Re[2]: NTPsec reports excessive jitter from GPSD/KPPS SHM

[Nick Burkitt]

Hi Gary.

I think I finally understand what the problem is. The order in which devices are specified on the gpsd command line is in fact critically important, and must be coordinated with the specification of SHM units in the ntpd configuration. 

Since the beginning of this discussion, ntpd has never actually seen both GPS and PPS at the same time, since one of them would always be SHM(2), which is not mentioned in my ntpd configuration.

With this command line:

gpsd -n /dev/ACM4 /dev/pps0

and this ntp.conf GPS/PPS configuration:

# GPS Serial data reference (NTP0)

refclock shm unit 0 refid GPS noselect

# GPS PPS reference (NTP1)

refclock shm unit 2 minpoll 3 maxpoll 3 refid PPS prefer 

I finally see ntpmon output that makes sense (to me):

     remote           refid      st t when poll reach   delay   offset   jitter
 SHM(0)          .GPS.            0 l    1   64  377      0ns -67.60ms 1.9155ms
*SHM(2)          .PPS.            0 l    1    8  377      0ns 306.33us 20.710us
-t2.time.bf1.yah 98.138.17.15     3 u    2   64  377 100.54ms 247.84us 1.7718ms
-vps5.ctyme.com  216.218.254.202  2 u    2   64  377 22.541ms -1.585ms 1.7352ms
-time.cloudflare 10.4.0.197       3 u   11   64  377 22.042ms -3.304ms 1.2201ms
+mis.wci.com     192.12.19.20     2 u   14   64  377 42.174ms -671.7us 1.4036ms
 pool.ntp.org    .POOL.          16 p    -  256    0      0ns      0ns  1.907us
-66.85.78.80     172.16.23.153    2 u   31   64  377 63.698ms -162.2us 5.8518ms
-ns1.backplanedn 173.162.192.156  2 u   30   64  377 92.035ms 6.9021ms 4.5433ms
+t2.time.gq1.yah 208.71.46.33     2 u   27   64  377 45.646ms 728.41us 1.1335ms
-static-72-87-88 128.59.0.245     2 u   28   64  363 91.383ms -2.542ms 1.2817ms
ntpd ntpsec-1.1.8+ 2020-03-04T20:55:58Z        Updated: 2020-03-09T16:22:31 (4)
 lstint avgint rstr r m v  count rport remote address
      0     61    0 . 6 2   2187 58711 localhost
      2   8342   d0 . 4 4     17   123 vps5.ctyme.com
      3   8342   d0 . 4 4     17   123 t2.time.bf1.yahoo.com
     11   7851   d0 . 4 4     18   123 time.cloudflare.com
     14   7851   d0 . 4 4     18   123 mis.wci.com
     27   7850   d0 . 4 4     18   123 t2.time.gq1.yahoo.com
     29  10265   d0 . 4 4     14   123 static-72-87-88-202.prvdri.fios.verizon.
     31   8341   d0 . 4 4     17   123 ns1.backplanedns.org
     32   7850   d0 . 4 4     18   123 66.85.78.80

-Nick

------ Original Message ------

From: "Gary E. Miller" <address@hidden>

To: address@hidden

Sent: 3/8/2020 3:43:29 PM

Subject: Re: NTPsec reports excessive jitter from GPSD/KPPS SHM

Yo Nick!

 

On Sun, 08 Mar 2020 03:33:17 +0000

"Nick Burkitt" <address@hidden> wrote:

 

Hi Gary.

 

Sorry, I'm having a little trouble keeping up.

 

remote refid st t when poll reach delay

offset jitter

SHM(0) .GPS. 0 l 8 64 377 0.0000

-65.3223 1.4824

SHM(1) .PPS. 0 l - 8 0 0.0000

0.0000 0.0019

 

SHM(0) shows a 1.4 millisecond, not microsecond, jitter.

 

Yeah, you are right. Give it some time. And remember, last you sent

an ntp.conf, did you not have noselect on SHM(0?

 

SHM(1) has a reach of 0, so no information.

 

Correct, gpsd is not sending a SHM(1) when started with:

gpsd -n /dev/pps0 /dev/ttyACMx

 

Where does SHM(2) come into the picture?

 

As I said, it is the GPS serial time. Not the order of your

command line. /dev/pps0 is first, so it shows as SHM(0). /dev/pps0

has no secondary time, so SHM(1) is unused. /dev/ttyACMx is second

on the command line, thus becomes SHM(2). It has no secondary (PP@) so

SHM(3) is unused.

 

Here's what ntpshmmon says.

 

root@MPM-4006:~# ntpshmmon -n20

ntpshmmon: version 3.20

# Name Seen@ Clock Real

L Prc

sample NTP2 1583636734.047149990 1583636733.998375333

1583636734.000000000 0 -20

sample NTP0 1583636734.080901499 1583636734.077254944

1583636733.999650744 0 -20

 

Yup. NTP2 is PGS, and NTP0 is PPS.

 

Be sure to noselect SHM(2) and prefer SHM(0)

 

RGDS

GARY

---------------------------------------------------------------------------

Gary E. Miller Rellim 109 NW Wilmington Ave., Suite E, Bend, OR 97703

address@hidden Tel:+1 541 382 8588

 

Veritas liberabit vos. -- Quid est veritas?

"If you can't measure it, you can't improve it." - Lord Kelvin