Empathy List Archives

usrp-users@lists.ettus.com

Discussion and technical support related to USRP, UHD, RFNoC

View all threads

Re: [USRP-users] mimo sync and acceracy

iftah giladi

Sun, Sep 7, 2014 7:08 PM

Hi Marcus and all,

Thanks for the quick answer.

In your answer you wrote:

```
  What is the accuracy of this:
```

tod = uhd::time_spec_t::get_system_time();

usrp->set_time_now(tod);

The two USRPs should then be sample-synchronous. However, setting the

time without any trigger is not going to be very exact, give or take

10ms, because the USRP just sets its time when it receives the command.

If you want more accuracy, you need a pulse that will go up at the exact

time you want to set at the PPS input and use set_time_unknown_pps or

one of its sister methods. To be honest, though, when using the system

time, 10ms isn't that much, so "accuracy" is always relative.

And my new question now is:

If I don't have an external source, and I don't use a GPSDO, so were will
that pps will come from?

Actually what you saying if I got you right is that If I want both usrp to
have the same time with accuracy butter then 10msec, then I HAVE to either
use to GPSDO PPS or an external PPS

Or I won't get the accuracy in time between the two usrp, right?

And what if I won't use either? They will still be sample synced right? Only
the time base will be biased, right?

Plus and that has nothing to do with the time issue the two channel will
have a phase difference dew to the different PLL who created the A2D clock,
right?

Thanks again,

iftah

Hi Marcus and all, Thanks for the quick answer. In your answer you wrote: >1. What is the accuracy of this: >tod = uhd::time_spec_t::get_system_time(); >usrp->set_time_now(tod); The two USRPs should then be sample-synchronous. However, setting the time without any trigger is not going to be very exact, give or take 10ms, because the USRP just sets its time when it receives the command. If you want more accuracy, you need a pulse that will go up at the exact time you want to set at the PPS input and use set_time_unknown_pps or one of its sister methods. To be honest, though, when using the system time, 10ms isn't that much, so "accuracy" is always relative. And my new question now is: If I don't have an external source, and I don't use a GPSDO, so were will that pps will come from? Actually what you saying if I got you right is that If I want both usrp to have the same time with accuracy butter then 10msec, then I HAVE to either use to GPSDO PPS or an external PPS Or I won't get the accuracy in time between the two usrp, right? And what if I won't use either? They will still be sample synced right? Only the time base will be biased, right? Plus and that has nothing to do with the time issue the two channel will have a phase difference dew to the different PLL who created the A2D clock, right? Thanks again, iftah

Marcus Müller

Tue, Sep 9, 2014 12:42 PM

Hi Iftah,

On 07.09.2014 21:08, iftah giladi via USRP-users wrote:

And my new question now is:

If I don't have an external source, and I don't use a GPSDO, so were will
that pps will come from?

The master N2x0.

Actually what you saying if I got you right is that If I want both usrp to
have the same time with accuracy butter then 10msec, then I HAVE to either
use to GPSDO PPS or an external PPS

The time will be exactly the same on both USRPs, but it will have an
offset to the "real" time.

You can of course go the RF communication route and correlate an
arbitrary signal [1] that is sent on a specific time and find the
maximum to estimate the temporal offset and account for it in software.

And what if I won't use either? They will still be sample synced right? Only
the time base will be biased, right?

yep, the time will be biased. Now, the harsh truth is: if you want
GPS-synchronous absolute time, you'll have to get it from GPS or
something technically comparable[2].

Plus and that has nothing to do with the time issue the two channel will
have a phase difference dew to the different PLL who created the A2D clock,
right?

Yes, aside from very few daughterboards where the tuner has a mechanism
to avoid that.

Greetings,
Marcus

[1] well, it must not be wide-sense cyclostationary within your
observation length, otherwise you'd get ambiguities. Also, the temporal
resolution of any signal, be it pulse, chirp, pseudorandom noise or
whatever, is inversely proportional to its bandwidth.
[2] GPS employs knowledge about multiple satellites to eliminate the
time the waves travel from satellite to GPSDO. As a completely
unnecessary but fun side effect, you can use this knowledge to estimate
your location in space if you see enough satellites.

Hi Iftah, On 07.09.2014 21:08, iftah giladi via USRP-users wrote: > And my new question now is: > > If I don't have an external source, and I don't use a GPSDO, so were will > that pps will come from? > The master N2x0. > Actually what you saying if I got you right is that If I want both usrp to > have the same time with accuracy butter then 10msec, then I HAVE to either > use to GPSDO PPS or an external PPS The time will be exactly the same on both USRPs, but it will have an offset to the "real" time. You can of course go the RF communication route and correlate an arbitrary signal [1] that is sent on a specific time and find the maximum to *estimate* the temporal offset and account for it in software. > And what if I won't use either? They will still be sample synced right? Only > the time base will be biased, right? yep, the time will be biased. Now, the harsh truth is: if you want GPS-synchronous absolute time, you'll have to get it from GPS or something technically comparable[2]. > > Plus and that has nothing to do with the time issue the two channel will > have a phase difference dew to the different PLL who created the A2D clock, > right? Yes, aside from very few daughterboards where the tuner has a mechanism to avoid that. Greetings, Marcus [1] well, it must not be wide-sense cyclostationary within your observation length, otherwise you'd get ambiguities. Also, the temporal resolution of any signal, be it pulse, chirp, pseudorandom noise or whatever, is inversely proportional to its bandwidth. [2] GPS employs knowledge about multiple satellites to eliminate the time the waves travel from satellite to GPSDO. As a completely unnecessary but fun side effect, you can use this knowledge to estimate your location in space if you see enough satellites.