I am working on this topic under the guidence of Ramiro Debbe.

We are looking weather the efficiency of TPC in reconstruction of tracks is changing or not with the anode voltage. For this purpose we are using the pp data from Run 183 of 2009 with different anode voltage. The different voltages used are:

Inner / Outer

1000 / 1390

1020 / 1390

1040 / 1390

1060 / 1390

1080 / 1390

1100 / 1390

1170 / 1390

Number of primary tracks for each event are shown in the figures below

              Inner Anode Voltage = 1000                                           Inner Anode Voltage = 1020                                           Inner Anode Voltage = 1040

       Inner Anode Voltage = 1060                     Inner Anode Voltage = 1080                      Inner Anode Voltage = 110                      Inner Anode Voltage = 1170

From the figure we can see clearly that the data with anode voltage 1040 and 1060 volt have lots of events with zero primary tracks. At the same time the global tracks have the same distribution and nothing wrong in the data set with voltage 1040 and 1060 volts. The mean number of tracks per event plotted with inner anode voltages for both the global and primary tracks. In the plot of mean number of primary tracks per events with the inner anode voltage I dropped the two points with inner anode voltage 1040 and 1060 volts. We have tried to find the problem with these two voltage data but we could not find anything. I looked for dead sectors with track end points and with the calculated azimuthal angle phi. I also looked at track quality and couldn't find a good answer. We suspected the vertex finder to be connected to this problem but did not find the actual problem.
 The dependency of mean number of tracks with inner anode voltage is shown in figure below :

          Mean number of global tracks per events Vs. Inner anode voltage                               Mean number of global tracks per events Vs. Inner anode voltage

From the figure we can see the flat region from 1080 to 1170 volts with small positive slope that is consistent with what Yuri Fisyak and Ramiro Debbe see with other methods.