GPS NTP reference clock

Hej,
I do have a problem on my OSX 10.6.6 Server - I cannot successfully connect a GPS receiver for my NTP service.
What I did works perfectly on OSX 10.6.6 client, but on the server it fails with
fcntl(F_SETOWN) fails for clock I/O: Inappropriate ioctl for device
The device is a HOLUX GM-210, connected via USB (PL2303 driver from Prolific)
works fine with cu, but not with the bundled ntpd.
I looked around, but only found one with the exact same problem back in 2004, but his question was never answered.
So what might be the reason for this behavior, and how to fix it?

MrHoffman wrote:
To clarify my understanding of the current configuration: setting the 127.127.20.0 driver doesn't get accepted via Server Admin, or that reference clock driver value is accepted but no synchronization happens? (I'd tend to expect it to be accepted, but that you're not getting a lock?)
Yes, serveradmin accepts that happily, but it does not show up as a valid peer.
if you're messing with the ntp configuration files (and adding stuff), try putting your additional lines after a # line or two. (The comment characters will reportedly keep Server Admin from stepping on your changes. This is one of the differences between Mac OS X and Linux distributions.)
Thanks for this valuable tip.
What does +ntpq -p+ show?
schneekoenig:~ oskar$ ntpq -p
remote refid st t when poll reach delay offset jitter
==============================================================================
*time.euro.apple 17.72.133.54 2 u 772 68m 377 56.646 9.836 0.607
so, basically, the 127.127.20.0 does not show up.
Given the pl2303 stuff usually instantiates itself as a USB TTY device (/dev/ttyUSBxyz something or other? I run with another series of USB-to-serial widgets, so I will not see the same device names here as you'll get), and you might (will?) need to add a link to instantiate an appropriately-named a gps device for the NTP software to find. (Get a full disk backup first!) Something akin to this +ln -s /dev/whatever /dev/gps0+ command, given that the ntp daemon expects gps devices such as the /dev/gps0, gps1 or gps2 devices. Given you're using 127.127.20.0, the target device ntp will look for here is /dev/gps0, while 127.127.20.1 would seek /dev/gps1, etc. (Again, please make a full disk external backup before you go working with links in /dev given mistakes here can be deadly to OS stability. I'd suggest rebooting immediately after adding the link into /dev, too, to ensure ntp and the rest of the box reboots with happy-bits and not sad-bits.)
done all that before asking here
Once you reboot, try that +ntpq -p+ command, and see if you get non-zero values associated with your selected time base.
I'd guess you might have to play with the mode setting, too. (I'd likely try getting this to work with a one-line server command first, no mode, no other options, etc.)
I guess the problem is more likely in the way OSX Server handles the usbserial device. The exact same configuration works just fine on SnowLeopard 10.6.6 Client. Hmm, the client runs a 32 bit kernel, whereas the Server is in native 64 bit. Maybe the Prolific driver is not complete, in that fcntl(F_SETOWN) does not return the expected results in 64 bit mode. As stated before, ntpd finds the correct serial port, opens it successfully but then fails on the fcntl system call.
Jan 20 09:45:36 schneekoenig org.ntp.ntpd[25567]: refclock_setup fd 5 modem status: 0x6Jan 20 09:45:36 schneekoenig org.ntp.ntpd[25567]: refclock_ioctl: fd 5 flags 0x1
Jan 20 09:45:36 schneekoenig org.ntp.ntpd[25567]: addto_syslog: fcntl(F_SETOWN) fails for clock I/O: Inappropriate ioctl for device
Jan 20 09:45:36 schneekoenig org.ntp.ntpd[25567]: addto_syslog: configuration of 127.127.20.0 failed

Similar Messages

  • Using an external reference clock for counter measuremen​ts

    Hi all,
    I have several PXI chassis
    configured for testing, each of which uses an NI-DAQ-6221 for
    conducting period and frequency measurements among other things.
    However each of my setups give me a slightly different reading when conducting identical measurements due to
    the tolerance of the DAQ on board reference
    clocks.
    The chassis' have a 10MHz GPS calibrated reference fed into them via the reference input and this is connected to the backplane. I have tried and failed to try to route and use this as my reference
    base for my counter measurements so that each of my cards will give me
    the same reading every time. I have tried the routing method using the
    DAQ connect terminals vi but could not get this to work. I have looked through the labview examples and cant find anything thta shows me what i need to do to take frequency and period measurments using my external reference as my base clock. There are some examples showing the use of the DAQmx timing vi set to sample clock mode which allows a clock source to be selected. However these examples are only for counting digital events and when i apply the same vi to a frequency or period measurement an error is thrown saying 'invalid timing type for this channel', 'you can select: implicit, on demand'.
    I do not
    understand the working of the DAQ hardware very well so was hoping
    someone could help explain to me how to do this if it is even possible
    or give me any guidance as to how else to overcome my problem. Any vis
    showing how to use my external 10MHz reference
    as the base clock for these kind of measurements would be greatly appreciated.
    Regards,
    Dan

