How Do Atomic Clocks Work?

[Bill Hammack] aka [Engineerguy] is back again with another fantastic informational video. This time around he’s describing exactly how an atomic clock does what it does. He starts off with a great analogy of jello jiggling when poked. He explains how this is similar to the quartz crystal inside the clock oscillating due to the electrical “poke” we give it.  He goes on to explain how GPS satellites rely on this accuracy when determining physical locations on the ground.

As usual, [Bill] does a fantastic job of delivering the information quickly and packed full of detail, while still keeping it simple enough that even those unfamiliar with the technology can follow along.

23 thoughts on “How Do Atomic Clocks Work?

  1. What amazes me most is the miniaturization and commoditization we have achieved with such a complex device. You can buy a rubidium standard (I think that GPS satellite use that now instead of cesium standards) on eBay for about 60€.

    1. the price !
      Your clocks are synchronized over the main’s frequency. It’s not that precise but it does the job.
      You can hack them with a quartz oscillator if you want.

      1. Indeed, mains is synced to atomic clocks and damn accurate.

        As for losing accuracy from alarm/oven clocks, if they are digital they are spot on not just theoretically but in my RL experience too.
        So if you have issues you either have electric motor clocks or you have some serious wiring interference issues and might call your powercompany about that.

  2. I have a rather unusual question regarding GPS and atomic clocks. If GPS relies on atomic clocks to be accurate and over time, with plate tectonics shifting out location imperceptibly, how is it that GPS remains accurate?
    I know that our clocks are adjusted (leap year, leap seconds, etc.) for the error in our time system and the time it takes to complete both a rotation (day) and revolution around the sun (year)..

    So the question: Over time, wouldn’t the accuracy of GPS dwindle? I know it’s by a small amount so just tell me that it’s not enough to affect our GPS system over the course of human history and I’ll shut up. Maybe it’s just my German genes that encourages me to think this way.


    1. I think you’re mistaken the absolute position on a sphere given a set of defined coordinates, and the juxtaposition of a real map over those coordinates. To navigate securely in 100 years, you just have to offset your map by a l meter.

      At least that’s how I understand it.

    2. Jason,
      Actual time isn’t what matters, relative time is. The GPS receiver doesn’t have to know that it is exactly 6pm, just that the signal it received from the satellite left the satellite exactly 3.4562ms ago. So our imperfect method of keeping time doesn’t affect it.

      As far as moving land masses go, that is more a question of MAP accuracy, not position information. GPS tells you your latitude and longitude, not what street you are on. The software running on your cell phone, handheld GPS, or in your car converts that to a location on a map.

      1. Thank you both for the follow up. I guess with all the recent earthquakes and realization that we are on a moving, dynamic planet, I imagined myself 10,000 years into the future. If the GPS were to pinpoint my location on the globe at any given point, I’m guessing (by then) that maps would be updated to reflect both the movement of the tectonic plates as well as the location of the north and south poles relative to the GPS satellites. I guess the question was more or less, how much deviation over time can we expect our GPS systems to be out in 10,000 years if both the GPS systems weren’t corrected for the changes in the planet and if the mapping system that overlays onto the sphere isn’t updated?

        As far as understanding that small handheld gps units translates to a street address… yes, I get that… it’s approximated and usually if you are travelling down a road, the GPS unit snaps to the road you are on. My GNex does it all the time in google maps.

      2. Jason,

        Considering that GPS satellite last about 10 years, (and need continual guidance updates, etc.) there would be a lot of updating going on in those 10,000 years.

        The Paleomap project has a bunch of maps relating to continental drift, with projections for continental motions for the next 50 million+ years.

        For a really rough guess in short term like 10,000 years, might get an estimate from looking at relative motions of particular plates (e.g. 2.5cm/year mid Atlantic ridge spreading x 10,000 )

    3. According to the U.S. Airforce, the first generation of GPS satellites had a design lifetime of 7.5 years, which has been well exceeded. The next generation GPS IIF satellite have a longer design life, but I don’t know what that is.

      If it was NASA instead of the military putting these up, Congress would have already cut the funding. So I don’t think you have to worry about GPS in 10,000 years, ’cause none of it will be working.

  3. My favorites of Bill’s videos are the 2 early ones on whiffletrees and the IBM selectric

    Dave Jones did a teardown on one of the rubidium standards on EEVblog

    @ zuul
    There are a number of clocks watches and other time keeping devices that use the shortwave broadcasts, as well as a number of stations broadcasting. +/-1 millisecond is good enough for a lot of applications. Rubidium and Cesium standards are very expensive new, these cheap ones on ebay are mostly torn out of old cellphone towers. So unless your atomic clock cost several thousand dollars, it probably won’t have a physics package of it’s own.

    40 kHz JJY Japan
    50 kHz RTZ Russia
    60 kHz JJY Japan
    60 kHz WWVB United States
    60 kHz MSF United Kingdom
    66.66 kHz RBU Russia
    68.50 kHz BPC China
    75.00 kHz HBG Switzerland – Ceased operation this January
    77.50 kHz DCF77 Germany
    162.0 kHz TDF France
    2.5 MHz BPM China
    2.5 MHz WWV United States
    2.5 MHz WWVH United States
    3.33 MHz CHU Canada
    4.996 MHz RWM Russia
    5 MHz BSF Taiwan
    5 MHz BPM China
    5 MHz WWV United States
    5 MHz WWVH United States
    5 MHz HLA South Korea
    5 MHz LOL1 Argentina
    5 MHz YVTO Venezuela
    7.85 MHz CHU Canada
    9.996 MHz RWM Russia
    10 MHz BPM China
    10 MHz WWV United States
    10 MHz WWVH United States
    10 MHz LOL1 Argentina
    10 MHz PPE[6] Brazil
    11 MHz ATA India
    14.67 MHz CHU Canada
    14.996 MHz RWM Russia
    15 MHz BPM China
    15 MHz BSF Taiwan
    15 MHz WWV United States
    15 MHz WWVH United States
    20 MHz WWV United States

  4. Maybe I’m nitpicking, but three satellites gives you two points of intersection of which one can be discarded(if you assume that you are on the surface of the earth).

  5. Last I checked, atomic clocks don’t use quartz, as by definition they rely on the atomic oscillation of a single given element, rather than a vibrating crystal of SiO4 molecules. Not having watched the video, I’m going to put this error down to the editor.

    1. he also messed up gps, 3 sats does not narrow it down to one point, it narrows it down to 2 points (one very high in the sky)
      you need 4 satalites to reduce it to a single exact point

  6. Whether one uses atomic or quartz clocks doesn’t seem to help me and other time wasters to keep them on time. Do you think the time may come when someone invents a personal ‘buzzer’ attached to an atomic clock to give people like me a ‘jig along’?

    I loved your analogies BTW…

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.