Getting Started with DSTs

DST format:

•    Int_t     EventNo;         // event counter 

•    Int_t     RunNo;           // run number (!)

•    Int_t     MustardCnt;      // mustard event number 

•    Int_t     tdc;             // tdc counts (monitoring delay between trigger and next clock rising edge 

•    Float_t   ScanValue;       // scan value (generally threshold)(!) 

•    Int_t     OKFlag;          // general purpose flag (currently used to flag out-of-sync events) 

•    Int_t     NTracks;         // Number of tracks reconstructed (upto 4) 

•    Int_t     NTelHits;        // Number of clusters in analog devices (up to 100) 

•    Int_t     NBinHits;        // Number of clusters in binary devices (up to 1000) 

For each of the reconstructed tracks, the following information is available

•    Float_t   x[4];            // track X value at Z=0 

•    Float_t   y[4];            // track Y value at Z=0 

•    Float_t   gx[4];           // track dX/dZ 

•    Float_t   gy[4];           // track dY/dZ 

The cluster information of the telescopes is also available (to allow for cross-checks and custom alignments)

•    Int_t     plane_id[100];   // Identifier of plane, following the old naming scheme, ie 16021 is telescope 2, plane 1 

•    Int_t     width[100];      // number of channels above clustering threshold 

•    Int_t     signal[100];     // total signal in cluster 

•    Float_t   xlocal[100];     // position (in cm) of the center of gravity of the cluster, in the local frame 

Finally, for each binary cluster:

•    Int_t     plane_id[1500];  // identifier of plane (ie 13110 for module 11, plane 0) 

•    Int_t     time_bin[1500];  // time bin of the cluster 

•    Int_t     chan[1500];      // lowest channel in cluster 

•    Int_t     width[1500];     // number of channels involved 

•    Float_t   xlocal[1500];    // position of cluster in local frame 

Hits only

Some users might be interested in the (binary and telescope) hits only and prefer to do their own alignment. Those know all they have to know and can generate their analysis macro skeleton by executing the following ROOT commands: TFile f("mydst.root"); DST->MakeCode("mymacro.C"); The result should look like this . On execution of the macro, a warning will appear:

Warning in : Unknow class: rootTrack. Cannot read BranchClones: Tracks

Warning in : Unknow class: outBinHit. Cannot read BranchClones: BinHits

Warning in : Unknow class: outTelHit. Cannot read BranchClones: TelHits

Warning in : Cannot find class:outEvent

This can be ignored, the macro will continue without any problem.

Using the tracks

Those who want to make full use of the DST - including the tracks - need the alignment parameters to be able to extrapolate the tracks from the global system to each DUT's local system. An example of a macro that opens the DST, reads in the alignment parameters and extrapolates the track is available here .

Advanced DST use

The code generated by MakeCode is flexible, but not very elegant. In the offline program (DST) events contained in a class that consists of some simple types and TClonesArrays of 3 types of objects: tracks, binary hits and telescope hits (or analog hits in general). The same class can be used to load a whole event from the DST into one "Event" object. This method is not as flexible as the previous, but results in far more compact code. The class definitions can be loaded into ROOT as a dynamic library. Source code, a Makefile and a macro filling a hitmap are available:

• class definition sources etc
• macro

Summary Ntuple

We are working on a summary Ntuple that will contain a number of important parameters/results for each plane and run. Most likely these will be bias voltage, threshold, angle and magnetic field for the run parameters and efficiency, noise, average multiplicity, etc with their errors. In the first place this is a convenient way to represent all results (rather than generating hundreds of PS files). Secondly, it might be a useful tool to get a quick insight into the general behaviour of X as a function of Y and thus facilitate a more profound analysis.