I made this excel file that should be able to compute the bits for you.
You should only edit the yellow cells to suit your needs, then start with bit 0 of UMCTL0 and choose either 1 or 0 so that the absolute error in percent is minimum. Continue down with the following bits.