NavList:
A Community Devoted to the Preservation and Practice of Celestial Navigation and Other Methods of Traditional Wayfinding
Re: sextant precision.
From: Bill B
Date: 2005 Jun 23, 17:45 -0500
From: Bill B
Date: 2005 Jun 23, 17:45 -0500
Bill you wrote: "My contention is that a optically perfect shade with faces parallel will still offset the ray if not perfectly perpendicular to the ray path. You are probably correct that the offset is negligible in practice, but nonetheless does exist." Frank responded: "So experiment! Get a piece of flat clear glass, and hold it in the light path. Rotate it. Tilt it 45 degrees to the light path. Then try perpendicular to the rays. Does it change the alignment of direct and reflected images in your sextant? There's nothing like experimental evidence to constrain one's theoretical musings. Here's another: No sextant involved. Hold a piece of flat glass close to some text on a page at normal reading distance. Rotate it. Does the text appear to shift from refraction (it should)? Now hold that same piece of flat glass up in front of some distant scenery. When you rotate the glass as before, do objects shift back and forth (they shouldn't)? TRY this experiment and see if you agree with my should/shouldn't comments. Then work out the theory... Evidence constrains theory." Frank Tried the "no sextant" version of the experiment you proposed. I used a 1/2"-thick piece of float glass with polished edges (I generally us it with sandpaper adhered to it to lap plane blades and bottoms, and chisel backs). The glass thickness near the edges was checked with a micrometer to verify that they were parallel. I then viewed a table edge at about 3 feet, a sidewalk edge about 10 feet away, a rooftop about 100 yards away, and later the moon. I used both the left and right vertical edges of the glass, then flipped the glass end for end and repeated for each viewing. The glass was rotated from approx. perpendicular to the line of sight to approx. 15d to the line of sight. I attempted to keep the position on the glass where the images were equidistant from my eye, and used the dominant eye only. I preface the following with the caution that there were no reasonable measurement practices in place. Displacement was judged by eye, and the angle of the glass to the line of sight was hand held and at best approximate. Trial 1, Wednesday eve In all cases, from the table edge to the moon, there was marked displacement of the line/object halves viewed through the glass vs. the direct view (table edge, sidewalk edge, and moon) as the glass approached a severe angle (closer to parallel than perpendicular) to the line of sight. The moon halves were displaced by about 1/4 of its diameter (approx. 7') as the glass reached a severe angle (approx. 15-20d off the line of sight). Gauging displacement by the difference between the glass and air images by distance along the glass edge, the amount of displacement appeared to be about the same for far and near objects. I again caution that angles were approximate, and foreshortening was compensated for by eye. Trial 2, Thursday afternoon Same methodology as trial 1, but paid closer attention to the final angle of the glass in relation to the line of sight. A number of objects from 3 feet to 200 yards were observed, including a table edge, patio edge, roof lines, and power lines. In these observations the displacement appeared to diminish with distance, but was still easily visible at all distances. All I can safely conclude is that refraction is still with us (thank goodness, I have a lot invested in camera lenses ;-) Currently I am considering how I can use photo clamps, articulated arms, tripods etc. to set the angle of the glass to the line of sight consistently for subjects at different distance, as well as a method to measure the offset. So at the moment, my muse, the "should" does, and the "shouldn't" does (to a magnitude related to the "should" yet be be determined). Theory will follow experiments with greater controls in place. Bill