Fix float/double inconsistency in spin/jet model

One of the auxilliary functions, j_BH, which calculates the angular momentum of a BH given its spin, was using fabsf instead of fabs while taking in an argument of type double, as mentioned in #885 (closed). This fixes the inconsistency. Note that the function still outputs a float, because j_BH is never low or high enough to require a double in internal units (unlike some of the other accretion-related quantities).

Closes #885 (closed).