FlightAware Discussions

Discussing best 978 MHz hardware setup

Can someone pls make a heatmap with:

rtl_power -f 977M:979M:125 -i 15 -e 30m scan.csv

and share it? Will take 30 minutes to collect data. I’m very curious if other guys have the same line at 978.1

In progress. Will post results after about 30 minutes.
75mm Magmount antenna on Cookie can, Flightaware Filter (light blue color), FA ProStick (orange color)

pi@raspberrypi:~ $ sudo systemctl stop piaware
pi@raspberrypi:~ $ sudo systemctl stop dump978-fa
pi@raspberrypi:~ $ rtl_power -d 00000978 -f 977M:979M:125 -i 15 -e 30m scan.csv
-bash: rtl_power: command not found
pi@raspberrypi:~ $ sudo apt install rtl-sdr

pi@raspberrypi:~ $ rtl_power -d 00000978 -f 977M:979M:125 -i 15 -e 30m scan.csv
Found 2 device(s):
  0:  Realtek, RTL2832UFA, SN: 00000978
  1:  Realtek, RTL2838UHIDIR, SN: 00001090

Using device 0: Generic RTL2832U
Number of frequency hops: 1
Dongle bandwidth: 2000000Hz
Downsampling by: 1x
Cropping by: 0.00%
Total FFT bins: 16384
Logged FFT bins: 16384
FFT bin size: 122.07Hz
Buffer size: 32768 bytes (8.19ms)
Reporting every 15 seconds
Found Rafael Micro R820T tuner
Tuner gain set to automatic.
Exact sample rate is: 2000000.052982 Hz
[R82XX] PLL not locked!

Upon completion of scan, restored system to normal by Rebooting

pi@raspberrypi:~ $ sudo reboot


Installed heatmap generation tools

pi@raspberrypi:~ $ sudo apt-get install python-imaging 
pi@raspberrypi:~ $ sudo wget https://raw.githubusercontent.com/keenerd/rtl-sdr-misc/master/heatmap/heatmap.py 
pi@raspberrypi:~ $ sudo chmod +x heatmap.py 

Generated Heatmap

pi@raspberrypi:~ $ ./heatmap.py scan.csv scan.png
x: 16385, y: 120, z: (-40.010000, -9.940000)
pi@raspberrypi:~ $


The thick black line below is scanned IMAGE

Uploaded file “scan.csv” to Dropbox. You can download if you want to generate image by another tool



Thanks! It is significantly different from mine, you’ve got a lot of green lines, where I basically have none. No idea why this is like that. Maybe because I’m not in the city. Maybe because my setup is bad.

I used this to make heatmap: https://github.com/dhogborg/rtl-gopow/releases

Don’t run it on Pi - download CSV to big computer and run it there.

FWIW, UAT is 2-FSK so it transmits on a pair of frequencies, nominally 978.3125MHz and 977.6875MHz.

The uplink signals might be long enough to show up on a realtime waterfall display (they are ~4ms long), but I wouldn’t expect much to be visible on a longer duration scan like rtl_power.

1 Like

Here’s an example waterfall with UAT data.
The steady vertical lines are most likely internal noise within the dongle itself.

The periodic bursts centered on ~977.687MHz will be mostly-empty uplink messages - almost all the bits will be zeroes, which is why it’s centered on only one frequency.

The wider bursts that stretch up above 978MHz will be uplink messages that actually contain some non-zero data.

To give you an idea of the timescale, these bursts are roughly 1 second apart.


Thanks for the details. I’m more worried about any interference around that may be harmful for my setup.
But congratulate me, I got my first UAT aircraft today. Also telescopic antenna will arrive today.

Experimenting with telescopic antenna. Not sure right now, but looks like if it is extended to 1/2 wavelength (153 mm) it works better than 1/1 - 307 mm.

Funny 978 wavelength amazingly coincides with 12 inches = 1 foot, giving 6 inches for 1/2 and 3 inches for 1/4, US folk will be happy!


Read some of that.

Basically 1/2 wavelength dipole is optimal.
With longer antennas you need impedance matching.

