Thinking back, I don't think I've ever worked on an INS whose basic design was later than about 1975 so that may explain the difference between my theory and anybody else's. Having said that, I'm not sure the amplitude (a few hundred metres at worst I believe) of the Schuler cycle is such as to cause unacceptable errors in any INS - it's not used as an instrument landing system is it ! I wonder (and need to get some old course notes out when I've time) whether the issue of Schuler errors, that may need something doing about, in a working INS may be cumulative rather than absolute? In other words that the cycles can lead to a "divergent position error" that could steadily increase to peaks big enough to be a problem.
G
Last edited by Genghis the Engineer; 17th Feb 2004 at 15:45.