[USRP-users] usrp b210 and mimo

Juha Vierinen x at mit.edu
Fri Dec 5 14:15:38 EST 2014

I tried again with the latest images and uhd driver. I also am using
external power. I also checked that the signal level on the 10 MHz is 13
dBm and the PPS is a clean TTL signal. The relative time delay between the
N200 and B210 is still variable.

The relative time delay between multiple N200 devices isn't variable
between restarts, which makes me believe there is something wrong with the
the implementation of usrp.set_start_time() on the b210. This could be
easily tested with a b210 + n200 mimo setup and a split pulse generator
signal feeding into both systems.

I was hoping to use the b210 for a radar application as the dynamic range
of the tuner is better than what we currently have. Unfortunately the
absolute timing capability doesn't work now. Is this a capability that is
even promised by the specs? Am I even supposed to be able to align a N200
and a B210 to within a microsecond of each other?


On Fri, Dec 5, 2014 at 5:25 PM, Juha Vierinen <x at mit.edu> wrote:

> > Is the B200 timing offset always before (or after) the N200, or does it
> randomly bracket the N200?
> The b200 signal starts consistently about 20 ms +/- 100 us earlier than
> the n200 signal. The standard deviation of the offset is about 100 us.
> I'm using the noise injection of our radar signal, and various power line
> interference spikes seen by a n200 and a b210 simultaneously as an external
> timing reference.
> > There are a couple of things I'll mention up front that come to mind
> because of what you are doing…but they don't explain your issue:
> > 1)N2x0 will always sample the PPS with a 100MHz clock, B200 will sample
> PPS with whatever rate is configured as the master clock rate. If that
> value is 50MHz then there's a 10nS ambiguity introduced …if 100/master_rate
> is not an integer then the sampling clock edges are going to constantly
> shift phase w.r.t each other.
> I'm using 1 MHz. The B210 is using 32 MHz I think. Both are integer
> decimation rates. Once the B210 streaming is started, there is no relative
> drift as long as the samples keep coming in.
> > 2) Group delay through the DSP's is quite different…it will in general
> be larger for B200 I suspect…but it will depend heavily on what the
> configured decimation is.
> I kind of figured there would be a constant group delay offset, and I
> think it is about 20 ms for 1 MHz.
> juha
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ettus.com/pipermail/usrp-users_lists.ettus.com/attachments/20141205/3d465ebb/attachment-0002.html>

More information about the USRP-users mailing list