08/02/2016, LAr week hilum meeting Status of the root trees combining BSEC, PSEC, IC and gadc data and some plots obtained from them A. Maslennikov (Novosibirsk BINP) As it was announced at the 13 /01/2016 phone meeting, I have produced a first version of ntuples and some short instruction how to use it. Please see /afs/cern.ch/user/a/amaslenn/public/hilum/bsec_gadc/ directory, in particular the file “Bsec_Gadc_Str.usage”. The structure and the meaning of data is described in comment lines of the header file Bsec_Gadc_Str.h, as well as in the E-mail which I have sent at January 14th. For the moment I have produced 2 files for “single bunch runs” (1140-41) and 2 files with “long train + one single bunch” (1169, 1193). Also since the phone meeting, I have made some debugging and applied absolute calibration of intensity. 08/02/2016, LAr week hilum meeting
For gadc calibration, I think I keep using what John R For gadc calibration, I think I keep using what John R. is calling “Kozelov Jul 2014” coefficients (see J.R. presentation at the 13 jan 2016 phone meeting, page 16 – it differs from “Kozelov Jan 2014”, page 17, by factor of 1.5 – as far as I could see, these two sets have similar uniformity over number of protons per bunch). For Protvino SEC I use arbitrary factor 0.1 (0.167) for Y (respectively, X) sum to convert from ADC counts to intensity in 10^9 protons. Similarly for ionization chamber (IC) I use a factor 4.e-5 to convert from ADC counts to intensity in 10^9 protons (tuned for lower intensity runs, should be rechecked) For Novosibirsk SEC I use factor 1.862e-2 (1.076e-2) for Y (respectively, X) integral (flat pedestal term is NOT included in integral) to convert from ADC counts to intensity in 10^9 protons. These preliminary numbers were obtained from comparison with foils irradiation in 2013 (), see my presentation at the 22 sep 2015 meeting. So now for all our devices, measuring the intensity, there are variables normalized to 10^9 protons (for integrals) or to protons per second for rates. “Raw” (ADC) values are also kept for cross-check.
Run 1141, 172 protons/gadc, t_spill 1s Small ratio BSEC X/ BSEC Y IC works up to 2.510^10
There are spills with big and not so big drift in X (bottom left)
Run 1140, 496 protons/gadc, t_spill 0.35 s, almost no gadc ovf
There are spills with big drift in X (bottom left) – but spills are short in this run (0.35 s)
Run 1169, 250 protons/gadc, t_spill 0.65 s IC is “anti-saturated” Very unusual : gadc is > BSEC at small and intermediate intensity (top left) PSEC and BSEC agree well (bottom left)
There are spills with big and not so big drift in X (bottom left)
Run 1193, 162 protons/gadc, t_spill 0.75 s Good agreement Gadc – BSEC Y PSEC is broken
Run 1193: the beam drift in X seems to be much smaller than in other runs
Synchronization of gadc and BSEC – gadc is integrated with tau = 50 ms and profile is stored with a step of 1 ms GADC overflows don’t have strong effect here – 156 ovf in the left spill, 11941 in the right
I’ll continue to check and improve reconstruction procedures and produce new files. So please try to use ntuples (technically it is NOT difficult, really) and send your comments, suggestions and complaints. I think we anyway need some iterations to optimize the format and debug the content.
Instructions from the file “Bsec_Gadc_Str.usage”: Create your new work directory and copy files from my CERN afs dir: /afs/cern.ch/user/a/amaslenn/public/hilum/bsec_gadc // Then for the very first time only do two actions: 1) to create dictionaries, do from a linux command line rootcint -f Bsec_Gadc_Str_Dict.cxx -c Bsec_Gadc_Str.h Bsec_Gadc_Str_LinkDef.h // 2) to create shared library .so, // 2a) open a root session root -l // 2b) issue the command in root .L Bsec_Gadc_Str.h+ // 2c) close root session .q
Then start a root session and simultaneously open file with bsec-gadc data : root –l bsec_gadc_v1_run_1141.root // The pointer bsec_gadc_tree is present, you may do e.g. bsec_gadc_tree->Print() or ->Scan() // then issue a command to create pointer to Bsec_Gadc_Str structure // (it is called bg1 ) and associate the tree content with it: .x load_bsec_gadc_tree.C // then you may load entry number ie (starting from 0), corresponding to one spill bsec_gadc_tree->GetEntry(ie); // and you have filled in the bg1 structure all variables, described in Bsec_Gadc_Str.h file for(int k=0; k<7; k++) printf(" %f", bg1->fpar_y[15][k]); printf("\n"); // or you may plot some hists or graphs bsec_gadc_tree->Draw("gadc_intens[15]:fpar_y[0][0]","","*");