Yb clock - stability estimation procedure?
Hello time-nuts,
The NIST paper describes the estimation of the stability of one Yb clock
by simply comparing two equivalent clocks, and dividing by sqrt(2).
This is obviously a common Metrological Practise, every time if
"something better" is not existent or not available.
This practise can be found everywhere in metrology: the comparison of Cs
clocks of all National Metrological Institutes, the comparison of two
Josephson Junctions in situ, claiming 1e-19 stability, the comparison of
the old Weston Cells, comparison of the Primary Kilograms and stating a
deviation of 1e-8, and so on.
Those comparisons and stability estimations later become fixed
definitions of the new definition of the unit, accompanied by setting
the uncertainty to the stability estimation found before.
That means, the next definition of the second, based on the Yb optical
clock would be provided by a new value and definition for the frequency
of the optical excitation, with an uncertainty of something like 1e-18,
or the Allan deviation given in the paper.
I wonder, what is the validity of this stability estimation, as the
number of the different standards is very limited, and as there's always
the probability, that two different clocks /standards may drift in the
same direction.
Also, there are always some physical effects left, which may (in alinear
manner) shift the realization of the unit, let it be the magnetic field
for a Cs clock, or an electrical filed for optical clocks.
Does anybody know, where I can find the suitable standardized
metrological regulation for that problem, i.e. under which circumstances
such a logical step from estimation to specification is valid, and the
associated statistical calculation framework?
I have naively transferred this procedure to my artefact standards, i.e.
5 Vishay precision resistor , and 4 volt references.
As those groups have very small annual drift and as I don't see a
logical difference in comparison to the stability estimations of those
quantum references, I also claim the stability of each artefact to be in
the order of the found drift within the observed group.
Now I would like to know, if I have overlooked something, and how to
make a serious stability estimation by correct metrological/statistical
calculations.
Thanks Frank
Hi
The most common approach is to assume that the two devices are not correlated. SInce it's a negative, you really can't prove it. What you can do is to disprove it by finding and documenting a correlation.
ADEV it's self has a confidence level based on the number of samples taken. What is normally reported is the calculated number, not the number plus the uncertainty. The same carries over to the square root of 2. It's simply the best estimate of what's going on. As long as they say what the do / do what they say, it's not really a problem.
The next step is typically to build a couple more devices and start doing a simultaneous N way comparison. That will let you play with math and better estimate the performance of each of your devices. The best case would be to compare devices made by different labs using different approaches. That usually lets you spot the correlation issues between devices.
Bob
On Aug 25, 2013, at 12:47 PM, Frank Stellmach frank.stellmach@freenet.de wrote:
Hello time-nuts,
The NIST paper describes the estimation of the stability of one Yb clock by simply comparing two equivalent clocks, and dividing by sqrt(2).
This is obviously a common Metrological Practise, every time if "something better" is not existent or not available.
This practise can be found everywhere in metrology: the comparison of Cs clocks of all National Metrological Institutes, the comparison of two Josephson Junctions in situ, claiming 1e-19 stability, the comparison of the old Weston Cells, comparison of the Primary Kilograms and stating a deviation of 1e-8, and so on.
Those comparisons and stability estimations later become fixed definitions of the new definition of the unit, accompanied by setting the uncertainty to the stability estimation found before.
That means, the next definition of the second, based on the Yb optical clock would be provided by a new value and definition for the frequency of the optical excitation, with an uncertainty of something like 1e-18, or the Allan deviation given in the paper.
I wonder, what is the validity of this stability estimation, as the number of the different standards is very limited, and as there's always the probability, that two different clocks /standards may drift in the same direction.
Also, there are always some physical effects left, which may (in alinear manner) shift the realization of the unit, let it be the magnetic field for a Cs clock, or an electrical filed for optical clocks.
Does anybody know, where I can find the suitable standardized metrological regulation for that problem, i.e. under which circumstances such a logical step from estimation to specification is valid, and the associated statistical calculation framework?
I have naively transferred this procedure to my artefact standards, i.e. 5 Vishay precision resistor , and 4 volt references.
As those groups have very small annual drift and as I don't see a logical difference in comparison to the stability estimations of those quantum references, I also claim the stability of each artefact to be in the order of the found drift within the observed group.
Now I would like to know, if I have overlooked something, and how to make a serious stability estimation by correct metrological/statistical calculations.
