[time-nuts] Changes in commercial GPS clocks over the decades

[The Fiber Guru]

This is a good example of the difference between metrology and operational terminology.  Yes, not traceable in the strict sense of a metrologist, but in this instance, “traceability” implies the reported Stratum level of the delivered clock signal.  For example, if I receive a clock signal with label of PRS, it is believed to originate from a clock that is using GPS to discipline the oscillator.  While that is no guarantee of its stability, it is at least an indicator to use when decisioning where the network element will derive synchronization.

When visiting NIST, they always tend to treat any clock source as inherently unstable.  This can be concerning until you understand they are thinking it terms of 23 places or more behind the decimal point, where as a network operators we are happy with magnitudes less than that.  All a degree of perspective.

db

> On Nov 1, 2020, at 9:57 AM, Magnus Danielson <magnus at rubidium.se> wrote:
> 
> Hi,
> 
>> On 2020-10-31 17:29, The Fiber Guru wrote:
>> Yes, prior to use of GPS to discipline the Local oscillator, telecom timing was a “trickle down” topology where a cesium source in Kansas City was distributed across the continent.
> In Europe for instance, we chose other towns than Kansas City, but ah well.
>> The cesium was the gold standard and as the timing signals cascaded to distant geographical regions, it was obviously less pure, so if cesium was at Stratum 1 (stability), the next element in line that mediated time of downstream was at Stratum 2.  As it arrived at your local telco central office, it was at Stratum 4 (ok, but useless in today sadly networks).
>> 
>> Enter GPS and instantly every local central office could achieve Stratum 1 traceability, and if GPS was lost the best Rubidium's could holdover at Stratum 2e (slightly better that Stratum 2).  If the clock had OCXO, holdover was Stratum 3 or 3e depending on the quality of the oscillator.
> 
> No, that's not traceability. Use of that term for meaning locked to is
> strongly discouraged. Traceability is about unbroken chain of
> calibrations to SI units and the paper trail that goes with it. You do
> not achieve that here.
> 
> Stratum 1, 2, 3 and 4 is ANSI T1.101 terminology, but not used in
> international standards. For instance, there is no real equivalence of
> Stratum 4 in the international standards, because it makes no real sense
> to synchronize to the line card oscillator.
> 
>> Stratum levels are reported to most network elements by embedding the Stratum value message in the Facility Data Link of an ESF T1 signal.  The network element would read the Stratum level from the incoming timing signal to determine if it should lock to the signal, or fall back to its internal clock, usually an ocxo at Stratum 3.
> 
> This is the SSM code.
> 
> The routing protocol was limited to source quality. It created great
> pains since routing-wise it is far from sufficient. It required separate
> monitoring and reconfiguration tool or very strict planning to be
> "safe". Good description for routing is found in ITU-T G.781. ETSI has a
> good overview/teaching document for which the ref number now slipped my
> mind.
> 
>> 
>> If the master GPS clock suffered loss of GPS and the backup oscillator deteriorated to a low stability, t