[time-nuts] Re: TimeProvider 2700 PTP UTC offset issue.
Andrew Back
andrew at carrierdetect.com
Mon Jun 27 13:02:44 UTC 2022
Hi Markus,
I've pasted output from some pmc GET commands below, along with show
clock on the GM. Seems like the grandmaster is getting 37 from GPS,
while announcing 35?
Also not sure why TIME_PROPERTIES_DATA_SET has currentUtcOffsetValid and
timeTraceable both set to true, but GRANDMASTER_SETTINGS_NP has these
set to false.
Best,
Andrew
//
andrew at bbb:~$ sudo pmc -u -b 0 'get TIME_PROPERTIES_DATA_SET'
sending: GET TIME_PROPERTIES_DATA_SET
   94e36d.fffe.d2695e-0 seq 0 RESPONSE MANAGEMENT
TIME_PROPERTIES_DATA_SET
      currentUtcOffset     35
      leap61               0
      leap59               0
      currentUtcOffsetValid 1
      ptpTimescale         1
      timeTraceable        1
      frequencyTraceable   1
      timeSource           0x20
andrew at bbb:~$ sudo pmc -u -b 0 'get PARENT_DATA_SET'
sending: GET PARENT_DATA_SET
   94e36d.fffe.d2695e-0 seq 0 RESPONSE MANAGEMENT PARENT_DATA_SET
      parentPortIdentity                   00b0ae.fffe.0357ef-2
      parentStats                          0
      observedParentOffsetScaledLogVariance 0xffff
      observedParentClockPhaseChangeRate   0x7fffffff
      grandmasterPriority1                 128
      gm.ClockClass                        6
      gm.ClockAccuracy                     0x21
      gm.OffsetScaledLogVariance           0x6400
      grandmasterPriority2                 128
      grandmasterIdentity                  00b0ae.fffe.0357ef
andrew at bbb:~$ sudo pmc -u -b 0 'GET GRANDMASTER_SETTINGS_NP'
sending: GET GRANDMASTER_SETTINGS_NP
   94e36d.fffe.d2695e-0 seq 0 RESPONSE MANAGEMENT GRANDMASTER_SETTINGS_NP
      clockClass             255
      clockAccuracy          0xfe
      offsetScaledLogVariance 0xffff
      currentUtcOffset       35
      leap61                 0
      leap59                 0
      currentUtcOffsetValid  0
      ptpTimescale           1
      timeTraceable          0
      frequencyTraceable     0
      timeSource             0xa0
tp2700> show clock
System time    : 2022-06-27 12:43:58
Leap Seconds   : 37
Leap pending   : +0
On 27/06/2022 12:22, Markus Kleinhenz via time-nuts wrote:
> Hi Andrew,
>
> have you checked the PTP-Messages themselves? There is a field
> currentUtcOffset in the PTP announce message used to transfer leapsecond
> info. If thats 35, then it seems like a firmware issue.
>
> Regards
> Markus
>
> Am 25.06.2022 um 10:30 schrieb Andrew Back via time-nuts:
>> Hoping that someone may be able to shed some light on a PTP issue I'm experiencing, which could well be a simple configuration issue or, as I suspect, somehow related to the grandmaster firmware.
>>
>> I have a Symmetricom TimeProvider 2700, with firmware which I think dates from ~2014, connected to a BeagleBoneBlack running Debian and Linux PTP. A direct Ethernet connection with no switches in between. The TP2700 has the 1588 Annex J Default profile active, with all default configuration. Linux PTP client is configured for utc_offset 37. The GM is referenced to GPS and "show clock" returns "Leap Seconds : 37". However, when ptp4l is started on the client I get the running in a temporal vortex message and "updating UTC offset to 35", despite not seeing an offset of 35 configured anywhere.
>>
>> I did also run "pmc -u -b 0 'get TIME_PROPERTIES_DATA_SET'" on the client, which similarly confirmed a UTC offset of 35.
>>
>> I don't seem to be able to manually configure UTC offset on the TP2700 and when I try, it complains that this is not possible with the current mode, since it's referenced to GPS I guess.
>>
>> Other than this perhaps being a grandmaster firmware issue, I'm out of ideas.
>>
>> Andrew
>> _______________________________________________
>> time-nuts mailing list -- time-nuts at lists.febo.com
>> To unsubscribe send an email to time-nuts-leave at lists.febo.com
>>
> _______________________________________________
> time-nuts mailing list -- time-nuts at lists.febo.com
> To unsubscribe send an email to time-nuts-leave at lists.febo.com
More information about the Time-nuts_lists.febo.com
mailing list