Thanks Frank
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Hi Frank,
Correct, there is the underlying assumption that common mode [frequency] drift isn't present. That's been true for Cs standards (and the definition of the SI second) as well. If you can't prove it wrong, at least you can try to measure the lower bounds of drift, if such drift does occur.
This is one reason why there have been so many clock comparisons the past decade. National labs love to compare clock type X against clock type Y to see if they can be the first to detect something strange in the universe. For example, it might be the case that hydrogen-based clocks drift at a different rate than Cs or Hg or Al or Ca or Yb, etc. Google for phrases like Fine Structure Constant Variation or Variations in Fundamental Constants. Some papers and talks to get you started:
Improved Limits on Variation of the Fine Structure Constant and Violation of Local Position Invariance
http://tf.nist.gov/general/pdf/2238.pdf
Laboratory Tests on Variations of Fundamental Constants
https://www.kvi.nl/ssp2012/material/36-peik/slides/36-1-SSP-Groningen-Peik-2.pdf
New limits on variation of the fine-structure constant using atomic dysprosium
http://arxiv.org/pdf/1304.6940v1.pdf
Searching for temporal variation of the fine-structure "constant" in radio-frequency transitions of Dy
http://www.int.washington.edu/talks/WorkShops/int_07_1/People/Budker_D/Budker.pdf
/tvb
----- Original Message -----
From: "Frank Stellmach" frank.stellmach@freenet.de
To: time-nuts@febo.com
Sent: Sunday, August 25, 2013 9:47 AM
Subject: [time-nuts] Yb clock - stability estimation procedure?
Hello time-nuts,
The NIST paper describes the estimation of the stability of one Yb clock
by simply comparing two equivalent clocks, and dividing by sqrt(2).
This is obviously a common Metrological Practise, every time if
"something better" is not existent or not available.
This practise can be found everywhere in metrology: the comparison of Cs
clocks of all National Metrological Institutes, the comparison of two
Josephson Junctions in situ, claiming 1e-19 stability, the comparison of
the old Weston Cells, comparison of the Primary Kilograms and stating a
deviation of 1e-8, and so on.
Those comparisons and stability estimations later become fixed
definitions of the new definition of the unit, accompanied by setting
the uncertainty to the stability estimation found before.
That means, the next definition of the second, based on the Yb optical
clock would be provided by a new value and definition for the frequency
of the optical excitation, with an uncertainty of something like 1e-18,
or the Allan deviation given in the paper.
I wonder, what is the validity of this stability estimation, as the
number of the different standards is very limited, and as there's always
the probability, that two different clocks /standards may drift in the
same direction.
Also, there are always some physical effects left, which may (in alinear
manner) shift the realization of the unit, let it be the magnetic field
for a Cs clock, or an electrical filed for optical clocks.
Does anybody know, where I can find the suitable standardized
metrological regulation for that problem, i.e. under which circumstances
such a logical step from estimation to specification is valid, and the
associated statistical calculation framework?
I have naively transferred this procedure to my artefact standards, i.e.
5 Vishay precision resistor , and 4 volt references.
As those groups have very small annual drift and as I don't see a
logical difference in comparison to the stability estimations of those
quantum references, I also claim the stability of each artefact to be in
the order of the found drift within the observed group.
Now I would like to know, if I have overlooked something, and how to
make a serious stability estimation by correct metrological/statistical
calculations.
Thanks Frank
On 08/25/2013 08:25 PM, Bob Camp wrote:
Hi
The most common approach is to assume that the two devices are not correlated. SInce it's a negative, you really can't prove it. What you can do is to disprove it by finding and documenting a correlation.
ADEV it's self has a confidence level based on the number of samples taken. What is normally reported is the calculated number, not the number plus the uncertainty. The same carries over to the square root of 2. It's simply the best estimate of what's going on. As long as they say what the do / do what they say, it's not really a problem.
The next step is typically to build a couple more devices and start doing a simultaneous N way comparison. That will let you play with math and better estimate the performance of each of your devices. The best case would be to compare devices made by different labs using different approaches. That usually lets you spot the correlation issues between devices.
Exactly. As you have three devices, measuring them pair-wise you get
three measures and three un-knowns, and you can untangle the stability
of each individual. If you have yet more, you can get some confidence
levels also as it becomes overdetermined.
