This Text Inserted from File doc/html/overview.html ============================= RecHitD ============================= author: Stoyan Stoynev, NU This rechit builder uses much of the code (a copy) from RecHitB. The substantial difference is that no Gatti fitting is applied for strip hit coordinate (and error) estimation. However the Gatti distribution is virtually used. Step one: A position estimator called X_Florida is used for initial position estimation. Three charges from the three adjacent strips are the input. These are charges from three time bins summed. Besides, every charge in a time bin is XTalk corrected (see CSCXonStrip_MatchGatti::setupMatrix()). No negative charges are allowed after the XTalk correction! X_Florida = (RightCharge - LeftCharge)/(CentralCharge - min(Left, Right)) Step two: A correction to X_Florida is made. It depends on the strip width and X_Florida itself. These corrections are parametrized in HardCodedCorrectionInitialization.cc. A linear interpolation is applied between points. The corrections are obtained in the following way. The gatti distribution (induced charge distribution) is generated for specific strip width. Then for a given offset (simulating the hit offset from the center of the strip) the charges for the three strips are extracted and X_Florida is calculated. The initial offset is in fact X_Gatti (the maximum of the distribution is at X_Gatti by definition), i.e. the position within the strip as predicted by the Gatti distribution. The correction is Correction = X_Gatti - X_Florida (there are no ambiguities) or in other words - the correction that should be applied to X_Florida in order to obtain X_Gatti. Thus the corrected X_Florida is effectively X_Gatti. In reality various systematic factors make this only an approximation (which is also true when we use a real Gatti fitting!). Step three: The uncertainty of the position is calculated by error propagation. Studies show that three factors play a role: Noise, XTalk asymmetry (or similar effects), constant term. These three factors should be obtained by studies. Note: The theoretical Gatti distribution depends on chamber parameters (of course). For all but ME1/1 chambers the relevant parameters are practically the same. So there are two types of corrections: for ME1/1 chambers and for non-ME1/1 chambers. All are encoded in HardCodedCorrectionInitialization.cc.
Status :