# Question about caesium clocks. **URL:** \ **Category:** Factual Questions\ **Created:** [September 18, 2019, 5:29pm UTC](https://boards.straightdope.com/t/question-about-caesium-clocks/840447 "2019-09-18T17:29:47Z")\ **Posts on this page:** 7\ **Page:** 1
**Author:** ![KidCharlemagne](https://avatars.discourse-cdn.com/v4/letter/k/ed8c4c/32.png) [@KidCharlemagne](https://boards.straightdope.com/u/KidCharlemagne)\ **Post date:** [September 18, 2019, 5:29pm UTC](https://boards.straightdope.com/t/question-about-caesium-clocks/840447/1 "2019-09-18T17:29:47Z")
If the detector on a caesium clock requires a consistent stream of excited atoms, how is the rate of the stream itself held constant? That is, what regulates the rate of release of atoms from the caesium oven? I’m guessing the sheer volume of atoms is so great that it’s not a case of the detector getting atoms at all but whether they’re getting excited ones but I’d like to hear from the experts. Thanks in advance. ---
**Author:** ![KidCharlemagne](https://avatars.discourse-cdn.com/v4/letter/k/ed8c4c/32.png) [@KidCharlemagne](https://boards.straightdope.com/u/KidCharlemagne)\ **Post date:** [September 18, 2019, 5:35pm UTC](https://boards.straightdope.com/t/question-about-caesium-clocks/840447/2 "2019-09-18T17:35:20Z")
Whoops one other question - when the frequency exciting the caesium is off, how does the clock determine whether the quartz tuning fork thing is vibrating too fast or too slow? Thanks ---
**Author:** ![Napier](https://avatars.discourse-cdn.com/v4/letter/n/ce73a5/32.png) [@Napier](https://boards.straightdope.com/u/Napier)\ **Post date:** [September 18, 2019, 6:28pm UTC](https://boards.straightdope.com/t/question-about-caesium-clocks/840447/3 "2019-09-18T18:28:26Z")
I think the clock doesn’t require that consistent a stream, and ordinary control methods work just fine. It isn’t as though the rate of that stream sets the rate of the clock itself. It just establishes how much cesium is available to resonate. As to your other question, I’ve always wondered that too. They might always keep it slightly to one side of the ideal frequency, or they might bounce back and forth. They could even use bulk flow of the stream to create a doppler effect to get a split. Or maybe it’s something I’m not thinking of or wouldn’t know at all. ---
**Author:** ![Bear\_Nenno](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/bear_nenno/32/3358_2.png) [@Bear\_Nenno](https://boards.straightdope.com/u/Bear_Nenno)\ **Post date:** [September 18, 2019, 6:34pm UTC](https://boards.straightdope.com/t/question-about-caesium-clocks/840447/4 "2019-09-18T18:34:04Z")
Magnets. > [@https://tycho.usno.navy.mil/cesium.html](#): > > In a cesium clock like these, liquid cesium is heated to a gaseous state in an oven. A hole in the oven allows the atoms to escape at high speed. These particles pass between two electromagnets whose field causes the atoms to separate into two beams, depending on which spin energy state they are in. Those in the lower energy state pass through the ends of a U-shaped cavity in which they are irradiated by microwaves of 3.26-cm wavelength. > The absorption of these microwaves excite transitions of many of the atoms from the lower to the higher energy state. The beam continues through another pair of electromagnets, whose field again divides up the beam. Those atoms in the higher energy state strike a hot wire, which ionizes them. Thereafter, a mass spectrometer selects only the cesium atoms from any impurities and directs them onto an electron multiplier. > > The frequency of the microwaves is adjusted until the electron multiplier output current is maximized, constituting the measurement of the atoms’ resonance frequency. This frequency is electronically divided down and used in a feedback control circuit (“servo-loop”) to keep a quartz crystal oscillator locked to a frequency of 5 megahertz (MHz), which is the actual output of the clock, along with a one-pulse-per-second signal. The entire apparatus is shielded from external magnetic fields. ---
**Author:** ![DPRK](https://avatars.discourse-cdn.com/v4/letter/d/4491bb/32.png) [@DPRK](https://boards.straightdope.com/u/DPRK)\ **Post date:** [September 18, 2019, 6:46pm UTC](https://boards.straightdope.com/t/question-about-caesium-clocks/840447/5 "2019-09-18T18:46:27Z")
> [@KidCharlemagne](#): > > If the detector on a caesium clock requires a consistent stream of excited atoms, how is the rate of the stream itself held constant? That is, what regulates the rate of release of atoms from the caesium oven? I’m guessing the sheer volume of atoms is so great that it’s not a case of the detector getting atoms at all but whether they’re getting excited ones but I’d like to hear from the experts. Thanks in advance. [This page](https://tycho.usno.navy.mil/cesium.html) describes an old-school caesium clock where the caesium is vaporized in an oven, then escapes from there through a hole, separated (according to energy