On Wed, 9 Jul 2014 16:56:05 +0000 Malcolm Robb <[email protected]> wrote:
> > Hi, > First post, hope I don't break any rules. I've a question about the > rtlsdr_set_freq_correction() function. > > I've used SDRSharp to fine tune my dongle to a local Airfield ATIS > broadcast, and like many others in our group come up with an error of > around 50ppm. So I can use rtlsdr_set_freq_correction() in my > application to fine tune the RF center frequency. However, there is > only one crystal in the dongle (28.8MHz) and that crystal generates > both the RF frequency and the digital sampling frequency. So > presumably if the RF frequency is 50ppm off, then so is the digital > sampling frequency. > > My question is : Does the rtlsdr_set_freq_correction() function > correct both the RF frequency (set by rtlsdr_set_center_freq()) and > the sampling frequency (set by rtlsdr_set_sample_rate()) , or just > the RF? > > The reason this matters is that I use the sampling frequency to > timestamp signals recovered from the raw samples, and 50ppm error is > enough to mess up the results. > > Cheers > Malcolm > > > > http://www.evrytania.com/lte-tools/78-default-category/77-lte-cell-scanner This program uses the local LTE cell as a reference, good to 5ppm, well if you have LTE. It runs, though I never used it to tweak the rtlsdr since I do mode-s (with the Malcom Robb version of dump1090 no less) and the error in mode-s center frequency is far worse than the dongle. The program works, albeit slowly, but you are only going to find the 700MHz LTE cells due to the limited range of the dongle. Note I ran this on my notebook PC, not a Beagle or PI. This doesn't mean it won't work on a small arm computer, but rather I never tried it. You would probably have to temperature test this to see if correction makes any sense. That is, run the LTE program with the dongle at different temperatues and see how the correction factor changes.
