GPS Epoch Issues?

swatson999

Well-Known Member
Joined
Oct 6, 2010
Messages
1,513
Are you talking about the GPS week rollover? That's not an "epoch", but yes, it happens every 19 years or so (1023 weeks).
 

swatson999

Well-Known Member
Joined
Oct 6, 2010
Messages
1,513
Not to get too technical, but there are different meanings of the word "epoch". ICD-200 defines each week as an epoch, and each 1023-week period as an epoch, and the GPS epoch as starting January 5/6, 1980.

But, as noted, all GPS manufacturers are aware of the week rollover at week 1024.
 

jnmeade

Active Member
Joined
Oct 9, 2011
Messages
308
Location
Eastern Iowa
Dynon, thanks for your prompt and useful response.

Steve. Thanks for your further elucidation on the use of the term "epoch". One is always sure to learn something on this forum. It's kind of why I read a lot and post almost never.
 

swatson999

Well-Known Member
Joined
Oct 6, 2010
Messages
1,513
No worries! GPS stuff can definitely be confusing, and the terminology within the system is pretty arcane.

You did raise a good point, though, about the week rollover that I'm sure a lot of people will be asking about in the next few months.

Now, don't ask me why the original system specs only allowed 1023 weeks, instead of using just a few more bits to make it last into the next century and beyond. :)
 

Dynon

Dynon Staff
Staff member
Joined
Jan 14, 2013
Messages
14,222
Location
Woodinville, WA
Well an update on this - we don't know of any issues here, but we're going to be doing a comprehensive engineering review to be sure. In the case that there is any exposure in any of the GPSes that we sell, we expect that we could address them via a software update. More to come as we learn more.
 

thibault

Member
Joined
Oct 25, 2009
Messages
191
Wow,  history really does repeat itself.   25 years ago I was the Software Lead Engineer for the Tomahawk Cruise Missile at Raytheon Missile Sysrems.  We were already looking at potential Y2K issues in every weapon system and GPS week rollover was also a real threat.   We found that the software would certainly fail and the Tomahawk missile would crash if the rollover event occurred during flight.  That situation and all Y2K like problems were prevented in our  systems, but only through the diligent efforts of many SW engineers.

Dyson,  please do a very deep dive into the SW to ensure this does not repeat.   Back in the day,  our hardware based GPS satellite simulators could be set to any date/time desired so they were the proof of the error and the demonstration of the correctness of the SW fix.
 
Top