On 03/10/2010 01:11 AM, Per Zetterberg wrote:
I can just tell you about my experience (for what is worth :-). I synchronized the TX and RX using a 10MHz reference from a signal generator. I filtered the 10MHz using a band-pass filter from mini-circuits and then splitted with a mini-circuits splitter. Then I connected a step function to the pps inputs which I manually triggered. I modified the apps for transmitting and receiving samples which are somewhere under usrp2/apps or something like that. I had a ten second sleep in the program to be able to start both USRP2s and manually trigger the step function. It seemed to work great with very repeatable results (basic db-boards cable between TX and RX). I can send my ugly code (which uses the VRT branch). The next step for me is to use a real GPS but other duties has taken over...
This is correct, you do not need to use a GPSDO, you can lock to any 10 MHz reference. And any digital signal source will suffice, it does not need to be a pulse per second source.
One thing more. When I cold start the USRP2s it seems to take some seconds before they are firmly locked to the 10MHz reference.
At power up the clock generator is not attempting to lock to the reference. It does not attempt that until the processor tells it to do so, which only happens after the processor is running, roughly 3 seconds after powerup. If you connect the reference to the USRP2 after it is powered up, it should lock in less than a second.
Matt _______________________________________________ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