state) by an electromagnet, zapped by microwaves in a cavity, divided up again by magnets (so you get only atoms that have changed from the lower to the higher energy state), ionized by a hot wire, and finally selected by a mass spectrometer onto the detector. So, it’s not getting random crap straight from the oven, but rather a carefully prepared beam of atoms that have undergone the right state transition. > [@KidCharlemagne](#): > > Whoops one other question - when the frequency exciting the caesium is off, how does the clock determine whether the quartz tuning fork thing is vibrating too fast or too slow? Thanks The exact mechanism is where most of the complexity of the clock comes in, but the same page explains that the microwave oscillator frequency is swept across a narrow range (producing a modulated signal) and adjusted via feedback so as to maximize the output current. ---
**Author:** ![KidCharlemagne](https://avatars.discourse-cdn.com/v4/letter/k/ed8c4c/32.png) [@KidCharlemagne](https://boards.straightdope.com/u/KidCharlemagne)\ **Post date:** [September 18, 2019, 11:16pm UTC](https://boards.straightdope.com/t/question-about-caesium-clocks/840447/6 "2019-09-18T23:16:11Z")
> [@DPRK](#): > > [This page](https://tycho.usno.navy.mil/cesium.html) describes an old-school caesium clock where the caesium is vaporized in an oven, then escapes from there through a hole, separated (according to energy state) by an electromagnet, zapped by microwaves in a cavity, divided up again by magnets (so you get only atoms that have changed from the lower to the higher energy state), ionized by a hot wire, and finally selected by a mass spectrometer onto the detector. So, it’s not getting random crap straight from the oven, but rather a carefully prepared beam of atoms that have undergone the right state transition. > The exact mechanism is where most of the complexity of the clock comes in, but the same page explains that the microwave oscillator frequency is swept across a narrow range (producing a modulated signal) and adjusted via feedback so as to maximize the output current. The magnets are only separating excited from unexcited atoms as far as I can tell, not regulating the overall release from the source. The first set filters out any previously excited atoms - which makes sense since you want to start with unexcited atoms, and the second set filters out the unexcited atoms so that only the excited ones get “counted.” At least that’s my reading of it. But what regulates the number of atoms leaving the oven in the first place? The reason that seemed important to me is that it appears it isn’t an all/nothing prospect for the atoms to be excited (perhaps because of the wiggle room in absorption frequency?) - meaning that at a frequency slightly off from the ideal, some of the atoms will still get excited which would mean you’d have to have a constant stream of atoms to accurately detect the percentage that are excited (unless it doesn’t matter for the reasons in my OP). ---
**Author:** ![DPRK](https://avatars.discourse-cdn.com/v4/letter/d/4491bb/32.png) [@DPRK](https://boards.straightdope.com/u/DPRK)\ **Post date:** [September 19, 2019, 12:00am UTC](https://boards.straightdope.com/t/question-about-caesium-clocks/840447/7 "2019-09-19T00:00:16Z")
> [@KidCharlemagne](#): > > The magnets are only separating excited from unexcited atoms as far as I can tell, not regulating the overall release from the source. The first set filters out any previously excited atoms - which makes sense since you want to start with unexcited atoms, and the second set filters out the unexcited atoms so that only the excited ones get “counted.” At least that’s my reading of it. But what regulates the number of atoms leaving the oven in the first place? The reason that seemed important to me is that it appears it isn’t an all/nothing prospect for the atoms to be excited (perhaps because of the wiggle room in absorption frequency?) - meaning that at a frequency slightly off from the ideal, some of the atoms will still get excited which would mean you’d have to have a constant stream of atoms to accurately detect the percentage that are excited (unless it doesn’t matter for the reasons in my OP). In real life, you don’t get a stable signal from this type of clock as soon as you turn it on; it takes days, at the very least, for things to settle down. The caesium oven has a thermistor so you can level it off at the right temperature, and similarly components like the ion pump and hot-wire ionizer have a certain specification. Presumably, after a while the system is in a steady state, including the number of atoms in your beam, but ISTM that the exact number is not important (just that the beam density be low enough to minimize self-interaction) and not directly measured. You are correct that the resonance is not infinitely sharp; the line width is 450 Hz or so. That still gives you a pretty sharp “peak” that your electronics can lock on to.