1: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2: \section{Data analysis\label{sec:data_analysis}}
3: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
4:
5: \subsection{Event selection\label{sec:event_selection}}
6:
7: The channel noise and pedestal levels are measured with dedicated
8: random trigger runs taken during the spill gaps.
9: The signal amplitudes of each pixel cell is reconstructed with
10: the following procedure~\cite{Dorokhov:2005}:
11: \begin{enumerate}
12: \item
13: Pixel pedestals stored in the data file are subtracted from
14: the amplitudes cell by cell.
15: \item
16: The amplitudes of six unconnected pixel rows
17: are averaged and subtracted to reduce common mode fluctuation.
18: \item
19: Hit pixels above threshold (which depends on the irradiation fluence),
20: are further analyzed. The threshold is optimized for each sample and ranges
21: between 35 -- 50 ADC counts (corresponding to 1000 -- 1500 electrons).
22: \item
23: %The pixels are checked for the expected hit frequency.
24: If the hit frequency is below 0.1 times or above 10 times the expected one,
25: the pixel is marked as noisy or dead and excluded from the amplitude reconstruction.
26: \item
27: Improved pedestals and common mode values derived from the good pixels
28: with amplitudes below threshold, are calculated and subtracted from the
29: pixel amplitudes.
30: \end{enumerate}
31: Steps (iii), (iv) and (v) are repeated for four iterations.
32: %until the mean value of the amplitude
33: %distribution is smaller than 0.1\% of the standard deviation or until
34: %four iterations have occurred.
35: The signal amplitudes of the telescope strip sensors are reconstructed
36: and corrected in a similar way.
37:
38: Events used for the
39: alignment of the pixel detector with the beam telescope are selected.
40: The selection procedure discards the events with multiple tracks
41: and tracks which do not fit to straight lines.
42: %The event selection proceeds as follows:
43: %\begin{enumerate}
44: %\item
45: % Up to four isolated strips in the beam telescope planes with amplitudes
46: % from 30 to 350 ADC counts are selected in each plane. The selected strips
47: % must be separated by at least five strips with signals below 30 ADC counts.
48: % %The strip numbers are used in the following steps.
49: %\item
50: % All unique pair combinations of the strips found in step 1 and belonging
51: % to different planes are selected. The procedure is done separately for the horizontal
52: % and vertical planes.
53: %\item
54: % The differences of the strip numbers is calculated for all the pairs
55: % found in step 2 and the most probable value of the difference is determined.
56: %\end{enumerate}
57: %The procedure is repeated with a wider cut on the amplitude range, between 10 and 400
58: %ADC counts. All the pairs with differences larger than one strip from the most probable
59: %value found in step 3 are discarded. The event is selected if each plane contains only one
60: %strip not previously rejected.
61:
62: \subsection{Alignment of the pixel sensor with the beam telescope\label{sec:alignment}}
63:
64: In each beam telescope plane the position of the impact point is reconstructed
65: with a center of gravity algorithm for clusters of two adjacent strips.
66: The main strip in the cluster must have a signal between 30 and 350 ADC counts and the
67: second strip is the next neighbour with the highest amplitude.
68: The hit position is then corrected using the $\eta$ algorithm described
69: in~\cite{Belau:1983eh}.
70:
71: The determination of the particle impact position in the pixel sensor
72: depends on the track incident angle $\alpha$ with respect to the pixel plane
73: (see Fig.~\ref{FIG:15deg_position_reconstruction}).
74: For $\alpha>60^\circ$ the hit position is measured using a center of gravity
75: algorithm. The pixel with the highest signal above threshold is located
76: and the surrounding eight pixels are included in the cluster. The cluster
77: is projected along the two coordinates $x$ and $y$. The weighted averages of
78: the hit coordinates $x$ and $y$ are calculated.
79:
80: %and the center of gravity is calculated in both cases.
81: \begin{figure}[hbt]
82: \begin{center}
83: \epsfig{file=15deg_position.eps,width=\linewidth}
84: \caption{Reconstruction of the impact position in the pixel sensor.}
85: \label{FIG:15deg_position_reconstruction}
86: \end{center}
87: \end{figure}
88:
89: For small angles, e.g. $\alpha=15^\circ$, the cluster size along $x$ is in
90: general larger than three pixels and a different approach is used.
