[HN Gopher] New atomic fountain clock joins group that keeps the...
       ___________________________________________________________________
        
       New atomic fountain clock joins group that keeps the world on time
        
       Author : austinallegro
       Score  : 87 points
       Date   : 2025-04-29 12:39 UTC (1 days ago)
        
 (HTM) web link (www.nist.gov)
 (TXT) w3m dump (www.nist.gov)
        
       | throw0101a wrote:
       | What are 'limits' on how accurate enough clocks 'should' be?
       | 
       | Presumably there's diminishing returns, but as the article says
       | we're at one part in 2.2e-16, are there practical application of
       | going further?
        
         | 0_____0 wrote:
         | With radio astronomy, where you're measuring phase of incoming
         | radiation, I think "more is more" applies. Would be interesting
         | to hear from someone who actually has experience in that domain
         | though (not me!)
        
       | move-on-by wrote:
       | How many atomic clocks are in operation in Colorado now? It would
       | be nice if they could be spread around a bit. I suppose there are
       | logistical issues that keep them centralized?
        
         | algorithmsRcool wrote:
         | The Naval Observatory in Washington DC has quite a few also
        
         | CamperBob2 wrote:
         | Commercial atomic clocks of various types aren't that rare.
         | Every cell site has a rubidium standard and/or GPS timing, many
         | data centers probably have a cesium standard, and radio
         | astronomers use H-masers for interferometry.
         | 
         | Everybody with a GPS-disciplined oscillator has access to time
         | and frequency from the Naval Observatory at the sub-100 ns
         | level, optionally augmented to +/- 1 ns with reasonably
         | affordable gear like https://www.sparkfun.com/sparkpnt-gnss-
         | disciplined-oscillato... .
         | 
         | A fountain clock is on a whole different level than any of
         | these. The same researchers who build fountains also work on
         | even better optical lattice clocks, none of which you can buy
         | from Sparkfun. These are research tools that don't have a
         | market, at least not yet.
         | 
         | The SI second definition will likely move from Cs-133 at 9 GHz
         | to Sr-87 at 400 THz before too long ( https://www.nist.gov/si-
         | redefinition/second-future ), but that probably won't shake up
         | the existing market too much.
        
         | fanf2 wrote:
         | As well as NIST there is Schriever spave force base
         | https://en.m.wikipedia.org/wiki/Schriever_Space_Force_Base
         | which is the ground operations centre for GPS. They have the
         | USNO alternate master clock, which maintains a copy of the USNO
         | time scale based on caesium beam and rubidium fountain clocks.
        
       | algorithmsRcool wrote:
       | I recall reading that our ability to measure time accurately
       | exceeds that of any other quantity. According to the NIST, the
       | newest Optical Lattice clocks would drift by less than 1 second
       | if they were started 13 billion years ago at the big bang. What
       | else can we measure down past 1 part per 10e18?
        
         | analog31 wrote:
         | Curiously, there's also a contender for the worst, which I
         | think at present is the gravitational constant.
        
           | teraflop wrote:
           | Yup. And an interesting detail is that we know the _product_
           | G*M_e (where G is the gravitational constant, and M_e is the
           | mass of the earth) to much higher precision than we know
           | either of its factors. And the same goes for the sun and most
           | of the other planets.
           | 
           | This is because the motion of celestial bodies and spacecraft
           | is dominated by gravitational forces which depend only on
           | G*M, and that motion can be measured extremely accurately
           | with e.g. Doppler radar.
        
         | ipdashc wrote:
         | > our ability to measure time accurately exceeds that of any
         | other quantity
         | 
         | TIL. I guess maybe that explains why the second is used as the
         | base of the SI
         | (https://en.wikipedia.org/wiki/2019_revision_of_the_SI)
         | post-2019, if my understanding of it is correct?
        
         | AlotOfReading wrote:
         | That's around the same sensitivity that LIGO operates at.
        
       | CGMthrowaway wrote:
       | Can't mention a fountain clock without calling to mind one the
       | Lion Clock in Alhambra Palace, of my favorite things ever:
       | 
       | https://www.iflscience.com/this-incredible-islamic-fountain-...
        
       | RetroTechie wrote:
       | Some years back there was talk of an atomic clock that came in a
       | very small package (matchbox / cigarette pack size?). Iirc, saw a
       | price indication of $1500 somewhere.
       | 
       | Something like that still around, and/or available? Any updated
       | designs?
       | 
       | Personally I have no need for ultra-accurate timekeeping. But
       | hey... an atomic clock is way cooler than a Nixie clock or oven-
       | controlled Xtal oscillator. And no... huge 2nd hand atomic clock
       | found on eBay etc doesn't cut it. Too big /heavy / power-hungry.
        
