without major issues and with some "air" between it and metal parts of the car.

Another possible option is the use of two shorter dipoles at 90 deg to each other. I'm wondering if the directionality of the horizontal antenna is going to be a problem, and this is an attempt to explore a solution to that.

I've seen a few antenna combiners on the interwebs. I'm wondering if combining the two antennas close to the mount and then running 50 ohm coax to the antenna input would be a good option.

I am at the point where I haven't tested anything but am looking for what I should be testing ... the short list if you will.

The radio that will be installed is period-consistent with the car and has an excellent FM front end and broadly speaking is constructed to a higher standard than the typical off-the-shelf aftermarket deck. However machines of this era don't deal with dual antennas internally as some newer decks, especially OEM decks, seem to do.

Newer vehicles, in particular those from upmarket European makers, seem to be using multi-band low-profile antennas mounted on the roof, back deck, or in at least one case under the fenders. My guess is these units are using amplifier chips to compensate for the small size (eg Atmel ATR4251).

Any comments are welcome.

 

[VEBill]

Generally, when you passively combine two antennas, the signals will add constructively in certain directions and produce nulls in other directions. Cross-dipoles can be an exception, but they're not going to fit.

What about purchasing an active antenna (the $20+ sort that require DC power for the preamplifier) and install one of those in the space under the nose? The short element should reduce directionality, and perhaps it'll even be short enough to install vertically.

Whatever you decide, do a trial installation (if feasible) to make sure it works to your satisfaction.

 

[gordguide]

I'll look into the active antenna a bit further. I've tried them in the home (Radio Shack model) and found them to be substandard to a "real" antenna, but perhaps I should look into it further. It definitely is compact.

The combiner I was thinking of using is a Winegard SD-3700 unit designed to reduce or eliminate the need for an antenna rotator. It appears to be purpose-built for this application but I don't have any experience with them. But, it's clearly a passive combiner.

It's my understanding that the newer automotive radios that do use two antennas have an active circuit that selects just the one with the best signal and switches as required. I don't know if that mitigates the cross-coupling you mention or not, but had I assumed it did. However, I haven't been able to find any circuit diagrams or chips that describe or perform this function.

 

[MikeHalloran]

The stock whip antenna is most likely a monopole with a reflector (the fender). It's sort of half a dipole.

A dipole comprises two poles, almost literally. A proper dipole for FM is about six feet long, maybe a bit more, fed in the center with twinlead. I.e., two wires about three feet long, colinear but not adjacent and not connected, with a ~10mm gap between them, fed by two wires spaced ~10mm apart.

For a couple of decades, some American cars had dipoles laminated into the windshield. They were not as effective as whips, but much more vandal-resistant. I think they had a few inches of the ends folded back, making them folded dipoles, which makes tuning a little less critical.

A dipole requires a 'balanced' feed, which may be matched to an unbalanced output with a simple passive device called a 'balun'. The name comes from 'bal(anced)' and 'un(balanced)'.

In order for an antenna to work effectively, it should be tuned to resonate with the input circuit of the receiver. Combining two antennas will not usually produce good results, because they will in general not both resonate with the receiver at the same frequency, and they will interact with each other.

I have used 'combiners' to use the stock whip antenna with a CB and a regular radio, i.e., two radios at very different frequencies and one antenna, not what you are proposing. They work just well enough to say they work, and not nearly as well as one antenna per radio.

You have a decent chance of laminating a dipole into the front fascia and getting halfway decent reception. Use a balun. Prototype the dipole by using paper tape to attach it to the exterior surface of the fascia. Metallic paint on the fascia may interfere with operation of the antenna.



Mike Halloran
Pembroke Pines, FL, USA

 

[gordguide]

Yes, I'm aware that conventional 1/2 wave verticals are effectively dipoles with the use of a ground plane. Any antenna I build will be tuned to just above the midpoint of the FM band.

When you punch in the values into the math for an FM band antenna you get:
88 Mhz 1/2 wave 63.81 inches; 1/4 wave 31.90 inches.
98 Mhz 1/2 wave 57.30 inches; 1/4 wave 28.65 inches.
108 Mhz 1/2 wave 52 inches; 1/4 wave 26 inches.

