Re: [gpsd-dev] Very basic PPS question:

[Andy Walls]

I'm no expert, but that is likely device dependant.

For the few devices I have worked with: the PPS pulse is for the top of the second that the device then reports via the serial stream.  

I verified this once against NTP servers, for the device I am currently working with.  (This is also how I found my device uses the wrong leapsecond correction after poweron, before it gets the almanac.  I guess I'm a mild time-nut.)

I also found, on average over a 4 hour data collection, that the serial timestamp gpsd reports from my device, is ~450 ms delayed from the PPS.  I use that value for the ntpd.conf fudge for my gpsd refclock.

Regards,
Andy

"Eric S. Raymond" <address@hidden> wrote:

When a GPS asserts PPS, is it top of second for the in-stream 
data *preceding* or *following*?  

I presume following, but I want an expert to confirm before I
put it in the HOWTO.

-- 
Sent from my Android phone with K-9 Mail. Please excuse my brevity.