Another issue you'll have is to start both dongles simultaneously...

with the sampling rate you'll have on the dongles this becomes quite tricky : you get blocks of samples from both dongles but you have no idea of how many samples they are shifted in time. One way to solve this is to search for cross correlation between channels, but this is not very easy in realtime. This can be done by post-processing offline.

I thaught of this a while ago, using a pin diode or something equivalent to shortcut both inputs to ground to find an easy way to have the synchro between channels but did not try it yet

sylvain

2013/8/25 Jiao Xianjun <putaoshu@gmail.com>

basically you are right.
But there is signal processing algorithm to track two dongles sampling/oscillator error separately and compensate them separately. Thus we can get them synchronized finally after signal processing algorithm.

What I need actually is that how to get two streams of sample from two dongles.

Actually I notice that in dump1090 it put rtlsdr_read_async in a new created background thread, then there is a possibility that I create two thread to operate two dongles. But in this method, there should be some mechanisms to synchronize two threads. It is complicated.

Anyone has simpler solution?

Thank you.

On Sun, Aug 25, 2013 at 9:49 PM, Thierry Leconte <thierry.leconte@laposte.net> wrote:

Le 25/08/2013 15:13, Jiao Xianjun a écrit :

Hi,

I want to use multiple rtl-sdr dongles to do some multi-antenna experiments.

Is it possible to read IQ samples from multiple rtl-sdr dongles in a synchronized manner?

Each dongle have its own clock, so it will never be synchronized at the sample level.