This is a legacy implementation written in Python and NumPy.
It is provided for reference reasons. Please use the new GNURadio based code which can be found at https://github.com/muccc/gr-iridium
The GNURadio based branch offers better performance with respect to computational effort and results.
Note: The rad1o has to be in hackrf-mode
hackrf_transfer -f 1625800000 -a 1 -l 40 -g 20 -s 2000000 -r /dev/fd3 3>&1 1>&2 | python2 extractor-python/extractor.py -c 1625800000 -r 2000000 -f hackrf --jobs 2 | fgrep "A:OK" >> output.bits
If you built/installed the rad1o branch of the hackrf tools, add -S 26214400
to the command line like this:
hackrf_transfer -f 1625800000 -a 1 -l 40 -g 20 -s 2000000 -S 26214400 -r /dev/fd3 3>&1 1>&2 | python2 extractor-python/extractor.py -c 1625800000 -r 2000000 -f hackrf --jobs 2 | fgrep "A:OK" >> output.bits
To capture and demodulate Iridium packets use extractor.py
. You can either process
a file offline or stream data into the tool.
By default, the extractor will drop samples if the computing power available is
not enough to keep up. If you have an already recorded file, use the -o
,--offline
option to not drop any samples. In this case the extractor will pause reading the
file (or input stream) until it can process more samples again.
The internal queue is filled with samples where the detector has detected activity in the file. By default it is 12000 elements long (roughly 4 GB at 2 Maps). You can tweak the length of the queue with this option
The center frequency of the samples data in Hz.
The sample rate of the samples in sps
File Format | extractor.py format option |
---|---|
complex uint8 (RTLSDR) | rtl |
complex int8 (hackrf, rad1o) | hackrf |
complex int16 (USRP with specrec from gr-analysis) | sc16 |
complex float (GNURadio, uhd_rx_cfile ) |
float |
The number of processes to spawn which demodulate packets. The detector runs in the main process.
detector-fft.py
Searches through the file in 1 ms steps to scan for activity
and copies these parts into snippets called <rawfilename>-<timestamp>.det
cut-and-downmix.py
Mixes the signal down to 0 Hz and cuts the beginning to match
the signal exactly. Output is <detfile>-f<frequency>.cut
demod.py
Does manual DQPSK demodulation of the signal to stdout.
If enabled inside demod.py
it also outputs
<cutfile>.peaks
(for debugging)
<cutfile>.data
the raw bit stream.