Hi everyone, thanks very much for the comments (and sorry for not answering every question)! Meanwhile we found the reason. =) Special thanks to Frank.
@ John: Yes, we use a crystal (Dallas-chip) RTC which works fine. When I query it directly for getting the current time it is always correct (no delay after many hours). But since access is (too) slow (I2C) we use that mentioned "trick"... So the problem must have been on the target.. It was a wrong CPU I/O bus speed, it is not 266 but 264 MHz. Since this value is used for calculating the real-time clock period we got 66500 instead of 66000. This was exactly the range of the time drift. So everything was fine with interrupts. (The timer should be a bit more precise since our machines can have uptimes of months or more - and that delay would lead to an error of nearly 2,5 days per year.) Thanks again! cu Iris -- Iris Lindner Software Development Industrial Print and Apply Labelling www.Logopak.com -- Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss
