Grahame,
That's a nice little experiment. The error may be due to the fuzziness
of the cancellation.
I once helped my boss and his son measure the speed of light with an old
laser pointer and a fiber-to-Ethernet converter's photodiode module,
using a telescope and a mirror in the neighbor's window. It was accurate
to a few tenths of a percent!
--
David Forbes, Tucson AZ