manfred jeitler, hephy vienna cms level-1 trigger mgu, 11 july 2011 1 the level-1 trigger of the cms...

Manfred Jeitler, HEPHY Vienna CMS Level-1 Trigger MGU, 11 July 2011 1

The Level-1 Trigger of the CMS Experiment

at LHCDesign, challenges and performance

Manfred Jeitler

MGU, 11 July 2011Московский государственный университетIn

stit

ute

of H

igh

En

ergy

Ph

ysic

s of

th

e A

ust

rian

Aca

dem

y of

Sci

ence

s`


The Large Hadron Collider


LHC p-p CollisionsOperations 2010-2011

2010: √s = 7 TeV – Peak Instantaneous

Luminosity:> 2×1032 cm-2s-1

2011: √s = 7 TeV– Peak Instantaneous

Luminosity:> 1.25 ×1033 cm-2s-1

– And increasing!

– Current Maximum of ~1400 bunches/beam in LHC

– Possible to gain another factor of 4

2808


CMS Detector


first particle physics experiments needed no trigger

were looking for most frequent events

people observed all events and then saw which of them occurred at which frequency

what is a trigger?


later physicists started to look for rare events

– “frequent” events were known already

searching “good” events among thousands of “background” events was partly done by auxiliary staff

– “scanning girls” for bubble chamber photographs

what is a trigger?


Получше — в горшочек, похуже — в зобочек!


Higgs -> 4 +30 MinBias


find the needle!


due to the extremely small cross sections of processes now under investigation it is impossible to check all events “by hand”

– ~ 1013 background events to one signal event

it would not even be possible to record all data in computer memories

we need a fast, automated decision (“trigger”) if an event is to be recorded or not

X-s

ecti

on

X

-secti

on


detectors yielding electrical output signals allow to select events to be recorded by electronic devices

– thresholds (discriminators)

– logical combinations (AND, OR, NOT)

– delays

– available in commercial “modules”

– connections by cables (“LEMO” cables)


because of the enormous amounts of data at major modern experiments electronic processing by such individual modules is impractical

– too big

– too expensive

– too error-prone

– too long signal propagation times

use custom-made highly integrated electronic components (“chips”)

400 x 1 x

~ 10 logical operations / module ~ 40000 logical operations in one chip


example: trigger logic of the L1-trigger of the CMS experiment


When do we trigger ?

„bunch” structure of the LHC collider– „bunches” of particles– 40 MHz

» a bunch arrives every 25 ns » bunches are spaced at 7.5 meters from each other» bunch spacing of 125 ns for heavy-ion operation

at nominal luminosity of the LHC collider (1034 cm-2 s-1) one expects about 20 proton-proton interactions for each collision of two bunches for ATLAS and CMS– only a small fraction of these “bunch crossings” contains at least one collision

event which is potentially interesting for searching for “new physics”– in this case all information for this bunch crossing is recorded for subsequent

data analysis and background suppression


Event size (bytes)

trigger: first level high levelATLAS, CMS 40 MHz 100 kHz 200 Hz LHCb 40 MHz 1 MHz 2 kHzALICE 10 kHz 1 kHz 100 Hz


How do we trigger ?

use as much information about the event as possible – allows for the best separation of signal and background

– ideal case: “complete analysis” using all the data supplied by the detector

problem: at a rate of 40 MHz it is impossible to read out all detector data

– (at sensible cost)

have to take preliminary decision based on part of the event data only be quick

– in case of positive trigger decision all detector data must still be available

– the data are stored temporarily in a “pipeline” in the detector electronics» “short term memory” of the detector

» “ring buffer”

» in hardware, can only afford a few μs

how to reconcile these contradictory requirements ?write

read


multi-level trigger

first stage takes preliminary decision based on part of the data

– rate is already strongly reduced at this stage– ~1 GHz of events (= 40 MHz bunch crossings) ~100 kHz– only for these bunch crossings are all the detector data read out of the

pipelines– still it would not be possible (with reasonable effort and cost) to write

all these data to tape for subsequent analysis and permanent storage

the second stage can use all detector data and perform a “complete analysis” of events

