[time-nuts] LPRO-GPSDO

Magnus Danielson magnus at rubidium.dyndns.org
Sat Jan 9 00:18:03 UTC 2010


Robert,

Robert Atkinson wrote:
> Hi Bob,
> Indeed! I must admit I thought at first that the LPRO was one of unit with built-in discipling to an external 1PPS. If this guy was selling in the UK he would be breaking advertising rules with his claims. I can see no validity for his claim of NIST traceability. NIST specifically say that the GPS signal (let alone the output of a GPSDO) is only traceable after the fact by analysing the monitoring data.

Which is the correct answer within the context of traceability, such as 
ISO 17025 (previously ISO 25). Many confuse the issue of adjustment with 
that of calibration and traceability. Adjustment may be used in a 
calibration procedure, but is not necessary for traceability. The 
calibration procedure however involves measuring and recording the 
deviations in accordance to some strict procedure, and this is done with 
reference to some reference standard, calculate the tolerances and 
uncertainties of the measure and also include the tolerance and 
uncertainty of the reference standard. Traceability is achieved by the 
combined records of all the transfer standards to the national standard 
which in itself should be traceable to the international standard.

You can have a clock being all wrong, but with propper traceability and 
no adjustment, you could make readings within certain limits traceable 
to the national standard. You can adjust all you want on a clock, but 
have no evidence of its current or previous reading.

Disciplining does not give traceability, but you can get adjustments and 
for some applications sufficient precission. Traceability is thus a 
heavy word to swing.

Disciplining an LRPO to GPS seems like a good idea thought. Can get you 
very nice hold-over properties.

Cheers,
Magnus



More information about the time-nuts mailing list