[LinuxPPS] Re: PLL sensitivity

[Remco den Besten]

The Linux systems run 2.6.27.x (x=5,6) with ntp-dev-4.2.5p113 and
FreeBSD is rel 7.0 with PPS_SYNC and also runs ntp-dev-4.2.5p113.

(I stuck at p113 because I discovered that 'newer' versions resulted in 
problems
with my scripts, making the graphs)

Helium and FreeBSD are approx. 4" next to each other in the same room
and use their own GPS receiver (i.e. not a splitted PPS source).
Helium an Oncore UT+ and FreeBSD a Rockwell Jupiter. Both receivers share 
the same antenna.
Phenomena are the same when both systems use Oncores or when I reverse the 
receivers.

Reg wrote:
> When the system is 'disturbed' e.g. by temperature or CPU load, FreeBSD
> shows one overshoot and is
> almost immediately 'back into business' (whereas Linux needs a lot of time
> to recuperate

Both FreeBSD and Linux run ntp.

I frankly dont see how the PPS implementation can cause a difference like
you see.  Rather it could be the version of ntp.

Can you report the versions of ntp that both the FreeBSD and the Linux
system are running.