– further reduction of rate: ~100 kHz ~100 Hz– only the events thus selected (twice filtered) are permanently recorded

40 MHz

100 kHz

300 Hz

to tape


How does the trigger actually select events ?

the first trigger stage has to process a limited amount of data within a very short time

– relatively simple algorithms– special electronic components

» ASICs (Application Specific Integrated Circuits)» FPGAs (Field Programmable Gate Arrays)

– something in between “hardware” and “software”: “firmware”» written in programming language (“VHDL”) and compiled » fast (uses always same number of clock cycles)» can be modified at any time when using FPGAs

the second stage (“High-Level Trigger”) has to use complex algorithms

– not time-critical any more (all detector data have already been retrieved)– uses a “computer farm” (large number of PCs)– programmed in high-level language (C++)


signals used by the first-level trigger

muons– tracks– several types of detectors (different requirements for barrel and endcaps):– in ATLAS:

» RPC (Resistive Plate Chambers): barrel» TGC (“Thin Gap Chambers”): endcaps» not in trigger: MDT (“Monitored Drift Tubes”)

– in CMS:» DT (Drift Tubes): barrel» CSC (Cathode Strip Chambers): endcaps» RPC (Resistive Plate Chambers): barrel + endcaps

calorimeters– clusters– electrons, jets, transverse energy, missing transverse energy– electromagnetic calorimeter– hadron calorimeter

only in high-level trigger: tracker detectors – silicon strip and pixel detectors, in ATLAS also straw tubes – cannot be read out quickly enough


how to find muon tracks in the CMS solenoidal field?(Drift Tubes)


RPC ComparatorRPC Comparator

Pattern matching Pattern matching

barrel: 3 layers/4 or 4/6barrel: 3 layers/4 or 4/6

endcaps: 3/3endcaps: 3/3

how to find muon tracks in the CMS solenoidal field?(Resistive Plate Chambers)


Muon Trigger

Muon Trigger•3 muon detectors to ||<2.4

• Drift Tubes

• Track Segment ID and Track Finder

• Cathode Strip Chambers

• Track Segment ID and Track Finder

• Resistive Plate Chambers

• Pattern Matching

•4 candidates per subsystem to Global Muon Trigger

•Global Muon Trigger sorts, removes duplicates, 4 top candidates to Global Trigger

•Track building at 40 MHz


Global Muon Trigger

match & merge barrel:

DT-RPC endcap:

CSC-RPC cancel duplicates

overlap region:DT-CSC

sort by momentum and quality


calorimeter system

Hadronic calorimeter (HCAL)Hadronic calorimeter (HCAL)

Electromagnetic (ECAL)Electromagnetic (ECAL)

HFHF

Hadron Forward CalorimeterHadron Forward Calorimeter


calorimeter trigger


Photon and electron trigger objects

L1 e/gam L1 e/gam HLT e/gam HLT e/gam

L2 Step

Spatial matching of ECAL clusters with

e/γ candidates at L1

Superclusters are formed

ET cut applied

Calorimetric (ECAL+HCAL) isolation

L3 Photons

Tight track isolation

L3 Electrons

Electron track reconstruction

Spatial matching of ECAL cluster Spatial matching of ECAL cluster and pixel trackand pixel track

Loose track isolation in a “hollow” cone


e/and Jet Algorithms

4x4 Tower sums from RCT to GCTJet or ET

• 12x12 trig. tower ET sliding in 4x4 steps w/central 4x4 ET > others

: isolated narrow energy deposits• Energy spread outside veto pattern

sets veto• Jet if all 9 4x4 region vetoes off

GCT uses Tower sums used for ET,MET jets for HT, MHT

Electron (Hit Tower + Max)• 2-tower ET + Hit tower H/E • Hit tower 2x5-crystal strips >90% of ET in 5x5 (Fine Grain)

Isolated Electron (3x3 Tower)• Quiet neighbors: all towers

pass Fine Grain & H/E• One “L” of 5 EM ET < Thr.


