Change in Phase on a USRP2 configured as a VNA

Hello Everyone,

I have put together a simple vna setup using a USRP2. For right now Im
using
the loop back cable that ettus sells (SMA-SMA) and a 3dB pad to connect
the
RX to the TX port. While running the attached setup in GRC I see a
couple of
things that concearn me and I was hoping to get some help on them.

The first thing is the startup phase. If Im not mistaken without
calibration
there is no way to determine what this should be even though the cable
thru
does not change substantially each time I restart the grc file. From
what I
read I believe this just has to do with the phase of the internal clock
in
the usrp when I execute a file in grc. I think I can live with this
since I
ll have a reference in my measurement that I can set the phase off of.

The second thing that I see though that is more troublesome is that when
I
start the file and just allow it to run the phase changes over time. It
will
be stable for like .5 minute or so and then jump in value and then stay
stable again for another interval. This pattern repeats. Im runnig at
200k
sample frequency and 500 decimation so I don’t think Im pushing the
bounds
of gig-e by any means. Plus not that Im sure that this has anything to
do
with it but the decimation is a multiple of 4 so Im getting the full
benefit
of the hardware. Am I wrong in thinking this should be stable over time
while the hardware is running? Is this indicative of a drop in
communication
to the USRP2 which causes it to shut down and then its merely the first
problem again, that is a random intial startup phase. I don’t see any U
of
O s in the log so I don’t think Im having that problem.

That third thing that I m curious about is the affects of changing
transmit
frequency (LO Frequency) , at run time, of the USRP2, and its affect on
the
measurement. Im assuming that there is some sort of settling time before
I
should assume the measurment is valid what is this? I m guessing this
will
also change the phase value. But how. I saw that matt asked in a
previous
thread why the user didn’t just transmit the bandwidth they needed to
the
USRP2 instead of using the LO. In this case since Im using the WBX board
and
I can’t use an LO of 0 MHz (I assume) it makes sense to use the LO of
the
WBX board as the RF source. It would allow me to use the full bandwidth
of
the WBX board 68MHZ to 2.2GHz without the gig-e limitations.

The final thing is which is better as a source. A constant or a
sinusoid? I
settled on a sin function by trial and error. But is there a better
rationale for selection?

Thanks,
Justin

This forum is not affiliated to the Ruby language, Ruby on Rails framework, nor any Ruby applications discussed here.

| Privacy Policy | Terms of Service | Remote Ruby Jobs