|
"Time is God's way of keeping everything from happening at once." --Anon.
How accurate can a "sundial" be? This is, of course, a subject that we have given some thought to. There are several components to this question, and the answers we have arrived at. The first component is the type of time we are measuring. Heliochronometers and sundials, by definition, measure "mean solar time" or "solar time". With the advent of long voyage navigation, another standard was clearly needed, 'Greenwich Mean Time' (GMT) was instituted in the late 17th Century, continued until the late 1920's (with some variations) and was then designated 'Universal time' (UT). Soon, it was noted that there were variations in UT. To address these variations, scales named UTO, UT1 and UT2, were introduced. See the links section for a more comprehensive discussion, including the variants UT0R, UT1R and UT2R. UT1 is the most important and usable for navigation, though there are tidal harmonics that must be corrected for, that vary the earth's rotation on a daily basis. UT1 also corrects for a slight wobble (about 15 meters) of the earth's axis as it spins. Confused yet? Sidereal time can also be used. This is time based on observation of an astronomical object sufficiently distant to make the slightly irregular orbit of the Earth insignificant. Very bright, distant stars are ideal candidates. In the late 1950's atomic clocks were introduced. I won't bore you with the machinations from then until 1972, but one modification included several different seconds that were not 1 second long! In 1972, the current UTC system was settled on, with "leap" seconds, of standard weight and size, added as needed, to keep things within reason (they are added at around 15 month intervals, on average). There are also leap years, in any set of 4 years at least 3 to 4 second variations are possible between UTC and Solar time, this is after the standard EOT is applied. * In order to smooth out the variations, our analemmas are averages, derived from solar observations over several decades. The second component is the actual "repeatability" of the geometry of the instrument. The relationship of the sight, analemma, vernier scale and hour ring dial. Through experimentation, we have determined the most critical aspects of our geometry are the central pivot holes. They must be perfectly perpendicular and on center relative to the hour ring dial. These holes are drilled, then bored in both parts during the same operation, as are the locating holes for the ring dial. The allowance for the "stack" or true position tolerances for these critical components is +/- .003" (.075mm). This along with the clearance for the pivot bolt translates to about +/- 5 seconds, worst case.
The third component is the width of the solar disk coupled with prevalent atmospheric conditions. The solar disk covers an apparent .5 degree of the sky and this accounts for 2 minutes. The use of a shadow sharpener or pinhole aperture, using the correct focal plane, gives a fairly focused solar image. By splitting this focused image with the proper part of the analemma, the problem of the width of the disk is minimized, as well as the penumbra effect, which is a combination of this width, along with atmospheric effects. Of course, atmospheric conditions can render the best solar instrument marginally or completely unusable. On a clear day with low humidity and the sun well above the horizon we would assign a +/- 5 to 7 second window for this component. The eyesight and judgment of the individual user is the other governing factor. In conclusion; a well made heliochronometer, of reasonable size (smaller than, say, a Honda) , set up perfectly, should return around a +/- 15 to 20 second accuracy. Here at 32 North latitude, we have bettered this with our instruments, but it takes occasional tweaking to maintain anything much more precise. Barring "exotic" technology (ccd's, mirrors, prisms, mirrored prisms, resolvers, computers etc.) or very large size, this seems a practical limit. The Indian observatory / sundial at Jaipur claims better than 3 second accuracy, we feel our products, while not quite as accurate, have a substantial advantage in price and portability, not to mention shipping and setup costs.
* A side note of interest here is that the recent tsunami in SE Asia seems to have "sped up" the rotation of the Earth, we may have to subtract a "leap" second. This is due to the tremendous weight of materiel that has fallen towards the center of Earth's mass, much like an ice skater in a spin, when their arms are brought closer to the body. |
Note: The above information applies to the Aten models. The Classic uses captive ball
bearings, running in V-grooves under pre-load, to locate the pivot.
