facing problem using Lewit J-Funtion

I tried J-function approach and got one uniform J-function curve versus water saturation from core data. I generalized J function as J(Sw) = A+B(Sw-SWi)^C. here C is negative. I got values of A , B and C from it. Now I applied the same to my actual permeability and porosity ( corrected for reservoir) data in order to get Sw.I used bristich units and got a good trend of calculated Sw with that calculated by software (powerlog), but my calculatd Sw is a bit shifter by say o.1. looking at data i feel its beccuase of in conistency in units, i tried a lot but still cannot solve
Note for original J ftn calaulation
J= Pc * (k/Po)^(-1/2) where pc = capillary pressure in Psi, K is permeability in mD, Po is porosity in %.

from it i generalized J - ftn as
J(Sw) = A+B(Sw-SWi)^C where c is negative
from it i got one J function representative of whole PC data.
I got A, B and c.

Now i used my original data of permeablility and porosity
calculated J ftn as
J=((Dw-D0)*g*h)*((k/Po)^(-1/2))/surface tension* cos theta.

where
Dw = density of water in lb/in3, g/gc= 1, h = ft ( if i use it in ft then it works but if i use it in inches then not works , I dont understand why!!), surface tension is in lbf/in.

by using value of this j-ftn then i calulated sw as

sw=swi+((j-A)/B)^1/c.


I tried to keep consitent units as was while geting a, b and C.

my question is why it works if i take h in fts and not in inches??