Re: [gpsd-users] Ublox EVK-M8T

[Denny Page]

> On Sep 21, 2017, at 16:18, Gary E. Miller <address@hidden> wrote:
> 
> Did you try NTPsec?  Then you can apply an offset.  And how does it
> even know which 100 milli sec unless it has PPS time?

I haven’t done anything with NTPsec. It’s on my interest list however.

As to how I knew which was the correct edge… with the “polarity” set to 
“rising” (the default), I can easily view a shift in the kernel’s view of the 
pps assert by varying the width of the pulse sent by the M8T. For instance, if 
I change the pulse width from 100ms to 400ms, I see an immediate 300ms shift in 
the pps assert against the (otherwise disciplined) system clock. Even with an 
undisciplined clock, this would be an obvious indication that the assert was 
occurring on the wrong edge. With the “polarity” set to “falling”, lengthening 
the pulse width has no effect on the pps assert.

In addition, I’ve compared it to other clocks. in the local network I have 5 
hardware based stratum 1 NTP servers, 2 PTP servers, and a GR-701W. It’s a 
hobby. 

Denny