Slashdot is powered by your submissions, so send in your scoop

 



Forgot your password?
typodupeerror
Science

How a Power Outage In Colorado Caused US Official Time To Be 4.8 Microseconds (npr.org) 63

Tony Isaac shares a report from NPR: The U.S. government calculates the country's official time using more than a dozen atomic clocks at a federal facility northwest of Denver. But when a destructive windstorm knocked out power to the National Institute of Standards and Technology (NIST) laboratory in Boulder on Wednesday and a backup generator subsequently failed, time ever so slightly slowed down. The lapse "resulted in NIST UTC [universal coordinated time] being 4.8 microseconds slower than it should have been," NIST spokesperson Rebecca Jacobson said in an email. [...]

Since 2007, the official time of the U.S. has been determined by the commerce secretary, who oversees NIST, along with the U.S. Navy. The national time standard is known as NIST UTC. (Somewhat confusingly, UTC itself is a separate, global time standard to which the U.S. and other countries contribute measurements.) NIST currently calculates the standard using a weighted average of the readings of 16 atomic clocks situated across the Boulder campus. Atomic clocks, including hydrogen masers and cesium beam clocks, rely on the natural resonant frequencies of atoms to tell time with extremely high accuracy.

All of the atomic clocks continued ticking through the power outage last week thanks to their battery backup systems, according to NIST supervisory research physicist Jeff Sherman. What failed was the connection between some of the clocks and NIST's measurement and distribution systems, he said. Some critical operations staff who were still on site following the severe weather were able to restore backup power by activating a diesel generator the team had kept in reserve, Sherman said.

How a Power Outage In Colorado Caused US Official Time To Be 4.8 Microseconds

Comments Filter:
  • by geekmux ( 1040042 ) on Tuesday December 23, 2025 @02:17AM (#65876757)

    But when a destructive windstorm knocked out power to the National Institute of Standards and Technology (NIST) laboratory in Boulder on Wednesday and a backup generator subsequently failed, time ever so slightly slowed down.

    Its not that I’m chastising NIST backup capability as much as I’m shocked that is the steps it takes to hobble one of the country’s most critical services.

    Then again, what exactly was the impact? Good time to bolster recovery processes if there is a measurable one I guess. Doesn’t happen very often.

    • Re:Cause and Effect. (Score:4, Interesting)

      by martin-boundary ( 547041 ) on Tuesday December 23, 2025 @03:24AM (#65876835)

      I am actually shocked. As claimed by the article, it's not the clocks themselves that were compromised, it's actually the network outage that caused the problem.

      That makes sense, because the time protocol requires estimating routing delays, so when the outage happened there may have been a change to the estimated delays in the network. And that seems frankly ridiculous, if true, since there's no reason that the network relaying the atomic clock readings should be tampered with on an ordinary operating day.

      So the fact that the reference clock is off by this much either suggests a flaw in the protocol (trying to do something that doesn't make sense, like re-estimating the delays continuously through another route) or a flaw in the system design (does the computer which averages the 16 atomic clock ticks actually have an atomic clock itself? It would be stupid to use a cheap commercial time crystal as a fallback in case of a complete outage).

      4 microseconds is an eternity in some applications. The idea that it is so easy to tamper with the reference time by messing with the network is unacceptable. NIST has some work to do here.

    • I'm not sure if this is related, but I know that during those windstorms Boulder intentionally cut power to many areas in order to reduce risk of wildfire if a power line went down. This is in response to a devastating wind-driven wildfire last year. I'm wondering if some of the NIST clocks were subject to this.

      • Yes. I still get NIST alerts: it was the wind, campus was closed as a result. Iâ(TM)ve also seen the clock (well, its late-aughts version) in person, and was told that they âoesteerâ time, which is what I think this 4 microsecond difference stems from: the steering wasnâ(TM)t happen so there was drift from the gold standard reference
    • by necro81 ( 917438 )

      Then again, what exactly was the impact?

