It has nothing to do with floating point precision, which you can't configure in Mathcad anyway. Restrict the number of points in Signal to 34115. Change your program so that it returns dSumDen (I'm not sure why you have it return zero at the moment, that seems kind of pointless). You will see that it is -9.152*10^307. If you increase the range to 34116 it fails because you have exceeded the range of the floating point processor. If you return Array from the program you can see that the values grow extremely fast. I think you have a bug somewhere. Richard