4
$\begingroup$

Using a RTL2832U USB dongle, I set up GNU Radio to produce a constellation plot of APRS traffic on 144.8 MHz. The result is what I expected: 144.8 MHz APRS Qt constellation on Gnu Radio Companion

So far, so good. Why not try the same thing using the wx GUI instead of the Qt version I had used ? Here are the results... The constellation plot does not look the same at all and it is nothing like the clean QAM constellation I had expected and obtained using the Qt GUI:

144.8 MHz APRS wx constellation on Gnu Radio Companion 144.8 MHz APRS wx constellation on Gnu Radio Companion

Can anyone explain why the wx GUI's constellation plot does not behave as I expect ? Is it simply a case of different default settings and naive user expectations ?

$\endgroup$
  • $\begingroup$ Why do you expect to get anything meaningful from a constellation plot? 2m APRS is usually modulated as AFSK on analog FM. But then, I haven't learned a lot about digital modulation yet, so I could be missing a way in which there's a useful result. $\endgroup$ – Kevin Reid AG6YO Jul 14 '14 at 0:14
  • $\begingroup$ It's definitely the case that the two sink blocks are different, incidentally: I took a look and the WX GUI one incorporates clock and phase recovery (hence all the options it has) whereas the Qt one appears to be (not sure) just dumping samples onto the screen with no analysis. $\endgroup$ – Kevin Reid AG6YO Jul 14 '14 at 0:16
  • $\begingroup$ @KevinReidAG6YO Thanks for taking a look at them - so they are different and I shouldn't expect the same result... phil-frost mentions that the WX sink's clock recovery is probably not doing anything useful... I don't know. I guess I'll have to learn about filtering and parameters - I probably expected magic... $\endgroup$ – Jean-Marc Liotier Jul 14 '14 at 7:11
6
$\begingroup$

There are a lot of things wrong here. Neither plot looks correct.

  • There is no way that Qt plot is realistic for anything but a signal generator. Where's the noise?
  • Where are the three missing constellation points in the Qt plot?
  • APRS isn't QAM (it's AFSK over FM), so I'm not sure why you are expecting QAM.
  • You don't have any filters, clock recovery, or really demodulation of any kind between the source and the sink. Normally you do all that stuff before the constellation such that each sample (and thus, each dot) corresponds to one symbol.
  • From the parameters, it looks like WX sink has some clock recovery, but I doubt it's doing anything useful. I doubt it's syncing up with your intended signal. Usually you still need to do some filtering and decimation to at least get it in the right ballpark. The clock recovery is only intended to compensate for slow phase drift. It doesn't know nearly enough to demodulate the signal. For example, how does it know the symbol rate?
$\endgroup$
  • $\begingroup$ No idea about where the (plentiful) noise went in the Qt plot - it does look miraculously clean. There are of course missing points, which appear from time to time - this is a screenshot of an animated plot and the points have low remanence. This is my first time demodulating a signal - my level of confusion is stellar. From what I have read, I am guessing that I should be doing some filtering, but I don't know where to start... More reading to do ! $\endgroup$ – Jean-Marc Liotier Jul 14 '14 at 7:04
  • $\begingroup$ Reputation so low I can't even vote you up... Consider yourself upvoted - thanks for your answer ! $\endgroup$ – Jean-Marc Liotier Jul 14 '14 at 7:07

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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