The pointing model includes elevation corrections for the temperature of the antenna; both the absolute temperataure, through the TEL parameter temp_mean_slope, and the front-to-back differential, through the parameter temp_slope. The temperatures involved are read from probes buried in the (thick) antenna legs, and they have relatively long response times. There are circumstances, then, when even this duplicity is insufficient to correct for rapid thermal changes. An example is shown here :

Note that the only significant systematic in the pointing is in the plot of del -vs- time, at centre-right; the elevation residual varying strongly with time in the early evening (prior to the normal start of the evening shift), and 'relaxing' afterwards. This phraseology should not prejudice the search for the underlying cause, but I am asuming that the likely uncontrolled parameter affecting pointing is the temperature. What we have available is the antenna leg temperatures and we shall see that in this case they fail to provide an algorithmic correction of any significance.

We plot subsets of these elevation data against each version of the antenna temperature and fit lines of the form del = A * X + B in order to derive adjustments to current values of temp_mean_slope and temp_slope. We note the rms scatter of the data about the fitted line and the range of temperatures covered in order to judge the usefulness of the suggested adjustments :

  Data subset     -vs- mean  temperature      -vs- front-back
                     A   rms   range           A   rms    range

    first 6        -37   2.2  5.8 to 6.0       88   1.4   0.4 to 0.5
                 +-  6                      +- 26

    latter 19     -1.8   1.2    4 to 6        -4.5  1.4  -0.5 to 0.5
                +- 0.3                      +- 1.8

'Significant' (>3sigma) trends are seen in 3 of the 4 fits, although in all three cases the ranges of temperatures covered are too small to validate their adoption. (The suggested changes are also rather excessively large and indicative of their uselessness).

Conclusion
It must be accepted that the antenna leg temperatures cannot always correct elevation pointing for severe thermal changes.