Since broadcast power is more effectively coupled (is that the right term?) as you move lower in the band, we can ignore the midpoint and shoot for a little higher, say 101 Mhz, which gives us 55.60 inches and 27.80 inches. That should help give us a bit of a boost to stations higher in the band, which are throwing more watts out and getting less range for the effort. The stations down low don't really need the help assuming the antenna is reasonably close.

Most car mast antennas are pretty close to 27.80' ... many are a bit shorter ... which makes sense, since with the ground plane that's effectively a 55.6' 1/2 wave dipole. Some might be tuned to 5/8 wave, not sure there.

For reasons that I haven't found properly explained it's recommended to build an antenna that is about 5% shorter than the math suggests, and people who do this report good results.

I haven't confirmed it but I suspect it's because of the velocity factor of the material used to construct the antenna (difference between speed of light and actual material used). If that's true, we can look up the velocity factor of any wire used from the manufacturer's specification, and punch that in to the equation. Eg 300 ohm twinlead is 0.82.

I was hoping to fit a 1/4 wave dipole there, so we're looking at about 26", which might fit but if necessary I'll go 1/8 wave at 13" (or so) if I use two antennas. I might try to fit a single 1/8 wave vertical there too ... apparently the ground plane does not have to be that large for FM so I could add that with a square of metal at the base, perhaps?

I have to look into what issues an 1/8 wave presents over 1/4 or 1/2 wave ... I'm sure there is something I'll need to consider.

I'm no expert on antennas by any means (far from it) ... but from what I've read there's got to be some impedance issue to look at when changing to a different fraction of the target wavelength. Baluns don't scare me although I might need some direction on finding the correct impedances and turns ratio.

I understand that using two antennas is hardly trivial. Apparently I need to carefully watch impedance (including carefully match lead lengths of the two antennas) and there could easily be interference with regard to spacing.

I was hoping that orienting them 90 degrees would mitigate that, since we are no longer trying to couple them to increase directionality and obtain an in-phase +3 dB as you would if you were trying to shoot for a distant transmitter with dual or quad setups.

Now that I think about it a bit I might just look at a vertical 1/8 wave alone.

That leaves me with the following to check out:
Active antenna
2x 1/8 wave w/combiner
1/8 wave vertical with ground plane
1/2 wave dipole across the nose

Your comments are welcome and helpful. It's got me thinking a bit on other options to explore. If any of my assumptions/conclusions so far are dead wrong, I'd love to hear about it.

 

[VEBill]

"2x 1/8 wave w/combiner"

You may have seen rural homes with towers holding multiple TV antennas; sometimes stacked facing the same direction, sometimes facing in different directions. Combining those sorts of antennas is the intended purpose of the Winegard antenna combiner. It has no application to automobile FM (primarily because cars move and turn).

When designing a passive antenna for ground mobile platforms (such as your car), your goal is to achieve the lowest possible directivity (antenna gain, not including the loss) in the horizontal plane. This will provide an antenna pattern that is as circular as possible.

If you were going to combine multiple antennas, then those antennas should be stacked in a co-linear vertical array that will provide increased gain towards the horizontal, while maintaining a circular pattern in the horizontal plane. But then it's taller than a quarter wavelength and provides less performance. So you wouldn't bother.

Also keep in mind that commercial FM stations provide very strong signals. A 'wet noodle' (or 'damp string') may work just fine in the city and nearby areas.

 

[gordguide]

Yes, I understand that stacked arrays are used to increase directionality and, done properly, increase antenna gain, in fringe reception areas.

I also understand that the Winegard is a passive device.

I have no intention of building a co-linear array as it increases, not decreases, directionality. The only reason I looked at two antennas is to reduce directionality inherent in the single horizontal dipole.

In my research I was led to believe that if you paid careful attention to the phase of each antenna element (proper impedances, strict cable length and routing, careful element tuning so they are identical, etc) you could avoid out-of-phase attenuation.

In existing crossed dipoles one element is 90 deg phased to the other at the feedpoint, which results in an omnidirectional pattern.

A strong FM broadcast signal is always handy, but my experience with indoor FM tuners suggests that a poor tuner + a great antenna outperforms a megabuck tuner with a poor antenna.

