See my post over on the other site:
http://www.cleanmpg.com/forums/showt...?t=5747&page=6

It isn't a limitation of the math on the Scangauge, it is an error in the 2's complement math. Now that I know the error, I have communicated this to Ron so maybe he can fix it. Also, the error factor is really 1999 not 1995.

You are absolutely correct regarding the number of bits however, it is definitely 16 bits. I just couldn't see why it turned into 1995 (or 1999) but after playing with it for a couple of hours I finally found the problem.