Global Trigger

Receives Trigger Objects:• 4 forward and 4 central jets, 4 -jets, 4 isolated and 4 non-isolated e/, total ET,

missing ET, HT and position information from GCT• 4 with position, sign, and quality information from GMT

128 different conditions (thresholds, topological cuts) can be combined to make 128 physics triggers

Forwards Level-1 Accept Trigger, Timing and Control (TTC) system for read-out of the detector front-end electronics


Global Trigger and Trigger Control System

Every bunch crossing for which a potential Level-1 accept is inhibited is counted as “dead”.– needed to calculate recorded

integrated luminosity

CMS

regionaltriggers

GCTGMT

40 MHz pipeline

calo & muonobjects

GLOBALTRIGGER

LOGIC1. Logic OR2. Prescale

3. Rate Counters

128xalgos

technical trigger signals(beam monitoring systems etc.)

64xtechnical

GLOBAL TRIGGER

FINAL DECISIONLOGIC

physics

calib

random

rand calibTRIGGERCONTROLSYSTEM

L1A candidate

Level-1 Accept (via TTC system to detector)

backpressure,trigger rules

deadtimecounters


L1 Trigger Custom Hardware

– Hundreds of boards

– Thousands of:» ASICs

» FPGAs

» Copper Cables

» Optical Fibers

» (Wo)man hours

Global Muon Trigger&

Global Trigger(Vienna)

RCT(Wisconsin)

CSCTF(Florida)DTTF

(Vienna)

GCT(Imperial)

RPC PaC(Warsaw)


The CMS Level-1 Trigger Only calorimeter and muon systems participate in CMS L1

0.9<||<2.4

4 4 4 4 4+4 4+4

4 4

MIP+ISO bitsMIP+

ISO bits

e, J, ET, HT, ETmisse, J, ET, HT, ET

miss

L1A L1A

40 M

Hz

pipe

line

40 M

Hz

pipe

line

Calorimeter TriggerCalorimeter Trigger

ECALTrigger

Primitives

ECALTrigger

Primitives

HCAL/HFTrigger

Primitives

HCAL/HFTrigger

Primitives

RegionalCalorimeter

Trigger

RegionalCalorimeter

Trigger

GlobalCalorimeter

Trigger

GlobalCalorimeter

Trigger

Muon TriggerMuon Trigger

RPC hitsRPC hits CSC hitsCSC hits DT hitsDT hits

Segment finder

Segment finder

Track finderTrack finder

Pattern Comparator

Pattern Comparator

Segment finder

Segment finder

Track finderTrack finder

Global Muon TriggerGlobal Muon Trigger

Global TriggerGlobal Trigger

TTC systemTTC system TTS systemTTS system

Detector FrontendDetector Frontend

StatusStatus

Link systemLink

system

32 partitions32 partitions

CMS experiment

0<||<5||<3 ||<1.2||<3


Global Trigger interface


ATLAS+CMS:what’s the difference ? similar task

similar conditions similar technology


ATLAS+CMS:what is common ?

same physics objectives same input rate

– 40 MHz bunch crossing frequency similar rate after Level-1 trigger

– 50 .. 100 kHz similar final event rate

– 100 .. 200 Hz to tape similar allowed latency

– pipeline length– within this time, Level-1 trigger decision must be taken and detectors must be read

out– ~ 3 μs

» 2.5 μs for ATLAS, 3.2 μs for CMS


ATLAS+CMS:what is different ?

different magnetic field– toroidal field in ATLAS (plus central solenoid)

» get track momentum from η (pseudorapidity)

– solenoidal field only in CMS» get track momentum from φ (azimuth)

number of trigger stages– two stages (“Level-1” and “High-Level Trigger”) in CMS– ATLAS has intermediate stage between the two: “Level-2”

