Figure 8a. The cooling of the right, bottom secondary support rod is shown for a night in May of 1996. An accurate time of the enclosure opening is not known, but is thought to have been within 15 minutes of the first data point shown in this graph. The squares show the temperature of the ambient air at the time of the rod measurements. The circles show the rod temperature. An exponential fit to the circles is shown by the line. The fit parameters are shown in the inset. M3 is the time constant of the exponential fit in units of inverse seconds. Converting to hours gives a time constant of 1.03 hours

This time constant is only an upper limit to the true thermal time constant of the secondary rods. As can be seen from the air temperature measurements in the graph above, the driving function of the thermal changes in the rod is significantly slower than a delta function. Note that the rod temperatures follow the air temperatures quite closely. Also note that the time of opening (m4) and the amplitude of the temperature change (m2) are poorly determined by these data. These two parameters are strongly coupled.

The measurements on August 15 were started 1.1 hours after the enclosure was opened. 1.1 hours after the enclosure was opened, these measurements show a rod cooling rate of -1.44 C/hr. In the absence of rod data on the night of August 15, I will treat this cooling rate as typical of the conditions at the time of those defocus measurements.

Figure 8b. This shows the temperature differentials measured between the top and bottom rods on the night of May 7, 1996. These data show that differences on the order of 0.5љ C can be expected while the telescope cools. The solid line shows the temperature differences between the right top and bottom rods and the dashed line shows the differentials between the left top and bottom rods. Note that only half of the rod temperatures were measured at this time. It is not currently know how typical these cooling curves are.