I'm trying to demodulate a qpsk signal with gnu radio. The signal is from the meteor satellite on 137MHz https://www.sigidwiki.com/wiki/Low_Rate_Picture_Transmission_(LRPT). I recorded the signal with gqrx without any special settings.

This is my flow graph:

After filtering the signal looks like this which doesn't look like a qpsk signal:

My input signal looks weird but the fft plot seems to be fine:

I already got pictures from it with other software (glrpt) so too much noise shouldn't be the problem. Does anyone know what i'm doing wrong?

  • $\begingroup$ shouldn't the input signal have more, many more samples per cycle? $\endgroup$ Oct 5, 2022 at 20:10

1 Answer 1


Does anyone know what i'm doing wrong?

Sad, honest answer: no. This could be incorrect assumptions on symbol rate or frequency offset, an incorrect receive filter, a lack of equalization, too much noise, or the lack of timing recovery paired with bad luck. In any case, choosing a decimation where every symbol has 1.6 samples is ... not helping visualizing this, at all.

I honestly don't see anything that resembles the constant envelope that QPSK would have, though. So, if adjusting the decimation of your RRC such that there's an integer number of samples per symbol doesn't help, you should check your assumptions on your signal.

  • $\begingroup$ Already tried many things but nothing really changed anything. The recording is from the meteor satellite, from which i got pictures with other programs fine, so too much noise shouldn't be the problem. Could recording with a higher sample rate fix it? $\endgroup$
    – fjsdlkfj
    Jul 3, 2022 at 15:35
  • $\begingroup$ I recommended somerthing specific. Did you try that? $\endgroup$ Jul 3, 2022 at 15:42
  • $\begingroup$ Yes. Looks about the same $\endgroup$
    – fjsdlkfj
    Jul 3, 2022 at 15:54
  • $\begingroup$ hm, then you'll have to simply check your assumptions, they are necessarily wrong. $\endgroup$ Jul 3, 2022 at 16:16
  • $\begingroup$ Checked everything. I noticed my input signal looking weird (added picture in post) but the fft plot seems to display everything fine. Maybe there is some weird stuff going on with the recording? $\endgroup$
    – fjsdlkfj
    Jul 3, 2022 at 17:02

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.