Diary: Testing homebrew DIY ADS-B antennas - Feb 2017

Comparing several homemade DIY ADS-B antennas against a pro FlightAware antenna.
Why? The antenna is a key element in an ADS-B setup, but cost can be a barrier of entry. Undoubtedly, a pro antenna (comparison of five commercial antennas here) is the easiest and a homemade antenna is significantly cheaper than most ready-made solutions, but I haven't came across a systematic and published side-by-side performance comparison under real-life conditions for ADS-B antennas.
Sort of a diary, really, I hate excessive technical information whenever unwarranted, but in this case, describing minuscule tidbits are justified as circumstances will influence reception performance.
There's no "assume", simulations, infinite ground planes or conjecture here - I've built antennas, placed them on four mounts on a window ledge, and documented results. Including my errors too, because detailing mishaps might help others trying to run a similar test, something I'd love to see and then share results to the public domain - better approximation of a homemade antenna performance can be only achieved with more data at different locations with different setups.
Images open full-screen when you click or tap on them, right-click and save images to evaluate data. Links open in a new tab, some product links are Amazon affiliate links (you pay the same, full disclaimer here).
If you want the best ADS-B performance easily: buy a FlightAware antenna (Amazon USA) and an Uputronics 1090 Mhz filtered preamp (distributors for UK and USA). $100 well spent, but 1) that's a lot of money for most including yours truly, and 2) we're discussing the other end of financial existence, so none of the homebrew DIY antennas described here cost more than $1, utilize soldering or require expensive equipment.
This is an ongoing test, I'll update with further testing results, follow on Twitter (blue button top left of screen) to get notified.

Special thanks to

I could not have conducted this test without manufacturer support, therefore would like to express my heartfelt thanks to the following companies in alphabetical order: FlightAware for reference antenna and Pro Stick Pro dongle, ModMyPi for two Pi3 Starter Kits, Nooelec for RTL-SDR dongles and connectors, rtl-sdr.com for dongles, Uputronics for 1090 MHz filtered preamp and wideband preamp.
This test is not sanctioned, endorsed nor approved by any of the manufacturers listed above, opinions here and on this webpage are not influenced by them (further details in my Manifesto), but I wanted to make it clear that I could not have done a test involving hundreds of dollars of equipment without their help.
Please support further tests by buying equipment from manufacturers linked above.
Many thanks!

Testing environment

I live in the city centre of Cork, which is the second-largest city in the Republic of Ireland, official population ~ 120,000.
Maximum theoretical range of an aircraft flying at 40,000 feet indicated by blue circle, forum post link how to calculate for your location:

Antennas are 18 meters above sea level, house blocks line of sight in a semicircle, restricted direct line of sight to horizon from NW to SE - in short, conditions most city dwellers will face.
I get several WiFi signals on a smartphone and on a tablet at antenna location, multiple mobile phone and local utility towers within a five-mile radius. In a recent test, a $50 1090 MHz cavity filter did not improve reception compared to an FA filter, but a filtered Uputronics preamp did compared to an unfiltered one, so I'd say my location is experiencing medium to high levels of electrical noise. It's probably not as bad as Times Square, but far from the Australian outback.

Back-end computing

Results are publicly verifyable from FlighAware station numbers, I refer to stations by IP number in this text: IP 17 is #29213, IP 16 is #29214, IP 08 is #37296, IP 03 is #37301.

Polar plots from planefinder, IP 17 is #94562, IP 16 is #94561, IP 08 is #95661, IP 03 is #95662.
Four ADS-B stations based on Raspberry Pi microcomputers, consisting of two identical Starter Kits supplied by ModMyPi (IP 17 and IP 16, pink ones in images), and two of my own, a Pi 2 (IP 08) and Pi 3 (IP 03). Two 2.5A power supplies coming with ModMyPi Starter Kit, wired 2A power supply for Pi 3, 5V 2A smartphone charger with USB to mictoUSB cable for Pi 2.
RTL-SDRs: Nooelec SMArts, and rtl-sdr.com v.2 and v.3 used interchangeably, previous testing revealed that there's no major difference in ADS-B performance. All testing carried out unless otherwise noted with unmodified, unamped and unfiltered dongles.
Nooelec's F type female to SMA Male pigtails used with F type connectors, transmission line and homemade antennas are from the same 75 Ohm RG 59 B/U cable off the same drum: double-shielded coax with external diameter of 5.76mm, center conductor diameter 0.58mm (AWG 23). Length of coax is 5 meters / 15 foot.

