[ntp:questions] Fw: Re: rf links in the LAN

Mark Newman markn_46 at yahoo.com
Mon Jan 12 19:56:33 UTC 2009



--- On Mon, 1/12/09, marknmbox-88 at yahoo.com <marknmbox-88 at yahoo.com> wrote:

From: marknmbox-88 at yahoo.com <marknmbox-88 at yahoo.com>
Subject: Re: [ntp:questions] rf links in the LAN
To: rgilbert88 at comcast.net
Date: Monday, January 12, 2009, 2:54 PM

Sorry this should have the CR/LFs.

The numbers that are typically quoted are 100 microseconds for a LAN and greater 
than 10 mills for a WAN using internet servers.

This is a LAN with multiple STRATUMs the top level STRATUM uses something like
 a rubidium clock.  Most of the network is implemented using some form of ethernet.  
The "catch" is that some of the lower level stratums communicate via RF (Radio).  
Since this is a LAN with connections to a precision source and predictable routing 
I am assuming that I will see an accuracy in the 100s of microseconds range,

I am hoping that said radios hops will have no effect on the accuracy of the lower 
level clocks.



--- On Mon, 1/12/09, Richard B. Gilbert <rgilbert88 at comcast.net> wrote:

From: Richard B. Gilbert <rgilbert88 at comcast.net>
Subject: Re: [ntp:questions] rf links in the LAN
To: questions at lists.ntp.org
Date: Monday, January 12, 2009, 1:02 PM

Mark Newman wrote:
> If a system consists of a large LAN which consists of small fiber optic LANs connected together via an RF network what precision can be expected?  One of the subLANs contains a STRATUM 1 server sync'ed via a precision clock.  The other subLANs contain level 2 STRATUM servers.
> 
> What precision should I expect to see?

I believe that "precision" is the wrong word here.  I think you want to 
know the "accuracy"!

"Precision" is the smallest possible difference between two consecutive 
readings of the system clock and is usually expressed as a negative 
power of 2; e.g. 2^-20.

"Accuracy" is the
 difference between the clock's time and the correct time!

BTW, it's customary to use your carriage-return/enter key after typing 
about 70 characters.  Posting ~300 characters without a single carriage 
return makes your message extremely difficult for most people to read!

If you are getting your time from an internet server that has the 
correct time, the uncertainty at your end is one half the round trip delay.

If you have a hardware reference clock; e.g. a GPS timing receiver, your 
accuracy might be as good as +/- 100 microseconds.  The GPS receiver 
will generally know the time +/- 50 nanoseconds but considerable 
uncertainty is introduced in getting the time from the GPS receiver into 
your computer which generally must be done via a serial port!

_______________________________________________
questions mailing list
questions at lists.ntp.org
https://lists.ntp.org/mailman/listinfo/questions



More information about the questions mailing list