data : 

• pedestal : /star/data03/daq/2014/ssdped/15154028/st_pedestal_adc_15154028_raw_0000001.daq
• physics : /star/data03/daq/2014/156/15156024/st_physics_15156024_raw_1000011.daq

The offline code is doing (per wafer) :

1. scan strips
2. if signal/rms > 5 (of the current strip), current strip is chosen as a cluster seed
3. scan the left and right of the cluster seed in order to add neighboring strips to the seed --> cluster

goal : check first the raw distribution of signal, s/rms to see if s/rms>5 is suited

A) raw distribution after pedestal and zero subtraction : 1) per ladder 2) per wafer 
Fig.1  : <signal> per wafers


B) <pedestal> and <rms> per wafer 
comments :

• side P has <rms> ~3 ADC and increases for the wafers on the 2nd half of each ladders
• side N has a higher <rms> ~ 5 but seems quite homogeneous
• the increase in <pedestal> along a ladder is shown here (for example only, ladder 1)

Fig. 2 : <pedestal> per wafers


Fig. 3 : <RMS> per wafers


C) signal/rms distribution : last page has the summary <signal/rms> per wafer side P and N : here
comments :

• <signal/rms> side P ~ 3
• <signal/rms> side N shows a split for all second half ladders (form wafer 8 to 16) ; 1st half <signal/rms>~ 12 ; 2nd half : <signal/rms>~ 3 (similar to side P)

Fig. 4 : <s/rms> per wafers


D) test
the change was to increase the cut on the cluster seed search to 10 for the N-strips side, P-side strips cut is unchanged (5). Initially this cut is set to 5 for both sides.

(code is in StSsdWafer::doFindCluster())

The next 2 plots show the charge matching before and after the change for 2 different wafers : one without a deviation from the perfect charge matching (Fig. 5) and the other with a deviation in the charge matching plot (Fig. 6)
Each figure shows the charge matching before (after) change on the left (right)
Fig. 5 :

Fig. 6 :

The other wafers are there