Monitoring DT trigger rates using online lumi Luminosity - PowerPoint PPT Presentation

Monitoring DT trigger rates using online lumi Luminosity monitoring using DT trigger rates? 24/09/2012 Ignacio Redondo Fernández & Jessica Turner I.Redondo (CIEMAT) 1

Monitoring DT trigger rates • Scaler data is stored in DB with 5 min granularity. – flashlist granularity is 30 s (I believe). – Larger granularity requires online SW changes – An ntuple including other beam related variables is being produced daily for easy analysis • Observed early last year (as soon as collision muons started to dominate rate) that DT triger rates were remarkably linear with online luminosity. – See https://indico.cern.ch/getFile.py/access?contribId=5&resId=0&materialId=slides&confId=135091 • Estimate saturation effects <1% at lumi< LHC _design – Double BS muon rates or a pt threshold could be implemented above 24/09/2012 I.Redondo (CIEMAT) 2

Monitoring • Data >1 year ago 24/09/2012 I.Redondo (CIEMAT) 3

Monitoring DT trigger rates • Fitted parameters (per chamber) up to 250E30 have been used as an online monitoring tool – Predict chamber rates from online HFlumi – Plot ratio of measured over expected – Chamber rates typically well within 5% Ratio of Measured over Extrapolated (from early 2011 fit) MB2 MB3 MB4 MB1 24/09/2012 I.Redondo (CIEMAT) 4 2012: Castor removal reduce neutron rates

BS rate: counting muon tracks • Let´s concentrate in the Barrel Sorter rate: – input to GMT from DTTF – coincidence of triggers from at least 2 out of 4 chambers in sector – Match DTTF algorithm – Survive sorters ghost-killing • Relatively high level – no time dependent pt-cut (just geometry and magnetic field) – does not depend on RPC, midly on CSC – tipically within 3% of extrapolation based on online HF lumi (fit <250E30) up to – Precision limited by bad vacuum fills which results in halo muon rate in the barrel.(~KHz) • This can be substracted IF we would sample the rate before collisions (and after DT HV is turned on at squeeze) • In fact, potential for vacuum monitor. Q: List of bad vacuum fills? • Further effects worth checking would be dependence with pressure via vdrift 24/09/2012 I.Redondo (CIEMAT) 5

2011: DT rate vs. HFonline lumi expected at same time stamp (from DB) “Cosmics” 305 Hz Excellent linear correlation (<1%) over 3 orders of magnitude up to 3.5E33 DT fitted lumi minus HFonline lumi Histogram is projection of Fit-Point 24/09/2012 I.Redondo (CIEMAT) 6

2011: DT rate vs. HFoffline lumi(Calc2) expected at same time stamp “Cosmics” 238 Hz Excellent correlation (<1%) up to 4E33 Narrower and more gaussian residuals DT fitted lumi minus HFoffline lumi 24/09/2012 I.Redondo (CIEMAT) 7

2011: DT rate vs. HFoffline lumi(Calc2) expected at same time stamp “Cosmics” 238 Hz Excellent correlation (<1%) up to 4E33 Narrower and more gaussian residuals DT fitted lumi minus HFoffline lumi DT fitted lumi minus HFonline lumi 24/09/2012 I.Redondo (CIEMAT) 8

2011: DT rate vs. Pixel offline lumi expected at same time stamp “Cosmics” 284 Hz Excellent correlation (<1%) up to 4E33 Slope vs. HFoffline and vs. pixelLumi equal (+-0.5%) DT fitted lumi minus Pixel offline lumi Larger scatter from Pixel lumi 24/09/2012 I.Redondo (CIEMAT) 9

2011 Lumi_DT (pixel calibrated) vs. time Ratio to pixelLumi stable over 2011 During TS 1. Major HF redefinition 8.5 % effect 2. Loss of DT eff due to 2. 21/5/2012 Vdrif change ~<2% DT Resynchronization ~+3 % 20/Oct/2011 11/June/2012 6/April/2011 24/09/2012 I.Redondo (CIEMAT) 10

Summary ca Summer • BS rate correlates linearly very well with pixel lumi, HF online& offline lumi over a few orders of magnitud. • A BS rate based lumi estimator: – would probably need less recalibrations than HF online lumi. – seems to suffer less from statistical fluctuations than pixel lumi and lives naturally online. • Q: What to do with it is up to Lumi POG – Could automatize to produce weekly (daily if DB values) crosscheck plot – Could provide pixel calibrated lumi online with time granularity : • 5 min (reading from DB) => small job, available in short time • 30 s (build xdaq app. that reads flash list) => need help,,available in longer time • 23.35 s (modifying publishing in DTTF supervisor) => not in my hands • Correlation with pixel 2012 pending to fully clarify 2012 data 24/09/2012 I.Redondo (CIEMAT) 11

