Intro :

The quality of SSD hits is calculated as the ratio of the sum of strips fired by the GEANT hit over the charge of the reconstructed cluster.

One possibility of why quality of hits may be greater than 1 could be how the charge from GEANT is converted into electrons, then ADC values.

The charge of GEANT particle are shared between strips (on P and N side) according few criterias :

1. ionisation
2. diffusion of holes and electrons
3. coupling between adjacents strips on a same side

1. and 2. are simulated using a parameter sigma_part^2 = sigma_ionisation^2 + sigma_diffusion^2

3. depends on P or N-strips and has 2 values, according the strips ionized is on the left, or on the right.

(see thesis of Fabrice Retière for details)

In the code, we find the charge on the main strip.  

Then the charge of the 3 others adjacents strips are calculated.

Once it is done, we-recalculate the small amount of charge induced by those neighboring strips to the main strip.

This could be an explanation of why I see that the sum of the charge collected by the strips is higher than the initial charge of the GEANT hit (Fig 1-a)

Fig 1-a : ratio between the sum of the charge of fired strips and and the charge of GEANT hit

Fig 1-b : ratio between the sum of the charge of fired strips and and the charge of GEANT hit when removing induction from the neighboring strips

 

 

In this example, I have 103 hits from the ssdHit table (GEANT).

We see that all entries are not here, because my histos have a range [0.5 ;1.5].

There are actually 1 hit where this ratio is greater than 1.5

Next figure (Fig 2-a) is the number of SSD hit according to each GEANT track (using the trackTable)

Fig 2-a : number of SSD per GEANT track

From 362 GEANT tracks, I have actually :

• 259 tracks without SSD hits
• 101 tracks with 1 SSD
• 1 track with 4 SSD hit
• 1 track with more than 5 SSD hits

Fig 2-b : number of SSD and SVT per GEANT track

 

Fig 2-b-2 :  number of SSD and SVT per GEANT track with and without a cut on TPC hits

cuts :

• |eta|<1.0 : for both with and without the cut on the number of TPC hits
• N_{TPC} >15

 

Fig 2-c : number of SVT(SSD)hits assuming SSD(SVT) hits per tracks

Fig2-c shows :

• (blue) : number of SSD hits per tracks when this the current track has more than 3 SVT hits
• (red)  : number of SVT hits per tracks when this the current track has more than 1 SSD hit.

So there is no correlation between tracks having more than expecting silicon hits (1 for SSD, 3 for SVT)

 

Fig3 : p_{T} of Tracks

 

 

Fig-3 shows the p_T of tGEANT tracks :

• (black) : all tracks
• (red) : for tracks having more than 3 SVT hits
• (blue) : for tracks havind more than 1 SSD hit (actually the p_T is not defined for this case, e.g p_T = -999) 

 So tracks with more than expected silicon hits seems not to be low p_T tracks

 

Following is the position of the vertex used in this event.

The question is to know why we have more than half of the tracks without a SSD (SVT) hit(s) :

StPrepEmbedMaker:INFO  - StPrepEmbedMaker::Do(gkine 10 8 0.100000 5.000000 -1.300000 1.300000 0.000000 6.280000 -1.528597 -1.508597;);

StPrepEmbedMaker:INFO  - StPrepEmbedMaker::Do(vxyz -0.830740 0.237042 -1.518597);

primary vertex set to  -0.830739975  0.237041995 -1.51859701

StPrepEmbedMaker:INFO  - StPrepEmbedMaker::Do(vsig 0 0;);

 

II Parameters set in the simulation

sigma_part,sigma_ionisation and sigma_diffusion are set according this tab :

This tab show results from tests of wafers manufactured by Canberra and Eurysis Companies.

The values used in simulation are the ones from Canberra.

From the db :

 

mysql> select parDiffP,parDiffN,parIndRightP,parIndRightN,parIndLeftP,parIndLeftN from slsCtrl;

+--------------+--------------+--------------+--------------+--------------+--------------+

| parDiffP             | parDiffN             | parIndRightP     | parIndRightN      | parIndLeftP        | parIndLeftN       |

+--------------+--------------+--------------+--------------+--------------+--------------+

| 0.0012300000   | 0.0009400000   | 0.0210000000   | 0.0260000000   | 0.0130000000  | 0.0100000000   |

+--------------+--------------+--------------+--------------+--------------+--------------+

 

 III Charge sharing 

The charge from the GEANT hit is shared using a diffusion factor and a induction factor (due to strip-to-strip capacitance).

Step 1 : the GEANT particle cross a wafer

Step 2 : it ionizes 2 strips : the one which is the closest to the GEANT hit should collect the main part of the charge. (This strip is found by taking the local position of the hit and by converting  it into strip Id).

The opposite strip collects the remaining charge.

a is a coefficient depending on parDiffP and parDiffN (different ionization factor for P and N-side strips)

Step 3 : the strip-to-strip capacitance : each strip will induce a fraction of their collected signal to their neighboring, depending if they are on their right or left.

Values from previous Tab. are used

Step 4 : Finally the 2 first strips have also to ionize each others.

 

Fig 4 :