Fractional PPM correction & issue #200, pull #201
Robin Gape
The following report applies to an installation of GQRX from
the PPA repositories, running 64 bit Ubuntu 16.04 on a Gigabyte
Brik. The symptoms appear under the Unity desktop, and under the
Gnome3 Flashback desktop.
Using a Funcube Dongle Pro, the frequency correction may
be entered as a ⅒ ppm (x.n ppm), but the value applied is
the integral value to the left of the decimal point (x
ppm). The ⅒ ppm (.n ppm) makes no difference, from x.0 to
x.9. The frequency correction is only applied in integral values
of ppm.
The same syndrome applies with Funcube Dongle Pro Plus.
This issue should have been resolved by issue 200 (https://github.com/csete/gqrx/issues/200) and its resolution in pull 201 (https://github.com/csete/gqrx/pull/201) which should permit correction entries as ⅒ ppm (x.0, x.1, x.2…x.9). However, pull 35 (code at https://github.com/csete/gqrx/pull/35/commits/6180f8b7befcd40364d9fd77f23fbe7753d2d779) contains a line (and several similar):
status set_freq_corr(int ppm);
which suggests where the issue might be coming from.
Why is this an issue? At 1296MHz, the correction is applied in steps of ~1.25kHz, which is quite coarse for beacon monitoring &c.
Alex: What's the best way of taking this forward? Should I:
a) raise a bug report, or
b) raise a feature request, or
c) attempt to hack a patch (noting that I've not compiled GQRX for quite some time!)
Cheers,
Robin, G8DQX
Alexandru Csete
Good catch! I think it would be a good idea to file a bug report to
have it on the list.
As for fixing it, the code that you point out in pull 35 is no longer
in gqrx but in external libraries gr-fcd and gr-fcdproplus:
https://github.com/gnuradio/gnuradio/blob/master/gr-fcd/include/gnuradio/fcd/source_c.h#L126
https://github.com/dl1ksv/gr-fcdproplus/blob/master/include/fcdproplus/fcdproplus.h#L90
Changing the API here might not be a good idea since it would break
binary compatibility. One could work around this in the gr-osmosdr
wrapper by applying the frequency correction here as floating point
number instead of calling the gr-fcd / gr-fcdpoplus functions.
This is just an idea in case somebody wants to try. In the long run we
need a new "libfcd" driver library anyway.
Alex
> You received this message because you are subscribed to the Google Groups
> "Gqrx SDR" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to gqrx+uns...@googlegroups.com.
> To post to this group, send email to gq...@googlegroups.com.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/gqrx/084b2f4c-a76c-e0db-d562-fc384e85533b%40googlemail.com.
> For more options, visit https://groups.google.com/d/optout.