Noise run analysis

1. General Information

During this run the ZS threshold was set to 3σ.

The reconstruction was performed with AliRoot v4-14-Release. The official OCDB files were used from AliEn. These files were produced from previous calibration runs.

The analysis of the SSD clusters revealed that we have the following number of golden clusters over all the events for the two runs:

2. Results from the run 55633

Number of clusters per module/number of events [%]

Module number with clusters

Local coordinates x-y-z

Global coordinates x-y-z

Cluster type - Charge ratio - Cluster charge

Cluster charge (P-Side) - Cluster charge (N-Side) - qN/qP

Cluster charge map - Cluster charge correlation

3. Results from the run 55634

Number of clusters per module/number of events [%]

Module number with clusters

Local coordinates x-y-z

Global coordinates x-y-z

Cluster type - Charge ratio - Cluster charge

Cluster charge (P-Side) - Cluster charge (N-Side) - qN/qP

Cluster charge map - Cluster charge correlation

4. Results from the run 55635

Number of clusters per module/number of events [%]

Module number with clusters

Local coordinates x-y-z

Global coordinates x-y-z

Cluster type - Charge ratio - Cluster charge

Cluster charge (P-Side) - Cluster charge (N-Side) - qN/qP

Cluster charge map - Cluster charge correlation

-- PanosChristakoglou - 02 Sep 2008

Frequency distribution points per module/per event (golden points only)

There is kind of an exponential distribution. About 90% of the modules have no points (as it should be)!

I also looked at x,y distributions in modules. In many cases, most hits are again at 1 or two strips in a module. So it seems that by masking out a limited number of strips, we could clean up this distribution. The question is whether this set of strips is stable in time.

It would be interesting to look for non-gaussian tails in pedestal runs.