[ntp:questions] Meinberg NTP monitor, silly question

David J Taylor david-taylor at blueyonder.delete-this-bit.and-this-part.co.uk.invalid
Sun Dec 20 12:35:45 UTC 2009


> Device Mangler just reports the standard MS driver for that port, also
> not reflecting the actual currently in use baud rate settings either, so
> nothing new there.  (I've never yet seen that tool actualy reflect what
> the actual conditions are, on 2k or XP!)

Then the device driver isn't installed correctly.  On my Windows-7 system, 
in Device Manager selecting the COM port, Driver tab, Driver details 
button, I see three entries:

  <directory>serenum.sys
  <directory>serial.sys
  <directory>serialpps.sys

where the first two have a certificate icon against them, and the third 
not, and the details display if you click the third are obviously 
different.

> SerialPPS and the ppsapi-provider DLL are installed and pointed to in
> the appropriate way, that I can tell, and the machine has been rebooted
> several time since then, with no errors.   I will poke about and check
> again though, as there has been one Windows Update since then.

I haven't found Windows update destroying the updated serial drivers.

> If the SerialPPS driver was not installed, accessable and working, would
> not the Meinberg NTP status panel show an error?

The event log on the startup of NTP might show something, and the program 
continues in a "degraded" mode (i.e. with the user-mode timestamps).

> I'll let things run like this for a while, before messing some more.
> But for now, for Faros use, it's a *HUGE* improvement on the NTP access
> I was getting via my ISP (Demon) so, so far, so good.
>
> I'm going to start to write all this up soon, for others who also run
> Faros, and have timing issues.   I was always told the best way to learn
> something, is to have to teach others about it.  How true!
>
> Seasons Greetings etc.
>
> Dave Baxter.


What I see on PC Feenix (Windows XP) is this, trimmed:

12/12/2009,07:49:09,NTP,Information,None,3,N/A,FEENIX,Using user-mode PPS 
timestamp for GPS_NMEA(1)
12/12/2009,07:49:07,NTP,Information,None,3,N/A,FEENIX,GPS_NMEA(1) serial 
/dev/gps1 open at 4800 bps
12/12/2009,07:49:07,NTP,Information,None,3,N/A,FEENIX,proto: precision = 
1.900 usec
12/12/2009,07:49:04,NTP,Information,None,3,N/A,FEENIX,HZ 64.000 using 43 
msec timer 23.256 Hz 64 deep
12/12/2009,07:49:04,NTP,Information,None,3,N/A,FEENIX,"Windows clock 
precision 15.625 msec, min. slew 6.400 ppm/s   "
12/12/2009,07:49:04,NTP,Information,None,3,N/A,FEENIX,Clock interrupt 
period 15.625 msec
12/12/2009,07:49:04,NTP,Information,None,3,N/A,FEENIX,Performance counter 
frequency 3.580 MHz
12/12/2009,07:49:04,NTP,Information,None,3,N/A,FEENIX,"MM timer 
resolution: 1..1000000 msec, set to 1 msec   "
12/12/2009,07:49:04,NTP,Information,None,3,N/A,FEENIX,Raised to realtime 
priority class
12/12/2009,07:49:04,NTP,Information,None,3,N/A,FEENIX,ntpd 4.2.6-o Dec 09 
11:48:30.27 (UTC-00:00) 2009  (1)


.. and what I see at the start of the ntpq -p display is:

oPPS(1)          .PPS.
*GPS_NMEA(1)     .GPS.


So I am now confused, in that NTP has said it is using the user-mode and 
not the kernel-mode timestamps, and yet I have "o" tally entry.  Let me 
check device manager on that PC.....   Yes, just like the Windows-7 
system, the serialpps.sys is listed in the port drivers for COM1.  Looking 
at the jitter it looks more like kernel-mode than user-mode.

Dave Hart - any chance that 4.2.6 doesn't report using kernel-mode any 
more?

73,
David 




More information about the questions mailing list