How come you didnt just multiply it by say .5 from the original signal (or whatever the appropriate value would be) so that then if altering the original signal in the chip the dwell on the coilpack could have been altered too?

Surely the way you have done it now makes it impossible for a mapper to optimise?