91: The pixel corresponding to the track exit point is defined by the first
92: pixel above threshold along the $x$ axis and with the highest signal along
93: the $y$ axis. A $3\times3$ cluster is constructed by including the following two
94: pixels along the $x$ axis, which are required to be over threshold, and the
95: corresponding three pixels in the upper and lower rows along the $y$ axis.
96: %The resulting cluster size is three by three pixels.
97: The $y$ coordinate of the cluster is determined with the center of gravity
98: of the signal collected in the three $x$-rows. The $x$ coordinate of the
99: track exit point is determined using the signal in the three cluster $y$-columns.
100: We define the signals $A_0,A_1,A_2$ and the $x$ coordinates $x_0,x_1,x_2$
101: of the three pixel columns, respectively. The signal sums $S_i=\sum_{j=0}^i A_j$ with $i=(0,1,2)$
102: vs. their $x$ coordinates are fitted with a straight line.
103: The intercept of this line with the $x$ axis gives the coordinate of the
104: track exit point.
105:
106: The alignment is performed by minimizing the residuals in the following system of equations
107: \begin{eqnarray}
108: \sum_{i=1}^{8} \left[ p_{ij} C_{xi} \right] + C_{x0} - x_{j} & = & R_{xj}, \\
109: \sum_{i=1}^{8} \left[ p_{ij} C_{yi} \right] + C_{y0} - y_{j} & = & R_{yj},
110: \end{eqnarray}
111: where $p_{ij}$ is the position in the $i$-th beam telescope plane\footnote{The indexes $i=[1,...,4]$ correspond
112: to the position in the $x$ planes, while $i=[5,...,8]$ to the positions in $y$ telescope planes.}
113: for the $j$-th event, $x_{j}$ and $y_{j}$ are the hit coordinates in the pixel coordinate system,
114: $R_{xj}$ and $R_{yj}$ are the residuals in the $x$ and $y$ coordinates,
115: respectively, $C_{xi}$ and $C_{yi}$ are the fit parameters for the $x$ and $y$ coordinates,
116: respectively.
117:
118: \subsection{Gain calibration\label{sec:gain_calibration}}
119:
120: A calibration of the readout chain can be performed by injecting a known charge
121: at the input of the chip preamplifier via a 1.7 pF capacitor. The average signal
122: as a function of the time interval between the calibration and the trigger pulses
123: is show in Fig.~\ref{FIG:delay_scan}. The average signal measured with straight minimum
124: ionizing tracks and an unirradiated sensor is 710 ADC counts. This implies that the sampling
125: time of our measurements is about 55 ns.
126: \begin{figure}[hbt]
127: \begin{center}
128: \mbox{
129: \subfigure[]{\scalebox{0.90}{
130: \epsfig{file=Dot1_delay.eps,width=\linewidth}
131: \label{FIG:delay_scan}
132: }}
133: }
134: \mbox{
135: \subfigure[]{\scalebox{0.90}{
136: \epsfig{file=Dot1_Vcal.eps,width=\linewidth}
137: \label{FIG:gain_calibration}
138: }}
139: }
140: \caption{(a) Average signal amplitude as function of the time interval between the
141: calibration and trigger pulses for different values of the injected charge.
142: (b) Average signal amplitude as function of injected charge. The delay is set to
143: 55 nsec.}
144: \end{center}
145: \end{figure}
146: The gain calibration is performed by setting the delay to this value and measuring
147: the average signal amplitude as function of the charge injected in the preamplifier
148: (see Fig.~\ref{FIG:gain_calibration}). The charge released by a minimum ionizing particle (m.i.p.)
149: is 22400 electrons for a silicon thickness of 285 $\mu$m.
150: The response of the frontend electronics is fairly linear up
151: to about 1.2 m.i.p. and saturates for higher values of the injected charge.
152: Figure~\ref{FIG:gain_calibration} is used hereafter to translate the ADC
153: values into charge in electrons. More details about the gain calibration
154: can be found in~\cite{Dorokhov:2005}.
155: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%