F5 Fit¶
Note: some of the following text was taken from Walter Ogburn's doctoral thesis.
Overview¶
The ionization channels are also fit with an alternative, time-domain routine called “F5”. Like the OptimalFilterChargeX routine, this is a joint fit of the Qinner and Qouter channels. The fit for each detector uses four templates: one each for the inner and outer electrodes, and two to represent crosstalk between them. These are the same templates used for the OptimalFilterChargeX routine.
The fit has five parameters: the amplitudes of the pulses in the inner and outer channels; the pre-pulse baseline in the two channels; and a common start time for both pulses. Of these five, three are not actually free parameters of the fit, but are calculated beforehand: the pre-pulse baselines are found by simply averaging the part of the trace before the pulse starts; and the start time is taken from the OptimalFilterChargeX fit result.
Saturating ionization pulses (those above 1 MeV) are analyzed only with the F5 algorithm. Although not used for WIMP searches, the saturated pulses are important for studying backgrounds from radioactivity and muons at energies above ~1 MeV. Because the optimal filter requires an accurate template, the distortion of saturating pulses produces a poor fit with the optimal filter routine. In the case of saturating pulses, another start time must be used. For these pulses, the start time input to the template fit is found by walking backwards from the first saturated sample of the trace to find the last sample that's within five standard deviations from the pre-pulse baseline (this sentence needs some verification).
Future Improvements / Extensions¶
The F5 was originally ported from Darkpipe into BatRoot in 2008. It runs on ZIPs and mZIPs, but needs to be updated to run on iZIPs. It also has yet to be modified to work on detectors that have only 1 charge channel (such as endcap detectors).