Units of measurements throughout this text are metric units, millimeters and meters from now on: antenna building requires millimeter-level precision, and the majority of the world uses metric, 2.54 mm equals 1/10th of an inch if you're more familiar with imperial.
Antenna elements are expressed in wavelength, full-wave (1 lambda, also written as 1/1) is 300/1090=2752 mm or 27.52 cm. 1/2 is 13.76 cm, 1/4 is 68.8 mm.
Antenna building accuracy to the nearest millimeter, measured by digital caliper. Cutting an antenna to sub-mm precision level is a challenge: blade is 0.6 mm wide, so a quarter-wave element can be 68.5 to 69.5 mm in real life. A household match is ~ 1mm.

Distance unit is the nautical mile, which is 1.852 kilometers, equals 1.15078 land mile. Polar plot references use azimuth, where 000 is North, 090 is East, 180 degrees is South, 270 is West, you do the math for interim cardinal points. For "max is 164.33 @290" read maximum range of aircraft was 164.33 nautical miles on 290 degrees azimuth. I usually get maximum range with aircraft flying in the Europe - USA flight corridor around 290, maximum range ever for my location was ~ 175 nm.
Receive performance measured against a $45 FlightAware antenna (Amazon USA), expressed as percentage of total position reports achieved with FA antenna (=100%). I honestly think if you have spare $50 and never built an antenna, stop reading and save yourself time, effort and unnecessary pain and order one right now with an appropriate pigtail according to your receiver standard - e.g. this N Male to SMA Male for v.3s and SMArts.
Latest PiAware 3.3.0 used, my installation steps here or official link, standard image without any adjustments, did not change gain manually, simultaneously feeding FlightFeeder for polar plots. Guide how to feed multiple sites here.
PiAware starts a new day at 0000 GMT, which is local midnight in Ireland, so I strive to change antennas as close to 2350 as feasible. There's barely any traffic after 2300 in my part of the woods, on a a normal day I usually get between 20,000 and 25,000 position reports, 30,000+ is rarely seen:

Changing antennas at midnight is not always possible due a small hindrance called having a life, but all testing is carried out for at least 20 hours unless otherwise noted.

Your (aeronautical) mileage may vary

No two locations are the same, electrical noise, WiFi, day-to-day weather and aircraft density variations, impedance mismatches, different receivers, antenna materials and construction errors, variation in coax cables WILL have an effect on ADS-B performance (list is not exhaustive).
I built antennas, placed them outdoors next to each other, took images and recorded results. My sole purpose was to test antennas against a reference antenna on the same day, to see what works and what does not.
I don't have expensive lab testing equipment or an SWR meter, because $45 buys a purpose-designed and excellent antenna with a sturdy mount. However, I like building and testing antennas, and firmly believe that a homemade solution can be nearly, and hopefully as good as a commercial product.
Multiply four stations with power supply, microSD card, Pi, LAN cable, dongle, pigtail, F-connector(s), coax, antenna connection, then antenna. That's minimum 40 possible points of error just in hardware, than you got antenna positioning differences and weather, so results are exact, but not absolute, far from it.
Building an antenna and expecting the exact percentage difference versus a FA antenna at a different location will just not happen. As you will see below, on otherwise identical testing setups my figures varied from day to day, and from antenna mount to mount. Same antenna at a different location in a different setup had less performance, and I had different end results with the same antenna as an other tester.
Once again: Your results will NOT be the same.
Links to inspirational ADS-B antenna building posts are included, but not only those referring to antenna designs that are supposed to work according to simulations, I tried designs that "should not work". Limiting antenna choices to tried and tested designs is counterproductive, when progress is made when enthusiasts stray off the beaten path.
Credit and acknowledgement to all fellow maniacs who published their results for the benefit of others.

