How Would We Measure a Second Without Clocks?

[CAH]

hey!
What if we got rid of all the clocks around the universe, how would be remesure a second?

 

[nuuskur]

It was something about a Cesium atom flashing 9 billion something times and that period is defined to be a second.

 

[jbriggs444]

The definition of a second is: "The second is 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 cesium 133 atom."

That is the definition because it is what an atomic clock measures. So the answer to the question is: You would measure a second by building a new atomic clock.

 

[Buckleymanor]

The definition of a second is: "The second is 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 cesium 133 atom."

That is the definition because it is what an atomic clock measures. So the answer to the question is: You would measure a second by building a new atomic clock.

Eventualy, first you would have to decide what to base your measurement on.So back to basic principles a year the time it takes for the Earth to travel once around the Sun then a month, a week followed by a day (one Earth rotation)..Then divide your your day up into usefull portions let's say 24 and then call them hours divide them into minutes and the minutes into seconds.You then find that fluctuations in the Earth's rotation make the measurement unreliable for certain applications so you then build a new atomic clock based as close as possible on one of Earth's rotation and it's sub-divisions.

 

[Blackberg]

Sundials. Hourglasses. Stargazing. Divisions by 24, and 60. You'd get there.

 

[Nugatory]

We may be risking confusing people by bringing the movement of the heavenly bodies into this discussion. There was a time long ago when we derived our units of time from the Earth's rotation, the sun rising and setting, the seasons coming and going, and the Earth circling the sun, but those days are long gone. The modern definition of the second is, as jbriggs444 says above, "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 cesium 133 atom." That definition works anywhere that you can find a cesium atom; we could communicate it to an alien civilization in a distant galaxy that has never heard of the Earth and knows nothing of its orbital and rotational periods.

 

[Drakkith]

hey!
What if we got rid of all the clocks around the universe, how would be remesure a second?

Without some kind of clock, we wouldn't be able to measure a second. This doesn't change how a second is defined, as explained above, but without some device to refer to we couldn't accurately measure a second. (Unless you consider going "one one-thousand, two one-thousand..." accurate enough)

 

[Buckleymanor]

We may be risking confusing people by bringing the movement of the heavenly bodies into this discussion. There was a time long ago when we derived our units of time from the Earth's rotation, the sun rising and setting, the seasons coming and going, and the Earth circling the sun, but those days are long gone. The modern definition of the second is, as jbriggs444 says above, "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 cesium 133 atom." That definition works anywhere that you can find a cesium atom; we could communicate it to an alien civilization in a distant galaxy that has never heard of the Earth and knows nothing of its orbital and rotational periods.

Totally disagree, 9 192 631 770 might well the period of transition between two hyperfine levels of the ground state of the cesium 133 atom.That in itself means nothing though without putting it into context . So you communicate to an alien civilisation that 1 second is equal to 9 192 631 770 vibrations of a cesium atom and they reply why!

 

[sophiecentaur]

Totally disagree, 9 192 631 770 might well the period of transition between two hyperfine levels of the ground state of the cesium 133 atom.That in itself means nothing though without putting it into context . So you communicate to an alien civilisation that 1 second is equal to 9 192 631 770 vibrations of a cesium atom and they reply why!

Not sure I agree with looking at it that way. It would be more likely to communicate 'time' in any message we were to send in terms of cycles of a particular atomic transition. There would be little point in introducing the second into the message. It would be like talking about mass and length on this forum using pounds and feet.

 

[Buckleymanor]

Not sure I agree with looking at it that way. It would be more likely to communicate 'time' in any message we were to send in terms of cycles of a particular atomic transition. There would be little point in introducing the second into the message. It would be like talking about mass and length on this forum using pounds and feet.

Can see your point but in the beginning there was pounds and feet it wasn't till metres and kilo's became a more usefull tool that they were adopted.You could argue that if it were not for pounds and feet there would be no metres and kilo's and then again there might have been kilos and metres without the invention of pounds and feet .However it's all part of the rich tapestry of the history of physics and I suppose we would be all be that much poorer if we ignored it and came to the conclusion that it is what it is because it is.

