I don't think the solution is ntp when it is over 15 seconds per day.
You use ntp once per day to maintain within 1 second. I like to have a
goal of being within 1/4 of a second to meet a spec of 1 second.
Really, you would like a remote system not attached to the network to
stay within a second or two a week. My cheap quartz wall clock is
probably within a second or two a week. I never set it except for
daylight savings time and when the battery is changed.
Here are the results from less than 24 hours:
$ hwclock ; date
Tue Apr 18 09:50:59 2006 0.000000 seconds
Tue Apr 18 09:50:44 UTC 2006
$
The hardware clock is still within a second, maybe a second slow, and
the system clock over 15 seconds slow. That would be two minutes a
week, or almost 10 minutes per month.
Oh, I checked after re-booting, and it does correctly read the
hardware clock on re-boot, and the system clock is back in step with
the battery-backed clock.
Tue Apr 18 10:10:09 2006 0.000000 seconds
Tue Apr 18 10:10:09 UTC 2006
$
The temporary solution I will try is to run "hwclock --hctosys" every
hour or half hour.
Don.
On 4/17/06, Yan Seiner <> wrote:
>
> --- In "Don W. Carr" <> wrote:
> >
> > I have been running my 7260 now for about 8 hours since setting both the
> > system clock with date, and the battery backed real-time clock with
> hwclock
> > to my atomic clock that synchronizes every night. It appears that the
> > battery-backed real-time clock is right on, but the system clock has
> lost
> > about 4 seconds in 8 hours (not good). I have an application where I
> need to
> > know with max 1 second error when a train crosses a sensor.
> >
> > Below you can see I ran hwclock and date at the same time, right at
> the top
> > of the minute.
> >
> > $ hwclock ; date
> > Mon Apr 17 20:10:00 2006 0.000000 seconds
> > Mon Apr 17 20:09:56 UTC 2006
> > $
> >
> > When I started the test, they both read exactly the same.
> >
> > Any ideas on how to keep the system clock in better sync? I would
> like these
> > systems to stay within 1 second even if they are disconnected from
> the net
> > for 24 hours.
> >
>
> Like I said, the only thing that comes to mind is to write a driver
> for the RTC clock that interfaces with something like ntpd....
>
> I don't have time for this right now, and won't have for another month
> or two. I'm in a panic to get the basic UI written at the moment...
>
> At some point I'll have to deal with it you haven't... :-)
>
> --Yan
>
> --Yan
>
>
>
>
>
>
> ________________________________
> YAHOO! GROUPS LINKS
>
>
> Visit your group "ts-7000" on the web.
>
> To unsubscribe from this group, send an email to:
>
>
> Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service.
>
> ________________________________
>
--
Dr. Don W. Carr
J. G. Montenegro 2258
Guadalajara, Mexico
+52-333-630-0704
+52-333-836-4500 ext 2930
Yahoo! Groups Links
<*> To visit your group on the web, go to:
http://groups.yahoo.com/group/ts-7000/
<*> To unsubscribe from this group, send an email to:
<*> Your use of Yahoo! Groups is subject to:
http://docs.yahoo.com/info/terms/
|