I would post the solution but it basically avoids using any integer maths or floats at all. I instead use the detach interrupt function, so unfortunately it has nothing to do with my initial question of getting floats to work. If you still think it would be suitable, I could always post it.