 

[Nugatory]

Totally disagree, 9 192 631 770 might well the period of transition between two hyperfine levels of the ground state of the cesium 133 atom.That in itself means nothing though without putting it into context . So you communicate to an alien civilisation that 1 second is equal to 9 192 631 770 vibrations of a cesium atom and they reply why!

We wouldn't even mention the second or that 9172631770 number in the conversation. They'd ask "What's your unit of time?" and we'd reply "cesium atoms, one cycle is one unit", and we'd be done (almost - there are other issues to work through, but they're a digression here).

Original poster's question, however, is a bit different. He was asking how we would reestablish the second if every existing clock were to be destroyed or otherwise gotten rid of. Jbriggs444 answered that question above: we'd find a cesium atom and start counting cycles.

After hearing that answer, the original poster might reasonably ask "but why 9192631770?". The answer is that that value was chosen to minimize the amount of retooling that would be required. A wristwatch that sometimes ticks after 9192620000 cycles of a cesium atom and other times after 9192640000 cycles (which is pretty good accuracy for most practical purposes) didn't suddenly become obsolete when we adopted the cesium standard.

 

[sophiecentaur]

how we would reestablish the second if every existing clock were to be destroyed or otherwise gotten rid of.

Or, of course, if we happened to be on a distant planet who's rotation period we happened not to know and needed to start from scratch, with no basic equipment - just some ol' books.

 

[Buckleymanor]

We wouldn't even mention the second or that 9172631770 number in the conversation. They'd ask "What's your unit of time?" and we'd reply "cesium atoms, one cycle is one unit", and we'd be done (almost - there are other issues to work through, but they're a digression here).

Original poster's question, however, is a bit different. He was asking how we would reestablish the second if every existing clock were to be destroyed or otherwise gotten rid of. Jbriggs444 answered that question above: we'd find a cesium atom and start counting cycles.

After hearing that answer, the original poster might reasonably ask "but why 9192631770?". The answer is that that value was chosen to minimize the amount of retooling that would be required. A wristwatch that sometimes ticks after 9192620000 cycles of a cesium atom and other times after 9192640000 cycles (which is pretty good accuracy for most practical purposes) didn't suddenly become obsolete when we adopted the cesium standard.

That's all very good but why should you trust the 9192631770 cycles of a cesium atom surely you have only arrived at that standard from the breakdown of a day or a wristwatch that only ticks after 9192620000 cycles.So if you were starting from scratch you would have to check it.

 

[sophiecentaur]

That's all very good but why should you trust the 9192631770 cycles of a cesium atom surely you have only arrived at that standard from the breakdown of a day or a wristwatch that only ticks after 9192620000 cycles.So if you were starting from scratch you would have to check it.

It's not a matter of "trusting", surely. The only trust involving is to rely on your friend, the other end of the line, having the same conditions as you do for his reference atomic oscillator. That will give you both the same standard 'second'. The accuracy that you can work to is potentially far greater than just the ticks of your one second clock mechanism. You can measure tiny fractions of the tick interval and get the phase slope and, hence the frequency accuracy can be as good as your measurement time will allow. To get the best reliability, you would need to do the usual thing of averaging over a long period of time

 

[jbriggs444]

That's all very good but why should you trust the 9192631770 cycles of a cesium atom surely you have only arrived at that standard from the breakdown of a day or a wristwatch that only ticks after 9192620000 cycles.So if you were starting from scratch you would have to check it.

Check it? The second is 9192631770 cycles of the radiation corresponding to the hyperfine transition of the ground state of the cesium atom by definition. You don't check the second against the other (non-existent or broken by hypothesis) clocks. You check those other non-standard clocks against the standard definition.

If you were starting from scratch you could define a new unit of time, call it the "fuzzletynobbit" and say "to hell with the second". But that was not the challenge posed in post #1 above.

 

[dextercioby]

The question in the original post is not trivial, but rather interesting if reformulated: if one were to suddenly 'lose' all clocks in the universe, thus the ability to measure time at microseconds to years scale, how would one measure the energy of the photon released when the cesium atom undergoes a hyperfine transition for its chemical-bond electron?

