1 jones 1.1 % Hall C Howto Template
  2           % Rename this file to "something.tex" where something is a unique
  3           % filename.
  4           % To process this document, you also need the files:
  5           %    chowto.cls
  6           %    chowto.bib
  7           %    chowto.bst
  8           %
  9           % If you refer to publications not in chowto.bib, then create a file with the
 10           % same name as this with the ``.tex'' changed to ``.bib'' and put the
 11           % references into that file in BibTeX format.
 12           %
 13           % Instructions for downloading and using this template can be found at
 14           %
 15           %           http://www.jlab.org/~saw/howto/
 16           % 
 17           \documentclass{chowto}
 18           
 19           \title{Time of Flight Scintillator Hodoscope Calibration}
 20           \howtotype{user} % ``expert'', ``user'', ``reference''
 21           %\experiment{Name of experiment} % Optional
 22 jones 1.1 \author{Peter Bosted}
 23           \category{hms} % Subject area of this document
 24                   % Allowed choices are
 25                   %
 26                   % hms             HMS detectors and other HMS stuff (eg rotation)
 27                   % sos             SOS detectors and other SOS stuff i(eg doors)
 28                   % magnet          SOS/HMS or other magnet systems
 29                   % beamline        Beamline instrumentation
 30                   % daq             Data Acuisition and Analysis
 31                   % electronics     
 32                   % target   
 33                   % general         All other subjects
 34           
 35           %\maintainer{Name of person maintaining document} % Optional
 36           \date{\today} % Can use \today as the argument
 37           
 38           \begin{document}
 39           
 40           \begin{abstract}
 41           This document describes how to optimize the scintillator hodoscope
 42           parameters to obtain the most accurate possible timing information.
 43 jones 1.1 This in turn leads to the best determination of the time-of-flight
 44           between S1 and S2, which can be used to distinguish between
 45           particles of different mass (electrons, pions, kaons, and protons).
 46           \end{abstract}
 47           
 48           \section{Introduction}
 49           The scintillators in the HMS and SOS spectrometers come in two
 50           sets: S1 and S2. In each set, there are horizontal and vertical
 51           bars, labeled X and Y. Each bar has a photo-multiplier tybe (PMT)
 52            at each end. The PMT
 53           outputs go through a splitter: one output goes to an Analog to Digital
 54           Converter (ADC) and one
 55           to a discriminator, then to a multi-hit Time to Digital 
 56           Converter (TDC). To determine the
 57           time at which a particle passed through a scintillator, relative
 58           to the time in one of the paddles (used as the reference paddle),
 59           one needs to correct the TDC values (which measure time relative
 60           to the trigger time) for three effects:
 61           \begin{itemize}
 62           \item
 63           The average time delay between when light hits the photo-cathode
 64 jones 1.1 of the PMT, and when a pulse emerges from the anode (there are
 65           variations from PMT to PMT, and the offset also depends on the
 66           HV). Thus new calibrations are needed every time the HV is changed.
 67           Also, the length of cable for each PMT is slightly different going
 68           from the PMT eventually to the TDC: this is also taken into
 69           account by an overall time offset for each PMT.
 70           \item
 71           The time it takes for light produced in the paddle to get to the
 72           PMT. To first order, this is proportional to the distance from
 73           where the particle hit the paddle to the PMT, which requires
 74           information from the tracking code. If the photons went straight
 75           to the PMT, then the time just depends on the speed of light
 76           in plastic. In practice, the light bounces around, so takes about
 77           40% longer to reach the PMT that one might expect. This can
 78           depend on the PMT geometry (i.e. how thick it is), and how much
 79           is has aged. So we can fit an effective velocity of light for
 80           each paddle. 
 81           \item
 82           The TDC value depends on the time at which the pulse crosses the
 83           discriminator threshold, hence it depends on the pulse height.
 84           Since pulses are attenuated in the plastic, there is a correlation
 85 jones 1.1 of pulse height with distance from the PMT, so it is hard to
 86           distinguish this correction from the effective velocity one. 
 87           Several forms have been tried (no ADC correction, correction
 88           proportional to $\sqrt{ADC}$, to $1./\sqrt{ADC}$, and to
 89           ignoring the ADC but using a path length squared correction.