      The NIST timekeeping makes it way into GPS. In 4.8 microseconds light can travel 1439 m - nearly 1 mile.

      That is not meant to imply, at all, that GPS is suddenly going to be off by 1439 m (compared to ground truth, or other GNSSs). But there is some chance of an error propagating.

      That said, GPS satellites have their own precision timekeepers, which kept ticking away unaffected. Mostly for GPS we care about timing differences between satellites, anyway. The satellites' clocks are only periodically

      • by ceoyoyo ( 59147 )

        There are two parts to GPS data. There's the time signal from the satellite as well as the ephemeris data that lets you calculate where the satellite is. The former is from an on-board atomic clock but the latter is from ground observations that undoubtedly use one or another NIST time standards.

      • GPS uses coordinated UTC not NIST, NIST contributes to that so at worst their is a chance of some fraction of that error propagating in. Also as you mention the source is just to set them to be in sync with each other, so unless they were being synced with different sources at different times it should not be an issue.
  • by spatialguy ( 951355 ) on Tuesday December 23, 2025 @02:17AM (#65876759)
    It will probably take a a few years to catch up
  • by fahrbot-bot ( 874524 ) on Tuesday December 23, 2025 @02:30AM (#65876773)

    How a Power Outage In Colorado Caused US Official Time To Be 4.8 Microseconds

    4.8 microseconds what? Longer, shorter - wider, thinner? Unless... (*gasp*) Time is now 4.8 microseconds. /s :-)

  • The USA really is going back in time compared to the rest of the world. And here I thought that was just a joke we made about Trump and his policies.

  • by hadleyburg ( 823868 ) on Tuesday December 23, 2025 @03:26AM (#65876839)

    "Freedom time"

  • I just love that a bunch of people immediately just went to work when the power went out.

    "How's clock number five?"

    "Not sure, but eleven and twelve are holding. We seem to be losing synchronization."

    "We can't lose NIST UTC!! Don't we have that generator in the back?"

  • by Terje Mathisen ( 128806 ) on Tuesday December 23, 2025 @04:39AM (#65876899)

    NIST have always been the world leader in creating ever more accurate clocks, the current masers work in picoseconds and below, so allowing the reference to drift by 4.8 microseconds means that precision dropped by at least 6 orders of magnitude.

    If allowed to propagate to the GPS control clocks, this would have been enough to totally destroy the navigation system since a clock that is off by 4.8 us corresponds to a position error of 1500 kilometers. (OTOH, USNO has its own large ensemble of atomic clocks, so they don't depend short term on NIST updates.)

    Full disclosure: I worked with the NTP Hackers (network time protocol) team for 20+ years, so I'm probably a bit more interested in precise timekeeping than most. I have personally soldered together 4 or 5 GPS-based reference clocks that would deliver 25-35 ns RMS precision.

    • by AmiMoJo ( 196126 ) on Tuesday December 23, 2025 @06:47AM (#65877025) Homepage Journal

      It's an interesting scenario for GNSS receivers. Most are multi constellation these days. They will have one of three or four constellations giving them a very different time and location.

      I'm sure the better ones test for this scenario, but even there the proof is the real world application.

    • by russotto ( 537200 ) on Tuesday December 23, 2025 @09:17AM (#65877201) Journal

      It appears most of the high-precision users either detected or were informed that the NIST time was no longer healthy. NTP was affected, a bit more disturbing is that some other high precision users (but not GPS) were affected:

      On Dec 21, 2025, at 2:30 PM, âjeff.sâ¦@nist.govâ(TM) via Internet-time-service Internet-time-service@list.nist.gov wrote:

      Dear colleagues,

      Utility power was recently restored to the NIST Boulder campus. Assessment and repair activity is in progress, but I want to give a brief status update regarding Internet Time Services on the NIST Boulder campus. As usual, status notes per-server will be manually updated here:

      https://tf.nist.gov/tf-cgi/ser... [nist.gov]

      Clocks and time transfer services operated from the NIST WWV/Ft. Collins and Gaithersburg, MD campuses are independent and were unaffected throughout.

