phos4 studies and fit functions for timing
DESCRIPTION
Phos4 Studies and Fit functions for Timing. Michael Henke, Valerie Lang Kirchhoff-Institute for Physics , University of Heidelberg L1Calo Joint Meeting, 12.01.2010, Heidelberg. Phos4 Scan-Studies. Run 124774: 12.08.2009 2079 Events 20-40 Events per Phos4 block LAr pulsed - PowerPoint PPT PresentationTRANSCRIPT
![Page 1: Phos4 Studies and Fit functions for Timing](https://reader035.vdocuments.mx/reader035/viewer/2022070420/56815fdd550346895dcee49b/html5/thumbnails/1.jpg)
Phos4 Studies and Fit functions for Timing
Michael Henke, Valerie LangKirchhoff-Institute for Physics, University of Heidelberg
L1Calo Joint Meeting, 12.01.2010, Heidelberg
![Page 2: Phos4 Studies and Fit functions for Timing](https://reader035.vdocuments.mx/reader035/viewer/2022070420/56815fdd550346895dcee49b/html5/thumbnails/2.jpg)
Michael Henke, Valerie Lang 2
Phos4 Scan-Studies
• Run 124774:– 12.08.2009– 2079 Events– 20-40 Events
per Phos4 block– LAr pulsed– TriggerTower
0x00100501(emBarrel)
12.01.2010
![Page 3: Phos4 Studies and Fit functions for Timing](https://reader035.vdocuments.mx/reader035/viewer/2022070420/56815fdd550346895dcee49b/html5/thumbnails/3.jpg)
Michael Henke, Valerie Lang 3
Phos4 Scan-Studies
• Run 124774 Trigger Tower 0x00100501(similar to Taylor‘s work)
12.01.2010
Corrected for a shift of the pedestal with respect to the Phos4-Step
![Page 4: Phos4 Studies and Fit functions for Timing](https://reader035.vdocuments.mx/reader035/viewer/2022070420/56815fdd550346895dcee49b/html5/thumbnails/4.jpg)
Michael Henke, Valerie Lang 4
Close look at pulse top• Deviations from an ideal
peak form around the maximum
• This complicates the determination of the pulse maximum‘s position.
12.01.2010
Why bother?1. Maximum‘s position from
Phos4 histogram = reference value for quality determination of a certain fit function to Phos4 pulse
2. Calibration precision
![Page 5: Phos4 Studies and Fit functions for Timing](https://reader035.vdocuments.mx/reader035/viewer/2022070420/56815fdd550346895dcee49b/html5/thumbnails/5.jpg)
Michael Henke, Valerie Lang 5
Peak tops for other trigger towers
12.01.2010
Trigger tower 0x00100200 fairly ideal pulse
Trigger tower 0x00100901 Trigger tower 0x00100e01
Trigger tower 0x00100301
![Page 6: Phos4 Studies and Fit functions for Timing](https://reader035.vdocuments.mx/reader035/viewer/2022070420/56815fdd550346895dcee49b/html5/thumbnails/6.jpg)
Michael Henke, Valerie Lang 6
Number of channels affected• Rough estimate:
– Look at 500 channels for 3 partitions– Count affected channels– No information on Tile Calorimeter as this wasn‘t pulsed
• Results:
12.01.2010
Calorimeter partition
Crate number
Percentage of affected channels
Comments
emBarrel 0 37%
emEndCap 2 16% In general: smaller deformations
forwardCal & hadEndCap
4 13% than in emBarrel
In total: 22%
![Page 7: Phos4 Studies and Fit functions for Timing](https://reader035.vdocuments.mx/reader035/viewer/2022070420/56815fdd550346895dcee49b/html5/thumbnails/7.jpg)
Michael Henke, Valerie Lang 7
Causes of peak irregularities?• Data from Phos4-
Scan plotted without mean calculation over the individual Phos4-blocks(pedestal shift correction applied)
• Colour indicates number of times a certain ADC count occured
12.01.2010
Trigger tower 0x00100501:
![Page 8: Phos4 Studies and Fit functions for Timing](https://reader035.vdocuments.mx/reader035/viewer/2022070420/56815fdd550346895dcee49b/html5/thumbnails/8.jpg)
Michael Henke, Valerie Lang 8
Causes of peak irregularities?