    Hi Dan,
    Thanks for the reply. When trying to use the connect terminals i just tried connecting the PXI10 to the onboard clock in the hope that this would configure it to use the PXI10 signal to synchronise the onboard clocks via the PLL. Doing this didn't throw any errors as i recall, it just gave me a reading of infinity as it couldn't measure the signal.
    I have actually just managed to successfully do what i wanted thanks to help from Rob at NI tech support UK. There is a DAQmx timing property node which allows you to choose a reference clock source. By using the PXI_Clk10 signal this makes the DAQ synchronise its reference clocks to this signal. I have included an image of the code i used to achieve this to take period measurements.
    Here is a very useful link which Rob sent me which may be of help to anyone else with similar problems
    http://zone.ni.com/devzone/cda/tut/p/id/3615
    Best regards,
    Dan.
    Attachments:
    ExtRef Source Image.jpg ‏105 KB

  • NI5112 PFI Reference Clock Input Problem

    I'm writing C code to use an NI5112 O-Scope card as a digitizer and I would like to synchronize it with an HP 8648C signal generator (which has the high-stability timebase option) through PFI1. I want to use the signal generator for the reference clock as the NI5112 samples about 40 Hz high at 5 MSamples/s and because I need the two to drift together. When I use the signal generator as reference clock the result is a very "noisy", i.e. with an input sine wave the FFT is noisy, and I the timing is too bad for processing input GPS signals (which have stringent timing requirements). If I use an Agilent 33120A function generator as the reference clock, I get better results, but GPS processing is still compromised. When not using a
    n external reference everything looks clean and works fine except for 40 Hz error in the sample rate. The reference outputs from both the 8648C and 33120A work fine with each other and look good and clean with the o-scope, so I can't figure what the problem is. Maybe the card is just extremely picky about external reference clocks. The function call I'm using to initialize this is:
    handleErr(niScope_ConfigureClock(vi, NISCOPE_VAL_PFI_1,
    NISCOPE_VAL_NO_SOURCE,
    NISCOPE_VAL_NO_SOURCE,
    NISCOPE_VAL_FALSE));
    The program runs normally with no errors, it's just that the results are bad. Any help would be appreciated.

    Hello,
    If the data that you are acquiring is very noisy, I would start getting suspicious about the input impedance (1MOhm or 50Ohm). Does the impedance of the signal generator match the scope's? How does your signal look in the Scope Soft fron Panel? Does it make a difference if you use the internal clock?
    Another thing that I would consider is the following parameter from the niScope_ConfigureClock function: NISCOPE_ATTR_CLOCK_SYNC_PULSE_SOURCE. In order to synchronize your arb to the scope, not only they need to share the same clock (10MHz), but also the slave receives this sync pulse from the master. In you application, if the signal generator has the ability to send a sync pulse, you can route that signal through PFI2 to ensure synchronization.
    Hope this helps. Good luck with your application!
    Annette Perez

  • NTP Status: Clock is unsynchronized

    I have been having issues with a 3750X Stack that is not synchronizing with our NTP server. I have tried adding and removing the NTP server but it does not help. We are running Version 15.2(1)E. Any suggestions would be great! Thanks!

    Daniel,
    Thanks for the quick reply.  Here is the “sh status” and “sh ntp associations”
    Sh ntp status
    Clock is unsynchronized, stratum 16, no reference clock
    nominal freq is 119.2092 Hz, actual freq is 119.2092 Hz, precision is 2**19
    ntp uptime is 26320700 (1/100 of seconds), resolution is 8403
    reference time is 00000000.00000000 (18:00:00.000 CST Thu Dec 31 1899)
    clock offset is 0.0000 msec, root delay is 0.00 msec
    root dispersion is 0.16 msec, peer dispersion is 0.00 msec
    loopfilter state is 'FSET' (Drift set from file), drift is 0.000000000 s/s
    system poll interval is 64, never updated.
    sh ntp associations
      address         ref clock       st   when   poll reach  delay  offset   disp
    ~x.x.x.x    .LOCL.           1     40     64   377  2.409 2723270 16.508
    ~x.x.x.x     x.x.x.x     2      5     64   377  2.298 2723270 16.528
    * sys.peer, # selected, + candidate, - outlyer, x falseticker, ~ configured
    We are using our internal Windows Domain Controller as an NTP source.  I have multiple 3750 stacks looking at our DCs without an issue.  Thanks!

