27. CAST Detector Lab variations and fitting by run Appendix
The main difference between how the CAST detector lab data is treated in (Christoph Krieger 2018) and in this thesis is that here we treat each data taking run of the CDL data as one unit, instead of all data for one target/filter combination. This is mainly due to the strong detector variation in the gas gain during the CDL data taking. As such it is not possible to fit the spectra when combining all data from all runs for a single target/filter combination. In the extreme cases (\(\ce{Cu}-\ce{Ni}\) for example) this leads to visibly two main peaks in the charge spectrum. Each run is fit separately. The one question about this approach is whether the cluster properties are stable under gas gain changes. Sec. 27.1 shows that this is indeed the case.
27.1. Influence of gas gain variations on cluster properties
The following figures are ridgeline plots of all relevant cluster properties as introduced in sec. 9.4.2. For each plot and each property all CDL runs are shown as kernel density estimations. Outside the number of hits and total charge in a cluster (which are expected to vary with gas gain of course) the properties remain stable even in the cases that vary strongly.
