Top Banner
E906/Drell-Yan: Monte Carlo, Data Acquisition and Data Analysis Paul E. Reimer Expected Rates and Monte Carlo (WBS 2.5.1) Data Acquisition (WBS 2.4) CODA Latched data TDC data Data Analysis (WBS 2.5.2)
9

E906/Drell-Yan: Monte Carlo, Data Acquisition and Data Analysis

Mar 22, 2016

Download

Documents

Sasha

E906/Drell-Yan: Monte Carlo, Data Acquisition and Data Analysis. Paul E. Reimer Expected Rates and Monte Carlo (WBS 2.5.1) Data Acquisition (WBS 2.4) CODA Latched data TDC data Data Analysis (WBS 2.5.2). Data Sources. Station 2 & 3 Drift Chambers 1700 Channels Multi-hit TDC’s. - PowerPoint PPT Presentation
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Page 1: E906/Drell-Yan: Monte Carlo, Data Acquisition and Data Analysis

E906/Drell-Yan:Monte Carlo, Data Acquisition and Data AnalysisPaul E. Reimer

Expected Rates and Monte Carlo (WBS 2.5.1) Data Acquisition (WBS 2.4)

– CODA– Latched data– TDC data

Data Analysis (WBS 2.5.2)

Page 2: E906/Drell-Yan: Monte Carlo, Data Acquisition and Data Analysis

211 December 2006 E906/Drell-Yan Review: Data Acquisition and Analysis

Data Sources

256 Hodoscopes—bit latches

MWPC 5500 Channels

bit latches

Station 2 & 3Drift Chambers1700 ChannelsMulti-hit TDC’s

Station 4 PropTubes 400 Channels

Bit latches

Page 3: E906/Drell-Yan: Monte Carlo, Data Acquisition and Data Analysis

311 December 2006 E906/Drell-Yan Review: Data Acquisition and Analysis

Data Acquisition Framework: CODAWBS 2.4.1

The CODA (CEBAF Online Data Acquisition) software frame work will be used for Data Acquisition

CODA is a software toolkit from which data acquisition systems with varying degrees of complexity can be built.

A typical system might look

Single Board Computer

FASTBUS VME CAMAC

ROC rcServer

EB

ROC

Network: Ethernet FDDI ATM

UNIX/LINUX

DISK/ TAPEET

User Proc.

RunControl GUI

ER

cdev

dp_tcl

mSQL Database

msqld cmlogServer

cmlog Database

Take from D. Abbot, CODA Users Workshop 2004

Page 4: E906/Drell-Yan: Monte Carlo, Data Acquisition and Data Analysis

411 December 2006 E906/Drell-Yan Review: Data Acquisition and Analysis

DAQ: Latch Electronics Station 1 MWPC, Station 4 Prop. Tubes

and Hodoscope readout Readout will be based on the Nevis

system Equipment recovered from E866/NuSea

and E871/HyperCP (approx. 20k ch. )

Detector ChannelCount

Tracking Station 1MWPC

5500

Station 4Prop. Tubes

400

Hodoscopes All Stations 256

Total 6156– Ansley delay cables– Coincidence Registers (CR)– Encoder board (ENC) and – FIFO cards

Rate will not be a problem—HyperCP had 20k channels and a trigger rate of 75 kHz—deadtime of approx. 3s

From Y.C. Chen et al. NIM A 455 (2000) 424

Page 5: E906/Drell-Yan: Monte Carlo, Data Acquisition and Data Analysis

511 December 2006 E906/Drell-Yan Review: Data Acquisition and Analysis

DAQ: Multi Hit TDCWBS 2.4.2, 2.4.3

Station 2 and 3 drift chambers readout via multi-hit TDC (1700 channels plus spares)

Fermilab Prep has sufficient channels of LeCroy 3377 FERA TDC Dead time of 1.8s + 0.1s/hit

Fermilab PREP recommends testing these modules prior to experiment and has offered to supply test procedures and setup. E906 will supply labor

FERA readout options under exploration– LeCroy 1190 FERA to VME memory interface– Struck SIS3700 32-bit ECL input FIFO (VME module)

http://www.struck.de/sis3700.htm– Fermilab DYC3 FERA to RS485 interface

http://ppd.fnal.gov/elec/dyc3/dyc3nov.html

Page 6: E906/Drell-Yan: Monte Carlo, Data Acquisition and Data Analysis

611 December 2006 E906/Drell-Yan Review: Data Acquisition and Analysis

Data Storage

Total data set for 5.2£1018 protons is approximately 4 TB (1 kHz/2£1012 protons £ 1.5 kB event size)

Temporarily store data locally on disk Background spool data to Feynman Computing Center for long

term storage (discussed with Fermilab CD)– Sustained data rate of 125 kB/s (easily within network

capacity)

Page 7: E906/Drell-Yan: Monte Carlo, Data Acquisition and Data Analysis

711 December 2006 E906/Drell-Yan Review: Data Acquisition and Analysis

Analysis SoftwareWBS 2.5.2

Modification of existing E866 analysis codes– Track recognition, trace back and

energy loss algorithms are identical– E866 code already ported to Fermilab

computing farms—probably not necessary.

– New event decoding routines needed Computing requirements

– Minimal—E866 event analysis rate was approx. 20 ms/event on vintage 1997 HP PA8000 0.18 GHz

– Scale to today: Should be able to do real time analysis of all data on small cluster (1 to 4) Linux-based PC’s

– We do not anticipate heavy demand on computing resources

Page 8: E906/Drell-Yan: Monte Carlo, Data Acquisition and Data Analysis

811 December 2006 E906/Drell-Yan Review: Data Acquisition and Analysis

Monte Carlo simulations

GEANT-Based Monte Carlo– Simulation of detector through end of 1st magnet– Used for configuration of hadron absorber within aperture of 1st magnet and

particle rates at Station 1– Extend simulation to entire spectrometer for data analysis (WBS 2.5.1

Abilene, Los Alamos, Argonne)

FastMC– Modified version of E866/E772 “fast” Monte Carlo Code to include E906

geometry.– Only traces muons from Drell-Yan events, and the decays of J/, 0, , 0,

00, D (D-bar) and K decays.– Magnetic field is simplified, but muon energy loss and multiple scattering are

included.– Verified with

• E866 rate and resolution comparisons• full GEANT simulation of target, beam dump and absorber

Page 9: E906/Drell-Yan: Monte Carlo, Data Acquisition and Data Analysis

911 December 2006 E906/Drell-Yan Review: Data Acquisition and Analysis

Conclusions

DAQ – CODA framework from JLab– Latched readout via Nevis electronics– Drift Chambers via LeCroy 3377 TDC

Data Analysis – Code is straight forward modification of E866 Analysis– Data storage requirements are “small”– CPU requirements are small