[idea] Crossband and Dopplershift compensation

[rfamp]

Hi, thank you for this great software. It is great to have such an allrounder on linux. As you have invited us to share ideas, I have two ideas which maybe interesting for amateur radio users:

1. Adding the capability to compensate for doppler shifts on satellites using e.g. gr-gpredict-doppler and gpredict
2. Adding the possibilty to tune RX and TX independently (cloning the frequency selector), to add crossband capability.

Both in combination might be interesting for amateurradio satellites communications. Do you have something like this planned?
Regards

[kantooon]

Hello and welcome! Thanks for the kind words and regarding your two ideas:

1. I've thought about integration with GPredict myself before. The hard part is doing this without making the install process much harder for users. OOT modules are rarely packaged by Linux distributions, so there needs to be some process by which a) the OOT module is kept up to date with evolving GNU radio versions and b) installation is as seamless as possible on most modern distros. That said, I'm sure there can be an acceptable solution, even if that means I have to write and maintain some code separately.
2. This idea is intriguing, but I'm not sure I understand fully. Cossband is possible right now out of the box by setting the TX offset with the help of a text box in the "Setup -> Radio settings" page. The text box is called "TX frequency" and if you input a frequency in kHz and press Enter, the code will automatically populate the correct split offset also in kHz (I've used this for QO100 operation). If you had something else in mind, perhaps you could expand a bit on the use cases and how it would work, also what benefit it might have over a static defined split offset. If there's a clear path to improve the tunig interface, I'll be happy to do more work on it.

Have a great day!

[rfamp]

I'm honest, this is currently beyond my scope, but when I have time, I'll try to dive into the code and see if I can contribute in any way.
Thanks for your answer and interest.

[kantooon]

I've started to implement the GPredict integration. gr-gpredict-doppler is not usable because it's written in Python and only has RX functionality, but GPredict can control both uplink and downlink frequencies. The uplink frequency is a bit of a pain because of the way tuning works, but anyway, this is how settings will look in GPredict:

I thinkl this will take a while to have operational, but preliminary RX tests are good.

[kantooon]

Preliminary support is in the next branch now. There are two issues outstanding currently:

1. For sats with linear transponders, you can't sweep the transponder manually while GPredict correction is engaged since it will override any manual tuning requests.
2. Lime hardware has issues with sample artefacts which manifest as audio pops while automatic tuning is done. I've posted this on their forums and hope to get a solution eventually. The PlutoSDR on the other hand doesn't have this sort of issues.

This code should be considered highly experimemental at this point, I've done what is likely an incomplete protocol implementation and there may be issues including crashes. That said, if you want to give it a shot, use the next branch, and configure GPredict as shown in the image above, enable both Remote Control and GPredict control in QRadioLink and set the radio interface in Gpredict to issue tune requests less frequently (around 100-200 msec between requests). The port number has to be the same in both QRadioLink and GPredict but there is no special limitation for which port it is.

[rfamp]

Hi,
wow, I just looked through the commits you made, this looks very interesting, I will try the branch in the next days and report back to you. I'm sorry that I can not be more helpful at the moment due to time issues.
Regarding the first problem: What about using something like a the RIT to tune around the transponder and keep the rx frequency locked on gpredicts output. Something like a frequency dial which, if activated, adds a offset to the frequency from gpredict?
Have a nice day

[kantooon]

I'm a bit busy right now, but the tuning will be done in the future in software instead of changing the device frequency directly. That will make it possible to sweep a linear transponder instead of being fixed on the freq. But this code is complex and it will take a while to implement. Right now there's only a PoC implementation to check GPredict can tune the radio both on uplink and downlink.

[kantooon]

Apologies for ignoring this feature for so long, but I got sidetracked with other stuff, and in addition I had no hardware capability for crossband operation to be able to really test this. But now I can work crossband thanks to the LimeRFE, and I plan to start working again on the Doppler correction code this holiday season. Tuning the radio is the easy part, but you've got no idea how hard it has become to both update the user interface and keep a consistent internal state while doing this. Hopefully I'll be able to crack this nut soon.