Happy GPS Week Rollover!

It’s probably overstating things to compare the GPS week rollover to the Y2K bug (The Next Web, The Verge) but it’s hard not to see some parallels. In each case, it’s a function of too little memory assigned to timestamps.

One of the things that GPS satellites do is transmit precise timing information. It does this in part by stating the week as a 10-bit integer, counting from week one. That means the date number rolls over every 1,024 weeks. Every 19.7 years, then, the GPS week “rolls over” and a new GPS epoch begins. It’s already happened once: the GPS era started on 6 January 1980, and the first rollover occured in August 1999. The next one takes place—oh dear—tomorrow. Cue the mass hysteria.

How will our GPS receivers respond to that rollover? Because it’s happened before, and because consumer GPS tech doesn’t necessarily stay in use for long periods of time, it’s unlikely that your or my GPS receivers are affected. Anything released in the past two decades would have programmed after the last rollover. Also, receivers might have have had an offset to the 1,024-week limit programmed into their firmware, starting the clock from the date of compilation rather than August 1999, so devices affected may not be affected all at once. U.S. government agencies note that GPS receivers that conform to the IS-GPS-200 specification should not be affected: PC Mag pins that on devices manufactured in 2010 or later.

My own legacy GPS receivers—a Garmin eTrex Legend H and an Oregon 450t—date from just before 2010, and while I haven’t used them in years (when your smartphone has built-in GPS, dedicated receivers are superfluous in the most common use cases), I’m half tempted to fire them up and see what happens. Both TomTom and Garmin claim that the vast majority of their devices are unaffected, but neither go so far as to give us a list of affected devices. Firmware updates are apparently being issued for some of those affected devices—but again, a list would help.

In any event, it appears that using GPS for location, even on an affected device, will not be broken: at worst your tracklogs will have inaccurate timestamps. Receivers that use GPS for accurate timekeeping that have not been updated to handle the rollover might run into some trouble, though. And, like Y2K, any problems might be in industrial or embedded systems rather than consumer tech—and from what I’ve seen online they’ve been getting warnings about this for years.

In other words, as far as I can tell, the GPS world will not come crashing down tomorrow.