GPS week rollover due in less than 13 years

Discussion in 'Global Navigation Satellite Systems' started by ian_okey, Nov 9, 2006.

  1. ian_okey

    ian_okey Guest

    This may seem like a long time in the future, but in the lifespan of
    many electronic appliances this age is just reaching maturity. I have
    a number of calculators, radios and even a computer that are older than
    this and still going strong.

    My question is how many current GPS implementations can be expected to
    fail on 24 March 2019. There is a method, using the leap second count
    value, whereby the GPS engine can estimate which week cycle it is
    currently in but how many implementations is this used. The leap
    second algorithm is described at
    http://www.leapsecond.com/notes/gpswnro.htm

    I am specifically worried by the costs associated with updating
    potentially thousands of units that are embedded in equipment dispersed
    around the world if they are liable to fail in such a short timeframe.

    Ian Okey
     
    ian_okey, Nov 9, 2006
    #1
    1. Advertisements

  2. OT:
    Somebody somewhere told that there will be no more leap second adjustments.
    Those who need it have other means to fix it.
    Can anybony refere to a reliable source??


    Thomas
     
    Thomas Lindberg, Nov 9, 2006
    #2
    1. Advertisements

  3. ian_okey

    Marc Brett Guest

    It's still a matter of some debate:
    http://www.ucolick.org/~sla/leapsecs/onlinebib.html
     
    Marc Brett, Nov 9, 2006
    #3
  4. ian_okey

    KeesC Guest

    Come on, we already had a rollover in 1998 or 1999. While everybody was busy
    tackling the Y2K problem, the GPS rollover passed by without any problems I
    heard of.

    No reason to create panic, you will be able to find your supermarket after
    24 March 2019.

    Kees.
     
    KeesC, Nov 9, 2006
    #4
  5. Besides, by then there will be a whole new generation of GPS technology and
    devices out, and you will have just bought a new GPS (again) that will take
    any rollover into account..........
     
    Pieter Litchfield, Nov 10, 2006
    #5
  6. www.globalslowing.org
     
    John R. Copeland, Nov 10, 2006
    #6
  7. ian_okey

    peter Guest

    Probably very few. The last rollover happened on August 21st, 1999 and
    my GPS was unaffected. I expect it to be unaffected by the next one as
    well in the rather unlikely event that it's still functional then.
    AIRC, there were a few models that didn't handle the last rollover
    transparently, but the most common problem was that they reported the
    wrong date (typically 1024 weeks earlier than actual) while still
    navigating properly. Some did need to collect a full new almanac
    before continuing to provide position data.
     
    peter, Nov 10, 2006
    #7
  8. ian_okey

    pat_n_ed Guest

    Ian,

    Your concern is well founded, based on past history. See:

    http://catless.ncl.ac.uk/Risks/18.24.html#subj3.1

    http://catless.ncl.ac.uk/Risks/19.80.html#subj5

    http://catless.ncl.ac.uk/Risks/20.55.html#subj3.1

    http://catless.ncl.ac.uk/Risks/20.55.html#subj4.1

    http://catless.ncl.ac.uk/Risks/20.62.html#subj11.1


    By my count, the next "rollover" will occur at midnight on 6/7 April
    2019, but I may have counted wrong. If you get a 2 week head start on
    addressing the problem, all the better.

    "Rollover" is in quotes because the interface specification
    (IS-GPS-200D, para. 6.2.4) now says that the week number increments by
    1 each week "without ever resetting to zero". However, the next
    sentence recognizes that the Nav Message "may not reflect the current
    full GPS week number".

    The information in the leap seconds section of the navigation message
    (SF 4, pg. 18) shouldn't help, since it's just the 8 MSBs of the 10 bit
    week number.

    If you're involved with an actual product, here's a story about a good
    business approach to similar problems (see the leap second narrative
    about halfway down the page):

    http://www.5starsnews.com/email-marketing/1115607.htm

    If you have a specific receiver, or a specific embedded application, in
    mind, the best way to predict its performance is by testing with a
    signal simulator.


    Ed M.
     
    pat_n_ed, Nov 10, 2006
    #8
  9. ian_okey

    ian_okey Guest

    There were quite a few problems associated with the last rollover. I
    would expect that there will be a lot more on the next cycle because
    many more applications rely upon an embedded GPS to provide a cheap and
    accurate time reference. Examples are cellular base stations,
    satellite earth stations and remote asset monitoring equipment that use
    the GPS to synchronise their operation. The majority of these units
    are in fixed installations and so ignore the position data provided.

    My concern is that these embedded GPS chipset solutions that are then
    used for capital equipment products may not handle the rollover
    correctly. The typical life of many of these products can be more than
    20 years and will involve one or more week rollover cycles. The
    throw-away consumer handheld or car sat-nav units are not so important.
    Their lifespan is much more limited.

    Ian
     
    ian_okey, Nov 10, 2006
    #9
    1. Advertisements

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.