      Soon after the last notice, NIST facilities staff stationed on-site started a diesel generator held in reserve and activated a power transfer switch positioned to supply âoesecond backupâ power to the affected laboratory. The period without ac power (due to automatic âoefirst backupâ generator failure after 2 days of continuous operation) was about 2 hours. However, large battery banks kept all clocks and most measurement and distribution chains powered throughout. Additional quick action by NIST facility staff secured temperature control for the most sensitive clocks. We regained some monitoring ability showing that the disseminated UTC(NIST) signal likely did not deviate by more than 5 us (five millionths of a second) and appeared stable. Knowing this, I decided to keep the Boulder Internet Time Servers active until we lost monitoring or some other event caused the time scale deviation to increase significantly.

      To put a deviation of a few microseconds in context, the NIST time scale usually performs about five thousand times better than this at the nanosecond scale by composing a special statistical average of many clocks. Such precision is important for scientific applications, telecommunications, critical infrastructure, and integrity monitoring of positioning systems. But this precision is not achievable with time transfer over the public Internet; uncertainties on the order of 1 millisecond (one thousandth of one second) are more typical due to asymmetry and fluctuations in packet delay.

      NIST provides high-precision time transfer by other service arrangements; some direct fiber-optic links were affected and users will be contacted separately. However, the most popular method based on common-view time transfer using GPS satellites as âoetransfer standardsâ seamlessly transitioned to using the clocks at NISTâ(TM)s WWV/Ft. Collins campus as a reference standard. This design feature mitigated the impact to many users of the high-precision time signal.

      Best wishes,

      -Jeff Sherman

    • Agreed. Don't forget synchronization of power grids, telephone networks, and so on.
    • Why doesn't NIST have backup NTP servers in other parts of the country, where the weather is likely to be different and the same issues won't necessarily strike. From what I understand, there is one in Hawaii, but how about a few others - in Florida, Pennsylvania, Washington state and Arizona, in addition to Boulder? That way, even if Boulder had lost its power, the other stations would still enable the time to be available globally

    • by necro81 ( 917438 )

      a clock that is off by 4.8 us corresponds to a position error of 1500 kilometers

      1439 meters I think, not kilometers. (299792458 m/s * 4.8e-6 s)

      Still, a large amount for navigation.

    • How could this affect GPS? I was under the impression that GPS satellites have their own atomic clocks onboard and that any correction (relativistic or otherwise) is done on the receiving end. One can't just "update" a GPS clock from the ground can they?

      • by ceoyoyo ( 59147 )

        You also need to know where the satellites are. That's sent out in the ephemeris, which is derived from ground observations.

        You can also update GPS satellite clocks and this is done fairly regularly. The clocks in the satellites aren't nearly as good as ground based clocks and also encode the entire flight history of the satellite.

  • Ok, some clocks were disconnected. But why haven't the remaining clocks been accurate enough? Shouldn't even a single clock suffice? It sounds like the time is only accurate if an average of all clocks can be calculated.
  • Sorry for the rhetorical question.

    • you mean: Do slashdot editors not even read their

      or possibly: Do slashdot editors even read

      orrrrr.... is slashdot using LLMs to write the
  • by DrunkenTerror ( 561616 ) on Tuesday December 23, 2025 @08:06AM (#65877091) Homepage Journal

    i knew something felt off this past weekend.

  • by Teun ( 17872 )
    Isn't that (very) close to GMT?
    I assume the US would be greater again by using TUT, based in NY. OK, the orange man would not understand the implication of changing a world-wide standard but Trump Universal Time just sounds better.
  • Those servers are so overloaded that I remember having over 60 seconds of clock skew on my servers at some point (detected on my customer's machines due to 2fa codes failing).

Suburbia is where the developer bulldozes out the trees, then names the streets after them. -- Bill Vaughn

Working...