 

[sophiecentaur]

how would one measure the energy of the photon released

That's not how it's done- i.e. you don't just look at individual photons. You use a cell containing the gas and build an oscillator that interacts with a vast number of atoms in the cell at the appropriate transition frequency. It's effectively a resonance you are using (as with an atomic absorption line). By keeping the cell under the right conditions of temperature, pressure etc. the error in finding the frequency of this resonance can be reduced to the level that's acceptable for that standard.
I don't know the details of building a frequency standard or why particular elements have been used over the years but you could google something like 'frequency standard technology' and find some article to suit your level of understanding. Needless to say, you will find some articles that will totally flummox you (and me!) but there will be some that make sense to you.

 

[Buckleymanor]

Check it? The second is 9192631770 cycles of the radiation corresponding to the hyperfine transition of the ground state of the cesium atom by definition. You don't check the second against the other (non-existent or broken by hypothesis) clocks. You check those other non-standard clocks against the standard definition.

If you were starting from scratch you could define a new unit of time, call it the "fuzzletynobbit" and say "to hell with the second". But that was not the challenge posed in post #1 above.

As Sophie points out you have to average out the time period over a long period of time.The longer the more accurate the definition of the second until other influences make the reading less accurate.All this averageing though is tied indirectly to your broken hypothetical clocks.If it was not then your fuzzletynobbit might well end up having 1234567890 cycles and why not! Because your clock alarm would be waking you to go to work in the middle of the night when you are not a shift worker.Time and the second have been intricately linked to the rotation of the planet for good reasons.Inventing a new clock whatever it timeing mechanism has to improve and keep those links intact.

 

[Buckleymanor]

It's not a matter of "trusting", surely. The only trust involving is to rely on your friend, the other end of the line, having the same conditions as you do for his reference atomic oscillator. That will give you both the same standard 'second'. The accuracy that you can work to is potentially far greater than just the ticks of your one second clock mechanism. You can measure tiny fractions of the tick interval and get the phase slope and, hence the frequency accuracy can be as good as your measurement time will allow. To get the best reliability, you would need to do the usual thing of averaging over a long period of time

Trust is probably the wrong word.Reliant would be better.
No doubt about the accuracy of the cesium clock it's the reliability of it to wake me when I wantYou can't take these things for granted.

 

[jbriggs444]

As Sophie points out you have to average out the time period over a long period of time.The longer the more accurate the definition of the second until other influences make the reading less accurate.All this averageing though is tied indirectly to your broken hypothetical clocks.If it was not then your fuzzletynobbit might well end up having 1234567890 cycles and why not! Because your clock alarm would be waking you to go to work in the middle of the night when you are not a shift worker.Time and the second have been intricately linked to the rotation of the planet for good reasons.Inventing a new clock whatever it timeing mechanism has to improve and keep those links intact.

What are you talking about? An atomic clock is plenty accurate enough not to set off a shift worker's alarm in the middle of the night.

 

[sophiecentaur]

What are you talking about? An atomic clock is plenty accurate enough not to set off a shift worker's alarm in the middle of the night.

It's not the atomic standard that will give you a problem. It will be the way the 'diurnal time system' on that particular planet has been set up (referenced to the atomic standard) that can cause problems. They would still need to get the 'day' the right number of seconds long or clock midday and solar noon would steadily march apart. The shift worker's bedside clock (and all others) would need to be calibrated, not only to the second but to the rotation of the planet - as do our Earthbound clocks.
Highly accurate time is hardly relevant to everyday living but the coarse measurement and synchronisation of the days, over the centuries is very important. Good old GMT would have been and will always be fine for practical purposes.

 

[Buckleymanor]

What are you talking about? An atomic clock is plenty accurate enough not to set off a shift worker's alarm in the middle of the night.

Is it not clear I will try again.Why is the atomic clock plenty accurate enough?
The atomic clock is only as accurate as the callibration of it's cycle to the Earth's rotation.9192631770 cycles is not the number used because it looks nice.The reason why it's chosen is because it is the number calculated, that makes it as certain as possible, when measured against the length of day, that the shift worker's alarm does not go off in the middle of the night.That is until we find a more accurate way of measuring the second which is tied more closely to the Earth's rotation.

