With my wife manning the TX this afternoon and pointing it out of the landing window we managed a daylight speech test today with 100mm optics at both ends at 0.3km using the v.simple AM lightbeam TX described in the previous post. The LED was run with an 80mA standing current. Biggest problem was aligning the TX beam and the lack of talkback, but I managed to get 59+ copy of my wife saying, "I'm bored with this. Come home and have some tea" when I was aligned on the TX beam.
There is no doubt the simple circuit can be refined to improve audio quality when clipping, but it works. The RX is far more sensitive at night too. I am now finding a way of replacing the XYL for these tests by using another source of speech such as a radio receiver or MP3 player. She is good at cooking and other things, but not at doing optical tests!
Later this week I'll begin experiments with a simple single LED AM 481THz transceiver.