> Your method of using the sun to determine where south is will be inaccurate > except on the four days in the year when the sun's shadow crosses the > vertical axis of the analemma, because of the equation of time.
No, I'm sorry, this is not the case. The method described by Rolf is well known as the Indian Circles. It yields the south dircetion by means of the suns altitude (the sun always culminates in exact south when one is north of the northern tropic). The method described by me of course takes into account the equation of time, because the position in the horizontal system (azimuth and elevation) is derived from the suns coordinates in the equatorial system (rectascension and declination) and the sidereal time. To calculate the suns equatorial coordinates Kepler's equation is used which describes the position on an ellipse in a 2-body-problem. > My suggestion would be to ask one of your geological friends to borrow his > Brunton compass, then subtract off the magnetic deviation from north found > from an OS map. This will be accurate to about one or two degrees. This > error will not affect the accuracy of your dial as much as the fact that it > does not take account of the equation of time. In front of buildings the magnetic deviation can change significantly because of iron used for construction. > If you really want an accurate declination of your wall I suggest that you > measure it on a starry night. Make a rectangular sighting board. Place one > edge against the wall and look towards the west up the other edge until a > known star is occluded by the end of that edge. Note the exact GMT when > this occurs. Repeat for a few other other known stars. That's just quite the same what I described. It is also possible to take the time when the wall is in side-light (of the sun). - Daniel