» Level-2 receives “Regions of Interest (RoI)” information from Level-1» takes a closer look at these regions» reduces rate to 3.5 kHz» allows to reduce data traffic


data are stored in “pipelines” for a few microseconds– e.g. in CMS: 3.2 s = 128 clock cycles at 40 MHz

– question of cost

– decision must never take longer! (must be synchronous!) no iterative algorithms

decision based on “look-up tables”– all possible cases are provided for

– OR, AND, NOT can also be represented in this way

ATLAS+CMS:principle of the first-level trigger


principle of event selection: ATLAS vs. CMSCMS:

no “cuts” at individual low-level trigger systems

– only look for muons, electrons, jets, choose the “best” of these candidate objects and forward to Level-1 Global trigger

so, all possible kinds of events may be combined

selection is only made by “Level-1 Global Trigger”

– trigger information from all muon detectors and calorimeter systems available: completely flexible

High-Level trigger analyzes complete detector data for the whole detector

ATLAS:

Level-1 trigger delivers numbers of candidate objects

– muon tracks, electron candidates, jets over appropriate threshold

– no topological information used (no angular correlation between different objects)

Level-2 trigger carries out detailed analysis for “Regions of Interest”

– using complete detector information for these regions

High-Level trigger analyzes complete detector data for the whole detector


First Collisions Tuesday March 30, 2010 12:58 L~1027cm-2s-1 ~ 60 Hz Collision Rate

Time to optimize the trigger synchronization!


Minimum Bias trigger detectors

HF – Hadron Forward Calorimeter

BSC – Beam Scintillator Counters (hit and coincidence rate, MIP detection eff >96%)

BPTX – Beam Pick-up Timing for the eXperiments monitors

(precise info about bunch structure and timing of the incoming beam, res. 0.2 ns)


2010 Commissioning: Synchronizing the Trigger Collision products take longer to get

to outer part of detector– e.g. to reach CSC takes longer than

ECAL Take into account many different

cable lengths– Even within a sub-detector

First runs with only a Minimum-Bias Trigger– Take data with a number of delays

– Find the best alignment between subsystem (CSC shown) and Minimum-Bias Trigger

– Repeated for each Trigger subsystem

– Cross-checks available in DQM


LHC is design to deliver one bunch crossing (BX) every 25 ns

Trigger has to provide a correct BX assignment

Optimization of the trigger synchronization allows for an overall efficiency increase

ExampleExample:: L1 RPC trigger synchronization is in a very L1 RPC trigger synchronization is in a very

good shape good shape

no pre-no pre-

triggerstriggers

TRIGGERSTRIGGERSRPC HitsRPC Hits

BX

=0 :

99.9

99%

BX

=0 :

99.9

99%

2010 Commissioning: Synchronizing the Trigger


2010 Operational Challenge: Rates Final month of 2010 p-p running, max inst. lumi

increased by > O(4)!– No longer relying on Minimum Bias but a full

menu of physics object triggers– L1 works with the HLT to maintain a sustainable

HLT stored data rate» Up to 65 kHz in, ~400 Hz out

– New L1 & HLT configurations and menus:» Higher thresh. L1 Triggers for HLT seeds» Prescaling at L1 and HLT to reduce rates» Several prescale “sets” to allow as much data

to tape as possible• Lumi decreases -> change set of prescales,

keep rate about the same

– Possible to predict output rates with MC– Must take into account other factors

» Cosmic backgrounds, pile-up effects, detector effects, etc.


