NavList:
A Community Devoted to the Preservation and Practice of Celestial Navigation and Other Methods of Traditional Wayfinding
Re: Instumental error?
From: Bill B
Date: 2005 Apr 22, 16:02 -0500
From: Bill B
Date: 2005 Apr 22, 16:02 -0500
Fred wrote: > I find it hard to detect both images, even with > sun semi diameters, and I have yet to figure out what's up in that > regard. In a perfect world, we want to eliminate any affect of backlash (gear slack) by always turning the knob in the same direction. Which means (when measuring Sun diameter) one tangency is bringing the bodies together, the other telling when they separate. In comparing notes with Alex, we both noted that our separation (leave) STDEV were significantly better than our touching tangency STDEV. My tripod-mounted separations are in the 0.16' STDEV neighborhood, and the touching STDEV in the mid 0.34' neighborhood. Since my ability to find any effect of backlash with power lines when tripod mounted (a linear target, and approx. plus/minus 0.1' actual deviation) is better than my my nominal ability to see discs (0.15 to 0.2'), I just go ahead and use separation for on-the-arc and off-the-arc Sun diameters. Not cricket and bad for a plastic sextant, but for the moment it works for me and seems to make sense from a numbers standpoint. It is difficult for me to find the sweet spot and overlap between the glass and mirror for sun diameter shots--a real cat & mouse game. Both discs have to lie in the center of the vertical "gray area" both horizontally and vertically. Even then, just a twitch one way or the other and the brighter overlap is gone. More so for a quarter moon and Sun, even with a shade over the moon. With two sun images I find it helpful to get a rough alignment by placing both images on the glass side and using the reflection of the mirror image off the glass before taxing they old eyes with the fine adjustment. For moon and star/planet I prefer bringing them together. Remove the scope (a trick Alex passed on to me), and get rough alignment. Put the scope back on and use Frank's method. Look, move the drum 0.2'. Look again. Which if I get it quickly works well. If it takes me five or more tries, things go to heck in a hand basket quickly. My left eye turns a pinpoint into a fuzzy oval on end, my right a fuzzy oval on it's side. Keep going and I start seeing two star images. Tried my glasses and it solves that problem, but makes aligning the eye to the center of the scope an exercise (often in futility) on circus-people-contortionist shots. Watering eyes from 30F temps and 15 mph breezes didn't help the February/March shots much ;-) I debated whether to post the following (don't want to be a tease) without a link and reference to the study, but after 45 minutes of surfing I cannot relocate the information. I may have read it somewhere, but believe it was on David Burch's site: http://www.starpath.com/index.htm It was buried somewhere as a link in one the plastic sextant > using PDFs. One of the links appears to be no longer active. I do not recall if he named the source of the study, but do recall I was unable to locate the study via internet at the time. The upshot was a study indicating observational accuracy improved from 0 to 1000 observations. Surprise ;-) It continued to improve from 1000 to 2000, and again from 2000 to 3000. Truly wish I had been able to view the data, and apologize for hanging out bait for something that might be of interest. I would imagine it would reflect a standard learning curve, approaching a point of diminishing returns, but that is only conjecture. If any of the list members are familiar with the study and have access or a URL, please share. At any rate, it appears to be a long road ahead for me. I again highly recommend his article on fitting multiple observations to the Hc slope (and identifying possible outliers) for a time period) as opposed to using linear regression to fit the slope to the data points). http://www.starpath.com/online/celestial/sight_average.pdf > If you can get the standard deviation of Ho-Hc under 0.5' > of arc on dry land, that's good enough for use at sea, especially > on small vessels. I would like to get to a level where I can touch discs consistently at a 0.3' level or better. Have tried a few experiments with touching the Sun to a power line of roofline, and it seems easier. I have not yet done any controlled experiments and run numbers. Even if it does prove to be easier, on land I am not bobbing around and trying to catch the top of a wave. My water experience with my cardboard unit tells me the whole game changes in a light displacement boat in 6-8 foot waves. Bill