Thanks for your reply. You are absolutely right that lowest possible directivity is what I need to be thinking about.

 

[gordguide]

I think perhaps I should clarify where I seem to be leaning, with your help.

Originally I was thinking along the lines of modern cars with multiple FM antennas. Apparently, after some further research, these units are using what is called a Diversity Antenna system, whereby an electronic circuit selects dynamically the antenna that has the best reception. Delphi Fuba (automotive OEM supplier) is an example of such a circuit, some cars are now using 3 antennas with the Fuba system.

If I were to go forward along those lines, I would have to source some electronics from one of the usual suppliers. At this point I'm not thinking of going that way, so now it's a matter of what kind of antenna to use, rather than "how many".

I'm still looking into the dual dipole since this, properly constructed, is essentially a single antenna with two phased elements. The question I'm exploring now is whether it's essential that the two elements are physically constructed "like one antenna", ie in a cross shape "X", or whether I can separate them physically in a "T" shape for easier fitment.

 

[MikeHalloran]

I can see a lateral dipole in the fascia at either end, or in the windshield, but where would the other one go?
Maybe in an FRP fender, or an FRP hood.
Not in a convertible top.
Maybe in an FRP hardtop.

How about buying a larger car? ;-)



Mike Halloran
Pembroke Pines, FL, USA

 

[gordguide]

The front "fender" is thermoplastic, the hood is aluminum, other body parts steel except for rear bumper which is also thermoplastic.

If you've ever seen a photo of a miata, basically everything forward of the hood/headlight nacelles is air. The radiator is mounted pretty much between those headlights. There is quite a bit of room there to mount stuff under the nose.

As for buying a larger car, well, they're all larger.

I've found something interesting in the pile of antenna literature I've been digging into. Called a Capacitance Hat, if you add it to the ends of your antenna elements it effectively lengthens the element. Some say 30%, others say more is possible. It could be interesting to play with and see what it tunes to.

I'm going out to buy some copper pipe on Monday and start building some antennas. It's time to put some of this to a real-world test, I think.

 

[VEBill]

I own one. I have a crossed-dipole (folded dipoles) FM antenna stored away somewhere. I bought it from Radio Shack when they were clearing them out for a cheap price. The two elements are simply wired in parallel. It is an ideal antenna for omnidirectional reception.

But it's huge. You'd need a ten foot mast installed in the middle of the car.

It's very unlikely that you'll be able to use the same concept embedded within a tiny car.

 

[MikeHalloran]

Copper pipe might be overkill.

One advantage of building an antenna _in_ something is that it doesn't need to be self-supporting. I was thinking 16..22ga solid wire, tensioned with dial cord or fishing line.



Mike Halloran
Pembroke Pines, FL, USA

 

[gordguide]

I'm going to build a few examples and see what happens next.

It may be that I'm trying to overthink the problem; i.e. perhaps there are a number of options that all work fine. It's just that I know from experience that antennas matter.

I will build a Moxon type out of copper tubing and try to see if I can rig up a 1/4 wave vertical with a Capacitance Hat (to reduce overall height) and a ground plane (some piece of metal).

Those two represent pretty advanced versions of the common antennas we would normally buy in stores. The Moxon is a horizontal two element wire beam antenna with one driven element and a reflector. It is very flexible dimension-wise ... all it really needs is to be in a rectangle shape, the corners go where you want them. There should be no problem getting a 1/2 wave version in the nose.

It is a directional antenna with high rejection front/back. Now, a few posts back I said that was exactly what I didn't want, but there's more to the story now.

The factor that changes that is the ground plane ... literally the highway under the car. It's one of those pesky little things that is never in the literature and pattern charts, since it's assumed you will be far, far away from a ground plane unless the antenna specifically says you need to be on the ground plane, in which case it's assumed the ground plane is perfect. Neither is the case with this project.

The Moxon is fairly sensitive to the ground plane (20 feet up, you can still see it's effect in Ham Radio bands), and people who build them pretty much all comment on their results. Were it not for that, I may have missed the significance of the car nose mounting position being low to the ground.