 

[DrGreg]

That is until we find a more accurate way of measuring the second which is tied more closely to the Earth's rotation.

So long as we continue to define the second in terms of caesium, that isn't going to happen. The caesium definition is already slightly out-of-step with the rotation of the Earth, which is why we have leap seconds.

 

[sophiecentaur]

So long as we continue to define the second in terms of caesium, that isn't going to happen. The caesium definition is already slightly out-of-step with the rotation of the Earth, which is why we have leap seconds.

The fact is that the Earth's rotation and orbit time is not constant so as a fraction of a year, the second is not a fixed quantity. It's not the Caesium that's out of step with the Earth - it's the other way round. The vagaries of the movements of third planet from the Sun are of no interest to Scientists on other planets. They may, however, be interested in our measurements of the rates of some processes we have observed, relative to the Caesium clock frequency. They can repeat our experiments based on it.

 

[Nugatory]

The atomic clock is only as accurate as the calibration of its cycle to the Earth's rotation. 9192631770 cycles is not the number used because it looks nice.The reason why it's chosen is because it is the number calculated,]that makes it as certain as possible, when measured against the length of day, that the shift worker's alarm does not go off in the middle of the night. That is until we find a more accurate way of measuring the second which is tied more closely to the Earth's rotation.

That is simply incorrect. 9192631770 was chosen to minimize the amount of retooling and replacement of existing timepieces that would be necessary. The variation in the Earth's rotation is much greater than the one part in 10¹⁴ accuracy of the cesium clock, so it has never been and never will be possible to define a unit of time based on the Earth's rotation that will be as precise (in the sense of "your second" measuring the same amount of time as "my second") as the cesium clock definition. The 9192631770-cycle second falls comfortably in the middle of the range of variation between the seconds counted by existing timepieces, so by choosing that number people could continue to use the clocks they already had; the errors relative to the cesium clocks would be no greater than the errors relative to any other clock so would create no new problems. Back in the days of mechanical watches, two family members might find that their watches drifted apart by 60 seconds across a day. The cesium clock standard would allow them to conclude that one watch was, say, 35 seconds fast while the other was 25 seconds slow - but the watches were no less useful than before).

Part of the problem here is that you are confusing how we label points in time ("I want my alarm clock to go off at 8:00 AM every morning") and how we measure the passage of time ("How much time has passed between this event and that event?"). The label "8:00 AM 6 April 2015" is attached to a particular time that has a particular significance in terms of the Earth's diurnal rotation, as is the label "8:00 AM 7 April 2015". However there is no time interval, specified to one part in 10¹⁴, such that two successive "8:00 AM" times are separated by exactly 24*3600 of those intervals - that's why we have leap seconds. If you had a clock that ticked exactly 86400 times between "8:00 AM 6 April 2015" and "8:00 AM 7 April 2015" and never reset it, you would find that it falls increasingly out of step with the Earth's rotation.

 

[Buckleymanor]

That is simply incorrect. 9192631770 was chosen to minimize the amount of retooling and replacement of existing timepieces that would be necessary. The variation in the Earth's rotation is much greater than the one part in 10¹⁴ accuracy of the cesium clock, so it has never been and never will be possible to define a unit of time based on the Earth's rotation that will be as precise (in the sense of "your second" measuring the same amount of time as "my second") as the cesium clock definition. The 9192631770-cycle second falls comfortably in the middle of the range of variation between the seconds counted by existing timepieces, so by choosing that number people could continue to use the clocks they already had; the errors relative to the cesium clocks would be no greater than the errors relative to any other clock so would create no new problems. Back in the days of mechanical watches, two family members might find that their watches drifted apart by 60 seconds across a day. The cesium clock standard would allow them to conclude that one watch was, say, 35 seconds fast while the other was 25 seconds slow - but the watches were no less useful than before).

