Although a satellite dish is a fine instrument, in fact, it is exacting only as well as the installation; that is, if the feed horn is not perfectly point to the exact center of the dish, at the right focal point; all numbers of any other spot you measure it at, like where the dish mounts to the pivot point; will be off as much as the feed horn is off. With bolting anything, you will find, there is almost no tool that can exactly measure the angle within the allotted points it has to measure with. The true declination of a satellite dish, is "the flat plate of the dish"; that is the point in the sky the dish is receiving the signal from, measured at the feed horn and how flat it mounts to the dish. So, if you measure at the flatness of the scalar rings, then the "stringed dish plate" as a completely flat plate; these two measurements should be the same; and then your pivot point measurement, +- the dish plate/feed horn measurement; this is your declination that has been set. On your measurement tool, can you read any angle below one, that is, how fine is it to smaller than one degree? Fine tuning from there will make the declination finer and finer, but only if you move it very little. Same with zenith; how well can you read and compare where the "mount" is pointing. Applying most tactics will get you within a degree or better; which will be close enough to receive the signals from the satellites, but not very perfectly, because every tune must be sung; that is, everything must then be perfected from imperfect "close" settings, so that the entire assembly is within the lowest tolerances of measurement, say within .1 (one tenth of a degree); then you'll find that it works better as a complete polar "ARCED" assembly, and if you can, if you get the tolerances within .01; the wind will push hard enough to make it move off .5 degrees, and it will not matter....And there are not many satellite dishes that are strong enough not to flex, that are tight enough, steel strong enough, pole that won't flex; ETC; to make up for that .1 -> .5 degrees of differences, making it impossible to put a dish within the required tolerance required for reception of every satellite signal. A satellite dish of 6' also has about a 3 degree "beamwidth"; which negates many exactnesses your polar mount is being set to, including declination. This 3 degree's is basically "sloppiness"; and applys to all the satellite points it is trying to pick up; so even if you get the mount to perform within a degree of tolerance, it is still picking up signals from outside of that fine pointing, it does.
So, for now, it is always close by measurement, and exactly tracking Clarke's Arc of satellites is getting it as exact as possible using a polar mounting of a satellite dish, and it will track as perfect as Clarke's arc is using a satellite dish of about 40 foot in diameter (needs more exact arc numbers), smaller than that, it is close, and good enough...True south is another number that is not talked about enough, because truely, south and north is not magnetic, it is "true" only when you add or subtract the magnetic deviation for "you're just a spot on the map"; and deviation of the true south/north line therefore must be determined by moving the compasses reading a little degree or more east or west...