Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

I GPS and Cs atomic clock question

  1. Nov 27, 2017 #1
    Ok, so pardon my ignorance if this can found easily online but I have these few questions. I was actually reading quite a complicated article about GPS and relativity but it had too much maths in it for my abilty.

    Ok, first of all, apart from the complicated details regarding the workings of laser, cryogenic cooling etc, does a Cs atomic clock work by cooling Cs133 ions to a low (how low?) temperature and then inserting them inside a microwave cavity and as they get radiated by the microwaves they then fall into resonance but only if the frequency is the correct one, so by looking at the feedback they can then determine that the highest peak (electric current or what?) is where the resonance is at.
    Now I assume this can only work so precisely due to the fact that the resonant peak of the CS ions is a very well defined and precise one correct?
    But if we are speaking about precision rates into the millionth's of a second, then how can they achieve those if the rest of the atomic clock is usual electronics, I'm talking about the feedback system and the microwave generating system etc? They use ordinary components like quartz crystal frequency oscillators circuits and microwave emitters etc right? so what physical process here governs the extreme accuracy which one can never expect from an ordinary quartz crystal controlled simple electronic clock? given the fact that quart crystal degrades over time and use and so it resonant frequency changes.
    If possible I would love a more in depth explanation of the innards of a CS atomic clock, maybe someone here has done some work with them.

    As for the GPS, I read that it is very important to have precise timing which is synchronized so they use atomic clocks both here on earth for time keeping and also in satellites used for GPS, now ok I think I get the part where a GPS device like my smartphone can calculate my physical location based on the signal from atleast 3 satellites as it probably receives a signal from each of the satellites which has a "time stamp" as I read but most importantly which travels as all Em waves travel at the speed of light in vacuum, but here's my question , earth's atmosphere is not vacuum instead it has many obstacles like fog and rain etc, so the speed of EM waves is slower than c, but how do we then know precisely what that speed is, so how can the GPS then add or subtract the speed of the person here on earth using the Doppler effect on top of the signal speed itself if we cannot know the precise signal speed? Since I don't know I assume i'm then asking does the receiver device (my smartphone) calculate my speed, either based on the constantly changing timestamps from satellites with respect to the atomic time standard here on earth or by the signal that it gets from the satellite, enlighten me please.
    I understand they slow down a tiny bit the frequency of the atomic clock onboard the satellite to compensate for gravitational and relativistic effects due to high speeds involved in satellite motion.
    Oh also since satellites transmit their signal, what kind of power source they use for that as I imagine space and weight is limited and a radiosingal that is strong enough requires some energy.

    Thank you.
  2. jcsd
  3. Nov 27, 2017 #2


    Staff: Mentor

    https://en.wikipedia.org/wiki/Atomic_clock#Mechanism also check the references linked at the bottom of that ariticle.

    Ditto for GPS, the references linked at the bottom of the Wikipedia article include lots of material.
  4. Nov 27, 2017 #3


