Hi Olivier, as discussed, an FSK with 1 S/symbol is not really an FSK; so I can't really confirm that it does what you want (it does do some DPSK, in fact). 6 S/symbol mean that one symbol, ie. the duration for which the output frequency is held constant, is six samples, nothing more, nothing less.
When you decimate something by a factor of six, the resampler has to choose a filter that kills the upper 5/6 of the spectrum, inevitably damaging your signal, if your max deviation is more than 1/6 of the nyquist rate. So that's definitely not what you want to do! Even if the signal is all contained within 1/6 of the original bandwidth, you'd again end up with something that is not FSK. As usual, I recommend adding frequency/time sinks at practically every point in your flow graph and check whether the signal there looks correct, according to your signal model. It's still a bit hard to help you, not understanding what your signal *should* look like, and not being able to compare it to what it actually looks like – so I'd recommend testing every step as mentioned, and discussing the first point where things don't look like they should, using drawings (! this is important: it's most often much easier to discuss a technical problem with a minimal drawing) and screenshots of your displayed signals. Best regards, Marcus On 21.07.2016 21:15, Olivier Goyette wrote: > Hi all ! > > I've been working on this flowgraph for a while now and I've been able > to pinpoint what is not working, but I don't understand how to resolve > this so, that's why i'm asking for your guidance. > > Take a look at the flowgraph i've joined. The thing that doesn't seem > to be working on this setup is the samples/symbol in th CPFSK mod block. > > When I set this parameter to 1 i'm able to retrieve my message at the > file sink. So it tells me that what I did is okay in *_some way_*. > > What is not working is when I set the same parameter to 6. At the > beginning I wasn't using the rational resampler and it wasn't working. > So I thought I would need something to "divide" the number of samples > to retrieve each symbols. I tried it with the rational resampler block > and set decimation to 6, but then, Daaaaaah......not working again. > > There is something I can't work around here because I do not have > sufficient knowledge about this samples/symbol thing. > > I'd like to hear from you pals if you could give me a bit of > information about how to make this work properly. > > thank you ! > > > _______________________________________________ > Discuss-gnuradio mailing list > Discuss-gnuradio@gnu.org > https://lists.gnu.org/mailman/listinfo/discuss-gnuradio
_______________________________________________ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org https://lists.gnu.org/mailman/listinfo/discuss-gnuradio
