[time-nuts] Imagery position accuracy

[Miguel Barbosa Gonçalves]

On 07/05/2013, at 01:20, Chris Albertson <albertson.chris at gmail.com> wrote:

> On Mon, May 6, 2013 at 3:38 PM, Miguel Barbosa Gonçalves <m at mbg.pt> wrote:
> ..
>>
>> You'll have to make the driver talk to the configuration / monitoring port
>> (port B) as the timing port (port A) does not accept commands.
>>
>
> So this does not apply to the Thunderbolt?  It only has one port.  If
> it worked on all PALISADE type GPSes I'd do it.
>>
>> ....
>> With this configuration I can stop using the PALISADE driver and configure
>> the Acutime to send NMEA sentences and use the PPS for better accuracy.
>
> NEMA is only used to number the seconds.  It is the PPS that does the
> timing.  You don't gain accuracy with NEMA.

Chris,

I know NMEA is *usually* only used to number seconds. Usually, the
manufacturers don't care with the accuracy of the NMEA sentences but
in theory if they were output precisely they could be used with NTP.
For I time nut I believe that would not be enough.

But did you actually read what I wrote? ;)

I said that with my configuration I can use the PALISADE or NMEA and
PPS. Actually I am using PPS on both as the PALISADE only gives me too
much jitter on my serial port.

After several hours running some interesting data:

acutime# uptime; ntpq -p; ntptime
 6:32AM  up  6:20, 1 user, load averages: 0.03, 0.04, 0.01
     remote           refid      st t when poll reach   delay   offset  jitter
==============================================================================
-pfsense.localdo 193.62.22.90     2 u   27   64  377    1.342    0.089   0.259
*GPS_PALISADE(0) .GPS.            0 l   13   16  377    0.000    1.057   0.053
oPPS(0)          .PPS.            0 l   12   16  377    0.000    0.009   0.004
+ntp02.oal.ul.pt 194.117.9.138    2 u   23   64  377   10.456    0.584   0.312
-ntp04.oal.ul.pt 194.117.9.137    2 u    5   64  377   10.545   -1.559   6.550
+gatekeeper1.ist 193.204.114.232  2 u   22   64  377    9.480    0.673   0.228
-gatekeeper2.ist 192.93.2.20      2 u   61   64  377   11.570   -0.353   2.329
ntp_gettime() returns code 0 (OK)
  time d5331e7c.79288cbc  Tue, May  7 2013  6:32:28.473, (.473275743),
  maximum error 7000 us, estimated error 1 us, TAI offset 0
ntp_adjtime() returns code 0 (OK)
  modes 0x0 (),
  offset 9.366 us, frequency 144.542 ppm, interval 256 s,
  maximum error 7000 us, estimated error 1 us,
  status 0x2107 (PLL,PPSFREQ,PPSTIME,PPSSIGNAL,NANO),
  time constant 4, precision 0.001 us, tolerance 496 ppm,
  pps frequency 144.542 ppm, stability 0.004 ppm, jitter 2.774 us,
  intervals 111, jitter exceeded 535, stability exceeded 1, errors 1.

I still need to calibrate the PPS in order to account for the 200 feet
(60 m) of cable I have between the antenna and the server.

Cheers,
Miguel

> And interresting trick is to put two or more GPS units on the same
> computer.  then use NTP to compare.  You can test out what works best
> and measure the differences
>

That is a good idea. I find that in my LAN I can get two stratum 1
servers (with different brands of receivers) 5 us of each other
consistently. And for NTP I believe that is good enough.

Cheers,
Miguel