Part of the problem here is that you are confusing how we label points in time ("I want my alarm clock to go off at 8:00 AM every morning") and how we measure the passage of time ("How much time has passed between this event and that event?"). The label "8:00 AM 6 April 2015" is attached to a particular time that has a particular significance in terms of the Earth's diurnal rotation, as is the label "8:00 AM 7 April 2015". However there is no time interval, specified to one part in 10¹⁴, such that two successive "8:00 AM" times are separated by exactly 24*3600 of those intervals - that's why we have leap seconds. If you had a clock that ticked exactly 86400 times between "8:00 AM 6 April 2015" and "8:00 AM 7 April 2015" and never reset it, you would find that it falls increasingly out of step with the Earth's rotation.

As certain as possible was my reply adding a leap second or removeing one falls into that catergory.The point is and remains that a leap second here or there is not tantamount to a diversion away from linking atomic time to the Earth's rotation.It is necesary and will remain so.Even if the leap second was abolished there are methods to maintain the link with Earth's rotation.quote (Instead of inserting a leap second at the end of the day, Google servers implement a leap smear, extending seconds slightly over a time window prior to the leap second).So main institutions still find that real time applications need to maintain the link.Cesium might run like clockwork but the Earth rotation does not, you can't change one but you have to change the other for pratical purposes.

 

[sophiecentaur]

As certain as possible was my reply adding a leap second or removeing one falls into that catergory.The point is and remains that a leap second here or there is not tantamount to a diversion away from linking atomic time to the Earth's rotation.It is necesary and will remain so.Even if the leap second was abolished there are methods to maintain the link with Earth's rotation.quote (Instead of inserting a leap second at the end of the day, Google servers implement a leap smear, extending seconds slightly over a time window prior to the leap second).So main institutions still find that real time applications need to maintain the link.Cesium might run like clockwork but the Earth rotation does not, you can't change one but you have to change the other for pratical purposes.

Really, the leap second has nothing to do with the definition of a standard second in terms of Caesium oscillations. Fitting the way our Earth behaves is an entirely different issue. The Solar system is a very creaky thing to base time measurement on and has been superseded as a source of a time standard.
Incidentally, on the subject of oscillator accuracy, I was incredibly impressed, several years ago when h.p. had a Rubidium 'brick' that could be switched on and, within a matter of less than an hour or so, was within one part in 10^10. After a while it got down to one part in 10^12, after a day. It needed no special environment and could be fitted into a suitcase for mobile synchronisation measurements. It's one thing to make a high accuracy source that operates in a fancy lab environment but that little gizmo (around 2kGBP cost at the time) really was an excellent and rugged tool. I imagine things have got a lot better in the past twenty years.

 

[Buckleymanor]

Really, the leap second has nothing to do with the definition of a standard second in terms of Caesium oscillations. Fitting the way our Earth behaves is an entirely different issue. The Solar system is a very creaky thing to base time measurement on and has been superseded as a source of a time standard.
Incidentally, on the subject of oscillator accuracy, I was incredibly impressed, several years ago when h.p. had a Rubidium 'brick' that could be switched on and, within a matter of less than an hour or so, was within one part in 10^10. After a while it got down to one part in 10^12, after a day. It needed no special environment and could be fitted into a suitcase for mobile synchronisation measurements. It's one thing to make a high accuracy source that operates in a fancy lab environment but that little gizmo (around 2kGBP cost at the time) really was an excellent and rugged tool. I imagine things have got a lot better in the past twenty years.

I was just trying to keep on topic as far as re- measuring the second as the OP asked.I somehow don't imagine he expects atomic time to be divorced from GMT and a totally different issue.

 

[sophiecentaur]

I re-read the OP and it is a very short paragraph which doesn't make itself clear. It's clearly open ended but I guess the wording "measure the second" is the problem. It's more a matter of either 'defining the second' or 'measuring time' and this thread had gone down both routes. There is no conflict, once we split the original question into two answerable ones. It's been interesting, in any case and we have more or less avoided shouting between ourselves about it.

If we had been asked how to 'measure a metre', we would have been in a similar situation.

 

[Drakkith]

I re-read the OP and it is a very short paragraph which doesn't make itself clear.

Yes, I came to the same conclusion back when I first posted. I was hoping the OP would reply and elaborate on what they meant.