 90           The fitting code can easily be changed to accommodate all of
 91           these cases, but the current version uses $1./\sqrt{ADC}$
 92           \end{itemize}
 93           
 94           The goal of the TOF calibration is to find the above three
 95           correction factors for each PMT, such that the corrected
 96           times are all in as perfect agreement with each other as possible.
 97           The average time difference between S1 and S2 can then be
 98           used to find the velocity of the particle in question, and
 99           distinguish between electrons, pions, kaons, and protons (and
100           even deuterons and tritons).
101           
102           \section{Step by step Instructions}
103           Pick a run to analyze that has mostly one particle type making
104           triggers, and singles rates that aren't too high ($<$100 kHz). Then
105           \begin{itemize}
106 jones 1.1 \item Type ``newgrp c-inclusive'' (or whichever group you 
107           analyzing data in) if this is not your default group, so that
108           other in the group can collaborate with you.
109           \item
110           Include the text  'hdumptof=1'' on the command line that you
111           use to replay the run (next to ``grun=12345'', etc.). 
112           (for SOS, uses ``sdumptof=1''.
113           These variables can be more permantly changed in the
114           file  `hdebug.param'' in PARAM directory. 
115           (or something like PARAM/online07 directory) 
116           \item
117           Run the analyzer for the run you have chosen.
118           \item
119           Copy the output file fort.37 (fort.38 for SOS) to a file you will
120           cal ``tofcal.inp'' in the directory where you will do the calibration.
121           \item
122           Copy the source code:
123           /group/hallc/packages/tof\_calibration/tofcal.f to your directory.
124           Also copy the ``Make file''
125           /group/hallc/packages/tof\_calibration/Maketof to your directory.
126           Log on to a Linux machine and ``Make'' (compile and link) the
127 jones 1.1 code by typing ``Maketof tofcal''
128           \item
129           Run the code by typing ``tofcal''
130           \item
131           The new tof parameters will be found in the file ``tofcal.param''. 
132           See if the results make sense. The offsets, velocities, ADC corrections,
133           and sigmas will be set to 0, 50., 0., and 100. respectively for
134           PMTs where their weren't enough hits (100 minimum) within the
135           starting time window. Note that each parameter is listed in four
136           columns and 16 rows. The columns correspond to S1X, S1Y, S2X, and S2y,
137           and the rows correspond to the paddle numbers. Since some arrays
138           don't have 16 paddles, one expects the last 1, 7, 0, and 7 rows
139           of columns 1, 2, 3, 4 to not have hits. (For SOS, this would be
140           last 8, 8, 1, and 7 rows). Also, quite often there are not enough
141           hits in the first row. If any other paddles don't have enough hits,
142           trying running more events. Also, check that the online histograms
143           show normal-looking ADC and TDC spectra for that PMT.  This can
144           also be checked in the text-formatted histograms that are dumped
145           out, called ``tofcal.adchist'' and ``tofcal.tdchist'' (see tofcal.f
146           for format). 
147           \item
148 jones 1.1 Check that values for parameters seem reasonable.
149            Normal values of invadc\_offset are between -50 and 50,
150            normal values of invadc\_velocity are between 12 and 17 (again, 50 is
151            default if not enough data for the fit), and normal values of
152            hhodo\_pos\_invadc\_adc are 20 to 50. If not reasonable, investigate.
153           \item
154           If all is working well, the values of ``sigma'' should be below 0.5
155           (units are nsec). These are the widths of the differences of
156           the time for each paddle with the averaged time for all paddles. 
157           With the new PMTs in the HMS, we should strive for 0.25 nsec. 
158           \item
159           Once happy with the new values, insert the contents of
160           ``tofcal.param'' into the end of the file ``hhodo.param\_xxxxx'' in the
161           ``PARAM'' directory, where xxxxx is the run number you used
162           for the calibration. You can then point to this file in the
163           master setup file ??? to use the new calibrations.
164           \item 
165           Re-analyze the run with the new parameters and make sure the
166           spectra of ``beta'' and ``mass'' are now narrower than before. 
167           
168           \end{itemize}
169 jones 1.1 
170           
171           
172           \end{document}