    User Avatar
    Science Advisor
    Gold Member
    2017 Award

    I am not sure of the level of your understanding but this is a pretty Noddy version of how to produce frequency standards.
    BTW, the atoms are not cooled - they are heated to a known temperature (boiled off a hot filament) and a beam of them is sent through a cavity.
    The basic approach is the same as for all oscillators (including a pendulum clock!!) when you try to make them as accurate as possible. There are a lot of 'things' that naturally oscillate with a pretty well defined frequency but the problem is in getting a signal out of them and supplying them with power to maintain the oscillation. This interference from the outside world will degrade the accuracy of the frequency. It affects the Q of the oscillation and allows interfering signals (vibrations, RF interference etc.) to disturb the natural frequency. So you need very light coupling with the high quality oscillator. Isolated gas atoms have very well defined energy transitions so their natural frequencies can be relied upon - as long as you can get signals to and from them.
    The basic principle of an 'atomic clock' is not too hard but the details are very involved. You start with atoms which are in a container with a very well controlled temperature and in a well defined magnetic field. (Very very well controlled). These atoms (the basic frequency reference) are chosen to have a very well defined energy transition (accurate to better than the clock spec requires) for a given temperature and magnetic field. You then produce a very accurate RF frequency signal (pretty damn STABLE crystal oscillator, also in closely controlled conditions), You introduce a small sample of the RF (microwave) into the cavity with the atoms in it and you vary the crystal oscillator frequency until you can detect resonance absorption by the atoms. This method ensures very loose coupling into the atomic oscillators and when you have found the absorption frequency, you know that the RF frequency (a harmonic of the basic crystal oscillator) is the same as the atomic transition energy.
    Finding a good resonance can be done by modulating (wobbling) the RF signal slightly so that you get a variation in the absorption level and you can then identify the very peak of the resonance by equalising the deviations on either side of f0. It's much better than just hunting for a maximum as it gives you a signed feedback signal with which to adjust your RF oscillator. The technique of wobbling a variable so that you can detect the changes and locate a maximum is common in many servo systems.
    Needless to say, the slave oscillator has to be very highly stable so that the slowly varying control feedback voltage can actually keep its frequency on track.
    I have used a Rubidium frequency source which is the size of a small brick and it reaches a 'good' frequency (+- 1/1010 ish)after only a half hour or so after switch on. Now that is really really smart because you don't need a fancy lab with air con and all the rest. This box fits inside a simple bit of lab equipment and you can turn it on in the morning. To my mind that is as good a bit of engineering as achieving a few orders of magnitude better with all that surrounding gear.
  5. Nov 28, 2017 #4
    Ok, reading up on the subject I think I have understood such basic premises as: (please correct if they are wrong)

    The heart of a Cs133 atomic clock is basically vacuum tube combined with an RF cavity, and some detector instrument, I read that the energy given off by the Cs ions after RF irradiation is in the form of fluorescence, so are they measuring maximum peak "photoelectric current" from the detector to know that they have hit the Cs133 absorption/resonance peak?
    So I assume they can be sure that at resonance the transition count of the Cs ions amount to one second because those transitions between energy states are a physical law of nature and so they are fundamental and precise, so as long as the other peripherals like the RF frequency etc are kept stable the transitions will be precise?
    Ok another question, normally there are electrons emitted from a cathode, but it is said that the "shower" in the clock is made from ions, so what happens, is there a finely deposited Cs133 on the cathode and when heated it is evaporated and emitted off of the cathode, but then does it comes off as both ions and electrons?
    I assume they then employ magnets to separate the ions from electrons.
    I also read that they "cool" down the Cs ions before they enter the cavity for irradiation and they do this with lasers, now I may not know the precise terms to describe this process but from what I read the laser beams interact with the ions and this interaction force the ion to eject a photon which then takes away from it;s kinetic energy which is it's temperature.
    As I read the NIST website link it seems that earlier Cs atomic clocks simply used the vacuum tube method of hot Cs emission and then the irradiation of those ions as they passed the cavity but never models use the laser method to cool the atoms down near absolute zero to increase their time in the cavity and so I assume they get a sharper peak output current which is then more specific and so more precise than the previous method?

    So the RF oscillator which is made from usual electronics is made accurate by the very feedback from the ion transition peak within the cavity correct?
    As you say sophie that the coupling needs to be delicate in order not to disturb the clock primary mechanism, but if I understand here the coupling is simply made by electric current coming out from the photoelectric sensor which is hit by the Cs ions or rather photons they emit after they fall back from their excited energy state?? and resonance is simply the frequency at which maximum number of ions get excited?

    Ok, so far so good for now , i'll wait for comments and then ask more questions
  6. Nov 28, 2017 #5


    User Avatar
    Science Advisor
    Gold Member
    2017 Award

    There is an output signal but the 'coupling' is between the RF and the Atoms. As far I am aware, the output signal is the level of the RF, due to interaction with the atoms (absorption) and not the number of atoms arriving at a cathode.
    Last edited: Nov 28, 2017
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook

Have something to add?
Draft saved Draft deleted