Elektročas HH3 – the most accurate pendulum clock on the planet
109 points by tverbeure 11 days ago | 42 comments
  • jesperwe 7 days ago |
    I passed the "Not like we would not have 150 tons of liquid Helium here" statement off as a joke until I got to the end and realized the author was NOT just another clock renovating hobbyist :-)

    Thanks for an enjoyable read!

    • Aeolun 6 days ago |
      He made mention of his position at CERN before that came along I think?
  • saagarjha 6 days ago |
    Really interesting that the clock appears to have detected an actual earthquake and corrected for it!
  • MezzoDelCammin 6 days ago |
    I think it's the first time I see a Czech/Slovak accented "č" on the front page of Hacker News. Well done!
    • Squarex 6 days ago |
      Well another time was back in 2019 [0] with an article about Karel Čapek [1] :)

      [0] https://news.ycombinator.com/front?day=2019-08-26

      [1] https://news.ycombinator.com/item?id=20789755

    • urcite_ty_kokos 6 days ago |
      Also, did you notice Merkur?
    • inglor_cz 6 days ago |
      This was indeed somewhat startling, like meeting an old, well-known dog that just starts mewing at you.

      We are used to being the Leprechaun Folk of the West - present almost everywhere, but only rarely seen.

      • MezzoDelCammin 6 days ago |
        Especially if You're old enough to remember what ISO 8859-2 was for. One of the standards that everyone knew existed, yet few actually cared enough to correctly implement in practice. I still have a label printer that knows accented characters exist, yet the implementation is "shrink the accented letter to fit the label height with the accent"
        • inglor_cz 6 days ago |
          "shrink the accented letter to fit the label height with the accent"

          Ah, memories of youth... not to mention KEYBCS2.

  • schobi 6 days ago |
    I'm glad they only thought about it, but did not implement SF6 flooding of the chamber. No need to vent it to the atmosphere just for fun.

    But controlling the internal pressure? I would expect that only a small difference would be needed - far from exploding or imploding the clock! Maybe small enough to have a reservoir and control only inlet and outlet valves? This could be done purely without modifications of the mechanism.

    • TheOtherHobbes 6 days ago |
      The biggest danger of explosion/implosion is the change to/from summer time.
    • dtgriscom 6 days ago |
      The control force would be the product of the air density change and the pendulum bob volume. If he increased the pendulum bob volume with a light but fixed-size object (a foam-filled sphere?) he wouldn't need to adjust the density as much.
  • progbits 6 days ago |
    The contrast of the pendulum clock in a wooden cabinet and the sensor mounted on what appears to be Merkur (building toy from 1950s, still available today), with the fiber optic cables and maser is quite something :)
    • timeon 6 days ago |
      Ehm, I'm also using Merkur to extend my cabinets/shelves.
  • avidiax 6 days ago |
    This is in effect a really complicated way of getting an analog readout of his cesium fountain.

    The chain adjuster was interesting, though.

    Another very accurate pendulum clock (this one solar disciplined): https://en.wikipedia.org/wiki/Clock_of_the_Long_Now

    • wbl 6 days ago |
      Not really: look at the Alan curves. This is a disciplined PLL with a fairly slow loop. Most of the accuracy at short time scales is from the pendulum it's only the long term stability that the atomic clock impacts.
      • avidiax 6 days ago |
        All of the accuracy is from the pendulum. The author hasn't implemented a correction for the integral error yet.
  • cenamus 6 days ago |
    Does anyone know anything about the czech/slovak "flip" clock on the top right of the ~fifth last picture?

    I'd love to have something like that for myself

  • metalman 6 days ago |
    The plausible gravity compensation mechanism idea,got me but good. Which if implimented could achive the ideal of the clock remaining exactly as it left the factory, but with 1sec/158 million year accuracy. For the convieneince of keeping the clock in your office, would require incorporating the building into the gravity compensation mechanism.Building codes might prove to be the biggest challenge.
  • kristjank 6 days ago |
    The chain-and-cup adjustment is so simple, yet so effective. Such an elegant solution
  • withinboredom 6 days ago |
    I'm just thinking of the discussion in the IT department:

    A: He wants what?

