Reference clocks.

[Gary E. Miller]

Yo All!

I configured my Pi2 to peer with a local server (spidey).  Both hosts
use PPS over serial.  Both have had time to mostly converge, but these are
not as quite as good a numbers as I would expect from chronyd.

Chronyd would also show me more digits of precision, but this gets us in
the ballpark.

On the Pi2:

root at pi2:~# ntpq -p
     remote           refid      st t when poll reach   delay   offset  jitter
==============================================================================
-SHM(0)          .GPS.            0 l   25   64  377    0.000  -117.23  71.346
*SHM(1)          .SHM.            0 l   24   64  377    0.000    0.082   0.009
+catbert.rellim. .GPS1.           1 s   31   64  377    0.359    0.399   0.072
-spidey.rellim.c .GPS1.           1 s   22   64  377    0.259   -0.038   0.073
-kong.rellim.com 204.17.205.4     2 s   41   64  377    0.310    0.052   0.600
+backup.rellim.c 204.17.205.1     2 s   45   64  377    0.428    0.104   0.085

Note the 9 microSec jitter on the PPS, and only 73 microSec to spidey.
Half of what I would expect given the USB/100Base-T.

From Spidey:

spidey NTP # ntpq -p
     remote           refid      st t when poll reach   delay   offset  jitter
==============================================================================
-india.colorado. .NIST.           1 u   29   64  377   81.520  -18.372   1.920
-clepsydra.hpl.h .GPS.            1 u   25   64  377   26.502   -0.009   0.057
+catbert.rellim. .SHM.            1 s   11   64  377    0.193    0.351   0.072
-pi2.rellim.com  .SHM.            1 s   42   64  376    0.259    0.038   0.055
-kong.rellim.com 204.17.205.4     2 s   44   64  377    0.178   -0.584   0.116
+backup.rellim.c 204.17.205.1     2 s   26   64  377    0.253    0.197   0.034
-SHM(0)          .GPS.            0 l    5   64  377    0.000  150.399  11.193
*SHM(1)          .GPS1.           0 l    4   64  377    0.000    0.055   0.013
-131.107.13.100  .ACTS.           1 u   95   64  376   14.851    8.093   0.918

Here the PPS is 13 microSec jitter, after a while I should get 2.  Note
the Pi2 jitter is only 0.055 microSec.  I have seen this effect before, the
USB polling may be at 2x that, but since the USB poll is locked to the PPS on
the Pi2 the poll is stable to better than the polling rate.

The Pi2 jitter is comparable to that from catbert which is on GigE.  And even
comparable to clepsydra, 26 milliSec away.

So, for a lightly loaded network, the Pi2 ethernet is not a major problem.

RGDS
GARY
---------------------------------------------------------------------------
Gary E. Miller Rellim 109 NW Wilmington Ave., Suite E, Bend, OR 97703
	gem at rellim.com  Tel:+1 541 382 8588
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 473 bytes
Desc: OpenPGP digital signature
URL: <http://lists.ntpsec.org/pipermail/devel/attachments/20160505/daec5962/attachment.bin>