Bob Axtell cotse.net> writes: > > What's the energy penalty for pointing the PV array off Sun- > > dead-center, at say 1 arc-second, 1 arc-minute, etc...? > its not very bad for the standard PV arrays now being made, but more=20 > severe for the new ones which require a high degree of accuracy. I think that structural issues and drive issues will prevent such accuracy from being reached using the clock-only method. So use a reasonably accurat= e clock and adjust the final precise focus using direct optical sensors. This is how I did it, the clock points the output only in 6 minute increments, the final precision pointing is done optically, and only if so fitted, sinc= e 6 minute accurate pointing is plenty enough for flat panels and skylight passive reflectors (my application). And it's 6 minutes because I am cheap and 86400/60/6 < 256. The Sun moves about 0.25 degrees across the sky in 6 minutes. If you need to be that accurate you will likely have other problem= s besides knowing what clock to use. Building a quartz based clock which can stay 5 minutes accurate over 6 months or so is trivial and costs a few $ fo= r a tcxo at most. Also, moving motors only once every 6 minutes saves a lot o= n motor and mechanical wear. In fact, I had it decimated to move less frequently for that reason. Then, again, the Sun itself will be a better clock than any low cost Earth based quartz thing when integrated over days or weeks. You can set your clock by the Sun (literally). I did look into that with the project alluded to above, and the setting worked on paper, but it was not worth implementin= g and testing the necessary filters. Once implemented, however, it would have integrated the sun zenith data over a few months and worked out its own continuous error and much more, such as the analemma phase and so on. My code ran in a 12F675 and had a lot of flash space to spare when done. 6 months of observation every 6 minutes yields some 20,000 data points in daylight, excepting in Winter-only, at average latitudes, assuming 1 degree resolution in observation using a simple cheap sun sensor, one can expect t= o resolve 1/sqrt(20000) degrees from that eventually which is about 0.00707 degrees and also 2 seconds in Sun arc motion terms. Probably plenty enough for anything not astronomical deep space telescope pointing. As an aside, I read that ICBMs used and might still use star pointing cameras as backup navigation systems. now, that would probably mandate such an accurate clock and sun tracking (which can easily be repurposed as star tracking) initial setter, the final precise setting and lock acquisition being, again, done optically, by the camera itself. I read what was available about this when I did the 12F675 based simple tracker. Not much, but enough :) Speaking of telescopes and initial pointing, pointing a telescope with 1 degree FOV with 0.25 degrees of accuracy using a 6 minute accurate clock should allow confortable target star acquisition imho... but none of the mechanical drives I used with my tracker came anywhere near 1 degree poiting accuracy anyway... they were low cost simple pointing drives= .. -- Peter --=20 http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist .