>What i see with my application is that only 1 of these 4
>positions will work correctly. Not easy to explain the fault
>but in the other 3 positions I can see on the TV monitor
>that the vertical graticule i generate will have a kink in
>1 or two places of 100ns (1 instruction Cycle).

Are you sure this is not a 1 instruction variation in the interrupt latency?
How can you be sure.?

--
http://www.piclist.com hint: The list server can filter out subtopics
(like ads or off topics) for you. See http://www.piclist.com/#topics