Day 1 Wednesday

Shakedown to iron out possible problems, check that all stations reliably feed data, mount won't fall on neighbor's roof. Bottom of mount is 39 cm above sloping tile roof (top of FA antenna is 108 cm above), 3.56 meters above ground.
Restricted line of sight due to trees, parabolic antennas and surrounding buildings:

Horizontal separation of 24 cm between antennas, I refer to mount positions from now on as 1-2-3-4, from right to left in the image below, so FA antenna is on mount position 2.
Three homebuilt antennas, 1/4, 1/2, 5/8 monopoles, just strip off  insulation and braid from end of coax cable. Discone in foreground marked by red arrows is not part of this test.

Interesting results:

Surprisingly, half-wave monopole collected 64.56% of FA antenna's totals.
Polar plots:

FA outer range ring is at 150 nm, rest is at 125nm, shapes look similar, but mount psn 4 exhibits different maximum range and reception pattern, so will be used for experimental antennas.
Added elements by pushing appropriate length of wire between insulator and braid and put out antennas in a downpour, see what happens tomorrow.

Day 2 Thursday

1/2, Full-wave and 1.25 Dipoles.

1/2 dipole must have had an issue, suspect water ingress due to exposed connection. Full-wave dipole surprising, 91.7% of totals, FA max range 160.3 nm, full-wave range 140 nm.

Day 3 Friday

Same setup as yesterday, to my utter delight, FA antenna refused to cooperate, so no useful figures for comparison today. Full-wave dipole 19,759 Totals, max range 132.94@135.

To be sure that error is not in my setup, took in all antennas, went through connections with a fine comb, de- then reassembled FA antenna. All antennas in the same pigtail => F connector => coax configuration, only variable is the antenna in use. All antennas use the same cable with same pigtails, so the playing field is level for all contestants.
Built tomorrow's menu: half-wave groundplane with four half-wave radials (this is also a type that won't work), cantenna with 1/4 element by abcd and kept in high regard by lots of users, and abcd's recently promoted no soldering, no connectors 1/4 ground plane with eight 1/4 radials.
IP 08 was also acting up all day today, so placed 1/4 8 radials on it.

Day 4 Saturday

As suspected, IP 8 station either has an issue, or the 8 radial antenna is this bad, or placement is off - I strive to keep antenna tops on the same level, but then it's not possible to keep 8 radials equidistant.

Cantenna with 1/4 element was fantastic 88.84%, element is centered and measured from start of F female-female connector.

Cantenna without camouflage sticks out like a sore thumb in a window, so taped it with black tape.
Switched over to rtl-sdr.com dongles, as I use them less than SMArts, two v.3 in boss setup and two v.2 in secondary setup:

Took halfwave with halfwave radials out of circulation, 8 radials 1/4 wave remains for further testing on a different station and on a different mount, can't explain its behavior. It's very likely there's a construction error I made.

Day 5 Sunday

Cantenna performs on the same level at the same mount position, 88.84 % yesterday, 87.62% today, 8 radials a.k.a. 8-legged spider somewhat better, but still unacceptable.

Polar plots for 1/4 element cantenna vs FA antenna:

Swapped 1/4 cantenna to 1/2 top cantenna to see what happens.

Day 6 Monday

1/2 cantenna much worse than 1/4, 44.85% only, puny range.

Connected full-wave dipole (two 1/2 elements) to IP 08 with reimaged microSD card in the afternoon, really curious whether station works, or I'll have to troubleshoot again. Suspect radials are influenced by rain, will test this assumption later on.
Built two antennas: four 1/4 radials with 1/4 and 1/2 top elements.
Smothered top with hot glue, adjusted four droops to cardinals. Tops are slightly higher than FA antenna, but I wish to keep ideal shape, doubt a finger length in elevation will have an effect on performance.

