Alt-Az Scan comparisons

Alt-Az scans, for LUT building and Pointing Models, were taken on 2012/10/02,
2012/11/20, 2013/01/26 and more recently on 2013/03/25.

Below are plots comparing these 4 Scans for the Donut measured dx,dy,xt,yt

The zero point for each hexapod degree of freedom for each scan has been individually
set using the average of those images taken at Zenith. (Also for cognoscenti note that
these are the measured values, not the correction values.)

You can see that there is excellent reproducibility for xt and yt - the Serrurier truss
is doing its thing consistently. However, the decenter has varied considerably between
these scans, with changes of up to 3mm between last October and this March. In addition,
there does seem to be a trend for larger decenters as a function of time.

I did check the upper truss temperatures for these images. These were:
20130325 15.4
20130126 15.9
20121120 15.3
20121002 8.5
so no obvious dependence on temperature.

To my mind this is further evidence that the top end is moving in systematically over time
with respect to the primary.

One addition question we could ask is: has the optical wavefront changed. My standard donut analysis summary
plot addresses that, comparing each wavefront against the reference wavefront from 20120914. Those plots are also attached for images 153563 from 20121120, and 190407 from 20130325. Each aberration surface is offset and rotated
in the comparison against the reference, and the residuals shown in the 2nd and 4th rows of the plots. The image from Nov compares extremely well against the Sept. reference. The March 2013 image does show some modest changes
against the Sept one, mostly in a slope of the Coma, something previously noted. There is also some modest increase
in the RMS of the residuals indicating some shifts in the exact shape of the surface.