[Discuss-gnuradio] Re: Phase-lock across USRP-2 boxes and radar receiver development

[Josh Blum]

In the case of the gnuradio-USRP2 driver, the CORDIC is always spinning, 
and this guarantees that the phase between CORDICs will always be 
different among multiple USRP2s.

In the case of the UHD USRP2 driver, the CORDIC only spins when samples 
are being processed through it. In this manner, as long as all USRP2s 
are told to stream RX samples at the same time, the phase will be the 
same across all USRP2s. Also, the USRP MIMO interface automates the 
synchronization and synchronized streaming among multiple USRP2s.

Thanks,
-Josh

On 10/08/2010 05:57 AM, Mamoru Yamamoto wrote:
> Mr. Josh Blum, Mr. Matt Ettus, Friends,
>
> I think that you are following discussion on the mailing list that is
> "phase-lock problem across USRP-2 boxes". Mr. Juha Vierinen kindly
> helped us with the information that
>
> -- USRP-2 FPGA with GNU Radio original does not support "phase-lock
> across USRP-2 boxes". -- USRP-2 FPGA with UHD supports that. -- The
> "phase-lock across USRP-2 boxes" is the "NCO phase reset at PPS
> input".
>
> Are these true?  We are developing a radar receiver, and already have
> a working program with GNU Radio original FPGA.  For us, easiest way
> is to use a revised FPGA with "NCO reset at PPS" over GNU Radio. Is
> there such version of the FPGA image?  The other way may be for us to
> switch to the UHD.
>
> What do you suggest?
>
> As our application is a radar receiver.  Then in our program, we
> apply the TX pulse to PPS input, and reset the time counter at the TX
> pulses.  Is this possible with UHD?  This is one critical point for
> us to move to UHD.
>
> I hope your help.  Thanks.
>
> Regards,