Preparation for W-production in December 8, 2009

 A single daq file from every run used for pp500 W analysis was produced with the chain:

"pp2009c ITTF VFPPVnoCTB BEmcChkStat beamLine Corr4 OSpaceZ2 OGridLeak3D -dstout -evout"

library stardev=SL09g,  BFC production date 12-5-2009, stats=1/7 of run 9 data

Section A)  Chi2/dof for  global tracks

Long global tracks w pt>1 GeV were sampled. 

Fig 1. TOP: # of used tracks per run. run index on the X-axis - details in attachment A)
MIDDLE: Mean chi2/dof. Greay dashed lines marks Y=1.0
BOTTOM: Y=chi2/dof vs. X=run index. Greay dashed lines marks Y=1.0 . Solid histogram marks mean chi2/dof.

Conclusions:
* chi2/dof changes within fills , most likely due to the pileup level
* overall changes in chi2/dof are between [1,1.4], the average chi2/dof=1.2
Question: what is the best strategy for W :
* further improve TPC calibration (new SC & GL) and delay production as long as it takes
* ask Sti (tracker) errors to be inflated by 10% - it will also delay start of production- needs Gene, Yuri, & Jerome acceptance
* recommend start of production, using TPC calib as is (at the end it is just 2 days of the farm CPU)

Section B)  selected plots from  Walgo 

Selected plots from W-algo, full set is in attachment B)

Fig 2. Jacobian peak (1/7 of data)

 Section C)    Comparison of SC GL reference plots

Fig 3. TPC : Left: Q/PT for e+, e- passing all W-cuts. Right: eta-phi uniformity for global tracks with  PT>1 GeV and some hit QA

Conclusions:
* TPC does not look worse than w/ SL09g - it is good enough

Fig 4. SC & GL residua for SL09b (top) and SL09g (bottom). Magenta line marks our nominal goal for W-publication of 300 mu