<html><head><meta http-equiv="Content-Type" content="text/html; charset=utf-8"></head><body style="word-wrap: break-word; -webkit-nbsp-mode: space; line-break: after-white-space;" class="">Dear all,<div class="">together with Mariangela and Luca, we investigated the run 1349, where apparently the rate of anti-coincidence events with Etot>100 MeV was higher than expected, i.e. higher than the average of the other runs.</div><div class="">I attach to this message a PDF file with the measured waveforms for these events (only beam-off period). For each event, there are 4 PDF pages, each reporting the 16 waveforms from a FADC. The name of each histogram is:</div><div class=""><br class=""></div><div class="">h_runNumber_eventNumber_slot_channel</div><div class=""><br class=""></div><div class="">As you see, many of these events are due to some electrical noise - i.e. these are clearly fake events, that get reconstructed in a wrong way. We think that there are two possible ways to proceed.</div><div class=""><br class=""></div><div class=""><ul class=""><li class="">The first option is to look, a posteriori, to all the events with Etot>100 MeV and no veto activity and verify if they are real events or noisy events. Since the two topologies are TOTALLY distinct, there is almost no risk of introducing bias in the analysis.</li><ul class=""><li class="">This method was used, for example, in the original E137 analysis.</li></ul><li class="">The second option is to implement a filter in the reconstruction code BDXreco, that gets applied before waveform integration, calibration, ecc…</li><ul class=""><li class="">In principle, this method is perforable, since it guarantees that there is no bias at all. Any algorithm can be applied, from deterministic ones to ML-based.</li></ul></ul></div><div class="">A similar effect is also present in other runs, with less intensity - i.e. run 1349 is affected by more noisy events, but these are present also in other runs - therefore, we would need to check all the runs we taken.</div><div class=""><br class=""></div><div class="">@Marco: can you check that this message was also sent to the new colleague who expressed interest in joining BDX efforts? The filter implementation can be a good task to be involved!</div><div class=""><br class=""></div><div class="">Bests,</div><div class="">Andrea</div><div class=""></div></body></html>