         | philipkglass wrote:
         | You're thinking of the Chip Scale Atomic Clock, first
         | demonstrated by NIST in 2003 and commercialized in 2011:
         | 
         | https://www.nist.gov/noac/technology/time-and-frequency/chip...
         | 
         | https://spectrum.ieee.org/chipscale-atomic-clock
         | 
         | Microchip launched their latest version earlier this year:
         | 
         | https://www.electronicspecifier.com/products/frequency-contr...
        
           | Palomides wrote:
           | I've heard they're struggling to get good consistent
           | perfomance on them compared to other alternatives in similar
           | sizes
        
           | WarOnPrivacy wrote:
           | > Microchip launched their latest version earlier this year:
           | 
           | Spectratime's version (mRO-50) went into a $2M piece of eye-
           | candy that mechanically syncs a watch.
           | https://www.urwerk.com/collections/ur-chronometry/amc
           | https://www.urwerk.com/sites/default/files/press/docs/urwerk_
           | amc_eng.pdf
        
         | ooterness wrote:
         | You're probably thinking of chip-scale atomic clocks (CSAC).
         | There's at least two companies that make them [1][2].
         | 
         | [1] https://www.microchip.com/en-us/products/clock-and-
         | timing/co...
         | 
         | [2] https://www.teledyne-si.com/en-us/Products-and-
         | Services_/Pag...
        
           | ginko wrote:
           | Any idea what one of those would cost? (and where to buy
           | them?)
        
             | nullc wrote:
             | SA.45s (and the eval board) has been on mouser in the past,
             | I'm not sure if they carry it now because mouser
             | continually blocks me these days. It was previously around
             | $1500.
             | 
             | If you're interested in a small atomic clock and don't
             | absolutely require the very low power consumption of the
             | SA.45s and the very small package size you can get better
             | performance, reliability, and cost in something a big
             | larger and quite a bit higher power consumption.
             | 
             | The CSAC improved atomic oscillators a lot more in power
             | than size... but its timekeeping performance is so/so as
             | far as atomic clocks go.
             | 
             | In particular, the surplus market has a lot of telecom
             | rubidium that can be had quite inexpensively.
             | 
             | I'm personally a fan of the PRS-10 (which also exists in a
             | benchtop form, the SRS725). They seem to regularly sell on
             | ebay for about $300 with a little breakout board for power.
             | And unlike many small rubidiums they have very good phase
             | noise. You can sync them to GPS time using a 1pps input
             | (though I believe on the bare modules the 1pps sync is
             | optional so if you want to use it be sure to get one that
             | has it).
        
               | RetroTechie wrote:
               | _> The CSAC improved atomic oscillators a lot more in
               | power than size... but its timekeeping performance is
               | so/so as far as atomic clocks go._
               | 
               | Timekeeping of atomic clocks is still waaaayy better than
               | next-best technologies (like temp-compensated Xtal
               | oscillators). And size/power/$ constraints matter. So
               | CSACs for the win imho.
        
               | fsh wrote:
               | A $5 GPS receiver runs circles around a CSAC, so the
               | range of useful applications is quite limited.
        
               | nullc wrote:
               | The timing output from a $5 gps receiver isn't
               | particularly impressive. You have to go to pretty long
               | time intervals for a non-timing GPS to win out. Depending
               | on your application those intervals may be thoroughly
               | irrelevant, e.g. using the device as a frequency source
               | rather than a clock.
               | 
               | There is also, of course, the issue of infrastructure
               | dependence. Particularly since wireless telephony has
               | moved almost exclusively to GNSS time we're going to have
               | a really bad time if kessler syndrome takes out the GNSS
               | satellites.
               | 
               | Edit: Here is an example adev chart for a inexpensive
               | atomic clock vs what appears to be a pretty good timing
               | GPS receiver: https://www.thinksrs.com/images/instr/prs10
               | /PRS10diag2LG.gif
               | 
               | So in that case the GPS accuracy only beats out the free
               | running atomic clock at intervals greater than 200,000
               | seconds or so.
               | 
               | Here is a collection of older timing receivers:
               | http://www.leapsecond.com/pages/3gps/gps-adev-mdev.gif
               | 
               | A quick look didn't turn up any cheap non-timing
               | receivers, but my experience is they're pretty bad (I
               | mean relative to atomic standards, of course).
               | 
               | There are better timing receivers than the ones charted
               | above, of course, but they are not $5. Their cost is now
               | in the same general ballpark as surplus atomic clocks.
               | 
               | Of course, if you have both you can sync one to the other
               | with whatever time constant maximizes the composite
               | performance and have the best of both.
               | 
               | (and primary atomic clocks don't have this drift issue,
               | but sadly the days of the occasional sub $1000 5071
               | showing up on auction sites seem to be over. :P )
        
               | schoen wrote:
               | > SA.45s (and the eval board) has been on mouser in the
               | past, I'm not sure if they carry it now because mouser
               | continually blocks me these days. It was previously
               | around $1500.
               | 
               | I searched on Mouser and found that part no longer has a
               | listed price; there are fancier chip-scale atomic clocks
               | from the same company, the only one with a listed price
               | going for $3,528.23.
               | 
               | Others from a different manufacturer are just over
               | $2,000. A kind of awesome thing is that one of those
               | chips is marked
               | 
               | Frequency 10.000000 MHz
               | 
               | It's awesome that that's not just an estimate or some
               | kind of exaggeration!
        