Since that will be less than 1 wavelength away (much less) it will affect the beam pattern significantly, most importantly by reducing the F/R rejection, making any antenna placed at the chosen location in the nose more omnidirectional. It should still have some F/R rejection, which may be desireable since the engine and it's associated noise is there, but nowhere near the -30dB the theory and tests indicate. Nobody is putting their Moxon 2 feet from the ground like I am, so this is purely experimental and interesting to me. It may well end up in the -10 to -3dB range. We'll see.

It is inherently 50 ohms and does not necessarily need a balun, although I may use one anyway to try and match carefully the impedance of the coax to the radio. It's easy to build and can be made quite sturdy if necessary. I'm thinking it will be a good example of how a horizontal dipole-type will work in this area of the car ... in other words, perfect for testing.

The other one I'm going to try is more complex but no less interesting to me. With the discovery of the Capacitance Hat, I was able to revisit the vertical mast as a possible nose-mounted option. It's going to be ugly, so you would not want to use this on the outside of the car, but that's not the case here. The usually reliable sources say it reduces the length required by 30%. That means I can attempt fitting a 1/4 wave vertical in less than 20 inches. That may well fit in the nose. Some suggest the actual reduction possible is greater than 30% in practice.

There may also be further reductions possible by the use of a tuning coil at the base of the antenna. W're definitely in the "full on experimental" category with that one, but I may play with it and see what happens. Or maybe not.

The ground plane is a bit more problematic. The theory says it should be 1/4 wavelength, but an IEEE paper I found suggests maybe not: " ... We demonstrate that often, it is the location of the antenna and its feeding point on the ground plane, rather than the size of the ground plane itself, that primarily establishes the antenna's performance in terms of its impedance, bandwidth, and radiation mode. ..."

Again, it will be interesting to play with.

Thanks for everyone's help. This is not an easy subject to explore, not just because it's complex, but also because (as VE1BLL so succinctly put it) a 'wet noodle' will often work at some sub-optimum level, so mere success at getting a station is no indication you are actually on to anything.

I want to thank everyone for their comments. Whether it's obvious or not, each comment was welcome and helped me wade through the maze.

 

[gordguide]

VE1BLL: I haven't rejected the crossed dipole completely. We'll see what the Capacitance Hat experiment reveals. I can build it 1/4 wave (math says 27", which would mean 4 elements at 13.5") and with the Hats, that goes to [some number less than] 10" each. It was a paper on the crossed dipole that lead me to the Hats. I may also try one for a roof-mount for my house.

MikeHalloran: Copper pipe is probably "overkill" but it has advantages in this location. It's rigid, can be painted (with a non-metalic paint) for appearance, etc. The rigidity may be helpful for fab of standoffs, rather than tensioned wire or mounting the wire against something metal or dielectric which you're not supposed to do. It will also be easier to move around if necessary. I also have enough pipe + elbows to build what I'm interested in.

One of the things I learned in my research is that a dipole does not have to be horizontal ... the explanation was that it's just the way it's been drawn in the usual textbooks, so that's an assumption people have made over the years. Somebody mentioned the in-window embedded wire antennas ... they are bent simply for convenience of installation, to fit the required length (1/2 wave) in the available space in the glass. You can also make them in a host of shapes ... inverted "V", "U"-shaped, etc. There is apparently no performance penalty to the non "T" shaped versions.

 

[Higgler]

Regarding,
I haven't confirmed it but I suspect it's because of the velocity factor of the material used to construct the antenna (difference between speed of light and actual material used).

There is inherent capacitive coupling between the antenna and surrounding groundplane, this added capacitance makes the antenna look electrically longer, so one needs to shorten the overall height to get your effective length back to what you wish.

What would work well is a monopole on the inside front window of the vehicle. Use a triangle shape with mulitple thin wires or silver conducting epoxy after you prove the wires work. Need about 23" height (1/6 wavelength at lowest frequency). Combining more than one or switching between two or 4 antennas may give a little extra performance.

 

[Warpspeed]

Does your vehicle have a heated rear window for demisting ?
If it does, then there is your FM antenna.

Just fit a resonant series common mode choke that will carry the required (maybe) ten amp heating current.