[LinuxPPS] High values of jitter/estimated error.

[Stefano Lucetti]

On Mon, 30 Apr 2007 12:00:27 -0600
Reg Clemens <reg at dwf.com> wrote:

> 
> Unless you have run a program that prints out the ASSERT and CLEAR timestamps,
> and are sure that you want to be using ASSERT, I would try CLEAR just to be 
> sure.
...
> In any case, you large jitter may just be from using the wrong side of the 
> signal,
> try the other.
> 

Ok, I will try using the CLEAR edge and let you know. Indeed, if had looked the output of ppstest with more attention,  I would have noticed that the timings for the CLEAR signals have a lower variability over time (don't know exactly if it is relevant, but I guess so)...

nx9005:~# ppstest
found PPS source #0 "serial0" on "/dev/ttyS0"
ok, found 1 source(s), now start fetching data...
source 0 - assert 1178093574.000325207, sequence: 170735 - clear  1178093574.801542954, sequence: 170735
source 0 - assert 1178093575.000381915, sequence: 170736 - clear  1178093574.801542954, sequence: 170735
source 0 - assert 1178093575.000381915, sequence: 170736 - clear  1178093575.801541762, sequence: 170736
source 0 - assert 1178093576.000436503, sequence: 170737 - clear  1178093575.801541762, sequence: 170736
source 0 - assert 1178093576.000436503, sequence: 170737 - clear  1178093576.801541968, sequence: 170737
source 0 - assert 1178093577.000495927, sequence: 170738 - clear  1178093576.801541968, sequence: 170737
source 0 - assert 1178093577.000495927, sequence: 170738 - clear  1178093577.801540779, sequence: 170738
source 0 - assert 1178093578.000557587, sequence: 170739 - clear  1178093577.801540779, sequence: 170738


Thank you

Stefano