[ntp:questions] Sure serial - USB with PPS performs 16% better after Windows reboot

Ron Frazier (NTP) timekeepingntplist at c3energy.com
Sun Mar 25 22:14:12 UTC 2012

I see what you're saying.  However, all the internet servers are 
currently noselected.  So, everything on this graph represents NTPD 
polling the serial GPS through the serial - USB converter.

I'm not particularly worried about it.  I was just wondering why the 
system would apparently perform better after a reboot than after an NTPD 


On 3/25/2012 5:23 PM, Chris Albertson wrote:
> Ron,
> Yes, I see the graph has segments.
> I'm guessing that NTP's clock selection process switched the set of
> Internet servers it was using.  Would be easy to check the log.
> "offset" of course is the difference between two clocks.  My guess is
> that the GPS is stable and it's the Internet that is moving.
> At some point as you go much below 1mS error you will need some other
> reference in order to sort out if the error is in NTP, GPS or your
> Internet connection.
> On Sun, Mar 25, 2012 at 1:51 PM, Ron Frazier (NTP)
> <timekeepingntplist at c3energy.com>  wrote:
>> Hi all,
>> I think I've discovered something very interesting.  Just by chance, I had
>> to take the laptop computer to someone else's house the other day to do some
>> work.  So, I shut it down and unplugged the GPS and left.  When I came back,
>> I plugged the GPS back in and rebooted.  After a while, I checked my graphs.
>>   I found peak offsets in the 800 us range rather than in the 1.2 ms range I
>> had seen before.  I restarted NTPD and started another graph.  After a
>> while,  saw the 1.2 ms offsets I had seen in the past.  Then, after forming
>> a theory that the reboot may have some effect.  I rebooted again.  After
>> letting the graph run a while, I was again seeing offsets in the 800 us
>> range.  See the graph below.
>> http://dl.dropbox.com/u/9879631/2012-03-24-2327-loopstats-reboot-restart-reboot.png
>> There are three distinct sections.  First, is the beginning.  Here, as you
>> can see, the peak offsets are approximately + .8 / - .6 ms.  Then, there is
>> the middle.  As you can see, peak offsets are approximately + 1.2 / - 1.2
>> ms.  Then, there is the end.  There are two subsections here.  In the first,
>> the peak offsets are approximately + .8 / - .6 ms again.  In the second
>> subsection, the peak offsets are approximately + 1 / - .7 ms.  I don't know
>> why there is a variance between the subsections.  However, the beginning of
>> the graph is a time just after a reboot, the middle is a time just after an
>> NTPD restart, and the end is a time just after another reboot.
>> I can use the signal either way.  I just thought this was interesting.
>>   Anybody know what's going on?
>> Sincerely,
>> Ron


(PS - If you email me and don't get a quick response, don't be concerned.
I get about 300 emails per day from alternate energy mailing lists and
such.  I don't always see new messages very quickly.  If you need a
reply and have not heard from me in 1 - 2 weeks, send your message again.)

Ron Frazier
timekeepingdude AT c3energy.com

More information about the questions mailing list