[time-nuts] GPS week rollover

[Martin Burnicki]

Steve Olney wrote:
> Just curious - WNRO is not an issue for me even if my GPS receivers are
> affected - but I note that the time of rollover is being quoted as
> 23:59:59 UTC 6th April.
> 
> Wouldn't the time actually be 18 seconds earlier ? - i.e., 23:59:42 -
> due to leap seconds ?
> 
> BTW - I'm not clear about GPS time vs UTC (which is probably relevant to
> the above).  Any pointers to a clear explanation ?

The explanation is simple, IMO.

When the GPS system was put into operation in 1980 then the GPS time
matched UTC.

However, a number of leap seconds have been inserted into the UTC time
scale since, and each leap second delayed UTC by 1 s.

Yet 18 leap seconds have been inserted, so UTC is 18 s behind GPS time,
or the other way round, GPS time is 18 s ahead of UTC, so the WNRO
occurs 18 s before UTC midnight.

BTW, the effects of the rollover have already begun. For example, the
reference times of the almanac and clock correction parameters in the
the navigation messages received from the satellites refer to a time a
in the near future, so the week number of those time stamps has already
rolled over.

Martin