Forum: GNU Radio Can a host control an USRP1 and an USRP2 with the same clock?

Announcement (2017-05-07): www.ruby-forum.com is now read-only since I unfortunately do not have the time to support and maintain the forum any more. Please see rubyonrails.org/community and ruby-lang.org/en/community for other Rails- und Ruby-related community platforms.
18631ab8ac3a337d25f4192400719a5e?d=identicon&s=25 Jhon Lee (jhon)
on 2009-03-12 23:20
(Received via mailing list)
Hi all,

I would like to build a base station with multi-standards (WiFi, GSM,
WCDMA, etc.). For the first step, I want to build this base station
based on OpenBTS and BBN 802.11 projects. The OpenBTS works well with
USRP1 and people are making BBN 802.11 work with USRP2, so I want to
connect one USRP1 and one USRP2 to my host.
 I already read the wiki and the datasheets of the USRP1 and USRP2;
however, I still have some questions and hope to get some hints from you
guys, please!!

Is it possible to connect one USRP1 and one USRP2 to a single host PC at
the same time? Any gnuradio examples that can help me to figure out how
to do this?

If the USRP1 and the USRP2 can work on the same host at the same time,
can the USRP1 and  the USRP2 work synchronously (can they have the same
clock or common time reference?)?  Some discussions mentioned the
synchronization of two USRP2, but I cannot find discussions about 
synchronization of a USRP1 and a USRP2.


Thank you so much!

Jhon
72573690d4ab1e21a8c8a53b4654966b?d=identicon&s=25 Juha Vierinen (Guest)
on 2009-03-13 00:11
(Received via mailing list)
> If the USRP1 and the USRP2 can work on the same host at the same time, can
> the USRP1 and  the USRP2 work synchronously (can they have the same clock or
> common time reference?)?  Some discussions mentioned the synchronization of
> two USRP2, but I cannot find discussions about  synchronization of a USRP1
> and a USRP2.

You can timestamp USRP1 data by inserting a syncronization pulse
(e.g., GPS PPS) to a digital i/o pin. You need the gr-gpio firmware
for this. This will have approximately one sample accuracy with a very
small known constant offset. We have been using this for nearly a year
with good success. You can easily get less than 1 us absolute timing
accuracy this way. I'm guessing that similar functionality will soon
appear for USRP2 as well, so you can keep these two in sync to some
extent.

Regarding relative sample clock coherence, Martin DvH had a nice idea
about syncronizing the relative clock frequencies by inserting a PPS
or 10 MHz signal into the digital i/o pin. This allows you to estimate
relative clock frequency to a reference 10 MHz or a 1 PPS signal.

However, if you buy some more hardware, you can drive both USRP2 and
USRP1 from the same clock. The USRP2 takes a 10 MHz standard, but the
USRP2 needs a 64 MHz clock. I have used either a 6x multiplier  or a
cheap DDS board to generate the USRP1 clock (60 MHz and 64 MHz). I
don't really know a source for 6x clock multipliers (if somebody
knows, please tell me). The DDS I'm using to generate a 64 MHz clock
for USRP1 is this:

http://www.eskimo.com/~novatech/PDF_files/lpo400ads.pdf

It even fits nicely into a USRP1 box, taking the space of one
daughterboard slot.

Hope this helps.

BR,
juha
Fc3da6ed7be314cc1a8e662b6241cb28?d=identicon&s=25 Kyle Pearson (Guest)
on 2009-03-13 15:13
(Received via mailing list)
> However, if you buy some more hardware, you can drive both USRP2 and
> USRP1 from the same clock. The USRP2 takes a 10 MHz standard, but the
> USRP2 needs a 64 MHz clock.

You can actually use a common reference between the USRP1 and USRP2 by
modifying some of the USRP2 firmware code so that it can take a
reference other than 10MHz. See
http://gnuradio.org/trac/wiki/USRP2GenFAQ#Doesthee...
This code is located in usrp2/firmware/lib/clocks.c

-Kyle
This topic is locked and can not be replied to.