Re: [chrony-users] Chrony not taking SOCKET data from Application

[ Thread Index | Date Index | More Archives ]

I do not know what your setup is, but the problem is that the getting the
signal from the gps receiver, having it trigger an interrupt, then reading the
computer clock takes a while. A number of years ago when I tested this, it
took about 1 microsecond to carry all that out. That means that 100ns is
pretty tough without some sophisticated hardware, and without very careful
wiring . The antenna to the GPS receover and the signal from the gps to the
computer can easily smear out the edge of the signal to worse than 100ns.
Often that might be repeatable, and thus the variance of the time may be at
the 100ns level but the difference in time between true GPS time and the
computr clock could be out by a microsecond or more. Of course if you knew
what that was then you could subtract it, but how are you going to know?

So, yes, it is possible but hard.

I am not sure what you mean by using chrony as a measurement tool. Its memory
is the system clock. It does not have some other internal clock. It gets a
pulse at once per sec, it needs to compare that with a clock. You remember about 10 years ago a news flash came out that they had measured
the velocity of neutrinos and discovered it was faster than light. It
eventually turned out that there was a bad connection in the fibre optics link
from the gps antenna to the system, which smeared out the rising edge of
the pulse to make it late. Thus the time difference between receipt of the
neutrino and emission was too short, implying a faster speed.
Remember that 100ns is 100 ft of travel (30m of travel). That is not much.

William G. Unruh __| Canadian Institute for|____ Tel: +1(604)822-3273
Physics&Astronomy _|___ Advanced Research _|____ Fax: +1(604)822-5324
UBC, Vancouver,BC _|_ Program in Cosmology |____ unruh@xxxxxxxxxxxxxx
Canada V6T 1Z1 ____|____ and Gravity ______|_

On Thu, 16 Feb 2023, sarveshwar k wrote:

[CAUTION: Non-UBC Email]Hi Miroslav Lichvar,
Can you please provide below information:
Is it not possible to achieve synchronization accuracy less than 100ns with chrony (with GPS and
Can we configure chrony behaving as a measurement tool (not to make any corrections)?

Thanks & Regards

On Tue, Feb 7, 2023 at 2:05 PM Miroslav Lichvar <mlichvar@xxxxxxxxxx> wrote:
      On Tue, Feb 07, 2023 at 12:04:53PM +0530, sarveshwar k wrote:
      > > The reason for chronyc sources sometimes showing 2 as LastRx, even
      > > when no samples are lost, seems to be incorrect rounding of the sample
      > > time. I can fix that.
      > Will there be a release for this fix.
      > Current version I am using is 3.5 of chrony for which yocto recipe exists.

      It will be in the next release. Backport the commit dec07aa844f8 from
      the chrony repository to 3.5 shouldn't be too difficult. It's not
      actually a bug, but rather a data-minimization feature which is not
      necessary for reference clocks.

      Miroslav Lichvar

      To unsubscribe email chrony-users-request@xxxxxxxxxxxxxxxxxxxx
      with "unsubscribe" in the subject.
      For help email chrony-users-request@xxxxxxxxxxxxxxxxxxxx
      with "help" in the subject.
      Trouble?  Email listmaster@xxxxxxxxxxxxxxxxxxxx.

Mail converted by MHonArc 2.6.19+