Look at the performances of the SSD for the test production sample (P08ic)

There are FullField and Reversed FullField LowLuminosity data : 

• FullField : /star/data**/reco.2007LowLuminosity/FullField/P08ic/2007/120/81200[53,54,55,57]
• ReversedFullField : /star/data**/reco.2007LowLuminosity/FullField/P08ic/2007/159/81590[44,45,46,48]

I looked only at /120/8120054/ files (160 files) and /159/8159044/ files (258 files)

cuts of primary tracks:

• |z|<10 cm (vertex)
• TPC fitted > 20
• pT > 0.1
• |eta| in SSD acceptance

 

PART I FullField

Fig 1 : Efficiency vs momentum P 

Fig 2 : Efficiency vs (eta,phi)

The peaks at the edges of the eta range, where efficiency is ~ 100% are due to the fact that at these eta bins, there are really fews tracks.

For example : in this bin (1.2<eta<1.4) I have 14 tracks :

• 13 have SSDFitted Hits = SSDPossibleHits = 1
• only 1 has SSDFitted Hits = 0 and SSDPossibleHits = 1 
• then Efficiency = 13/14 = 93 % 

Fig 3 : sigma DCA in transverse plan :

Marker with TPC only is for tracks with no silicon hits : I should look at the data sample produced with TPC tracking only.

Marker with TPC+silicon(1) could be a SSD hit or 1 SVT hit but there are almost TPC+SSD.

At P = 1 GeV/c, the sigma_dca is :

• tpc : 2400 mu
• tpc+silicon(1) : 850 mu
• tpc+silicon(2) : 320 mu
• tpc+silicon(3) : 260 mu
• tpc+silicon(4) : 210 mu

These values are in the same order or a bit better than the Cu+Cu@200GeV

Fig 4: sigma DCA in Z direction :

At P = 1 GeV/c, the sigma_dca is :

• tpc : 2550 mu
• tpc+silicon(1) : 1650 mu
• tpc+silicon(2) : 500 mu
• tpc+silicon(3) : 380 mu
• tpc+silicon(4) : 310 mu

These values are higher than the Cu+Cu data, especially for the SSD where we had a sigma_dcaZ ~ 960 mu

One possible cause of the values seen for the SSD at lower P may be that FitSlicesY() (used to get the standard deviation of dca) is not correct in those cases since the dca distribution are not pure gaussian.

Following are the dca distribution in XY and Z for tpc+ssd and tpc+silicon(4) for 36 bins in P (delta_P =0.1 MeV/c, from 0.1 to 0.2 from the top left, etc ..  ) :

• tpc+ssd : dcaXY
• tpc+ssd : dcaZ
• tpc+silicon(4) : dcaXY
• tpc+silicon(4) : dcaZ

We see that for the low momentum tracks, the dca in XY and Z are wider than for the high momentum tracks.

PART II Reversed FullField

Fig 1 : Efficiency vs P 

Fig 2 : Efficiency vs (eta,phi)

Fig 3 : sigma DCA in transverse plan :

 

Fig 4 : sigma DCA in Z direction :

 

PART III : varying the eta and phi cut 

On the sample 8120053 from FullField, I changed :

• eta range from |eta|<1.2 to |eta|<0.7
• phi range by selecting good phi regions based on this plot : efficiency vs. phi

I rerun on a smaller datasample with these conditions and plot the sigma_dca in XY and Z.

• open symbols : no cuts
• filled circles and filled squares : cut only on eta
• filled triangles : cut on phi

 

Fig 1 : sigma dca in transverse plan

 

The cut on eta seems to have an effect (ie, decreased) the sigma_dca in transverse plan.

Cutting on eta remove wafers at the edges of the ladders

Fig 2 : sigma dca in Z direction

Cutting on eta does not improve the sigma_dca in Z direction (open squares vs. filled squares), but cutting on phi (ie. removing some ladders) seems to have a slight effect

(most visible for 1/P < 1) 

PART IV : tracks with 2 or more SSD hits fitted (done with 120_8120053 data : 320 files)

It appears that (as also shown for previous production), there are tracks with more than 1 SSD hit fitted.

They could come from the overlap between 2 adjacents ladders.

Fig 1 shows the (eta,phi) range of these tracks : it seems to have particular regions (ladders) where this happens

After the cuts, I get :

• 32474 tracks with NSSDFitted > 1
• 7471443 tracks with NSSDFitted = 1

So it only represents a low percentage of tracks.

Fig 2 : the chi^{2} of these tracks look the same

 

The only thing that differs from good tracks (NSSDFitted==1) is that the pT range seems to be lower.

Fig 3 : pT distribution for bad positive and negative tracks

Compare to

Fig 4 : pT distribution of good tracks