Coherent noise source identification in multi channel analysis
The evaluation of coherent noise can provide useful information in the study of detectors. The identification of coherent noise sources is also relevant for uncertainty calculations in analyse where several channels are combined. The study of the covariance matrix give information about coherent noises. Since covariance matrix of high dimension data could be difficult to analyse, the development of analysis tools is needed. Principal Component Analysis (PCA) is a powerful tool for such analysis. It has been shown that we can use PCA to find coherent noises in ATLAS calorimeter or the CALICE Si-W electromagnetic calorimeter physics prototype. However, if several coherent noise sources are combined, the interpretation of the PCA may become complicated. In this paper, we present another method based on the study of the covariance matrix to identify noise sources. This method has been developed for the study of front end ASICs dedicated to CALICE calorimeters. These calorimeters are designed and studied for experiments at the ILC. We also study the reliability of the method with simulations. Although this method has been developped for a specific application, it can be used for any multi channel analysis.