1. RIKEN Software Meeting
Software Update
Takashi Hachiya
RIKEN
2019/7/2
RIKEN Software Meeting

2. RIKEN Software Meeting
Things
Software Update Plan
Add SvxPixelHotDeadMap in SvxParManager and SvxDecode
Similar maps for Stripixel is not prepared.
Join to the CentralArm-Cluster analysis
Make Link of Cluster using Central Track.
Skipping dead area using Hot &dead map.
Making some additional documentation
2019/7/2
RIKEN Software Meeting

3. RIKEN Software Meeting
2019/7/2
RIKEN Software Meeting

4. Algorithm of Standalone Tracking
Track Finding
Track Fitting
QuickPt determination
Use first 3 points (2 pixel and 1 strip points)
Track Fitting W/O the primary vertex
to determine the momentum
Fit with multi-circles with R (pT).
Multiple scattering are taken into account

5. Algorithm for Track fitting
After Track finding using link method
Cluster list (link) includes more than 3 clusters
2 Pixel hits
More than 1 Stripixel hits.
QuickPt determination
Use first 3 points (2 pixel and 1 strip points)
Determine the center position (Cx, Cy) and radius (R) of the circle which pass through these 3 points.
The circle can be determined uniquely.
Fitting these cluster list (link) with multi-circles with pT.
Calculate Chi-square
Try finding the least Chi-Square
Using pT from quickPt funciton
Try to calculate chi2 using pT, pT+dpT, pT-dpT. (chi0, chi1, chi2)
Compare chi0 with chi1 and chi2.
If Chi0 > Chi1 -> Chi1 is good (1.5 pT is good)
else if Chi0 > Chi2-> Chi2 is good (0.5 pT) is good)
Repeat the chi2 comparison with changing pT and dpT until chi2 is minimum
Changing pT to pT + dpT
Changing dpT to dpT /2
7 times repeat as maximum.
If chi2 is minimum with less number of repeat, loop will be break. (chi0 > chi1 && chi0 > chi2 : this condition should never happen)
Calculate Log probability
Chi2 – determined above
NDF = (Ncluster – 2 ) * 2 – : Each point has 2 NDF (R and Z direction) what is -2 ?

6. QuickPt determination
Determine the center of the rotation and its radius
The center of rotation (cx, cy) should be on the perpendicular of both the lines of hit 1 and 2 and hit 2 and 3 which pass at the center of the lines.
The center position can be determined by sliding along with the red line from the center btw B-D)
A(x1, y1)
B(x2, y2)
D(x3, y3)
C(cx, cy) R
The radius is the length betwenn A and C
pT (momentum of this plane (x-y plane) is described as follows:
pT = * R * B. R=radius(cm), B=Field strength (0.9T)
, where is the speed of light . Constant B-Field is assumed.
pc = q * c *R *B [ j]
= 0.3 * R * B [GeV]
1[j] = 1/(q=1.6e-19) [eV]

7. Calculate Chi-Square Definition of Chi-Square X-Y plane NDF=(N-2)^2-1
Definition of Chi-Square
X-Y plane
NDF=(N-2)^2-1
N of defined theta for
r and z direction
p(pT, pz) vectors is defined along with
arc1 and arc2 at the crossing point.

8. Calculate Chi-Square 2 Lr12 is 2D distance btw p1 and p2
Lr12 is 2D distance btw p1 and p2
Lz12 is 3D distance btw p1 and p2
Numerator is unit pixel size in phi and z -direction
Sqrt(12) – might be Naverage hit
pixels in a cluster
Pixel
Strip
dr23
dr12
L23
L12

9. calculate Tangent at hit position
calculate_simpleTangent
Input :
2 hit positions (p1 and p2)
pT2 : pT squared determined by quickPT
Output
Momentum(p) – scalar value
Px-py vector at hit postion 1 and 2
Algorithm
Recalculate the center position of the circle where both points are on the circles with radius (R = b*pT, b=const)
Calculate the momentum vector (pT1 and pT2) at p1 and p2 along with the direction of tangent. P2 P1
P12