         | alnwlsn wrote:
         | How big is big? You can still get used FE-5680A rubidiums on
         | Ebay for a few hundred. They are about the size of a small book
         | and need ~20 watts.
         | 
         | Not sure how good/useful/not broken they are since I've never
         | had one, but over the last 10 years I've seen them in a good
         | number of hobbyist projects.
         | 
         | Or you could just go for GPS. That's technically still atomic
         | clocks, but in space!
        
       | WarOnPrivacy wrote:
       | NIST explainer: Cesium Fountain Clock [1999]
       | 
       | https://www.nist.gov/news-events/news/1999/12/nist-f1-cesium...
        
       | infogulch wrote:
       | Cesium-based atomic clocks ("fountain clocks") like these use the
       | natural resonance frequency of electron orbital energy states
       | under a microwave laser which can be counted to measure time.
       | Since there is a natural background noise of microwaves and many
       | frequencies can interact with orbitals it's important to isolate
       | the atoms from outside sources of electromagnetic radiation and
       | heat in order to maintain accuracy.
       | 
       | Earlier this year there was a big leap in so-called "nuclear
       | clocks" which uses the resonant frequency of energy states _of a
       | nucleus itself_ as opposed to electron orbitals around it.
       | Besides the  "more frequency = more better" factor that has
       | always driven clock accuracy -- thorium-229 nuclei excites in
       | ultraviolet wavelengths -- nuclear clocks are better isolated
       | than electron orbital-based clocks because the frequency band
       | where they interact is impossibly narrow. In fact, the reason why
       | it was only recently demonstrated is due to the difficulty of
       | producing the required frequency at a high enough precision to
       | interact reliably. This could lead to more accurate and more
       | compact and cheaper clocks.
       | 
       | Discussion 4 months ago:
       | https://news.ycombinator.com/item?id=42362215 | Major Leap for
       | Nuclear Clock Paves Way for Ultraprecise Timekeeping (nist.gov)
        
       | 0cf8612b2e1e wrote:
       | How do you sync atomic clocks? When the error rate is 1/1e16,
       | your error in propagating the time from one clock to another is
       | going to be off by many orders of magnitude vs the timekeeper
       | itself.
        
         | alnwlsn wrote:
         | Like this https://www.youtube.com/watch?v=RzOQHazjXvE
        
         | nullc wrote:
         | The modern approach is common view time transfer. Both clocks
         | can observe signal sources such as satellites and compute their
         | relative offsets. The signal doesn't need to be particularly
         | accurate (though it doesn't hurt), though you do need a good
         | model of its position and propagation. GNSS satellites are an
         | obvious choice though there are also specialized services for
         | CVTT.
         | 
         | Note that this is distinct from syncing _from_ GPS, which is a
         | thing people obviously do too, but CVTT can achieve much higher
         | accuracy.
         | 
         | Because you're synchronizing extremely stable clocks the
         | difference between them will primary be an offset (plus/minus a
         | slope from relativistic effects of different altitude). Because
         | of this you can average a large number of readings, so the only
         | major source of error will be systematic effects in
         | propagation/orbit/etc.
         | 
         | Historically, sync was obtained via traveling clocks-- e.g. you
         | sync one atomic clock up and load it, running, in a station
         | wagon... which is the same thing that is most often done for
         | voltage standards today (as atomic voltage standards remain
         | rare, compared to atomic clocks-- I think the least I've paid
         | for one is $15 excluding the ones that were free).
         | 
         | But vibration isn't great for anything with a crystal
         | oscillator in it, and the most modern atomic fountain clocks
         | don't work if they're accelerating in any direction except the
         | designed 'up' direction (gravity), because the little cloud of
         | cooled atoms will fall out of the measurement channel, which
         | makes sync by station wagon not viable.
         | 
         | Of course, once you talk about syncing there is always a
         | question of what you're syncing _to_. UTC doesn 't exist until
         | after the fact. Laboratories measure their offsets via CVTT and
         | UTC is calculated after the fact as past offsets to each of the
         | contributing clocks.
        
       | geerlingguy wrote:
       | New goal: upgrading from a Rubidium CSAC to NIST-F4.
       | 
       | I've also been reading about nuclear clocks[1]... skipping over
       | the uncertainty of the entire atom's chaotic oscillations
       | entirely!
       | 
       | [1] https://en.wikipedia.org/wiki/Nuclear_clock
        
         | nullc wrote:
         | Having to use vacuum UV is a bit of a bummer... just doing
         | anything inside a vacuum is a PITA.
        
       ___________________________________________________________________
       (page generated 2025-04-30 23:01 UTC)