12.01.2010
Trigger tower 0x00100501:
=> Irregularities don‘t stem from strongly deviating values in the mean calculation
![Page 9: Phos4 Studies and Fit functions for Timing](https://reader035.vdocuments.mx/reader035/viewer/2022070420/56815fdd550346895dcee49b/html5/thumbnails/9.jpg)
Michael Henke, Valerie Lang 9
Consider other runs• Run 124777:
– 5 minutes later than run 124774
– 2042 events– LAr pulsed– Unfortunately: many pulses
are saturated, i.e. no information about pulse peak form
– Compare 2 non-saturated channels
12.01.2010
Trigger tower 0x01100a03
![Page 10: Phos4 Studies and Fit functions for Timing](https://reader035.vdocuments.mx/reader035/viewer/2022070420/56815fdd550346895dcee49b/html5/thumbnails/10.jpg)
Michael Henke, Valerie Lang 10
Consider other runs• Run 124777:
– 5 minutes later than run 124774
– 2042 events– LAr pulsed– Unfortunately: many pulses
are saturated, i.e. no information about pulse peak form
– Compare 2 non-saturated channels
12.01.2010
Trigger tower 0x01100a03
![Page 11: Phos4 Studies and Fit functions for Timing](https://reader035.vdocuments.mx/reader035/viewer/2022070420/56815fdd550346895dcee49b/html5/thumbnails/11.jpg)
Michael Henke, Valerie Lang 11
Consider other runs• Run 124777:
– 5 minutes later than run 124774
– 2042 events– LAr pulsed
12.01.2010
Trigger tower 0x00110d03
Strong similarities of pulse forms between runs 124774 and 124777
![Page 12: Phos4 Studies and Fit functions for Timing](https://reader035.vdocuments.mx/reader035/viewer/2022070420/56815fdd550346895dcee49b/html5/thumbnails/12.jpg)
Michael Henke, Valerie Lang 12
Preliminary conclusions• Results:
– Pulse peak deformation non-negligible effect
– Causes need to be understood for reasonable testing of the fit functions and precise calibration
• Possible explanation:Addition of single calorimeter cell signals to a trigger tower is not perfectly aligned in time
12.01.2010
Next investigative steps:1. Look at newer runs Have timing updates fixed the problem?2. Take Phos4 runs with only parts of the trigger tower cells
activated see if effects disappear3. Look at Phos4 runs with the Tile Calorimeter pulsed Does
the same effect appear?
![Page 13: Phos4 Studies and Fit functions for Timing](https://reader035.vdocuments.mx/reader035/viewer/2022070420/56815fdd550346895dcee49b/html5/thumbnails/13.jpg)
Michael Henke, Valerie Lang 13
Fits to Phos4 pulses: Gauss-Landau fit
• Earlier studies:Simple Gauss-Landau fit works best for:– Tile calorimeter => pulse
doesn‘t drop below pedestal– 5 slice readout => small fitting
range– Sometimes missing Phos4
steps• Now:
– LAr Calorimeter– 15 slice readout– Improved control of Phos4chip
12.01.2010
Plot from Paul‘s diploma thesis
![Page 14: Phos4 Studies and Fit functions for Timing](https://reader035.vdocuments.mx/reader035/viewer/2022070420/56815fdd550346895dcee49b/html5/thumbnails/14.jpg)
Michael Henke, Valerie Lang 14
Fits to Phos4 pulses
12.01.2010
Gauss-Landau
Gauss-Landau with undershoot
Landau-Landau with undershoot
Lorentz-Landau with undershoot
Trigger tower 0x01140a02
![Page 15: Phos4 Studies and Fit functions for Timing](https://reader035.vdocuments.mx/reader035/viewer/2022070420/56815fdd550346895dcee49b/html5/thumbnails/15.jpg)
Michael Henke, Valerie Lang 15
Fits to Phos4 pulses• Gauss-Landau
fit with undershoot seems to work best– Largest
