On 04/25/2018 01:30 PM, Hojoon Yang via USRP-users wrote:
The questions were not about time_spec_t and get_real_secs()..
I wondered that "How can I execute the command usrp->get_time_now();
and usrp2->get_time_now(); simultaneously.
As I mentioned the previous question, I tried to use
set_command_time(), but it does not work.
I was responding to the second part of your question where you were
unable to get the fractional seconds part.
I'm a bit surprised that the get_time_now() cannot be wrapped in a
set_command_time(), but it is the case that
B2xx support only a subset of timed commands.
What you might want to do is start a timed streaming from both devices,
and inspect the timestamps as they arrive.
the code that I tried is
---------------------------------------------------------------
usrp->set_time_next_pps(uhd::time_spec_t(0.0));
usrp2->set_time_next_pps(uhd::time_spec_t(0.0));
boost::this_thread::sleep(boost::posix_time::seconds(2));
uhd::time_spec_t start_time = usrp->get_time_now() +
uhd::time_spec_t(0.3);
usrp->set_command_time(start_time);
uhd::time_spec_t usrp_t = usrp->get_time_now();
uhd::time_spec_t usrp2_t = usrp2->get_time_now();
usrp->clear_command_time();
std::cout << boost::format(
" Start : %f us\n"
" Response 1: %f us\n"
" Response 2: %f us"
) % (start_time.get_real_secs()*1e6) % (usrp_t.get_real_secs()*1e6) %
(usrp2_t.get_real_secs()*1e6) << std::endl;
std::cout << boost::format(
" Difference of response time 1: %f us"
) % ((usrp_t - start_time).get_real_secs()*1e6) << std::endl;
std::cout << boost::format(
" Difference of response time 2: %f us"
) % ((usrp2_t - start_time).get_real_secs()*1e6) << std::endl;
---------------------------------------------------------------
----------------------------OUTPUT-------------------------------------
Checking USRP devices for lock.
Waiting for reference lock....
* 30FDE72: LOCKED
Waiting for reference lock...
* 3125CB5: LOCKED
Querying Clock for time and setting USRP times...
Start : 1329026.312500 us
Response 1: 1329026.625000 us
Response 2: 1329261.625000 us
Difference of response time 1: 0.312500 us
Difference of response time 2: 235.312500 us
-----------------------------------------------------------------
What I expect was, in OUTPUT, Difference of response time 2 should be
0.312500 us, not 235.312500 us.
This result means that the command usrp->get_time_now(); and
usrp2->get_time_now(); were not executed simultaneously.
I also changed the code
-----------------------------------------------------------------
uhd::time_spec_t start_time = usrp->get_time_now() +
uhd::time_spec_t(0.3);
usrp->set_command_time(start_time);
uhd::time_spec_t usrp_t = usrp->get_time_now();
uhd::time_spec_t usrp2_t = usrp2->get_time_now();
usrp->clear_command_time();
-----------------------------------------------------------------
to
-----------------------------------------------------------------
uhd::time_spec_t start_time = usrp->get_time_now() +
uhd::time_spec_t(0.3);
usrp->set_command_time(start_time);
usrp2->set_command_time(start_time);
uhd::time_spec_t usrp_t = usrp->get_time_now();
uhd::time_spec_t usrp2_t = usrp2->get_time_now();
usrp->clear_command_time();
usrp2->clear_command_time();
-----------------------------------------------------------------
But it does not work.
Again, How can I execute the command usrp->get_time_now(); and
usrp2->get_time_now(); simultaneously?
usrp: USRP B210
usrp2: USRP B200
Hojoon
-----------------------원본메일-----------------------
보낸사람 : "Marcus D. Leech via USRP-users"
<[email protected]>
받는사람: [email protected]
보낸날짜: 2018-04-26 01:41:04 GMT +0900 (Asia/Seoul)
제목: Re: [USRP-users] synchronization between two usrps.
On 04/25/2018 10:53 AM, Hojoon Yang via USRP-users wrote:
Hi,
I connected one usrp b210 and one usrp b200 to octoclock for
synchronization.
I'm trying to verify that the usrps connected to octoclock is
"really" synchronized.
To check it, I did as follows.
usrp->set_time_next_pps(uhd::time_spec_t(0.0));
usrp2->set_time_next_pps(uhd::time_spec_t(0.0));
boost::this_thread::sleep(boost::posix_time::seconds(2));
uhd::time_spec_t start_time = usrp->get_time_now() +
uhd::time_spec_t(0.1);
usrp->set_command_time(start_time);
uhd::time_spec_t usrp_t = usrp->get_time_now();
uhd::time_spec_t usrp2_t = usrp2->get_time_now();
usrp->clear_command_time();
I think I did wrong, because the results are different from my
expectations.
Q1. Can I execute usrp->get_time_now(); and
usrp2->get_time_now(); simultaneously? I'm trying to use
set_command_time, but it didn't work well as above.
Q2. How can I verify that the two usrps connected to the
octoclock, is really synchronized? in terms of time..
I already used get_time_last_pps() command, this command works
perfect, the two usrps showed same value.
but this command only showed "the full_secs" which does not
include frac_secs.. I want to check "the frac_secs" between two
usrps.
I read this thread,
http://lists.ettus.com/pipermail/usrp-users_lists.ettus.com/2015-January/012227.html
But, I can't find the way to check frac_secs..
Hojoon
https://files.ettus.com/manual/classuhd_1_1usrp_1_1multi__usrp.html#afb4bffe3f969c11ee7c0a2cba5178780
Returns a time_spec_t, and there are multiple methods associated
with time_spec_t
https://files.ettus.com/manual/classuhd_1_1time__spec__t.html
Including get_real_secs()
Many of the example programs included with UHD source use this
method--you could look there usage examples.
_______________________________________________
USRP-users mailing list
[email protected]
http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
_______________________________________________
USRP-users mailing list
[email protected]
http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
