Created on Oct. 27
Update on Oct. 29


Created on Oct. 27

 I looked at the Run15 cosmic run production data produced in Sept. 2015. See Lidia's message here

http://www.star.bnl.gov/HyperNews-star/protected/get/starprod/665/1/1.html   (RFF)
http://www.star.bnl.gov/HyperNews-star/protected/get/starprod/665/1/1/1.html   (FF)

Production chain options:
 "DbV20150911 pp2015a mtd btof StiHftC pxlhit isthit NopxlIT NoistIT NosstIT BEmcChkStat CorrX -hitfilt KeepTpcHit"
PXL/IST hits are stored, IST hits are pretty clean even with no latest masking.

Here are some findings:

	RFF	FF
Residual Difference betwen TPC projection w.r.t. the IST hits
DCA position difference between two TPC track DcaGeometries (up_leg - bottom_leg)

They clearly show a difference the TPC projection in Z direction bteween two legs. The split is about 1 cm (or E,W +/- ~0.5cm). This could indicuate either the T0 or drift velocity calibration is not correct in this production.

(There could be a RFF/FF difference in Dca_XY difference, which may be due to other static calibrations.)

Update on Oct. 29


The above issue has been recognized by the TPC group, see the discussion thread here
http://www.star.bnl.gov/HyperNews-star/protected/get/startpc/1548/1/1.html
Per Yuri's suggestion, I re-ran some RFF cosmic production with DbV20150927. Here are the results.

	IST - TPC	IST - TPC (zoom in)	PXL2 (outer) - TPC

Once can see the E/W 1cm split is gone. Given the current statistics, E/W difference (if any) should be within a couple of hundre microns. To pin down precisely, we would need more statistics (probably the full sample). But I think this is not critical given the difference is well within the projection distribution band.

One may also notice that the dZ mean values are not centered at zero. After looking through the Run15 HFT geometry tables, I believe this is understood and HFT geometry tables will be updated soon.
1) For IST, the reason is the ids2tpc table so far for Run15 is an old version of the ids2tpc table obtained from Run14 calibration. The final Run14 calibration shows about 0.8-1 mm difference in dz from this table. We should update the Run15 IST ids2tpc table to be the final Run14 calibration one. Then the residual should be very close to zero, which should be expected as IST/TPC didn't move from Run14->Run15.
2) For PXL, as the whole detector was extracted out and a new detector was inserted for Run15. We initialized the Run15 all geometry table up as identity tables including the ids2tpc which from Run14 shows a 1mm shift (basically PXL center w.r.t the TPC center). This is also consistent with the distribution seen in the most right plot.

In addition, I also looked the DCAs bewteen two TPC track-legs for a single cosmic event. Here is the plot showing the comparison with the Run14 outcome.

Run15 Cosmic Ray	Run 14 Cosmic Ray (Calibration as of 01/29/2015)