Status now • BS rate correlates linearly very well with pixel lumi, HF online& offline lumi over a few orders of magnitud. • A BS rate based lumi estimator: – would probably need less recalibrations than HF online lumi. – seems to suffer less from statistical fluctuations than pixel lumi and lives naturally online. • Q: What to do with it is up to Lumi POG – Could automatize to produce weekly (daily if DB values) crosscheck plot – Could provide pixel calibrated lumi online with time granularity : • Find out GMT stores BS rate with lumi section granularity! • � � � � Use this to publish a BS based DT lumi in WBM 2011 calibration ok � waiting for WBM plots to be done Used End of Fill VdM scan 2710 to calibrated 2012 a+b data 2012 c data is off probably due to change in CSC input to BS (horns) J. Turner Study alternative estimators as, for instance, MB3 only rate . • Correlation with pixel 2012 pending to fully clarify 2012 data 24/09/2012 I.Redondo (CIEMAT) 12

• 5-7 % offest • Intriguing fill dependence wrt. HF online 24/09/2012 I.Redondo (CIEMAT) 13

Is BS rate still linear with lumi in 2012? • Bellow 4E33 at 7 TeV, yes, we all agreed. • Deviation after April 2012 TS seen – Effect in fact correlated with the correction (8.5 %) of the normalization of HF data in DB after the TS – Effect not present if: • the online data previous to the TS is corrected or if • the correlation is done vs. HF offline lumi (lumiCalc2.py) 24/09/2012 I.Redondo (CIEMAT) 14

Red: after April TS Black: before April TS Red: after April TS Black: before April TS/1.085 24/09/2012 I.Redondo (CIEMAT) 15

2012: DT rate vs. HFoffline lumi(Calc2) expected at same time stamp Larger spread but still Linear correlation (<3%) up to 7E33 DT fitted lumi minus HF offline lumi Low B fill 24/09/2012 I.Redondo (CIEMAT) 16

Using 2011 (vs. pixel calibration) on 2012 data During TS 1. Major HF redefinition 8.5 % effect 2. Loss of DT eff due to 2. 21/5/2012 Vdrif change ~<2% DT Resynchronization ~+3 % 11/June/2012 24/09/2012 I.Redondo (CIEMAT) 17 4/April/2012

Crosschecking Timestamp Matching • DB and offline lumi timestamps match: description of lumi timeline features (i.e. fills) • Pixel lumi (at lumi section level) show larger statistical(?) scatter • HFonline lower normalization CALC2_time-DB_time (s) September 7 2011 Sampling coincide within ~[-23,23] s 24/09/2012 I.Redondo (CIEMAT) 18 CALC2_time-DB_time (s)

Potential as Vacuum monitor: Fill 1901 18/07/2011 Between squeeze (DT HV on) and collisions Cosmics adter dump Azimutahl angle MB1s factor of ~10 Wheel Azimutahl angle MB2s factor of ~15 MB3s Factor of ~2, MB4s ~unaffected 24/09/2012 1.7 KHz BS rate vs. usual cosmic background (~260 Hz) I.Redondo (CIEMAT) 19

Saturating the rate (Wrong assumption: double muon uncorrelated) 20 x LHC design LHC design 24/09/2012 I.Redondo (CIEMAT) 20

Ratio of slopes 7TeV/8TeV = 0.895? However similar slope to offline lumi Expect more muon rate at a given lumi 24/09/2012 I.Redondo (CIEMAT) 21

24/09/2012 I.Redondo (CIEMAT) 22

24/09/2012 I.Redondo (CIEMAT) 23

24/09/2012 I.Redondo (CIEMAT) 24

2011 BS rate over prediceted BS rate Time since 01/01/2010 24/09/2012 I.Redondo (CIEMAT) 25

Monitoring DT trigger rates 24/09/2012 I.Redondo (CIEMAT) 26

BKG in external wheel MB1s (“punchthrough”) Monitoring DT trigger rates 24/09/2012 I.Redondo (CIEMAT) 27 BKG in MB4s (“neutrons”)

2012 HF online lumi 24/09/2012 I.Redondo (CIEMAT) 28

• Investigate off diagonal points • Quantify correlation Pixel offline lumi HF offline lumi 24/09/2012 I.Redondo (CIEMAT) 29

• Investigate off diagonal points • Quantify correlation Pixel offline lumi HF offline lumi 24/09/2012 I.Redondo (CIEMAT) 30