[time-nuts] x86 CPU Timekeeping and clock generation

[Javier Herrero]

Hello,

As pointed by others, probably the ntpd is applying a correction based 
on its last measured drift. If you have synchronized the test computer 
to the server via Ethernet, the drift value usually wanders somewhat (it 
should be perfectly zero since both test computer and server clocks are 
derived from the same atomic source, but the errors in the packet 
timestamping due to jitter in the network will make the drift to value 
to wander somewhat around zero).

You can try to do the initial synchronization, and after leaving the 
test computer to run free, stop the ntpd daemon, set a value of zero in 
the ntp.drift file, and restart the ntpd daemon. Then, the ntp will 
assume the drift as zero and no longer apply a drift correction, so the 
computer time should not drift.

Best regards,

Javier

On 6/1/21 6:35, Luiz Paulo Damaceno wrote:
> Hi all,
>
> I'm studying computer's timekeeping and i'm on level of remove the base
> crystal that feeds the entire PLL logic of the motherboard (24 MHz on
> motherboard that i'm using) and compare system's time with an NTP server.
>
> The 24 MHz comes from an synthesizer that is locked to an atomic clock, the
> clock of NTP server (also 24 MHz, but an embedded board (Tinkerboard)) also
> comes from the same Atomic clock that is feeding other synthesizer for
> generates 24 MHz to this board.
>
> The experiment is the following: 1- synchronize the computer's clock to NTP
> server then leave it running free (no periodic synchronization), 2 -
> acquire time difference between computer and time server. What i'm
> observing is that the computer time is drifting over time, but
> theoretically it cannot drift because its connected to same clock source
> (the atomic clock) of the time server.
>
> My question is: what i'm missing? Someone have some tips to me to research
> about timekeeping with base clock? I've already changed the base clock in
> some hertz (frequency correction based on the drift) at the synthesizer,
> but seems to have no effect, times keep drifting at constant rate, this
> means that the "1sec pulse" generated in system to count time is not truly
> 1 sec pulse, this because the system is in an different frequency of my ntp
> server time pulse.
>
> Thank you for your time and happy new year,
>
> Luiz
> _______________________________________________
> time-nuts mailing list -- time-nuts at lists.febo.com
> To unsubscribe, go to http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
> and follow the instructions there.