Cheers,
Magnus
Hi
On Aug 26, 2013, at 2:01 AM, Magnus Danielson magnus@rubidium.dyndns.org wrote:
On 08/25/2013 08:25 PM, Bob Camp wrote:
Hi
The most common approach is to assume that the two devices are not correlated. SInce it's a negative, you really can't prove it. What you can do is to disprove it by finding and documenting a correlation.
ADEV it's self has a confidence level based on the number of samples taken. What is normally reported is the calculated number, not the number plus the uncertainty. The same carries over to the square root of 2. It's simply the best estimate of what's going on. As long as they say what the do / do what they say, it's not really a problem.
The next step is typically to build a couple more devices and start doing a simultaneous N way comparison. That will let you play with math and better estimate the performance of each of your devices. The best case would be to compare devices made by different labs using different approaches. That usually lets you spot the correlation issues between devices.
Exactly. As you have three devices, measuring them pair-wise you get
three measures and three un-knowns, and you can untangle the stability
of each individual. If you have yet more, you can get some confidence
levels also as it becomes overdetermined.
The only real gotcha is common mode drift. If all your gizmos are made same place / same time / same parts then they may drift the same way. In this case "drift" could correlate to environment rather than just to time (aging). Back when HP pretty much made all the Cs standards this was a common thing to worry about when setting up an ensemble of them.
Of course you could always move your ensemble to an un-inhabited cave and …..
No, HP did not make the long tube Cs standards at NIST (as the NIST guys always love to point out) and they are very different animals than the ones you can buy. So, the international definition of the second has always been safe from manufacturer induced common mode.
Bob
Cheers,
Magnus
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Hi,
On 08/26/2013 01:56 PM, Bob Camp wrote:
Hi
On Aug 26, 2013, at 2:01 AM, Magnus Danielson magnus@rubidium.dyndns.org wrote:
Exactly. As you have three devices, measuring them pair-wise you get
three measures and three un-knowns, and you can untangle the
stability of each individual. If you have yet more, you can get some
confidence levels also as it becomes overdetermined.
The only real gotcha is common mode drift. If all your gizmos are made same place / same time / same parts then they may drift the same way. In this case "drift" could correlate to environment rather than just to time (aging). Back when HP pretty much made all the Cs standards this was a common thing to worry about when setting up an ensemble of them.
Indeed. An independent source make sense, like a different atomic
reference mechanism, or as you propose, a different location.
Of course you could always move your ensemble to an un-inhabited cave and …..
No, HP did not make the long tube Cs standards at NIST (as the NIST guys always love to point out) and they are very different animals than the ones you can buy. So, the international definition of the second has always been safe from manufacturer induced common mode.
Also, it wasn't as manufacture dominant as it has been. Not all the
commercial cesiums being used have been HP, even if HP have made a big
number of their dominance.
Cheers,
Magnus
Am Mon, 26 Aug 2013 07:56:51 -0400
schrieb Bob Camp lists@rtty.us:
Hi
On Aug 26, 2013, at 2:01 AM, Magnus Danielson
magnus@rubidium.dyndns.org wrote:
On 08/25/2013 08:25 PM, Bob Camp wrote:
Hi
The most common approach is to assume that the two devices are
not correlated. SInce it's a negative, you really can't prove it.
What you can do is to disprove it by finding and documenting a
correlation.
ADEV it's self has a confidence level based on the number of
samples taken. What is normally reported is the calculated number,
not the number plus the uncertainty. The same carries over to the
square root of 2. It's simply the best estimate of what's going
on. As long as they say what the do / do what they say, it's not
really a problem.
The next step is typically to build a couple more devices and
start doing a simultaneous N way comparison. That will let you
play with math and better estimate the performance of each of your
devices. The best case would be to compare devices made by
different labs using different approaches. That usually lets you
spot the correlation issues between devices.
Exactly. As you have three devices, measuring them pair-wise you get
three measures and three un-knowns, and you can untangle the
stability of each individual. If you have yet more, you can get
some confidence levels also as it becomes overdetermined.
The only real gotcha is common mode drift. If all your gizmos are
made same place / same time / same parts then they may drift the same
way. In this case "drift" could correlate to environment rather than
just to time (aging). Back when HP pretty much made all the Cs
standards this was a common thing to worry about when setting up an
ensemble of them.
