This is the first shot from the analysis of averaged laser tracks in order to understand the situation with statical and dynamical TPC distortions. This work has been initialized by Hejdar Agakishiev who proposed to use averaged over run laser cluster positions.

The data sample : it was used the following laser runs (local clock, no beam): 2003 Full Field (FF),  2003 Reverse Full Field (RFF), 2003 Zero Field (ZF), 2012 RFF, 2012 Reverse Half Field (RHF), and 2012 ZF. The data has been processed with all statical distortion correction included (Corr4 bfc option). The inner to outer sector misaligment corrections were switched off. The laser tracks were reconstructed as straight lines (charge = 0 => charge * mag. field = 0).

The cluster averaging allows to improve cluster resolution by ~ order of magnitude.  From two figures below we can see that diviation from straight line is 

• in pad direction resolution  has σ = 80 and 165 μm for inner and outer sectors, respectively.
• In time direction resolution has σ = 70 and 120 μm;
• 

   Below it is shown results of  straight line fits (separaterly for inner and outer sectors) for sector 4 in two projection xy  (dx is deviation from line, x - pad direction, y - row direction) and zy (z - drift direction). The fit was done by so called robust method (see TLinearFitter), which allows to reject up to 25% of bad measurement, in order to make the fit more stable fit.

For both 2003 and 2012 ZF data have very low efficiency for cluster reconstruction (cluster are too wide).    

The difference in slopes between Outer and Inner sectors for 2012 (RFF + RHF)  has demonstrated a dependence of side of TPC (west - east).

The difference in slopes between Outer and Inner sectors for 2003 (FF + RFF)  has demonstrated a dependence of side of TPC (west - east) and direction of magnetic field. (I don't understand why RFF for 2003 and 2012 have opposite sign. I will check it.)

Conclusion: 

• The first shot with averaged laser cluster still looks very promising. The cluster resultion has been improver by a order of magnitude.
• Comparison direction of inner tracks to outer ones from the TPC Central Membrane shows strong ExB effect which is not accounted by the static distortion corrections. 
• The reason for this note is appeal for attention to the problem. For me it is hard to believe that ~1 mrad direction off set between inner and outer sector due to ExB was missed during the analysis. Thus it could be a misinterpretation of the data. 
• The next step is to see  what is going on with other laser tracks. The most important is to see Z - dependence of this offset.