3 ms·

It's the combination of accuracy and precision that makes this remarkable. It's either the world's most accurate clock with a certain precision, or the world's

by peeters 13y ago

It's the combination of accuracy and precision that makes this remarkable. It's either the world's most accurate clock with a certain precision, or the world's most precise clock with a certain accuracy. I'm not sure which they actually claim. >They say their new clocks can keep time with an unprecedented precision of one part in 10^18. That seems to a measure of accuracy, not precision. I can make a clock that gives the time in trillions of years since the Big Bang, and it will be perfectly accurate for the distant future as long as it always returns "0". But that's not a great precision. I can also make a clock that gives the time in one 10^256th of a second, but it won't be accurate.

ISL 13y ago
This measurement states that it's possible to build two clocks that differ by less than two parts in 10^18. Accuracy: getting as close to the target as you can. Precision: hitting the same spot repeatedly. The paper is correct - these measurements hit the same spot repeatedly. Accuracy in this context is a tricky thing. The SI second is defined by the Cesium standard, but the best Cesium clocks are >100 times less precise. These two clocks are hitting the same tiny spot over and over again somewhere within the fuzzy bullseye that is the Cesium target. The SI standard is not defined as well as the new generation of clocks can measure. Once the world settles in on an optimal optical clock design (the previous leader has been an Aluminum clock, this one uses Ytterbium), the SI standard will move to the new technology. The impressive success of these clocks as a measure of gravitational redshift is also an Achilles heel. I'm not sure that anyone really knows how to do reliable time transfer with them yet, as one's depth in Earth's gravitational potential relative to another clock must be known at the centimeter scale. Another factor of 10-100 in precision will require that such metrology will improve by the same factor. It's a challenge.
- peeters 13y ago
  You're right. I was thinking that precision in a clock would be its claimed resolution. Similar to how the precision of a measuring tape would be the finest line markings on it, regardless of whether they were actually correct.
- jessriedel 13y ago
  I think "Achilles heel" is the wrong term. These clocks do not have an unusual sensitivity to something (e.g. vibrations) which could make them poor compared to some other hypothetical clock. Rather, all clocks of this accuracy must necessarily be sensitive to these gravitational effects because the gravitational field effects the passage of time itself.
  - ISL 13y ago
    Agreed - as an experimentalist specializing in precision gravitational measurements, their lives just got worlds harder. Now they're not fighting challenges within their instruments, but they have to ensure that they know each clock's depth in Earth's gravitational potential. Atomic clocks are small and elegant things with a pristine heart. They're now irreducibly sensitive to a very messy world. Earth tides (dirt, not water) are larger than centimeters. With further improved precision, they'll be sensitive to small changes in the locations of local gravitational sources, even big ones at long range. In our gravity lab in Seattle, we see the effects of annual variation in water table. All of these things must be properly characterized in order to compare two clocks that are not co-located. These are good problems to have, and are inherent to measuring time at this precision, but they're very daunting. Not insurmountable, but often frustrating.

by peeters 13y ago

ISL 13y ago

This measurement states that it's possible to build two clocks that differ by less than two parts in 10^18. Accuracy: getting as close to the target as you can. Precision: hitting the same spot repeatedly. The paper is correct - these measurements hit the same spot repeatedly. Accuracy in this context is a tricky thing. The SI second is defined by the Cesium standard, but the best Cesium clocks are >100 times less precise. These two clocks are hitting the same tiny spot over and over again somewhere within the fuzzy bullseye that is the Cesium target. The SI standard is not defined as well as the new generation of clocks can measure. Once the world settles in on an optimal optical clock design (the previous leader has been an Aluminum clock, this one uses Ytterbium), the SI standard will move to the new technology. The impressive success of these clocks as a measure of gravitational redshift is also an Achilles heel. I'm not sure that anyone really knows how to do reliable time transfer with them yet, as one's depth in Earth's gravitational potential relative to another clock must be known at the centimeter scale. Another factor of 10-100 in precision will require that such metrology will improve by the same factor. It's a challenge.

peeters 13y ago
You're right. I was thinking that precision in a clock would be its claimed resolution. Similar to how the precision of a measuring tape would be the finest line markings on it, regardless of whether they were actually correct.
jessriedel 13y ago
I think "Achilles heel" is the wrong term. These clocks do not have an unusual sensitivity to something (e.g. vibrations) which could make them poor compared to some other hypothetical clock. Rather, all clocks of this accuracy must necessarily be sensitive to these gravitational effects because the gravitational field effects the passage of time itself.
- ISL 13y ago
  Agreed - as an experimentalist specializing in precision gravitational measurements, their lives just got worlds harder. Now they're not fighting challenges within their instruments, but they have to ensure that they know each clock's depth in Earth's gravitational potential. Atomic clocks are small and elegant things with a pristine heart. They're now irreducibly sensitive to a very messy world. Earth tides (dirt, not water) are larger than centimeters. With further improved precision, they'll be sensitive to small changes in the locations of local gravitational sources, even big ones at long range. In our gravity lab in Seattle, we see the effects of annual variation in water table. All of these things must be properly characterized in order to compare two clocks that are not co-located. These are good problems to have, and are inherent to measuring time at this precision, but they're very daunting. Not insurmountable, but often frustrating.