[ntp:questions] Odd (mis)behavior when reference clock fails

Unruh unruh-spam at physics.ubc.ca
Fri Sep 26 17:14:44 UTC 2008


hal-usenet at ip-64-139-1-69.sjc.megapath.net (Hal Murray) writes:

>>Does it really make sense to get microsecond resolution on the glitches.

>I doubt if whatever the UPS is using for an A/D is good for microseconds.

>I've thought of using a wall-wart and some dividers to feed an audio
>input channel.  That would probably be good for 10s of microseconds.

>I'm saving that for when I have lots of spare time.


>>Try puttin in a microsleep into your routine to see if that solves the
>>clock problem. It is possible that the reading of the ups is done via some
>>sort of badly written interrupt routine which holds open the interrupts too
>>long. It is however a bit weird I will admit.

>I talk to the UPS over a serial port.  I suppose I could try
>a USB to serial gizmo.

That would be far worse. And serial is good maybe for 100s of milliseconds,
certainly not microseconds. 
Try decreasing the ups polling and see if makes a difference.


>Another idea is to make an echo plug and some software that sends
>stuff to itself and see what that does on various systems.

>-- 
>These are my opinions, not necessarily my employer's.  I hate spam.




More information about the questions mailing list