for D0 sample : files are :

/eliza3/starprod/embedding/2007ProductionMinBias/D0_[700,705]_20095002/P08ic.SL08c/st_physics_8125012_raw_*.MuDst.root 

(1780 files)

for D0 bar sample : files are :

/eliza3/starprod/embedding/2007ProductionMinBias/D0_bar_[700,705]_20095003/P08ic.SL08c/st_physics_8125012_raw_*.MuDst.root 

A. number of events : yellow area is a cut at |Vz|<10 cm

Unfortunately the primary vertex resolution is empty for all the files, so we cannot apply the same cuts as we applied for the real data ( primary vertex resolution in Z < 200 microns)

D0 sample :

before cut : # 24082

after cuts : # 17438

D0 bar sample :

before  cut : # 24606

after cut :# 17823

So the 2 samples make ~ 50k events before cut.

B. comparison of the number of silicon hits

The issue is that the distribution of silicon hits (3 for SVT , 1 for SSD) differ from real data ( dead wafers) and in simulation.

The simulation consist in single D0, geometry y2007g, reconstructed with BFC chain ( fixed vertex)

The real data consist in embedding data , so mostly real data from day 125

distributions are here : real data and simulation

C. Comparison of DCA plots

Distribution of sigmaDca^XY vs. number of silicon hits  : here 

comments :

• large DCA error for tracks with no silicon hits
• tracks with 1 silicon hit ( ie ssd ) have too large error value to be kept in analysis
• there are some slight differences btw simu and real data for tracks with same # of silicon hits (shape)

last point : the difference is better seen for the error vs. Pt : here  

It seems that the radius of first hit on track ( in this case the inner SVT hit ) is slightly different from simu and real data ( different geometry description)

radius of first hit on track for svt=3 and ssd=0 : here   

radius of first hit on track vs azimuthal angle for svt=3 and ssd=0 : here

comments :

• there is definitely a difference in the radius of the first hit : between 6.5 and 7.5 cm, the simulation shows [almost] no hit whereas the real data have a bunch of hits
• the phi dependance is interesting because we see the radius of the different pices ( rdo = inactive material in simu, but active in real data and silicon sensors = active material)

With a cut at radius < 6.4 cm, the 2 plots match fairly well : here 

Applying the same cut (r <6.4 cm, svt=3 hits) also gives a single band for the error of DCA in 3d (sqrt(dcaXY^2 +dcaZ^2) vs. momentum : here

D. geometry issues

Below are the plots for the radius of the first hit (SVT inner layer and SSD) vs. its azimuthal angle and the radius distribution.

The geometry in simulation was y2007g , SL08f lib.

The geometry in real data was y2007g (B2007g --> ry2007g --> y2007g) but uses SL08c lib.

SVT : here

SSD : here

June,7h  : update for the geomety plots

I was taking the azimuthal angle of track, instead of taking the azimuthal of hits.

Now there is no more broadening, but the SSD still presents the wave structure (alignement ?)

SVT: here

SSD:here