diy 2 axis PV tracker/heliostat - RFC

I would like to get a 2 axis tracker together suitable for pv or heliostat use, as described here:
http://www.solarelectronica.com/
The mechanical mountI imagine as easiest and most stable is an altazimuth (pan-tilt) as in
http://www.astronomyhints.com/mounts.html or http://en.wikipedia.org/wiki/Altazimuth_mount

For PV use you only have to get normal to the sun which is 'easy' in a sense , e.g. using a differential detector made of 4 small photodiodes/sensors.
The more challending part is heliostat, as the system must remember the target direction and
the mirror needs to be put at angles halfway between sun and target. So it needs
to know where the sun is, and remember where the target is.
In principle I'd like to do it without actually calculating where the sun should be , rather sensing where it in fact is. Then you don;t have to tell the system what time of day it is or where it is: calibration is just panning/tilting till the reflection is on the target, hit 'on target' button and that's it.

The guys at solarelectronica.com have solved this incredibly well if their claims are correct: 0.05 degrees accuracy. They seem to manage this with a single photocell...so how do they do it???
Mayb they are calculating sun position instead of sensing it?
From the site:

"The HelioCircuit includes an optical device (patent pending) that measures constantly the position of the sun and the target relative to the heliostat. The internal CPU generates all information needed in order to generate the electrical signals to drive the DC motors of the heliostat (Any Follower in the market with a mirror). "

Created

By:

jeremy