time-nuts@lists.febo.com

Discussion of precise time and frequency measurement

View all threads

new member with questions NTP, PRS, GPS, ocxo

CA
Chris Albertson
Sun, Aug 19, 2012 8:39 PM

On Sat, Aug 18, 2012 at 9:20 AM, David J Taylor <
david-taylor@blueyonder.co.uk> wrote:

From: Chris Albertson
Sent: Saturday, August 18, 2012 4:55 PM
[]

Power is a bigger issue.  It really does cost a bit to keep some
machines owered up 24x7 and NTP needs to run all the time.  It takes
NTP hours to stabilize and you want to wait 24 hours after to measure
performance.
[]

Chris Albertson
Redondo Beach, California
==============================**==================

I see this "hours" or "many hours" figure quoted, and yet here NTP is
sufficiently accurate within a few minutes.

For most uses "sufficiently" is about 0.1 seconds.  Maybe 0.01 seconds for
some critical data logging.  But this being a time nuts
list "sufficiently" means "unable, even with great effort to make it
better."

But also what is the point of a server it you don't run it 24x7?

Chris Albertson
Redondo Beach, California

On Sat, Aug 18, 2012 at 9:20 AM, David J Taylor < david-taylor@blueyonder.co.uk> wrote: > From: Chris Albertson > Sent: Saturday, August 18, 2012 4:55 PM > [] > > Power is a bigger issue. It really does cost a bit to keep some > machines owered up 24x7 and NTP needs to run all the time. It takes > NTP hours to stabilize and you want to wait 24 hours after to measure > performance. > [] > > Chris Albertson > Redondo Beach, California > ==============================**================== > > > I see this "hours" or "many hours" figure quoted, and yet here NTP is > sufficiently accurate within a few minutes. For most uses "sufficiently" is about 0.1 seconds. Maybe 0.01 seconds for some critical data logging. But this being a time nuts list "sufficiently" means "unable, even with great effort to make it better." But also what is the point of a server it you don't run it 24x7? Chris Albertson Redondo Beach, California