  • How to make PXI-6733 use PXI_Clk10 as the Reference Clock?

    I'm using PXI-6733 and it is put in the PXIe-1065 Chassis. As you know, the PXIe-1065 has a clock PXI_Clk10 and distributes it to all its slots. Now I want to use the PXI_Clk10 as the reference clock for PXI-6733, meaning I want to lock the PXI-6733 onboard clock to the Chassis clock PXI_Clk10.
    The question is that I can't select the PXI_Clk10 as a Ref.Clock. It is not in the list.And I have filter the channel names. Have you ever encounter this problem? Can PXI-6733 use PXI_Clk10?

    10MHz is routed via RTSI line. Unfortunately, 6733 doesn't have one.

  • 5600 Downconverter : 10 MHz Reference CLOCK

    Hi RFSA Guys,
    Is it possible to vary the  10 MHz Reference Clock and use  it ... for demodulation of RF Signals
    to a new value ... lets say 9.9992 MHz ?
    If this is possible, upto what resolution can I vary the clock..
    Can I vary it terms of 0.1 Hz or 0.01 Hz.. Is the  10 MHz Reference Clock truly that stable ?
    I have a competitor clock which is stable upto  0.01 Hz... but not in PXI Form.
    If this is not possible, Can anyone suggest a PXI Version of a Variable Stable External Reference Clock ?
    Basically, I am looking at varying the Reference Clock to vary my Symbol rate for a Device Under Test.
    Thanks for your help..
    Dharmendra

    Dharmendra,
    You might also run into problems when the downconverter tries to phase-lock to that reference clock, since it's expecting 10 MHz. Can you tell us a bit more about what you're trying to do? There might be a different way to accomplish it. Thanks!
    Chad B. » National Instruments » ni.com

  • How to get 10 MHz reference clock out of PCI-5922

    Hi,
    I am evaluating PCI-5922 for the my application.
    In my application, PCI-5922 will get signal from agilent 33250 function generator and both instruments will be triggered at the same time with the same trigger signal.
    Is there any way to take out reference clock (10 MHz) out of PCI-5922? If that is possible, synchronizing can be easier.
    I am expecting to have answers from experts!
    Zeehoon
    Solved!
    Go to Solution.

    Hi Zeehoon,
    There are two steps to export the 10MHz reference clock out of the PCI-5922.  
    1)Specify the reference clock source (to use the internal reference select "no source")
    2)Specify the reference clock output (a list of valid terminals can be found in the PCI-5922 Routing Matrix inside the High Speed Digitizers Help)
    Here is a LabVIEW snippet showing the clock exported to PFI 1:
    -Jennifer O.
    Message Edited by Jennifer O on 09-16-2009 04:00 PM
    Attachments:
    ExportReference.JPG ‏11 KB

  • External Reference Clock on pci-5640R

    Hi. I want to clock my IF-RIO 5640R board, with an external reference clock at 2MHz frequency and i would to generate a 200 MHz clock inside the board (as the vcxo).
    If I set properly the parameters inside the "configuration timebase" vi, using the pll on the cdc7005, can i achieve this feature?
    How can i do?
    Thanks....

    If you use the fixed-personality driver, and you are feeling a little bit brave, you should be able to use ni5640R Configure Timebase.vi. It can be found (at least on my computer) at:
    C:\Program Files\National Instruments\LabVIEW 8.2\instr.lib\ni5640R\Driver\NI-5640R VIs
    Leave the defaults as they are, except change as follows:
    Ref Divider (M) = 2    (for 2 MHz)
    SMB Ext Ref Enable = True
    VCXO Control = PLL
    CP Enable = True
    That should be everything, to the best of my memory. You may have to enable the Invert CP bit.
    Hope this helps,
    Ed

  • DDR3 sys_clk & reference clock

    Hi,
    I want to use one DDR3 MIG with 800Mhz.
    Can I use internal clock as MIG sys_clk and ref_clk instead of external clock source?
    For example, the input clock is 100MHz. I use MMCM generate 200Mhz internal clock as DDR3MIG sys_clk and ref_clk?
    Thanks
     

    Hi,
    Check this thread http://forums.xilinx.com/t5/MIG-Memory-Interface-Generator/Regrading-system-clock-generated-with-no-buffer-option-for-DDR3/td-p/464398
    The reference clock can be driven internally but system clock should come from CCIO pins.
    Thanks,
    Deepika.

