Where does one find Trimble Timekeeper software?  and documentation on
how to use it?

Thanks,
Paul

On 1/22/2012 3:06 PM, Dennis Czelusniak wrote:

If you are keeping a network of PCs in sync to time the software to
use is "NTP".
http://www.ntp.org/
NTP cansyncto your GPS(s) and to other NTP severs out on the Internet.

On Sun, Jan 22, 2012 at 1:06 PM, Dennis Czelusniak
czelusniakd@yahoo.com wrote:

--

Chris Albertson
Redondo Beach, California

.. with setup instructions and hints on using with Windows here:

http://www.satsignal.eu/ntp/setup.html

I have a few of my Windows PCs connected to a variety of GPS sources
(Garmin GPS 18/x LVC and Sure GPS evaluation board).

Cheers,
David

SatSignal software - quality software written to your requirements
Web:  http://www.satsignal.eu
Email:  david-taylor@blueyonder.co.uk

On Sun, Jan 22, 2012 at 11:18 PM, David J Taylor
david-taylor@blueyonder.co.uk wrote:

This is not an NTP list.  Best to ask about NTP at
questions@lists.ntp.org
Opinion on that list is that for best result, build the latest NTP
from source.  Executable files (.exe) that you find may not be recent
builds.)    Go the the above email for details

So back to time.  If the goal is keeping good time then it is best
not to use Microsoft Windows.  There are good technical reasons having
to do with the way MS Win. keeps and adjusts time.  The bottom line is
that you will never be able to do better then about the millisecond
level even with a directly attached GPS.  Using another OS, BSD or
Linux you can do almost three orders of magnitude better.  (Three
orders is huge.)  The OS and software is free and all you need is any
computer that has a physical serial port, not USB but a real DB9
connector.  This is a good use for a 10 year old notebook PC.  (The
computational load is trivial so even a 486 class computer is OK)

If you run BSD or a newer Linux then Pin-1 on the DB9 can be used for
time interval measurement at almost 1 uSec precision.  You can use
this for many things, NTP being just one of them.  If you have
multiple DB9 type serial ports then you can time tag multiple channels
to about a uSec or two. (certainly 3E-6 level)

Then if the Windows PCs need correct time they could use the server
running on the old BSD based notebook PC.  Zero added cost assuming
you can find and old computer.  Then you also have a tool that can
time tag any pulse at the uSec level.

Chris Albertson
Redondo Beach, California

El 23/01/2012 19:29, Chris Albertson escribió:

Hello,

I agree at all at not thinking on MS Windows for any precision timing
task (even low precision). We have some systems with a Windows based
workstation and Meinberg ntp distribution, and several linux embedded
processors taking the time from the workstation using ntp in a LAN (with
very very low traffic, also). This works well if you only need relative
precision (i.e. 1 second) time difference to the workstation, but once
we tried an application that required a synchronization between the
linux processors and the windows system in the order of 10ms, and this
never worked reliably. This is trivial to put into work in a LAN using
ntp or chrony. By orders of magnitude.

And that to not speak about the windows time server, not fully ntp
compatible but partially (to make things more confusing - the usual MS
way of taking an industry standard and pervert it on its own way -
rather imaginative also). You only need to have a look to this:
http://support.microsoft.com/kb/939322 amd particularly to the
paragraphs "We do not guarantee and we do not support the accuracy of
the W32Time service between nodes on a network" and "The W32Time service
cannot reliably maintain sync time to the range of 1 to 2 seconds". This
explains all...

Regards,

Javier

I agree that Windows is not the best OS for timekeeping, but with a
directly attached GPS/PPS device I consistently see better than the
millisecond level with Windows.  Please see:

http://www.satsignal.eu/mrtg/performance_ntp.php

PCs Feenix (XP), Stamsund (Win-7/32), Alta (Win-7/64) and Bacchus (Win
2000) demonstrate this.  All are connected to either GPS 18, GPS18x, or
Sure GPS devices.  There is also interpolation code available allowing you
to read time more precisely than 1 ms.

http://www.lochan.org/2005/keith-cl/useful/win32time.html

My own experience suggests that you need rather more than a PC running a
good timekeeping OS - such as FreeBSD.  Take a look at PC Pixie:

http://www.satsignal.eu/mrtg/performance_ntp.php

This is an Intel Atom box running FreeBSD 8.0, and yet in an open domestic
environment there are reported excursions of approaching 10 microseconds
when the heating is switched on in the morning.  So put that PC in some
sort of thermally-controlled enclosure if you want better than the 10
microsecond level.  The GPS puck is outside, by the way, unaffected by the
heating switch-on.

Comparing PCs Pixie and Feenix suggests that the FreeBSD system is only
two orders of magnitude better than Windows - still a great improvement,
of course!

Cheers,
David

On Tue, Jan 24, 2012 at 12:23 AM, David J Taylor
david-taylor@blueyonder.co.uk wrote:

Ok, about half of the Windows servers have 0.5 millisecond offsets and
some have 3 millisecond offsets.  I still call that "millisecond
level".    That is very different from "microsecond level".  To me the
terms "milli level", "nano level" and so on mean "round to the nearest
three orders of magnitude."  It is a VERY course level of rounding
even more so than"rough order of magnitude" type rounding.      So I
think "micro second level" means not "within one uSec but "you are
using uSec as your units when you tell people the number"

For most casual users even 100mS is better than they need.  No one
notices if you are 100mS late for a meeting.

Chris Albertson
Redondo Beach, California

Thanks, Chris.

http://www.satsignal.eu/mrtg/performance_ntp.php

The graphs may be confusing at first glance.  To overcome the limitation
of MRTG of only plotting positive numbers, I needed to add a fixed amount
to the reported offset so that it becomes a positive value.  I chose the
offset and scaling values so that a server keeping perfect time would be a
line right down the centre of the graph.  The offset reported by NTP is
the deviation from that centre-line.  The left axis annotation states this
offset.

FreeBSD PC Pixie shows deviations around 10 microseconds from nominal,
mostly caused by the heating switching on.

Windows XP PC Feenix shows maximum deviations of around 200 microseconds.

Windows-7/32 PC Stamsund shows maximum deviations which are likely under
50 microseconds.

The scale on the graphs for PCs Alta and Bacchus does not allow the
deviations to be estimated accurately (as they were not originally
intended as stratum-1 servers), but the deviation may well be under 0.3
ms.

A couple of these PCs have speaking clocks (software) running, and
multiple speakers speaking at differing times is rather distracting.  They
are also monitoring a satellite data broadcast, where event times are
compared across PCs, and across Europe, and having logs time stamped
reasonably accurately can help.  Other folk are using PCs to measure the
propagation delay of radio signals where, I understand, millisecond
accuracy is the level talked about.

Colloquially, I might say: FreeBSD tens of microseconds, Windows hundreds
of microseconds or perhaps Windows tenths of milliseconds.

Cheers,
David

SatSignal software - quality software written to your requirements
Web:  http://www.satsignal.eu
Email:  david-taylor@blueyonder.co.uk