Day 7 Tuesday

Quarter-wave ground plane 1/4 over four 1/4 radials excels, full-wave dipole is close, 1/2 over 1/4 four radials dismal. Antennas marked with red arrows, tops not on the same level:


Impressed by 1/4 ground plane 4 radials, takes a few minutes to make, only requires coax and a way to measure 69 mm, 94.51% of FA totals, max range 151.09@ vs FA max range of 149.88@250. Dipole Only 75% of FA today on mount psn 4, 115.56@250 and 127.33@290, still, the easiest to make of all antennas so far.
Plugged out stations at 2350, tomorrow's menu: manufacturer stock black antenna, same cut back to 1/4, telescopic supplied with many Nooelec dongles (Nano 2, 2+, Mini 2, 2+, XTR and XTR + among others) at 1/2, all on small magnetic mounts, all on identical small dog food / beans tins: diameter 7.4 cm, height 10.8 cm.
Slowly chewing through suggestions on coax collinears and Franklins, 2, 4, 6, 8, shorted top or not.

Day 8 Feb Wednesday

I'll have to rerun today's setup, stock and stock cut to 1/4 no data due to water pooling on tin.
Telescopic at 1/2 astounding, 5.42% better than FA:

Stock telescopic at 13.76 cm max range 162.31 nm@250 is better than FA antenna max range 148.54 nm @250, weather was horrendous all day, like standing under a shower. I've done a similar test earlier, stock telescopic was only 81.93% then in a different setup. I suspect more and more that antenna position on mount plays a part, so shifted 1/2 telescopic.

Day 09 Feb Thursday

No changes from yesterday, suspect two mounts are full of water, been raining heavily over the last few days.
1/2 Telescopic 88.98 % Totals - falls in line with previous test, note to self: outstanding result must be cross-checked at different mount positions.
Polar plots, FA 159.21@135, Telescopic 143.54@135.

Met up with local ADS-B feeder and enthusiast, given him a full-wave dipole, 1/2 arms, to replace FA antenna in his setup, consisting of orange Pro Stick and filter. Long-term results will tell the tale at an another and indoor location.

Day 10 Friday

IP 8 unfunctional, IP 16 FA seems reliable. IP 03 with SMArt's telescopic extended to 13.76 cm got 85.37 % of FA Totals in same mount psn, but less range.

Telescopic at 1/2 was 85.37% today and 88.98% yesterday - day-to-day variation of an antenna's performance at an outdoor location.

Day 11 Saturday

Day off, needed Pis. Looked at mounts and took them in, as suspected, filled with water, explaining earlier results.
Noise profiled all dongles used so far, nothing that would indicate an underlying problem influencing reception on the receiver side.

Day 12 Sunday

Mark as off, system, self and relationship maintenance day.
FA antenna is stable and provides data, new equal lengths of coax on stations 16, 03 and 08, same lengths as FA coax but with F connectors and F female-female barrel to save on coax when swapping antennas.

Monday 13th February

Made 8-element CoCo with velocity factor of .66 and .80 to be on the safe side (exact value unknown), then put out two CoCos and folded dipole. Then 08 has given up the ghost, again, so today's data will not be useful for comparison purposes as I had to dismantle all stations and retest components with SDR#. 0.66 CoCo simply does not work: I know that high gain is possible and some get excellent results, but without knowing the exact velocity factor, good performance is nearly impossible, or by pure luck.
Made loop with screw connectors with and without 4:1 balun to see whether it'll feed data.

Inability to allocate time to testing combined with equipment failures are very, very upsetting.

Tuesday 14 Feb

Finally, a full day's data without any kind of mishap.
Loop, Loop with 4:1 balun, .80 Velocity factor CoCo.

Max Ranges: FA 152.24  @ 250, Loop 115.33nm @ 290, Loop with balun 116 nm @ 195, CoCo 102 nm @ 250. CoCo dismal performance, I do not know the exact velocity factor, will do one more with longer elements, trial and error.
Loops are a dead-end street at the moment.