  • Synchronize Reference Clock error

    Hi,
    I am using 2 synchronized PCI (thet are of different models) with a RTSI cable and I am acquiring signals from both. In signal Express they work fine (and also the automatically generated code does) but in the examples such as "Analog Input Synchronizatio.vi" they do not.
    Particularly, the error I get is the -200452, regarding the RefClk.Src of DAQmx Timing but whatever I use as Reference Source Clock nothing changes.
    I am completely new to Labview: can you explain how to solve it and why I get this?
    Thanks in advance

    Hi,
    if you use two PCI, you must confiigure the Syncronization Type to "M Series (PCI)". Than you must set the RefClk.Src as OnboardClock. in this way it work fine.
    You can read several examples and the documentation in the following links:
    M Series Synchronization with LabVIEW and NI-DAQmx
    http://www.ni.com/white-paper/3615/en/
    Synchronous Analog Input Using RTSI on M Series with Counter Input Task
    https://decibel.ni.com/content/docs/DOC-24160
    Synchronize Continuous Acquisition of Multiple Analog Input Devices
    https://decibel.ni.com/content/docs/DOC-12182
    Best regards,
    A.

  • MATCHING GPS AND PC CLOCK

    Hello.
    I have a GPS Encore connected to my PC and I´m getting
    the time information of the GPS in LABVIEW. I want to match
    the PC clock to the reading I get from the GPS. How
    do I do it in LABVIEW?? Is there a command so I can set
    the PC clock time at my owne will??
    Thanks in advance
    Jony
    * Sent from RemarQ http://www.remarq.com The Internet's Discussion Network *
    The fastest and easiest way to search and participate in Usenet - Free!

    This is relatively straight forward:
    Get a Call library function (from the advanced menu).
    Right click and go to configure.
    A pannel will appear, starting at the top:
    Library: C:\windows\system\kernel32.DLL
    Function name : SetSystemTime
    Calling convention : StdCall (WINAPI)
    Below this there will be a parameter called Return - rename this ERROR.
    Set the parameter details to Numeric and Signed 32bit Integer. If the returned value is NON zero, the command was successfull.
    Click Add Parameter.
    A new parameter will appear called arg1 - rename this to SYSTEMTIME.
    Set the type to 'Adapt to type', and the value to 'Handles by Value'.
    Finally, set the function to be REENTRANT (top right listbox).
    You can now press OK.
    The Call Library function will have changed to refle
    ct the configuration.
    The SYSTEMTIME variable is a cluster of 8 signed 16bit integers. To set the time, the values should be as follows:
    Year (yyyy)
    month
    Day of week (ignored - windows decides this)
    Day
    Hours
    Minutes
    Seconds
    Milliseconds
    Happy programming
    Chris

  • 80mhz reference clock to be routed to ctr output or pxi_trigger output

    Hello Everybody,
    If possible, I would like to route 6124 internal 80Mhz clock to any counter or RTSI outs. When I check daq card device routes on NI-MAX, I'm seeing that 80MhzTimeBase can be indirectly routed. However, I can manage to route it to Ctr1Source, for example, and I can't go any further. Please help me to get 80mhz clock as is from any outputs.
    Thank you,

    You really should be serious with you shift key when it comes to units and metric prefixes.
    For the rest,  propper use of the shift key eases reading!!
    Have you tried a 1.2 MHz clock from a decent generator with a propper driver and cabling?  My guess would be that the DCM PLL can't lock because of signal degration... routing clock signals isn't just a wire! (At least on a PCB board, in LabVIEW it's a different thing )
    Greetings from Germany
    Henrik
    LV since v3.1
    “ground” is a convenient fantasy
    '˙˙˙˙uıɐƃɐ lɐıp puɐ °06 ǝuoɥd ɹnoʎ uɹnʇ ǝsɐǝld 'ʎɹɐuıƃɐɯı sı pǝlɐıp ǝʌɐɥ noʎ ɹǝqɯnu ǝɥʇ'

