Page History

...

This figure shows one such mapping.  It is the polar coordinate of the ACD, as seen from a point 600mm below the Z=0 plane of the LAT.

   Image Added
 
Given x,y,z the transformations to the new coordinates (u,v) are:

    double dist2 = x*x + y*y;
    double dist = sqrt(dist2);
    double zVal = z + 600.;
    double rad = atan2( dist , zVal );
    double u = rad * ( x / dist );
    double v = rad * ( y / dist );
 

 Then we divide the u,v plane into 20 segments in each direction, making 400 possible bins.

At the start of the reconstruction job we pre-compute the look-up table.   To build in tolerance we increase the size of all the detector elements by some amount (say 200mm), then we scan across the detector element in 10x10 steps (100 points per detector) for each point we get the (u,v) bin the point falls in a add this to the overall map.  Technically speaking the map is

 std std::map< AcdRecon::AcdUVBin, std::list< idents::AcdId > > >
 

 Then for each track in each event we

...

1. Continue to calculate and fill the AcdTkrHitPoca for all combinations of tracks and ACD tiles/ribbons with signals
2. Store the new information with the existing AcdTkrHitPoca structures in the recon file.  Those structures are designed to store all the distance information.
3. This means that the list of AcdTkrHitPocas will include both cases where the relevent detector has a signal and those where it does not.
4. Add new data fields to the AcdTkrHitPoca structure to store the MIP equivalent signal seen in the ACD detector element.  This will allow us to distiguish quickly between tiles near the track that have signal and those that don't.
5. Update AnalysisNtuple to ignore those AcdTkrHitPoca's that have no signal when filling merit variables, this will insure that the merit variables are identical to what we have now.
6.  Improve the logic in AnalysisNtuple to fill new merit variables that take into account signal size, distance of closest approach and track quality to find the tile/track combination most likely to VETO the event
   
   Here's a copy of the ACD Tile Map for posterity...