Best practices question

[Paul Theodoropoulos]

On 6/26/19 10:11, Paul Theodoropoulos wrote:
> On 6/26/19 10:03, Gary E. Miller wrote:
>> Yo Paul!
>>
>> On Wed, 26 Jun 2019 08:36:04 -0700
>> Paul Theodoropoulos<paul at anastrophe.com>  wrote:
>>
>>> refclock shm unit 1 refid PPS minpoll 0 maxpoll 0 flag4 1 prefer
>> Depending on many things, I find that poll 2 or 3 is often the best
>> choice.  One may give you better jitter and another better frequency
>> stability.
>>
>> ntpviz will help you see the differences.  I've done consecutive tests
>> stepping through the possible poll intervals.
>>
>
> For sure - I'll take a poke at it.
>
> So little time, so much experimenting to do!
>
wow, fascinating - I set my c-ntpsec PPS min/maxpoll to 2. The RMS time 
jitter rose a fairly significant amount, from hovering around 0.6µs  to 
now 1.0µs  -- while the RMS frequency jitter plummeted from hovering 
around 4ppb down to, well, barely discernably above 0ppb on the graph.

eh, saying it just doesn't tell the story -

Time jitter:  https://i.imgur.com/jNlaAAj.png

Frequency jitter: https://i.imgur.com/wAF2nu9.png


-- 
Paul Theodoropoulos
www.anastrophe.com

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.ntpsec.org/pipermail/users/attachments/20190626/a6bb80ca/attachment-0001.htm>