Threshold efficiency – probability that for reconstructed muon with momentum PT, track finder assigns momentum greater or equal to the threshold value

F it f unct ion:

p 0 E r f ( p1 x p 2 ) p 3

Eta region 1.2 <|η|

<2.1

CSC trigger performanceCSC trigger performance

DT trigger performanceDT trigger performance

DTTF PT > 10 GeV/c from J/PsiDTTF PT > 10 GeV/c from J/Psi

“Turn-on” expected to reach >“Turn-on” expected to reach >

90% efficiency at the nominal 90% efficiency at the nominal

cutcut

Eta region |η|<

1.1

2010 Trigger Efficiencies


2011: Operations: Controlling the Rates

LHC has been steadily increasing instantaneous luminosity delivered to the experiments in 2011– Well beyond what was delivered in 2010

– How do we best control rates while keeping as much physics as possible? Just a few examples of what the L1 Trigger can do:

– e/ triggers

» Energy corrections to improve resolution

» ECAL Spike Killing

» Isolation and H/E for e/ candidates

– Jet triggers

» Improve the resolution with jet energy corrections

– Muon triggers

» Improve ghost busting and pattern recognition

– Continue to develop the L1-HLT menus with prescale sets

» Must anticipate physics groups’ needs as conditions change


2011: Operations: Avoiding pre-triggers

Problem: – Triggers are pre-firing at about ~5% of the

time

– Forward hadron calorimeter Can we avoid this source of inefficiency?

– CMS BPTX_AND Trigger

» AND of two precision electrostatic beam pick-ups

» Provides colliding bunch structure in CMS for every fill

– Copy, reduce delay by one bunch-crossing unit (25 ns) and shorten to ~20ns

– Use it as veto on pre-triggers

– But it vetoes the slow particle (HSCP) trigger (BX+1) when the bunch spacing is 50 ns!

How to fix this? (Next Slide)

before

after


2011 Operations: Trigger on HSCPs

Heavy Stable Charged Particles (HSCPs) take as much as 2 BX the detector

– Look for late signal in muontriggers

– But preBPTX veto during 50 ns bunch spacing vetoes these!

RPC Trigger can extend detector signal to 2 BX

– Reduce RPC delay into GMT by 1 BX

– preBPTX will veto early in-time muons

– In-time muons still captured with collision BX

– HSCPs aligned with collision BX

– Create special HLT path to keep slow particles (the BXid will be +1 wrt to Tracker, for example).

HSCPs Triggering!

layer 6layer 5layer 4

Chamber hits extended hits

BX

Muon candidate

In-time muon


BPTX


Chamber hits Extended hits

BX

Muon candidate

HSCP


BPTX


Halfway through 2011

Middle of 2011 p-p run:

323 pb-1 data taken Maximum sustained L1 rate

of ~65 kHZ– 912b – 874 colliding at CMS, 50ns

spacing

– XX pb-1 recorded in a single fill


CMS Trigger & DAQ Systems

LHC beam crossing rate is 40 MHz & at full Luminosity of 1034 cm-2s-1 yields 109 collisions/s

Reduce to 100 kHz output to High Level Trigger and keep high-PT physics

Pipelined at 40 MHz for dead time free operationLatency of only 3.2 sec for collection, decision, propagation


... and where do we go from here?

upgrading the LHC to Super-LHC makes sense only if trigger systems are upgraded at the same time

ATLAS and CMS will have to use their trackers in the first-level trigger

but this is easier said than done probably, computers will get faster and more (all?) trigger

processing will be done in software


Conclusions

CMS L1 Trigger is performing well– 2010 and 2011 not without its challenges

– Complex system, in Hardware and Software

» Excellent pool of experts

» Problems addressed in a timely manner

LHC still increasing the luminosity – Will still stress system, both detectors and triggers

» Must remain diligent, even if luminosity stabilizes

– Expect L1 Menus to continue evolving

» Physics results may dictate changes instead of lumi

– Challenge to balance physics needs with rate limitations,

» L1T must continue to work with HLT

Looking forward to lots of interesting physics results!


Спасибо за приглашение!

[email protected] http://wwwhephy.oeaw.ac.at/u3w/j/jeitler/www

manfred jeitler, hephy vienna cms level-1 trigger mgu, 11 july 2011 1 the level-1 trigger of the cms...

Documents