On 12/15/2018 08:02 AM, Arun kumar Verma via USRP-users wrote:
Dear Users
I am facing problem of phase offset when i change center frequency
dynamically for TwinRX with X310 setup. Here is my C++ code,
Please advise
usrp_source1->set_rx_lo_source("internal",uhd::usrp::multi_usrp::ALL_LOS,0);
usrp_source1->set_rx_lo_source("companion",uhd::usrp::multi_usrp::ALL_LOS,1);
usrp_sourceDOA1->set_command_time(usrp_sourceDOA1->get_time_now() +
uhd::time_spec_t(0.01));
usrp_sourceDOA1->set_center_freq(m_CenterFrequency,0);
usrp_sourceDOA1->set_center_freq(m_CenterFrequency,1);
usrp_sourceDOA1->clear_command_time();
I feel set_command_time is not working properly. When i start the X310
my phase difference is almost zero but as i change my frequency pjhase
difference is random in nature.
Regards,
Arun Verma
Since you're using the "companion" configuration (where the 2nd-channel
LO is a "copy" of the first), you'll inevitably be dealing with
physical path-length issues between the output of the first-channel
synthesizer and the two-or-more mixers involved.
Since there's almost no way to make the path-lengths electrically the
same, given the array of switches, and circuit-board traces involved,
there'll be some non-zero difference in the electrical path lengths.
This will manifest as different phase offsets at different frequencies.
It will also change slightly with temperature, since circuit traces
that are non-zero-length change electrical length with temperature changes.
The good news is that it should be fairly reproducible, and can be
characterized for your configuration, and you can build a table and use
that for phase compensation.
This is very much like the "inside the radio" version of what inevitably
happens out in your antenna "plumbing":
(A) Getting *exact* phase matches even with apparently "identical"
pieces of coax and other bits of plumbing is near-impossible. This
gets worse with shorter wavelengths.
(B) Those phase lengths will inevitably change with temperature.
Further, it's even worse, because none of these transmission paths have
a dielectric constant of 1.0, which means that smaller physical
length differences are "ampified" by the slower velocity in these
lines--that is, wavelengths are physically shorter, so that a physical
difference affects more of a wavelength.
_______________________________________________
USRP-users mailing list
[email protected]
http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