As abcd567 just posted, you can also try 1.25 instead of 0.5 wavelength.

Gain of Dipole & Monopole Antennas

Note that for monopole with ground-plane, length is 1/2 of dipole.

Dipole Length in wavelengths Monopole length in wavelengths Directive gain (dBi) Notes
≪0.5 ≪0.25 1.76 Poor efficiency
0.5 0.25 2.15 Most common
1.0 0.5 4.0 Not used
1.25 0.625 5.2 Best gain (5/8)
1.5 0.75 3.5 Third harmonic
2.0 1.0 4.3 Not used


Dipole vs Monopole Antenna



A monopole can be visualized as being formed by replacing the bottom half of a vertical dipole antenna ( c ) with a conducting plane (ground plane) at right-angles to the remaining half. If the ground plane is large enough, the radio waves from the remaining upper half of the dipole (a) reflected from the ground plane will seem to come from an image antenna (b) forming the missing half of the dipole, which adds to the direct radiation to form a dipole radiation pattern. So the pattern of a monopole with a perfectly conducting, infinite ground plane is identical to the top half of a dipole pattern, with its maximum radiation in the horizontal direction, perpendicular to the antenna. Because it radiates only into the space above the ground plane, or half the space of a dipole antenna, a monopole antenna will have a gain of twice (3 dB greater than) the gain of a similar dipole antenna, and a radiation resistance half that of a dipole. Since a half-wave dipole has a gain of 2.19 dBi and a radiation resistance of 73 ohms, a quarter-wave monopole, the most common type, will have a gain of 2.19 + 3 = 5.19 dBi and a radiation resistance of about 36.8 ohms if it is mounted above a good ground plane.

The general effect of electrically small ground planes, as well as imperfectly conducting earth grounds, is to tilt the direction of maximum radiation up to higher elevation angles.

1 Like

Wow thanks for details, I was trying to google that but could not find anything like this about lengths of antennas. So 5/8 will likely be the best length for telescopic, that is 192 mm.

The impedance of a 5/8 monopole will be very high, around (75 - j425) ohms, so you will need an impedance matching device to match it to (50 + j 0) ohm. Without impedance matching, you will loose all benefit of high gain, and performance will be less than that of a 1/2 wavelength dipole or 1/4 wavelength monopole+groundplane.

Image: Wikipedia

Real (black) and imaginary (blue) parts of the dipole feedpoint impedance versus total length in wavelengths, assuming a conductor diameter of 0.001 wavelengths

Image: antenna-theory.com

Wow thanks. A rocket science. I just though bigger is better. Will try 1/4 spider soon.

1 Like

Best is a Flightaware 978 Mhz antenna:
Professional, high gain, impedance matched, sturdy and suitable for indoor/outdoor installation, and above all, at a very competetive price.

However if you want to have fun with DIY, here are two $1 DIY antennas:


Quick Spider


Did you ever try a Cantenna?
It is a wide-body “koaxialantenne” (coaxial dipole). Due to large dia of coke can instead of a narrow copper pipe, it performs much better than original koaxialantenne design. The added advantage of using coke can instead of a narrow copper pipe is that a groundplane disk (bottom of the can) is added.


(Translation in English in red color added by me)

That’s what i used before the spider.
Comparison was conducted in the attic but spider seemed better.

Anyway i’ll soon swap the spider on the roof for the ebay PCB antenna because i’m curious about the difference.
Also might move the rtl-sdr LNA onto the roof directly below the antenna and weatherproof it a bit.

1 Like

You mean this one?






I have not opened mine but that’s the one laying around :slight_smile:

(I don’t believe the 6 dBi but i can imagine it being slightly better than my spider which is probably not an optimal build but still)

1 Like

I have the 1090 version. I wonder how wide of a bandwidth does this antenna have? Will it work on 978 MHz?

I also ordered number of cheap snap-on ferrite beads and put them on all my cables. Definitely not a thing that can make anything worse.


Made this ugliness last night:

Set it up:

Works pretty well today! Recent aircraft.json from dump978:


Look ath the aircraft 2 - so many details!