What I don't understand is why such accurate timing is required. Is you sample data only valid every 60 seconds exactly? If so, then you need to be cafeful about the thing executing every 60 seconds. If not, then 60.5 is not so bad. Time is arbitrary. Some guy a long time ago said "Well, I think a second should be this long", and it has stuck. We've standardized it since then, but it's still arbitrary. Just something to think about...