  • NTP - clock unsynchronized

    Hello Guys,
    Why it is showing as unsynchronized? Could you please recommend me any free NTP server? Thx.
    I have also tried with :
    0.ch.pool.ntp.org
    VOICE_1#show ntp associations
      address         ref clock       st   when   poll reach  delay  offset   disp
      192.168.0.2     192.168.0.1      2     44     64     1  5.423 1046323 7937.5
     ~134.170.185.211 10.20.229.51     2     35     64     1 173.43 52356.2 7945.3
      192.168.0.252   192.168.0.1      2     19     64     1  1.157 1046324 7937.7
      192.168.0.253   192.168.0.1      2      7     64     1  1.849 1046319 7937.8
      192.168.0.254   192.168.0.1      2     25     64     1  2.910 1046322 7937.6
     * sys.peer, # selected, + candidate, - outlyer, x falseticker, ~ configured
    VOICE_1#show ntp status
    Clock is unsynchronized, stratum 16, no reference clock
    nominal freq is 250.0000 Hz, actual freq is 249.9933 Hz, precision is 2**24
    reference time is D899CF76.87BD1EC1 (17:59:34.530 GTM+1 Thu Feb 26 2015)
    clock offset is 0.0000 msec, root delay is 0.00 msec
    root dispersion is 4.47 msec, peer dispersion is 0.00 msec
    loopfilter state is 'CTRL' (Normal Controlled Loop), drift is 0.000026765 s/s
    system poll interval is 64, last update was -739 sec ago.
    VOICE_1#sh run | s ntp
    ntp server time.windows.com

    VOICE_1#sh run | s ntp
    ntp server time.windows.com
    Never use a Windows device as an NTP source.  Microsoft's implementation of NTP is not of the IEEE standard.
    If you want an outside NTP source, make sure your firewall allows you to.   Worldwide NTP pools can be found HERE. 

  • Importing/Exporting clock reference with 5133

    Hello-
    I am using SignalExpress for some simple DAQ.  I would like to either import or export a clock reference to lock to my signal source, but have been unsuccessful.
    Signal Express does have a tab in which I can choose the PFI1 as the reference clock, however I"m greeted with the following...
    Error -1074135022
    Possible reason(s):
    Attribute or property not supported.
    Device: Dev2
    Property: Reference Clock Source
    Status Code: -200197
    The 5133 manual suggest that importing/exporting the clock is in fact an option.
    Is it just not an option through Signal Express?  Do I need to make my own VI with the DAQ-mx VI's?  Maybe the reference I'm providing it is incorrect and it knows?
    Brandon

    After some more investigation...it appears that you cannot export the reference or sample clocks OUT on the PFI line of the 5133.
    However...it does appear that you can IMPORT a reference clock.
    As mentioned above...SignalExpress doesn't seem very happy when you try to do this.
    I found a program in this thread.... http://forums.ni.com/ni/board/message?board.id=250&message.id=41035&query.id=2207663#M41035
    ...where the author makes their own scope using the NI-SCOPE VI's.
    I modified the program using the ConfigureClock.VI to import a reference...however only with the same error message of this task not being supported by this device.
    What am I missing? 

  • Does Phone NTP Server reference change my licensed files, (CUCM)?

    Hi all,
    Good day.
    I just wanted to know of on my CUCM Admin web page, on PHONE NTP REFERENCE, if adding or deleting NTP's does make my CUCM to re-host my active licensed files.
    Because I changed the NTP server on the CLI call manager, (delete it and added one). And now all my licenses are invalid
    I have a few more questions.
    Which are the differences between of the Phone NTP Reference Server and NTP reference servers (CLI)?
    Which one I do not must change?
    If mi licenses say "Invalid", what is the worst case scenario could happen with my current configurations of IP Phones/GW/GK/Trunks/CSS?
    Thanks in advice.
    John

    I assume you are using a UCS server,,,
    The license MAC is generated by hashing the following user-configured parameters:
    ◊ Time zone
    ◊ NTP server 1 (or ?none?)
    ◊ NIC speed (or ?auto?)
    ◊ Hostname
    ◊ IP Address (or ?dhcp?)
    ◊ IP Mask (or ?dhcp?)
    ◊ Gateway Address (or ?dhcp?)
    ◊ Primary DNS (or ?dhcp?)
    ◊ SMTP server (or ?none?)
    ◊ Certificate Information (Organization, Unit, Location, State, Country)
    When you change any of the parameters that create the license MAC,  the license that you obtained with it becomes invalid. You must request a  rehosting of the license to obtain a valid license. The old license  continues to work for a 30-day grace period.
    To rehost your licenses, you must open a case with the licensing team. Contact the licensing team at [email protected]
    During  the grace period, you can change the settings back to the licensed  values to make your original license valid again. If you need more than  30 days of grace period, change your settings back to the licensed  values, then change them back to the new values that you want to use.  You will get another 30- day grace period.
    Please rate all useful posts
    "opportunity is a haughty goddess who waste no time with those who are unprepared"

Maybe you are looking for