NavList:
A Community Devoted to the Preservation and Practice of Celestial Navigation and Other Methods of Traditional Wayfinding
Re: SNO-T tests
From: Bill B
Date: 2005 Dec 14, 20:53 -0500
From: Bill B
Date: 2005 Dec 14, 20:53 -0500
Frank I had several concerns about your test setup, but have done the math where possible and am convinced that they would amount to only a few arcseconds. My next concern was using a CRT and graphics software. If I design the target with an application capable of 0.0001" resolution, lines are not uniform on the monitor. If export to a bit-mapped application, lines between it's resolution and input could be shifted either way. And can I be sure my graphics card is displaying what I want it to on the CRT? Another concern was that the maximum distance I can get from my CRT is 13 ft. If the display is 72 pixels per inch, a 1-pixel line is 18 arcseconds wide at 13 ft. I decided the best bet was to print a vector-graphics target to a PostScript printer. This allowed me the the luxury of being able to measure the distance between the lines on the print out to 0.005" with a graphic arts recticle ensuring uniform spacing, as well as using a thinner line--at a greater distance. The real plus is I was able to calculate the height of a 2 arcminutes target at 19.5 ft. I had the program create a 20 unit grid (horizontal lines) to that height, and extended the bottom line to the left as a base. I placed numbers from 0 (base) to 20 a bit in to the right--from the left side of the 20-unit grid. This is really nice. As one aligns and rocks the eye, the 0 replaces the number. (I may use degrees and minutes instead of 0-20 in the next version.) Even better, I am tripod mounted, so squaring up to the target is a breeze. The best part: I get a base-line alignment reading, dial up exactly 2 degrees, and if the 2 degree (20) line is not aligned, I can move the tripod forward or backward a smidgen and repeat until I get perfect alignment with a 120 second increase. Then each step represents exactly 6 arcminutes. Really sweet. Very little math to do--each step should be a multiple of base +6. (My base was 359d 20.3' at 19.5 ft, so parallax error of 39.7') My other major concern came from my sextant-as-rangefinder experiments. I made the faulty assumption that the leg of the oblique triangle between a spot on the index mirror and a spot on the horizon mirror/glass would remain the same length. (Triangle formed where the line of sight from the scope pierces the horizon glass, the object, and a spot on the index mirror.) This proved to be wrong with a front silvered index mirror. As a sanity check, once I had the distance of the sextant to target set I aligned the base (0) and 20 line (2 degree) then moved to the 10 line (1 degree) and alignment was still there. The reverse worked as well. So at least in this case of small angular change it does not significantly change readings. Repeatability of aligning the base line is plus/minus 0.1', so that is my level of confidence in other readings. I did one quick run through the 2 degree range and and they all align to predicted within 0.1'. Will let you know how it works out when I have a chance to do several rounds and record the readings. Thanks for the great idea. Bill