Slide 1Slide 2Slide 3Slide 4Slide 5Slide 6Slide 7Slide 8Slide 9Slide 10Slide 11Slide 12Slide 13Slide 14Slide 15Slide 16Slide 17Slide 18Slide 19Slide 20Slide 21Slide 22Slide 23Slide 24Slide 25Slide 26Slide 27Slide 28Slide 29Slide 308.882 LHC PhysicsExperimental Methods and MeasurementsB Physics Triggers and Lifetime Fitting[Lecture 19, April 15, 2009]C.Paus, LHC Physics: B Physics Trigger Strategies 2Organization Project 2●status: nothing yet received Project 3●B lifetime measurement is due May 2 Final Conference project●starts: 12:00 Kolker room on May 19●each student gives●17 minutes presentation●+3 minutes discussion●agenda to be publishedC.Paus, LHC Physics: B Physics Trigger Strategies 3Final Conference Project LHC Physics: “Experimental Methods and Measurements”●C.Paus: Welcome and LHC Overview●?: Search for Higgs in H→ZZ*●?: Search for Higgs in H→WW* ●?: Search for Higgs in qqH→qqWW*‘09The Physics Colloquium SeriesWednesday, April 15 at 4:15 pm in room 4-349 Sidney DrellSLAC National Accelerator Laboratory Steps Toward a World Free of Nuclear Weapons:$ Rekindling the Vision of Reagan and Gorbachev at ReykjavikSPRINGFor a full listing of this semester’s colloquia, please visit our website at web.mit.edu/physicsColloquium SeriesPhysicsC.Paus, LHC Physics: B Physics Trigger Strategies 5Lecture Outline B Trigger Strategies●CDF trigger system overview●high momentum displaced track triggers●experimental implications A bit of probabilities and fitting●probability density functions and expectation value●least square method●likelihood method●lifetime example●some technicalities for our applicationC.Paus, LHC Physics: B Physics Trigger Strategies 6Review: Hadron Collider Triggers Trigger Systems at Hadron Colliders come in levels●per level: time to analyse and data to analyse increase●Level 1 – implemented in hardware, custom electronic boards●rates at CDF: 3 MHz → 30 kHz or CMS: 25 MHz → 300 kHz??●Level 2 – implemented in programmable hardware (FPGAs*)●rates at CDF: 30 kHz → 1 kHz (CMS has no Level 2, merged with Level 3) ●Level 3 – implemented as PC farm, runs reconstruction program●rates at CDF: 1 kHz → 100 Hz (CMS has no Level 3)●CMS High Level Trigger (HLT) merges Level 2 and Level 3●rate: 100k → 100 Hz●dead timeless design by long trigger pipelines: parallelized* Field Programmable Gate ArrayC.Paus, LHC Physics: B Physics Trigger Strategies 7Trigger System at CDF Speed●trigger primitives are fast digitized objects●less precise but fast Trigger Primitives●muon stub ●electron cluster●track Triggers●combination of trigger primitives and application of certain cuts●ex. muon is already combined with track at level1C.Paus, LHC Physics: B Physics Trigger Strategies 8Outline of Track Trigger Tracking is 2 dimensional●look for high momentum Main idea: patterns/roads●granularity gives resolution Hardware used●four axial layers of COT●XFT/XTRP trigger boards Three step process●hit finding●segment finding●track findingC.Paus, LHC Physics: B Physics Trigger Strategies 9Hit and Segment FinderC.Paus, LHC Physics: B Physics Trigger Strategies 10Segment Finder and Linker Output of Segment Finder●inner two layers●phi position only●outer two layers●phi position●slope of segment (3 bits) Segment Linker●uses phi position in all layers●and slopes in the outer two●combines them to a road●288 linkers (per 1.25º) each linker uses a table of 1200 predefined roadsC.Paus, LHC Physics: B Physics Trigger Strategies 11XTRP Board Responsibility●select the best tracks (highest pT)●distribute XFT tracks to other systems●level1 muons trigger●level1 calorimeter trigger●level1 track trigger (pass info)●level2 track trigger, SVT (pass info)●level1 and level2 global triggers●extrapolate position to corresponding detector●extrapolate to muons chambers●extrapolate the calorimeterC.Paus, LHC Physics: B Physics Trigger Strategies 12Secondary Vertex Tracker (Trigger) Another trigger based on patterns●resolution and speed are astonishing: 47μm at 15 μs●32,000 roads to be searched●first such trigger at hadron collider eversingle hitsuper striproad4 detector layersXFT infoC.Paus, LHC Physics: B Physics Trigger Strategies 13Secondary Vertex Tracker (Trigger) Typical trigger requirements●two tracks with pT > 2 GeV●pT1+pT2 > 5.5 GeV●opening angle: 0º < Δφ < 135º (level 2: 12º < Δφ < 90º)●impact parameter: 120 μm < |d0| < 1 mm●upper cut due to pattern limitations (not enough memory/time to search through more pattern)●increased sensitivity to purely hadronic B decays by 4-5 orders of magnitude: this was the biggest upgrade for B physics in Run II of the Tevatron●B → Dπ (Bs mixing), etc.●B →ππ (direct CP violation, maybe α)C.Paus, LHC Physics: B Physics Trigger Strategies 14Secondary Vertex Tracker (Trigger) Problems you have to deal with●intrinsic bias of proper time distribution●sharp prompt peak disappears, reducing the background●also gets ride of some signal (bias)C.Paus, LHC Physics: B Physics Trigger Strategies 15Trigger Bias Curve Simulation of trigger bias●events we need to study are not in the data, to first order●needs careful Monte Carlo study: generate events and keep them●silicon hit occupancy needs to be reasonably well simulated●check hit resolution function carefully, too●Monte Carlo must also have the right lifetime .... The plot is for semileptonic events but looks similar for purely track based triggerC.Paus, LHC Physics: B Physics Trigger Strategies 16Likelihood Fits Have found a spectacularly nice talk from Colin Gay given at a Particle Physics students retreat. Please check it out, I am stealing a lot from there:●http://physics.bu.edu/NEPPSR/TALKS-2006/Likelihood_Gay.pdf You also might want to check the statistics talk on the same site:●http://physics.bu.edu/NEPPSR/TALKS-2006/Statistics_Blocker.pdfC.Paus, LHC Physics: B Physics Trigger Strategies 17Probability Assume X is a random variable, like for example our ct or m of an event●the probability to throw X into window [x,x+dx] is P(x)dx●P(x) is the probability density function (PDF) for X●if we look at the entire range of possible values for X like for example: we can be sure that X at any throw will take one of those values and this means the integral of the PDF should be one The expectation value of