Wednesday Feb 15

Full-wave dipole day today: 16 is parallel to coax, 17 FA, 08 terminal connectors with coax angling 90 degrees away from antenna vertical, 03 with 4:1 balun, coax at 90 degree angle.


Parallel probably had water ingress. Terminal connectors 64.25%. 4:1 Balun highest, 78.08%.
Ranges: FA 153.32@250, Full-wave with 4:1 Balun
Looked at results of independent site, antenna change at 7pm on the day (full-wave dipole parallel to coax) indicated by arrow, indoor amplified and filtered setup, raw data:

More data:

Results could be down to the presence of amplification and filtering, indoor location, or favorable terrain. Or all of them. Ultimately, 61.59 % average performance of FA antenna over 5 days.
Won't touch my antenna setup tonight, only shift them around to test if antenna location has any affect.

Thursday Feb 16

Same antennas as yesterday, but at different mount positions. Parallel to coax definitely had water ingress at one stage, rest working as they should: IP 17 FA max 155.80 @ 250 and 149.20 @ 130, IP 08 in terminal connectors 71.73%, max 150.88 @ 130, IP 03 with 4:1 balun on a full-wave 93.58%, max 143.54 @ 135, results at 2302:

Different results after swapping mounts, 71.73% vs 64.25%, 78.08% vs 93.58%. Today nearly 30% more traffic than yesteday so larger sample size.
Built two more CoCos for tomorrow.

Friday Feb 17

Looked at results at 0630 and CoCos don't work, despite triple-checking everything, meticulous watchmaker precision with digital caliper to mm level, continuity and shorts checked with a resounding beep after adding each element.
My life expectancy is just too short to spend hours measuring plus/minus 1 mm for a CoCo and to rebuild by guessing exact velocity factor.
Data for today: IP 08 and IP 16 with experimental antennas, 1/2 and 1/4 over two 1/4 radials, didn't really work out. IP 17 FA, IP 03 full-wave dipole with 4:1 Balun, day's results at 2340:

83.14% today. 78.08% and 93.58% earlier, illustrating that one comparison with outdoor antennas will not yield an exact result.
Made Cantenna from a drinks can as per inspirational link tonight, 1/4 top element with a twist: I'm not using an F female-female connector because I hate connectors and wanted to simplify the design at the same time:

I've never seen this solution anywhere, a cantenna in principle but easier to do, doubt it will work, it's too simple. Also made two 1/4 over four 1/4 groundplanes, one with 45 degrees, one with vertical radials.

Saturday Feb 18

Morning coffee, then jaw drops: cantenna from drinks can is better than FA at 1035. What's going on?

Raw data off FlightAware control panel included here for accountability, IP 17 is FA, IP 16 is drooped 1/4 over four 1/4 radials, IP 03 is cantenna from drinks can below in yellow, IP 08 is not on:

Today's data shows a cantenna from a drinks can was 28 % better than an FA antenna, but one observation is just one observation, need to replicate antennas tomorrow, will also test 1/2 can and 1/2 element. Haven't seen that design around, see what happens.

Sunday 19 February

IP 16 and IP 17 started the race on the gunshot, IP 08 acting up (had to be reimaged, no planefinder data for today, sorry), IP 03 1/2 over 1/2 can wasted time.
Swapped at 5am out to "Coketenna", and that's a new term, will refer to any antenna made from a drinks can with external diameter of 65 mm from now on as "Coketenna", primarily to distinguish them from Cantennas made from various diameter food tin cans, and because drinks can are quite uniform all around the world and readily available. Measured multiple soft drink cans and beer cans, and 65mm is a common denominator, aluminum wall thickness varies from 0.12 to 0.18 mm.
Three Coketennas against FA antenna, stations off at 2340, data shows that Coketennas are significantly better than FA antenna. All of them, but only IP16 is comparable as IP 03 did not feed from midnight, still they show higher totals:

Why? How? What's going on? All three Coketennas were built in less than 10 minutes, cost nothing, require only one measurement.
I've brought in FA antenna as there are way too much variables in the existing setup for my liking, including the remote possibility of FlightAware antenna problems - the Achilles heel of an FA antenna is water, and it's Noah's daydream here in the last few weeks.

