How long was the observation period between the initial and final
estimation?
Were the devices powered cycled, or GPSDO re-locked between each test?
On Thu, 7 Aug 2025 at 19:31, nirkicatal1@gmail.com wrote:
Hi,
We are observing a time difference of arrival (TDOA) variation between two
E320 devices using PCM signals.
-
The devices receive with zero hardware delay delta to the RF inputs.
-
Both devices have the latest filesystem installed (UHD 4.8).
-
Matched filtering is applied to estimate the time of arrival of the
pulses.
-
The time of arrivals are subtracted to compute the TDOA.
-
Internal GPSDO PPS and 10 MHz clocks are used.
-
cgps reports GDOP values around 2.0 for both devices.
-
Multiple 60-second test runs have been conducted.
-
A deviation of around 150 ns in TDOA between the initial and final
estimations across the radios is observed.
We know that the matched filtering estimation and other imperfections
follow a stationary Gaussian distribution with a 20 ns sigma (and a
constant mean per run). We verified this behavior on the same setup using
two B210 devices with external Ettus-supplied GPSDOs. The B210s report
similar GPS performance to the E320s.
What could be the cause of the non-stationary behavior observed on the
E320s?
USRP-users mailing list -- usrp-users@lists.ettus.com
To unsubscribe send an email to usrp-users-leave@lists.ettus.com
How long was the observation period between the initial and final
estimation?
Were the devices powered cycled, or GPSDO re-locked between each test?
On Thu, 7 Aug 2025 at 19:31, <nirkicatal1@gmail.com> wrote:
> Hi,
>
> We are observing a time difference of arrival (TDOA) variation between two
> E320 devices using PCM signals.
>
> -
>
> The devices receive with zero hardware delay delta to the RF inputs.
> -
>
> Both devices have the latest filesystem installed (UHD 4.8).
> -
>
> Matched filtering is applied to estimate the time of arrival of the
> pulses.
> -
>
> The time of arrivals are subtracted to compute the TDOA.
> -
>
> Internal GPSDO PPS and 10 MHz clocks are used.
> -
>
> cgps reports GDOP values around 2.0 for both devices.
> -
>
> Multiple 60-second test runs have been conducted.
> -
>
> A deviation of around 150 ns in TDOA between the initial and final
> estimations across the radios is observed.
>
> We know that the matched filtering estimation and other imperfections
> follow a stationary Gaussian distribution with a 20 ns sigma (and a
> constant mean per run). We verified this behavior on the same setup using
> two B210 devices with external Ettus-supplied GPSDOs. The B210s report
> similar GPS performance to the E320s.
>
> What could be the cause of the non-stationary behavior observed on the
> E320s?
> _______________________________________________
> USRP-users mailing list -- usrp-users@lists.ettus.com
> To unsubscribe send an email to usrp-users-leave@lists.ettus.com
>
