I am also on the SOC-i team. I agree, the frames posted look very similar to SOC-I’s signal. As Charlie said, the git repo is a great source of information, and has an example of a beacon .wav file. It should be able to at minimum convert any packet from SOC-i except for image packets into raw hex data, though if it gets a health beacon it can parse the data like battery charge etc. The only challenge is the fact that it does not like to work without a good signal to noise ratio. I am not quite sure how to process these observation recordings to achieve this. Does anyone have advice on extracting these packets into a low(er) noise IQ file? I’m also listening with my own antenna, I will pay extra attention to object F going forward.
Also about the frames, I believe we should see a packet about every 8 seconds. I’ve attached an image containing information about SOC-i’s transmitter. If you need any more information, let me know and I will get it for you.
… appears to be a 2400 baud sample using the ‘Inspectrum’ tool. Do I understand from your post that we should be expecting 9600 baud packets? Thanks!
Also, I’ve not been successful decoding this sample nor any of the SatNogs recordings using any of the normal deframers that are typically used w/ FSK - GMSK… so the exact format of these beacons remains a mystery to me.
Interesting. EIRSAT-1 is at 437.100 MHz, so this observation would still be a bit too high to be doppler shift, right? You have far more experience observing so I would love to get educated from you on how to interpret these observations, what makes you think these are a single sat rather than two sats passing at the same time?
For what it’s worth I’m not overly optimistic about this observation being SOC-i simply because of the beacon period - but I don’t want to rule anything out.
Remember that a LEO satellite in the 70cm band will have a total doppler shift in the region of 20kHz, so if you are listening to two satellites, one starting ascending and another ending at descending it will be the sum of those. The doppler correction for SOC-1 is at say -10kHz and EIRSAT-1 is at +10kHz, that leaves the at least 5kHz offset to the left in the flowgraph and matches the 25kHz difference.
I did a quick analysis of that just now, if I didn’t make any mistakes…
Here’s the adjusted wav file, I made it into a simple IQ-wav 16bit signed. Tuned 300.420kHz and downsampled to 48k. soci_narrow.zip (37.2 KB)
Carsons bandwidth for that is 14.4kHz and matches the spectrum in SDR++ (highlighted area, filter is 16kHz)
@K4KDR Did you measure the preamble ? that is likely not a 01010101 pattern but 00110011, explaing half the rate I measured.
00110011 shows up as the wider amplitude in the preamble waveform, and a 01010101 is the narrower one.
edit:
The only waveform that has looked like this in my previous measurements is Lilacsat-2 at 4k8 baud and 4000 deviation:
Looking at this strf obs, you see two signals following 58472 (EIRSAT-1) the main signals is at the expected frequency used by EIRSAT-1 and second signal around 437.125 (SOC-i frequency) seems like a artifact from the strong signal send by EIRSAT-1.
So at the moment my best guess is that SOC-i isn’t active and we found EIRSAT-1 artifacts around 437.125.
For the observations that i checked, EIRSAT-1 was indeed also passing at the time of the signals on 437.125 MHz
The search continues, but as even PI9CAM with there 25 meter dish aren’t receiving any signals, changes are slim.
I don’t speak matlab or octave, but looking at the code there seems to be no framing at all, nor checksum
It’s 4k8 with a 80bit preamble of 0011, then 18 byte blocks of telemetry.
I measured the lengths in Inspectrum and saw roughly 245bits preamble and 648 bits data.
That should be 4 blocks of 18 byte data, plus 9 more of something.
While monitoring an overhead CATSAT pass this morning over FN41iq, nothing was seen on 437.185 MHz. However, I did see signals with a visual pattern I had seen before for CATSAT on the waterfall at 437.2505 MHz.
Replay of the IQ decoded several 2K4 ax.100 Mode 5 FSK frames! The hex strings begin with 83 e7 80 01 01 05 01 02 9a or 83 e7 80 01 01 01 02 02 9a, which is what I’ve seen for CATSAT before. Perhaps others might consider checking this out further?
link to kss file, FM demod audio file snippet and IQ files made using SDRConsole here 10.38 MB folder on MEGA
Hi All, I thought I would introduce myself. I work with dspangler (Del) on CatSat! Thanks for finding us off frequency! We should be back on 437.185 now. I will try to be more active in this space.
