[Discuss-gnuradio] FM demodulation block performance ?

[jmfriedt]

I am still trying to receive NOAA APT signals using the RTL-SDR DVB dongle. 
What I did
so far:
1/ using a wideband discone antenna, I receive the NOAA APT characteristic 
signal during passes
on a AOR3000 scanner. I know the NFM mode of the AOR3000 is too narrow for full 
decoding of the
image (saturated whites), but at least I know my antenna is good enough to 
recover an audible and
usable signal,
2/ on the antenna I put a splitter (Minicircuits ZFDC-10-5) with 10 dB loss on 
the coupled output
(to the AOR3000) and 1.5 dB loss on the direct path (to the DVB receiver).
3/ I use http://picasaweb.google.com/lh/photo/2QDS5vwPra5Q7yQbSdUOng as an 
example of NBFM receiver
usage (extending the demodulation bandwidth to 30 kHz instead of the 5 kHz in 
the example), or I 
try a WBFM receiver module with a 30 to 40 kHz bandwidth, both of which were 
preceded by a 100-kHz
low-pass filter (either as in the screenshot using the FIR coefficients in a 
Xlating block,
or using the LPF block)

In all cases I get a clear signal on the AOR scanner (even after the 10 dB loss 
on the coupler) while
I cannot recover any usable signal on the DVB receiver, with RF gains in the 33 
to 39 dB range and
IF gains in the 25 to 40 dB range. A friend of mine got usable signals using 
the NFM demodulator from
SDRSharp (which I have not yet managed to get running) with the DVB dongle.

Any idea whether the FM demodulation blocks require some minimum signal to 
noise ratio to lock ?
more generally is there any other description/documentation on the blocks other 
than the doxygen
generated information and source codes ?

Thank you, Jean-Michel

-- 
JM Friedt, FEMTO-ST Time & Frequency/SENSeOR, 32 av. observatoire, 25044 
Besancon, France