[time-nuts] Re: Network interface cards that support timestamping

[James Clark]

On Fri, Feb 3, 2023 at 1:30 AM Matt Corallo via time-nuts
<time-nuts at lists.febo.com> wrote:
>
> Makes sense.
>
> Sadly, as far as I can tell, the i211 will mark both the rising and falling edge of the pulse, with
> no distinction between the two. While you'd expect chrony to be able to differentiate based on the
> time between them, the manual for chrony.conf seems to indicate otherwise, saying:
>
>  > width width
>  > This option specifies the width of the pulses (in seconds). It is used to filter PPS samples when
>  > the driver provides samples for both rising and falling edges. Note that it reduces the maximum
>  > allowed error of the time source which completes the PPS samples. If the duty cycle is
>  > configurable, 50% should be preferred in order to maximise the allowed error.
>
> which, maybe incorrectly, implied to me that there was no smarts for highly asymmetric
> pulse-to-not-pulsing time.

I think your analysis is correct, since
https://chrony.tuxfamily.org/examples.html#_server_using_reference_clock_on_nic
says

"The offset option of the SHM 0 refclock compensates for the delay of
messages received on the USB port. It needs to be measured carefully,
e.g. against a known good NTP server. A wrong offset could cause the
server to be off by an integer multiple of 62.5 milliseconds (1/16s)."

I've been writing a program to sync the PHC from a GPS and it includes
the obvious smarts to distinguish the rising and falling edges based
on the time between.  I have been testing with the i210 and it seems
to work fine. With these smarts, 50% duty cycle is the worst possible
configuration, since it prevents the smarts from working.

James