> If the GPS system goes down, getting lost is the least of your problems. I > can think of only two scenarios - hostile action (probably nuclear war) or a > Carrington event. In any case over the next few years we will be talking > around 100 GNSS satellites from several independent systems. GPS/Glonass > receivers are commonly available and GPS/Glonass/Galileo/Beidou/QZSS are > increasingly available. >
One of those rare times that I pipe up on something here these days. Here’s evidence of such a double failure in recent times, due to neither of your scenarios, in a real aircraft with a lot of internal redundancy: http://www.bushflyingdiaries.com/2013/10/double-gps-system-failure.html And I have another GPS failure mode to relate. It amounts to ‘human error in running the GPS network, triggering a software driven example of poor choice in system specification' For a week or two, a year or two back (I’d have to dig up the dates) the entire fleet of Pilatus PC-12’s in Australia had non-functional GPS systems, due to what turned out to be a stuffup in the configuration of a satellite on the northern edge of the Australian region, that started presenting invalid SBAS data that the PC12 Honeywell GPS systems decided was an attempt to subvert the GPS system. In response, in accordance with the (then) software specs (i.e. not a bug, but an intentional feature), the GPS systems on board shut down on the basis of not being able to trust the data being received. It arguably should have just shut down SBAS reception and reported that and kept right on going otherwise - but that’s not how the system designers had specified the outcome in the presence of bad SBAS data. It took Honeywell about a week to identify the cause and come up with the (obvious) workaround - disable SBAS at each aircraft start (its on by default) - until the Satellite itself got fixed (and it did then get fixed). In that week or so in the middle, the entire Australian fleet of PC-12NG’s (including mine) had no onboard GPS that worked. Amusingly, the less sophisticated GPS in my iPad kept running with AvPlan just fine (redundancy, redundancy, redundancy) The washup here is that another scenario exists in addition to hostile action or a Carrington Event - human error in the operation of the GPS system itself, combined with GPS receivers that are too smart for their own good (arguably also human error - in terms of specification of how they’re to work in the presence of doubtful input data). The root cause here appears to have been a botched configuration change on a *production* satellite somewhere in the vicinity of the equator. Simon _______________________________________________ Aus-soaring mailing list [email protected] http://lists.base64.com.au/listinfo/aus-soaring
