MINOS (Main Injector Neutrino Oscillation Search) is an experiment currently running in the US. A beam of neutrinos is created at Fermilab, Chicago, measured in the 1 kiloton 'Near Detector' and then travels 730km to the 5 kiloton 'Far Detector' in the Soudan Mine, Minnesota. In the intervening time, it is hoped that some of these neutrinos will change from one flavour to another. If this is observed, it is strong evidence for neutrino oscillations, the parameters of which can be measured to 10%. The MINOS experiment is a large project with a huge number of technical issues. Many aspects of the experiment were tested several years before the main experiment itself began to run, by employing a scaled down version of the detectors, known as the Calibration Detector (CalDet). This was placed in a test-beam at CERN and extensively studied, the data from which is analysed in this thesis. In this thesis, photomultiplier tube crosstalk is discusseed, a phenomenon which generates false signals in the MINOS detectors. It is studied and an algorithm pre sented to enable its removal. Particle identification via various methods at CalDet is also described. Various pieces of hardware are available to assist with this, and a comparison is made to software techniques which are used at the larger MINOS detectors. A study of the CalDet beamline simulation is carried out and the discrep ancies with data highlighted and explained. Finally, muon energy loss in CalDet is investigated. A comparison is made between published data and the observed data. NEMO-3 is an experiment that has been running for some time in the Frejus tunnel between France and Italy. It is a Ov(3(3 experiment, hoping to show that the neutrino is a Majorana particle and set limits on its mass. This experiment, like every, has backgrounds. The dangerous background signals that arise from the radioactive decay of Uranium and Thorium are discussed in this thesis, specifically the measurement of the quantity of 208T1 and 214Bi in the source foils of the detector. This is achieved by using Monte Carlo simulations of the contaminants behaviour in the detector, developing cuts on these events and applying them to the dataset.