applicable range– Best agreement
with histogram, especially around the peak (judged by eye)
12.01.2010
More quantitative approaches in work
Gauss-Landau with undershoot
Landau-Landau with undershoot
![Page 16: Phos4 Studies and Fit functions for Timing](https://reader035.vdocuments.mx/reader035/viewer/2022070420/56815fdd550346895dcee49b/html5/thumbnails/16.jpg)
Michael Henke, Valerie Lang 16
Fits to real pulse simulations• Motivation:
– Look at fits for all possible mistimings
• Procedure:– Read out every 25th value of
the Phos4 scan– Plot in a new histogram
12.01.2010
![Page 17: Phos4 Studies and Fit functions for Timing](https://reader035.vdocuments.mx/reader035/viewer/2022070420/56815fdd550346895dcee49b/html5/thumbnails/17.jpg)
Michael Henke, Valerie Lang 17
Fits to real pulse simulations• Motivation:
– Look at fits for all possible mistimings
• Procedure:– Read out every 25th value of
the Phos4 scan– Plot in a new histogram– Fit with Gauss-Landau and
Gauss-Landau with undershoot
– Move the readout point by 1ns in Phos4 histogram (Phos4 scan backwards)
12.01.2010
Gauss-Landau fitGauss-Landau fit with undershoot
![Page 18: Phos4 Studies and Fit functions for Timing](https://reader035.vdocuments.mx/reader035/viewer/2022070420/56815fdd550346895dcee49b/html5/thumbnails/18.jpg)
Michael Henke, Valerie Lang 18
Fits to real pulse simulations• Motivation:
– Look at fits for all possible mistimings
• Procedure:– Read out every 25th value of
the Phos4 scan– Plot in a new histogram– Fit with Gauss-Landau and
Gauss-Landau with undershoot
– Move the readout point by 1ns in Phos4 histogram (Phos4 scan backwards)
12.01.2010
Phos4 maximum‘s position: Determined by Gauss-Landau fit to the highest 10% of values
![Page 19: Phos4 Studies and Fit functions for Timing](https://reader035.vdocuments.mx/reader035/viewer/2022070420/56815fdd550346895dcee49b/html5/thumbnails/19.jpg)
Michael Henke, Valerie Lang 19
Fits to real pulse simulations
12.01.2010
Gauss-Landau fitGauss-Landau fit with undershoot
Better fit = Gauss-Landau with undershoot 7-10 slice readout is sufficient
![Page 20: Phos4 Studies and Fit functions for Timing](https://reader035.vdocuments.mx/reader035/viewer/2022070420/56815fdd550346895dcee49b/html5/thumbnails/20.jpg)
Michael Henke, Valerie Lang 20
Correlation plots• Definitions:
– mistimings = shift of digitalisation against perfect digitalisation
– fitdigi–fitphos4 = max. position from fit to real pulse – Phos4 reference
• Results:– Simple Gauss-Landau fit shows
some unwanted structure removed in new fit
– Comparable distribution widths in y-direction
12.01.2010
Gaus
s-La
ndau
Gaus
s-La
ndau
w
ith u
nder
shoo
t
LAr emEndCap A
![Page 21: Phos4 Studies and Fit functions for Timing](https://reader035.vdocuments.mx/reader035/viewer/2022070420/56815fdd550346895dcee49b/html5/thumbnails/21.jpg)
Michael Henke, Valerie Lang 21
Conclusions• Pulse peak deformations
– Not a rare phenomenon– To be further investigated (maybe already solved)
• Fit functions for fine timing– The Gauss-Landau function including the undershoot gives the
currently best fit for Phos4 and real pulses– 7-10 slice readout is sufficient for stable and precise fitting
• Outlook– Quantification of agreement of different fit functions with Phos4 scans– More detailed analysis of correlations regarding real pulses
12.01.2010