    B: A cable from the cesium to his office, for some ancient clock he got off ebay.

    A: These scientists man...

    • bayindirh 6 days ago |
      I work with some people at CERN on a project.

      What I got away is, the answer from IT would be “that’s a cool project. Just ask your department head, and if they say OK, consider it done”.

      CERN still operates with and fueled by curiosity, and that’s a good thing.

      • tgsovlerkhgsel 6 days ago |
        > Just ask your department head

        That'd already be a lot more bureaucracy than I'd have expected.

        • bayindirh 6 days ago |
          It’s a lot of bureaucracy, yes, but considering the planning they do and the scale of the projects they undertake, it’s understandable.

          They are massive in every aspect.

          • tgsovlerkhgsel 6 days ago |
            I don't think having to ask another person over a coffee before you get a fiber from an atomic clock for your hobby project is a lot of bureaucracy. I think it's very little bureaucracy, but at CERN I'd have expected even less.
  • zokier 6 days ago |
    It seems bit odd to call it most accurate pendelum clock if the accuracy stems from being disciplined by atomic clock. To me the interesting question would be how far can you push the freestanding accuracy of a pendelum clock.

    That is not to say this isn't a neat project in it's own right, it is, but just picking on the terminology.

    • stavros 6 days ago |
      I agree, it's not a clock if it can't keep time on its own. It's a really cool project, but the clickbait title disappoints unnecessarily. "I made an NTP pendulum clock" would have been accurate and still extremely cool.
    • jgalt212 6 days ago |
      better title

      Elektročas HH3 – the most accurate atomically assisted pendulum clock on the planet (cern.ch)

    • technothrasher 6 days ago |
      Pendulum clocks in general need to be periodically rated (length of pendulum adjusted) due to variations in temperature and pressure in order to achieve maximum isochronous movement. All this guy has done is created an automatic rating system. Otherwise, it operates the same way as most any other pendulum clock. So I think the "most accurate pendulum clock" description is fair (assuming it is the most accurate).

      I've had ideas for a while for similarly adjusting a pendulum clock from an NTP server, though I was going to add an extra gear to the motion train that would adjust the hands slightly every twelve hours. But that seems more like "cheating" that this rating system.

    • KaiserPro 6 days ago |
      So if you want the most accurate pendulum clock, then you'll need the vacuum flask one: https://www.rmg.co.uk/collections/objects/rmgc-object-79649 or its german counterpart: https://theregulatorclockcompany.com/riefler/
  • stavros 6 days ago |
    I wonder if it would have been easier to have a small electromagnet in the base to pull the pendulum down either before or after the lowest point, as necessary. Wouldn't that have been a much simpler solution?
    • phreeza 6 days ago |
      Mentioned in the article, this was considered too boring!
      • stavros 6 days ago |
        Ahh thanks, turns out I missed it because it was a small sentence right over the big equation that my eyes glazed over :P
  • mrlonglong 6 days ago |
    "Changing the 2, or π is difficult even for CERN". Brilliant!
  • hsuduebc2 6 days ago |
    I understant it only partially but still loving it!
  • blacksmith_tb 6 days ago |
    There's something very satisfying about electromechanical clocks, I think it may be the combination of math-y/steampunk-y gear-driven timekeeping with actual accuracy. I have always thought once I Made It Big I'd get myself a Seiko Spring Drive watch[1] - hasn't happened yet...

    1: https://www.seikowatches.com/us-en/customerservice/knowledge...

  • lakomen 5 days ago |
    Jeli je to hrvatski proizvod?
    • lubosm 5 days ago |
      Ne, nije. Autor je slovacki inzenjer u CERNu.