Geoscan X-band

[cyyyr]

Geoscan X-band transmitter will be placed on radio amateur cubesats and will follow amateur missions, like:

• Translate a Cubesat's ham-radio callsign
• Transmit CW, photos and videos with GFSK modulation to radio-amateurs around the World
• Will be used for amateurs to receive data in X-band on request through Geoscan Website and libre.space forum

Following information will be needed to demodulate a transmission:

• Modulation - GFSK
• Syncword - 0xF68D
• Preambule - fourty 1’s (1111111111111111111111111111111111111111)
• Fixed size 64 byte data-packet inside radio-packet
• Link with example IQ files (sample rate 2048000).

Info about our transmitter:

• frequency - 10450-10500 MHz
• baudrate - 125kbs
• bandwifth is around 250 kHz

[daniestevez]

Thanks for sharing this information. A few questions:

• Is this X-band transmitter available and usable in GEOSCAN-EDELVEIS, or will it only be available in future GEOSCAN satellites?
• Are the specifications of the transmitter already frozen or still open to change?
• Do the packets use any FEC, scrambling and/or CRC?

One remark is that the preamble doesn't serve to train the symbol synchronization loop, because it has no symbol transitions. Therefore, the closed-loop FSK demodulator in gr-satellites probably won't work well with these frames.

[daniestevez]

Can you give the full details for these? A document of several pages with all the definitions would be helpful.

[cyyyr]

Sure, I'll send it here as it will be ready! Thank you!

[cyyyr]

Hello, @daniestevez! Excuse me for a long reply, i've been away a few days! Here is our radio protocol document.
geoscan-x-trasmitter.odt

[daniestevez]

Hi @cyyyr, I'm afraid that documentation is very brief and ambiguous. I spotted at least the following problems:

• I am looking at IEEE 802.15.4-2020 Section 12: O-QPSK PHY, and I don't see any mention of convolutional coding. Section 21.3.6 SUN O-QPSK PHY; FEC describes a convolutional encoder, but it doesn't match the parameters you give (the constraint length is 7, not 5, for instance).
• Interleaving is not defined.
• Giving polynomials in hex or as polynomials can result ambiguous regarding bit endianness and bit order of LFSR.
• The scrambling parameters are somewhat ambiguous (it is not clear what output bit 0 means, for instance).
• The CRC parameters are insufficient to specify the code. There should be more endinanness parameters, as well as final XOR.

I don't want to spend more time iterating and pointing out what is missing in the documentation, so I would suggest that if you want to get support from the community to implement a decoder for this, you share all the documentation you have. In any case, I'm also happy to accept a pull request for the decoder if you implement it yourselves.

[cyyyr]

Sorry, if I did something wrong. We use high level functions to set transmitter up, so I am trying to figure out all nuances myself. It is not about that i am trying to hide something. I will provide any information that i could find and will help process as much as needed.
I will do my best to make documentation less ambiguous for community. Thank you very much! And again I didn't mean to take much of your time.