ADS-B Patch Antenna on RF Controlled Impedance PCB

In a moment of boredom/inspiration, I thought about making a patch antenna for my ADS-B aircraft tracking system, which is part of ADS-B Exchange. One can make one-off patch antennas from garden variety glass/epoxy FR4 board, but the result will have relatively high loss and it will not be repeatable. In order to simply order high quality batches of antennas from any PCB factory, one needs to use low loss, controlled impedance board stock.

With this antenna, the ADS-B receiver simply sits on my window sill and picks up aircraft 200 km away.

Cut Corners

The design is a typical rectangular patch antenna, with circular polarization. A circular polarized antenna allows more mounting options for a conformal antenna.

Slots

The side slots in this design increases the bandwidth, but it also reduces the operating frequency, so slots can be used to tune a patch antenna design.

Board Type

There are a variety of RF boards made by various manufacturers (Rogers, Isola, Panasonic...) and the RF parameters are different, therefore once you designed your spiffy new antenna for Rogers 4003C, then you cannot change your mind and manufacture it from Isola I-Tera MT40 - you need a new design.

I therefore made a simple design in KiCAD and submitted it to three PCB manufacturers (Eurocircuits, Ourpcb, Pcbonline), to see which board stock is the most cost effective. This showed me a number of things:

The board makers do not necessarily have the advertised stock, or
they may not be set up to actually use it, and
the prices vary dramatically, as I expected.

The most expensive board turned out to be Rogers 4003C, mainly because the manufacturers want to charge me hundreds of Euros to set up their machines to use it. Here is a little relative cost table with the RO4003C cost as 1:

RO4003C = 1
RO4350B = 0.44
I-Tera MT40 = 0.32
IS400 = 0.36
FR4 generic = 0.18

RF boards therefore cost two to four times more than garden variety boards, although the difference becomes smaller in larger quantities and when you order a hundred, it may not matter too much what you make it from.

The prototype patch antenna was made from Rogers 4350B, which has the following RF parameters:

Permittivity Dk = 3.66
Loss Df = 0.0031
Dielectric width Dw = 0.508 mm
Cu 1 Oz = 36 um

Garden variety FR4 board has a loss factor of 0.016 vs 0.0031 of the RO4350B board. This makes a big difference, since you want the antenna to radiate, not heat up like a dummy load. For the amount of time and effort required to design it, one can just as well use the best possible board and get the best possible antenna.

Conformal Aircraft Skin Mounting

This PCB antenna is superior (military) quality and optimized, suitable for use on an aircraft as a transmit antenna.

The patch is slightly flexible, so it could be mounted conformally with the aircraft skin and the weight is only 23 gram, so the antenna will weigh less than the RF coaxial cable. The antenna is circular polarized, so that one can mount it almost anywhere on an aircraft fuselage: top, bottom, wing or tail fin; or use it horizontally on the ground and get a good signal all around.

The best way to mount the antenna to an aircraft fuselage, is on the inside of a RF window, made from fibre glass reinforced epoxy. (Look for an existing small access hatch and redesign the hatch from fibreglass with the antenna bolted on the inside!). If the antenna is in contact with the glass, it will be detuned by about 500 kHz, which is not significant, so you do not need tall spacers. The board can be bent a bit in one direction - it will not affect the pattern significantly.

This way, you get an invisible antenna with no drag.

Tuning and Testing

I measured the prototype board and it was off by a mile and a half - as I expected - with RF, theory and reality never meet in public. So I trimmed it Ye Olde Fashioned Way with calipers, a steel ruler and scalpel, until it was spot on.

The VSWR is now a perfect 1.0 at 1090 MHz, and the 3 dB (half power) bandwidth (VSWR = 6) is about 30 MHz. A patch is very narrow band as you can see on the graph, so you may not need to use a bandpass filter with it - which could save you some money and complexity for an ADS-B receiving station. This is one of the reasons I opted for this design.

The impedance is almost perfect at 49.7 Ohm.

The VSWR is not too important for a receive antenna, but a good VSWR does indicate that the antenna is probably as good as it can be. In practice I can see a small difference between this patch and one that is 6 MHz off: Sitting in my office window, the good one picks up planes way past Brno (100 NM) and the sad one comes a few nautical miles short.

