I : Selection of good wafers

Goal : 

As mentionned in the collaboration meeting, a better estimation of the tracking efficiency may be obtained by looking at the wafer level, instead of evaluating an overall efficiency.

To do this, I will select only the good wafers, based on how the cluster reconstruction is.

A table (ssdWaferConfiguration) was created in order to do the reconstruction only in particular wafer (by default, it is done in all wafers)

Then I get a list of wafers to remove.

1. Look at the raw data
2. Select the number of raw clusters for which 2 < noise < 13 (Fig 1)
3. Evaluate the ratio of this number with all reconstructed clusters : ratio = #clusters (2<noise<13) / all clusters
4. if (ratio<0.5) for at least 1 side, remove this wafer
5. rerun the bfc with this waferConfiguration 
6. reevaluate the overall efficiency

Files used :

raw data : /star/data03/daq/2005/051/st_physics_6051015_raw_1040003.daq (February 20th)

pedestal file : ssdStripCalib.20050220.090439.root

bfc chain : P2005b,DbV20070518,MakeEvent,ITTF,ToF,ssddat,spt,SsdIt,SvtIt,pmdRaw,SCEbyE,OGridLeak,OShortR,OSpaceZ2,KeepSvtHit,hitfilt,skip1row (used  for P07ic)

starver SL08a

Fig 1 : <noise> per wafer for ladder 1 side P

Fig 2 : Distribution of the total signal of reconstructed clusters

• black : all reconstructed in ladder 2 side P
• red : all reconstructed clusters in ladder 2 side P but with 2<noise<13 ADC
• green : all matched clusters (used to identify hits)

With this cut, we see that the distribution of all clusters looks like the distribution for the matched clusters. This means that the cluster reconstruction efficiency may increase in this case.

Fig 3 : distribution of ratio (see above for the definition)

 Then the list of wafers for which ratio < 0.5 for one or the other side is obtained

Fig 4 : Distribution of the total signal of reconstructed clusters

This plot is the same as Fig 2 but with the list of 'good' wafers only

We see that now the distribution of the clusters than we can reconstruct match pretty well the matched clusters

II Efficiency : detail per wafers 

I showed at the last collaboration meeting the efficiency vs phi for a given number of runs (CuCu@200GeV).

Fig a : Efficiency vs phi 

As phi is the angle in the transverse direction, this plot describes the efficiency for a given ladder.

Then we find back ladder 7 and 8 off (-1 < phi < 0) and others ladders where the efficiency seems to be lower. 

To get a better overview, I did the same but also for the eta direction, which now describes a given wafer.

Then I get some kind of map in (eta,phi) ~ (wafer,ladder) where we can see which wafer is decreasing the overall efficiency

Fig b : efficiency (z-axis) vs phi (y-axis) and eta (x-axis)

Fig c : the same as Fig b. but in 3-dimensions