A Community Devoted to the Preservation and Practice of Celestial Navigation and Other Methods of Traditional Wayfinding
From: Frank Reed
Date: 2018 Sep 28, 16:53 -0700
A NavList lurker pointed out this interesting story...
The Geosat Follow-On mission was launched 20 years ago in February, 1998. It experienced a failure of its primary on-board GPS system. Its three redundant backup GPS receivers also failed, supposedly because they were mere duplicates of the flawed primary. Quoting this URL:
"The GPS receivers onboard were not fully operational (the GPS receivers failed in spite of 4 unit redundancy; this meant loss of primary orbit determination system and the loss of precision time tagging), and therefore orbit determination now relies on satellite laser ranging and Opnet Doppler."
The mission was salvaged only because an auxiliary (non-GPS) means of positioning was still possible - a backup to the backup - that utilized completely different technologies and methods, namely laser-ranging and Doppler off retro-reflectors.
The primary GPS and the backups all failed apparently because this was a new flavor of GPS navigation involving orbital speeds. It was a common bug in the algorithm. The backups shared the same software, which could not be easily replaced on-orbit. Potential Lessons: The backup should not be an exact duplicate of the primary, in case there is a systemic flaw with it (e.g., the backup GPS might utilize a different chipset from a different manufacturer). And a non-GPS-based means of navigation at a tertiary level of backup can have merit in extreme circumstances.
Thanks to the aforementioned lurker for pointing out this story!