|
|
From: Bill (no email)
Date: Thu Apr 20 2006 - 01:06:12 EDT
Alex
Thank you for a thorough look at the scopes.
> Remarks. Observations with |4SD-true.4SD|>.4 are discarded
> as SNO-T manual recommends.
As you know, I have some problems with 4SD comparisons where SD is taken
literally from the almanac. Discarding observations where |4SD-true.4SD|>.4
has to take into account mental interpolation of the almanac SD to be
workable IMHO.
For example:
Date SD 4SD
March 28 2006 16.1 64.4
March 29 2006 16.0 64.0
April 21 2006 16.0 64.0
April 22 2006 15.9 63.6
A change of nominally 0.005' in reality can cause a change of 0.1' SD in the
almanac. If identically observations are taken a day apart near a rounding
up or down, one might pass the test and the other not. If we interpolate
around March 28/29 we have a more realistic 4SD of 64.2 to compare our
observations against.
In the worst-case scenario using the|4SD-true.4SD|>.4 rule around March
28/29, +/- 0.2' error is already introduced into 4SD if using the almanac
figures directly. That leaves 0.2' operator and sextant error for two
tangency observations, or < 0.1' error per tangency. That's a tough mark to
hit, and otherwise usable observations might be discarded (false negative).
On the flip side of the coin--if it works the other way, some very sloppy
observations might pass the test resulting in a false positive.
Bill
|