Positioning of Wires on a Star

The two graphs below illustrate the comparison of micrometer measures which I made (observed measures) with accurate measures of the same stars made with speckle interferometers and by the Hipparcos satellite and referred to below as the reference measures. When making these comparisons it is vital that the epochs of measurement agree as closely as possible, otherwise the comparisons are not valid due to orbital motion (or proper motion) during the interval.

Figure 15.6a shows the differences between the observed and reference separations. In this case the sense is (observed-reference) so that for the closest pairs (below about 1'' or so) the measured separations are too large. This is not an uncommon feature of measurement by micrometer and it is particularly useful for anyone doing orbital analysis. Whilst the raw measures are published as they stand, in the case of a particularly careful orbit calculation, it pays to try and assess the "personal" error of the observers and then to apply correction to the observed positions. In practice this tends not to happen much because suitable refer-

Figure 15.6. a The error of a mean separation with separation. The solid line represents a running average. b The error of a mean position angle with position angle. The solid line represents a running average.

Figure 15.6. a The error of a mean separation with separation. The solid line represents a running average. b The error of a mean position angle with position angle. The solid line represents a running average.

ence measurements have not been available for comparison. This has changed for the better recently with the publication of the speckle results from the USNO (see the reference in Chapter 25) where suitably accurate and up-to-date measurements are available to enable observers to check their personal equation.

There is a large scatter at larger separations and this is due to a combination of the paucity of standards at these separations and fewer measures which I have made. Of the points in Figure 15.6a some 210 pairs below 2'' are compared, dropping to 69 pairs between 2 and 4'' and only 31 pairs between 4 and 10''.

What can be seen from the graph in Figure 15.6a is a tendency for me to measure the closest pairs (0.5-2'') as rather wider (about 10%) than they really are and from about 2 and wider there is not much systematic error to be seen.

In Figure 15.6b the graph shows the situation for the observed position angles for the same pairs as Figure 15.6a. Here there is clearly an anomaly at about PA 180°. This is where the two stars appear nearly vertical in the eyepiece. Although it is recommended to place the eyes either parallel to or at right angles to the line between the stars it is more uncomfortable to do the former so I conclude that using the eyes parallel to the line between the stars results in an error in position angle of about -0.5 to -1° when the stars are within about 40° of the vertical. Another way to avoid this is to use a prism in conjunction with the eyepiece to allow the field to be rotated by 180°. By making another set of position angle measures here the mean value should then be free of this particular bias.