Of course you could always move your ensemble to an un-inhabited cave
and …..
No, HP did not make the long tube Cs standards at NIST (as the NIST
guys always love to point out) and they are very different animals
than the ones you can buy. So, the international definition of the
second has always been safe from manufacturer induced common mode.
On top of that there's more variety. At the level of international
definition of the second, a number of large metrology labs are
involved. Any one of these, be it NIST, NICT, PTB, or the like, has
apparently their own unique set of machines, which is far from the
stuff one can actually buy. So, at an international level, there's no
such thing as manufacturer induced common mode. Even within the
national labs involved, quite often a number of standards of different
designs can be found, making common mode drift even less likely.
At the level of individual institutions, environment induced drift
however could play a role, as they usually have their standards located
in the same labs.
Florian
On 2013-08-27 13:35, Florian Teply wrote:
No, HP did not make the long tube Cs standards at NIST (as the NIST
guys always love to point out) and they are very different animals
than the ones you can buy. So, the international definition of the
second has always been safe from manufacturer induced common mode.
I vaguely remember the NIST guys wanting to copy some of the HP
electronics from the 5071A. So don't be so sure.
Regarding common mode environmental drift: When the 5071A was
introduced, an extensive set of tests was performed over the
better part of a year to determine common mode environmental
drift. The measurements were sensitive to parts in 10^15.
No environmental sensitivity whatsoever was measurable.
This was no accident, but rather the result of careful design.
OTOH, the 5061 could change frequency noticeably by just removing
the top cover.
Whatever the merits of the big iron Cs standards, the fact of
the matter is that after the 5071A was introduced and they got
into circulation, as a group, they accounted for upwards of
80% of the weight of the TAI.
Rick Karlquist
N6RK
Hi
There was a point in time where HP made a lot of the Cs standards out there.
Bob
On Aug 26, 2013, at 5:52 PM, Magnus Danielson magnus@rubidium.dyndns.org wrote:
Hi,
On 08/26/2013 01:56 PM, Bob Camp wrote:
Hi
On Aug 26, 2013, at 2:01 AM, Magnus Danielson magnus@rubidium.dyndns.org wrote:
Exactly. As you have three devices, measuring them pair-wise you get
three measures and three un-knowns, and you can untangle the
stability of each individual. If you have yet more, you can get some
confidence levels also as it becomes overdetermined.
The only real gotcha is common mode drift. If all your gizmos are made same place / same time / same parts then they may drift the same way. In this case "drift" could correlate to environment rather than just to time (aging). Back when HP pretty much made all the Cs standards this was a common thing to worry about when setting up an ensemble of them.
Indeed. An independent source make sense, like a different atomic
reference mechanism, or as you propose, a different location.
Of course you could always move your ensemble to an un-inhabited cave and …..
No, HP did not make the long tube Cs standards at NIST (as the NIST guys always love to point out) and they are very different animals than the ones you can buy. So, the international definition of the second has always been safe from manufacturer induced common mode.
Also, it wasn't as manufacture dominant as it has been. Not all the
commercial cesiums being used have been HP, even if HP have made a big
number of their dominance.
Cheers,
Magnus
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Hi
The commonly heard comments about common mode drift very much pre-date the 5071…
Bob
On Aug 27, 2013, at 5:57 PM, Richard Karlquist richard@karlquist.com wrote:
On 2013-08-27 13:35, Florian Teply wrote:
No, HP did not make the long tube Cs standards at NIST (as the NIST
guys always love to point out) and they are very different animals
than the ones you can buy. So, the international definition of the
second has always been safe from manufacturer induced common mode.
I vaguely remember the NIST guys wanting to copy some of the HP
electronics from the 5071A. So don't be so sure.
Regarding common mode environmental drift: When the 5071A was
introduced, an extensive set of tests was performed over the
better part of a year to determine common mode environmental
drift. The measurements were sensitive to parts in 10^15.
No environmental sensitivity whatsoever was measurable.
This was no accident, but rather the result of careful design.
OTOH, the 5061 could change frequency noticeably by just removing
the top cover.
Whatever the merits of the big iron Cs standards, the fact of
the matter is that after the 5071A was introduced and they got
into circulation, as a group, they accounted for upwards of
80% of the weight of the TAI.
Rick Karlquist
N6RK
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.