Monday February 20

Mark data today as OFF for comparison purposes.
Four factory-fresh SMArt bundles arrived, tested all four dongles before commissioning, as usual, very close performance by ear, by noise profiles and by measuring case temperatures with infrared thermometer.
Compared FA antenna's performance against local feeders to check for consistency:

Yesterday's FA antenna performance was not on the same level as during the week, therefore Coketenna performance yesterday could have been a one-off.
After all this, I only got 8 points of error for each station: antenna - mount - coax - dongle - Pi - microSD card - Ethernet cable - power supply. I cannot change, nor influence other factors, tomorrow's full day data will establish a better margin of error - I want a figure for system variance before continuing testing and cross-checking existing results.

Tuesday Feb 21

Data at 1628, all three stations with the same antennas, 15.33% difference in max/min Totals:

IP 03's antenna (at FA antenna's mount position 2) showing highest totals, but it was not on the same axis as other mounts (off by 10cm), used a different can as magnetic mount, and to top my misery, polar plots significantly differ, plus too short testing period. Damn, another day wasted.
Rebuilt mount with four identical cans, reimaged stations, reinforced mount. Then by pure luck, found that IP 08's Ethernet cable is intermittently unfunctional, and the microUSB power cable sheath is partially broken. Cannot express with words how much I hate connectors.
Right now, all four stations seem to reliably feed data to FA and planefinder. Only took three weeks.

Wednesday Feb 22

All stations with SMArt bundle's smallest antenna, all on same cans, with same mounts, horizontal position is the only difference for comparison purposes. Black bags over antennas to minimise the effect of weather - it's raining heavily as always, at least the sun comes out now and then:

Mount 1 was IP 08, Mount 2 IP 16, Mount 3 IP 17, Mount 4 IP 03.

12.63% difference in totals between Mount 2 and Mount 3, which are only 24 cm apart. Mount Position 4 plot is totally different from other mounts with less data as suspected.

Thursday Feb 23

Mark as OFF, tested other RTL-SDR dongles.

Friday Feb 24

Same antennas, but swapped stations around to minimize the effect of dongles and back-end setup , Mount 1 is IP 16, 2 is IP 17, 3 is IP 08, 4 is IP 03:

Mount 2 vs Mount 3 difference is 3.44% vs 12.63% two days ago, Mount 1 vs Mount 3 is 9.13%. Mount 4 is seriously lagging behind.
Polar plots comparison:

Mounts 1-2-3 exhibit quite similar reception patterns, maximum range was  ~ 144 nm @ 290 for all four mounts (plus-minus 0.3nm), which is the only azimuth on which all four antennas have an unrestricted line of sight to horizon. 1-2-3 will have three Pi 3s, mount position 4 will be reserved for experimental attempts with Pi 2.
Safe figure: 15% margin of error due to testing setup on three  mounts. I cannot lower this number without vacationing at a hilltop, where I could place antennas apart with 360 degrees of unrestricted line of sight - hmmm, sounds like a nice day out. On the other hand, it's astonishing that such a small horizontal difference clearly shows up in polar plots.
Met with independent tester armed with the very Coketenna I used for my testing, will see how his results compare to mine.

Saturday Feb 25

Manufacturer antennas: SMArt's smallest on IP 17, stock fixed black on IP 16, stock cut back to 1/4 on IP 03, IP 08 on mount 4 with v.3 telescopic extended to 1.5 but unfortunately tipped over in high winds.

Data at 2300:

2.31% difference in totals between 1/4 stock cut and SMArt's coiled whip, max range was 139.15@290, black stock cut to quarter-wave had 140.85@290, commonly shipped black whip 135.06@290. 27.74% more totals by snipping the top off.
Magnetic mounts are bone dry, despite horizontal rain combined with frequent and strong gusts; plastic bags work for a few days in a row. Look ugly as hell thought.

