Regarding Observation 1761812 … This satellite is listed in the database as FSK9600, but the bandwidth seems too narrow for that. The packet transmission rate is low. It seems that other recent “good” observations have not produced any decoded data. DK3WN lists this as FSK 1200, but also lacks any recent packet decodes.
I didn’t find any official source for 1200 FSK, however except DK3WN list, there are reports from JA0CAW and this article https://destevez.net/2018/12/decoding-astrocast-0-1/ by @EA4GPZ.
So I’ve added a new transmitter entry in SatNOGS DB that you can use to schedule Astrocast 0.2.
Thanks for the info and the notice!
Most likely this is one of the 1k2 FSK modems that Astrocast 0.1 uses. I have tried to decode the recording with the gr-satellites next branch Astrocast 0.1 decoder. The quality of the recording is too bad for a successful decode (there is clipping and lots of noise). However it seems that the signal is what I call “Astrocast FX.25 NRZ-I”. At least I am able to detect the 64-bit syncword with only 4 bit errors, though Reed-Solomon decoding fails.
My conclusion is that Astrocast 0.2 can most likely use the same modems as Astrocast 0.1, so I’m updating gr-satellites to reflect this.
It would be also interesting to know when the mode of Astrocast 0.2 changes, as we have observed in the past several times signal that looks like FSK9k6, for example in this observation https://network.satnogs.org/observations/1392321/.
I would like to play with some algorithms to detect the type of modulation. I’ve been trying to make a copy of raw IQ data for offline processing after a pass. I can see where the output files are stored (ie 48 kHz audio and waterfall). Of course the Doppler corrected IQ data would be even better! Is it easy to access the raw or Doppler corrected IQ data?
In satnogs-setup you can enable saving of IQ data, and define a place where it is saved. Note that the file is overwritten after each observation, so you will need to copy it out just after an observation is finished.
Be careful leaving this enabled and writing to a microSD card, as it’s a very good way of using up all the write cycles of the card. I would suggest mounting a USB drive and writing to that instead.
My Az/El station (#232) is saving IQ data by default and uploading it to AWS as per this guide: Uploading IQ files to S3
The last 2 weeks of my stations IQ data is available at https://satnogs-iq.s3-ap-southeast-2.amazonaws.com/ (just append the observation number and .raw to the URL, e.g. http://satnogs-iq.s3-ap-southeast-2.amazonaws.com/1331766.raw
If you have some particular sats you would like observed using my station I’d be happy to queue it up…
Thanks Mark - it’s very useful to access the IQ data! I’ve started some code to stream my IQ samples via UDP to another machine that can display them in (close to) real-time (or do replays). Of course the satnogs-monitor is great but it’s nice to explore the signal a bit more. So far my code only plots in the time domain - even that is useful to show the nature of some occasional interference.
Some more questions: do you know if the gr frontend can be started and just used to store the IQ data, without setting up a pass or doing Doppler correction? (eg so I can manually point at the local 70 cm beacon, then measure SNR etc?) Also (this might be a lot harder), where these’s uncertainty about the modulation (eg Astrocast above), IF it was possible to determine this early in the pass, could the satnogs mode be changed to suit the detected signal type?