[time-nuts] question about multi-way measurement

[David J Taylor]

From: Hal Murray

On the Raspberry Pi, the Ethernet is on USB so there is another source of
timing error.  It's full speed USB rather than low speed USB so the timing
noise isn't as bad as it could be - 1/8 ms vs 1 ms.  (I could be off on the
numbers or names, but the general idea is correct.)

If you think you can avoid that by averaging, be sure you understand hanging
bridges.

If you need good times on a Pi, consider a GPS hat.  The PPS goes in via a
GPIO pin with an interrupt so the timing can be pretty good.
=============================

I have some Raspberry Pi systems running with PPS sync from an indoor GPS 
receiver (on the upper floor of a two storey building) and in the table 
below RasPi-1 is in an unheated cupboard.  It plots the offset as reported 
by NTP every five minutes.

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

Avoid quick temperature changes for best performance (e.g RasPi-12 is 
located far too near a central heating radiator.

There are some notes here:

  https://ava.upuaut.net/?p=951
  https://www.satsignal.eu/ntp/Raspberry-Pi-quickstart.html

The Uputronics article says, incorrectly, that NTP won't work stand-alone, 
without the Internet.  I'll write to him about that!

Cheers,
David
-- 
SatSignal Software - Quality software for you
Web: http://www.satsignal.eu
Email: david-taylor at blueyonder.co.uk
Twitter: @gm8arv