[time-nuts] Rb Oscillator - rather fundamental question
jmfranke at cox.net
Tue Feb 23 21:21:28 UTC 2010
My two cents:
A primary standard is one that is directly compared with the accepted or
defined standard. In years past, an oscillator could only be called a
primary time or frequency standard if it had a clock in order to make
comparisons with the Earth's rotation. Note, there are no limits on the
accuracy or stability. A standard that is compared to a primary standard is
a secondary standard even if it is more stable.
Since 1967, the second has been defined to be the duration of 9,192,631,770
periods of the radiation corresponding to the transition between the two
hyperfine levels of the ground state of the caesium 133 atom.
Therefore, any standard which is compared directly with the transition is a
primary standard. Any standard compared to a primary standard is a
secondary standard. A standard more stable than a caesium standard is still
a secondary standard until the process the new standard is directly compared
to is designated the accepted or defined standard. A rubidium standard
compared to a caesium standard is still a secondary standard because it is
not directly compared the the caesium transition.
To summarize from the 1933 Bulletin 10 from General Radio on Frequency
Measurements at Radio Frequencies, "all frequency standards are compared
either directly or indirectly with a standard time interval derived from the
earth's rotation...It should be noted, however, that the accuracy depends on
how direct a comparison is made with the standard time interval."
Replace "derived from the earth's rotation" with "accepted or defined."
Direct comparison's yield primary standards and indirect yield secondary
From: "Rick Karlquist" <richard at karlquist.com>
Sent: Tuesday, February 23, 2010 3:00 PM
To: "Discussion of precise time and frequency measurement"
<time-nuts at febo.com>
Subject: Re: [time-nuts] Rb Oscillator - rather fundamental question
> WarrenS wrote:
>> All very informative and useful information for sure and good to know,
>> But I'm thinking the real difference between a primary and secondary
>> Has More to do with if there is anything else more accurate and
>> I'd guess a Rb would of made a great cave man Primary standard.
>> And sounds like it will NOT be long before the Freq and drift of a CS
>> Primary will be consider just another secondary standard that will have
>> be calibrated.
>> (to get the 1e-16 + or whatever accuracy/repeatability it is they are
>> working on.)
> Sorry but you have completely misunderstood the concept.
> It is admittedly a difficult concept to grasp; I know it took
> me a long time.
> A hydrogen maser with the wall shift servo'ed out will run rings
> around a compact Cs beam clock like the HP5062, used on submarines.
> (An interesting trivia item is that I don't believe the 5061 can
> fit through a submarine hatch). The 5062 is still a primary
> frequency standard and the hydrogen maser is still a secondary
> frequency standard.
> Regarding "drift" of primary cesium beam standards: the 5071A has
> unmeasurable drift, aging and tempco, down to a measurement limit
> of at least 1E-15. It has a typical *random* error of a few parts
> in 1E-13. The systematic error (average error of all 5071A's built)
> has been established to be below 1E-14. It will always be a
> primary standard even in the presence of longer reversible optically
> laboratory Cs beam standards of higher accuracy and better short term
> stability, or cesium fountains, etc. Even the 5061A/B is considered
> a primary standard, albeit with reduced accuracy, even though it
> has a measurable tempco. We were very proud of the E1938A crystal
> oscillator when it was able to meet the 5061 tempco spec. It
> is in no way a primary frequency standard regardless of that
> or any other accomplishment.
> Primary means that the clock will meet its spec without being
> "calibrated" against a better clock. Secondary means that
> calibration against a primary standard is necessary.
> Rick Karlquist N6RK
> time-nuts mailing list -- time-nuts at febo.com
> To unsubscribe, go to
> and follow the instructions there.
More information about the time-nuts