Re: Time precision in gps_fix_t

[Marco Spanghero]

Hi,

thanks for the reply. I am aware of it, but for the application I am working with now, the host system cannot access a PPS line. I can see the UBX decoding in the GPSD logs, and I would really like to be able to access that information.

When I do a gps_read and retrive the gps_data_t, for what I understand I get back the JSON object TPV.

Is there any way I can access the decoded UBX with gps_read?

As the driver decodes it, I assume it should be known to GPSD right? what I am interested in is to expose the decoding of UBX-GPSTIME. Any way I can achieve that?

Marco

On 22/01/2020 1:26 pm, Trygve Lunheim wrote:

Hi Marco,

Like Gary said, PPS is what you should be looking at for your application.

The ubx sentence with microsecond resolution is mostly useful for post-processing of position data.

The serial transmission and string processing is inherently non-deterministic, and will have jitter. And the clock you read will have latency issues.

From an u-blox manual:

  

(and that is before data goes through the uart, serial line and through the receiver end processing. Maybe a serial/usb interface with buffering as well)

You should read up on the Timepulse outputs (there should be 2 on the ubx-m8030), and how you can access them on your development card, or what you have available.

The pinouts don't seem to be publicly available, so you may need to contact u-blox.

If you are looking for references on high precision time synchronization I think this site has lots of information:

http://leapsecond.com/time-nuts.htm  

Kind regards,

Trygve

ons. 22. jan. 2020 kl. 04:09 skrev Marco Spanghero <address@hidden>:

Hi All,

I am using GPSD for research purposes about high precision time synchronization.

I wrote my own client in C that gets a structure gps_data_t that contains gps_fix_t. I am trying to access the time with a higher resolution than 1s.

The module I am using is a ublox ubx-m8030 that can output a ubx sentence with specific timing down to microseconds as I can see in ucenter software.

My questions are:

why the field time in gps_fix_t, even if it has second and subsecond field, the subseconds field is set to 0 in all measurements?

how to I access the ubx time sentence from gpsd?

where can I access a high resolution time information? I know GPS should do better than 1s, 100us should be reasonable.

Any help would be highly appreciated.

Marco

PhD Student

KTH Royal Institute of Technology

Stockholm

--

Marco Spanghero, PhD Student
Networked Systems Security Group
NSS Home Page