Now I got to measure the tuned antenna with calipers and have another prototype made. In your neck of the woods you may want to use someone else, but Eurocircuits is convenient, since one can simply drop a KiCAD pcb file on their web page to get a quote, you don't need to make Gerbers - they will do the file conversions for their machines.

In one off quantities, a small patch antenna will set you back about 150 Euros - sadly, it is not a cheap hobby...

Question Time

A commenter asked: How do you Tune a Patch Antenna???

Hmm...

The easiest antenna to tune is a telescopic, but when you bump it, it can go out of tune, so I use a permanent marker to draw rings around the joints. One can wipe it off again with alcohol solvent.

A holy antenna made from strips of rigid strapping bolted together, can also be tuned easily by moving the pieces one hole this way or that.

Patch antenna tuning requires some sharp craft tools: A big mug of coffee, steel ruler, scalpel and calipers, pencil and eraser (or sharpie and alcohol), plus a most secret material called Copper EMI Tape.

No matter how careful you are, a hand carved antenna will be off by half a mm, one way or the other and would therefore have to be tuned.

In general almost anything you do to a patch antenna will lower the centre frequency (Fc), but here are some tips:

Drilling a small hole anywhere, will lower the Fc by about 500 kHz.
Coating the antenna with conformal varnish will lower the Fc by about 500 kHz.
Cutting a 50 Ohm stripline feed into it, will raise the Fc by a several MHz.
Cutting slots into the sides will increase the bandwidth a little and lower the Fc by several MHz, depending on how deep you cut, about 2 to 3 MHz per mm.
Trimming 1 mm off the two sides will raise the Fc by 5 to 10%.
Cutting off two 20 mm triangles on opposing corners for circular polarization, will increase the Fc by about 60 MHz.

OK, so you cut off too much - that always happens:

You could resort to cutting the corners to move the Fc far up, or you can add a little bit of metal by sticking some EMI tape onto the antenna, but the effect is not exactly the same as not having cut off too much to begin with...

When is it good enough?

A Voltage Standing Wave Ratio (VSWR) below 2 means that less than 10% of the power is reflected in a transmit antenna and that is usually good enough. A VSWR of 6, is 50% reflected power and that is the 3 dB bandwidth level.

Remember however that a dummy load has a perfect match and transmits or receives nothing, so VSWR is not necessarily an indication of the antenna efficiency - you have to use low loss (expensive) PCB.

A practical lab bench radiation efficiency test, is to wave your hand in front of the antenna (15 cm or so) while watching the VSWR on a VNA. That will quickly give you an indication of the antenna radiation pattern.

As a dedicated home brewer, you can make a patch with a couple of small pieces of copper that you can slide over the slots with a toothpick and sweat solder with a hot soldering iron to tune it spot on.

The standard trick for a manufacturable antenna is to get some sample stock from a PCB manufacturer, design the prototype patch antenna for a frequency 5 to 10% above the desired Fc, then get busy with the ruler and scalpel at the ~~workshop~~kitchen table, to make the antenna the hard way using your mad craft skillz and trim it with side slots until it is on target. Then finally, you can measure it with calipers, set the KiCAD grid and line width to 100th of a mm, draw the footprint and have a board made. If you are very lucky, then that very first manufactured board will be good enough, without further tuning required.

Purchase Orders

If you are interested in buying this antenna, you could send me a message and once there are multiple takers, I'll have a batch made at Eurocircuits, next door in Hungary.

La voila!

Herman

Also see:

Parasitic Quadrifilar Helical Antenna

This article was reprinted in OSCAR News, March 2018: http://www.amsat-uk.org If you want to receive Satellite Weather Pictures , then you need a decent antenna, otherwise you will receive more noise than picture. For polar orbit satellites, one needs an antenna with a mushroom shaped radiation pattern . It needs to have strong gain towards the horizon where the satellites are distant, less gain upwards where they are close and as little as possible downwards, which would be wasted and a source of noise. Most satellites are spin stabilized and therefore the antenna also needs circular polarization, otherwise the received signal will flutter as the antennas rotate through nulls. The helical antenna, first proposed by Kraus in 1948, is the natural solution to circular polarized satellite communications. It is a simple twisted wire - there seems to be nothing to it. Various papers have been published on helix antennas, so the operation is pretty well understood. Therefore,

Aeronetworks sro - Airborne Linux

Search This Blog