Sunday Feb 26

Small stock black cut to 1/4 (69mm, testing the argument between ideal 1/4 dimension due to coax length inside magnetic mounts will wait) on 03 versus Nooelec's supplied telescopics, 1/2 with top sections extended on IP 17, 1/2 with bottom sections extended on IP 16, stock at full-wave (conveniently only have to collapse the top section) on IP 08.

Data at 2357:

Bottom sections extended seems to gather more data (13.25% more), but figures within 15% margin of error, noted that wind load is greater than with top sections extended as I had to adjust antenna to vertical orientation twice during the day. 1/1 (27.52cm)Telescopic on experimental mount 4.
Max ranges: 1/4 cut Common Stock 143.54@135, 1/2 Telescopic top section 134.81@290, bottom section 138.07@250 and 133.80@290, Telescopic full-wave 127.33@290, had 61.55% Totals vs other three mounts averaged.
Why full-wave telescopic? Because it's borderline usable for airband due to longer length.

Monday Feb 27

Same antennas as yesterday, day's results at 2351:

Bottom section extended had 6.74% more Totals. 1/1 Telescopic had 61.81% of other three mounts averaged, 61.55% yesterday.
1/4 stock cut had less totals two days in a row than 1/2 telescopic, but figures are so close that it's impossible to say which  one is better. One thing is certain: telescopics extended to 1/2 work quite well.

Tuesday Feb 28

Testing the difference between a stock RTL-SDR (v.3 i) versus an amplified RTL-SDR (FlightAware Pro Stick, $17 on Amazon USA) versus an amplified and filtered RTL-SDR (FlightAware Pro Stick Plus, $21 on Amazon USA, £25 at European distributor ModMyPi)
Mounts 1-2-3 with same small antennas (SMArt's small coiled whip), experimental mount 4 with SMArt telescopic extended to 60 cm e.g. 1/4 at 125MHz for airband (118 - 134 MHz) on an rtl-sdr.com v.3.

Data collected between 1330 - 2347:

Ranges: v.3 representing unamped RTL-SDRs 135.00@135 and 132.03@290, Pro Stick 147.64@135 and 13.59@290, Pro Stick Plus 135.43@135 and 140.98@290, caught a plane 156.66@250. 60cm antenna with unamped v.3 122.94@290.
Left antennas in peace, want to run setup for a full day, plus it's -2 degrees Celsius outside with horizontal rain and howling winds.

Wednesday Mar 1

Same config as yesterday, day's results at 2200:

Polar plots comparison:

On a medium-gain antenna. My location rarely exceeds 25,000 position reports with a FA antenna, no matter if a semi-pro Uputronics or consumer-grade FlightAware gear is used. Other feeders in my geographical area did not had such a bump, so it's clear that a Pro Stick Plus did benefit from a lower gain antenna at my location.

And on that happy note February's incessant failures, and this post is complete.
Testing goes on, March 2017 is in progress.


  1. In my opinion a like by like comparison cant be made as they are at different heights & not clear of the building which will affect their coverage, plus they will each affect each other, they would each need to be above the house in the clear with nothing around it on their own.

  2. All antennas at the same vertical height.

  3. Congrats man. Simply awesome work and detailed step-by-step explanation. Do you know if FR24 has similar statics plots to study the behaviour of diy antennas? I'm at the third diy antenna and still looking for best one.

    1. As far as I know it does not, however, I exclusively use PiAware with Planeplotter for stats.
      The best and easiest concept DIY design is the Coketenna me thinks :-)))

  4. This comment has been removed by the author.

  5. What program are you using and how are you getting those plots on the map for your antenna? I'd like to do the same kind of testing.

  6. A great read and i now know what i should be building ie a simple cantenna rather than a complicated collinear which would probably be worse than the FA. Thank you so posting the comprehensive results. Oxford.