Tsuneo, your work is great!!

Some years ago I also had to calculate the log, but of a division. So, I subtracted two logs. I used a very similar methode, the differencies were due to the different range of result of division. Also, I had to take a somewhat bigger look-up-table.

Sometimes it's also important to have a good monotonicity, means, that the result can take all possible values, so that a slowly changing signal will not present "lost values" on the display (e.g. "2.34", "2.35", "2.37", "2.38" for a very slowly rising signal). The very interesting result of my calculations and simulations was, that increasing the resolution of numbers not necessarily minimizes the monotonicity errors. I think it mainly has to do with the numberous and unavoidable rounding errors. As for my application good monotonicty was of superior interest, I finally had to calculate with a slightly coarser resolution (which was still precise enough, of course!) than I could have.

So, I would recommend to do a monotonicity check, when the result of your calculation is directly presented to the inclined operator via display.

Kai