Commissioning and Operation of the CMS Tracker analogue optical link system at TIF with CMSSW:

Commissioning and Operation of the CMS Tracker analogue optical link system

at TIF with CMSSW:

R.Bainbridge, A.Dos Santos Assis Jesus,

K.A.Gill, V. Radicci

Commissioning of the Optical Link:

• 5189 laser diodes were tested in the slice test at TIF• each laser was tuned, to determine the optimal gain and bias setting, during the standard commissioning

procedure with the CMS Tracker online software XDAQ and the analysis software RootAnalyser.• A commissioning run is performed for each temperature: 15,10,0,-10,-15 0C.

This analysis is performed using the offline commissioning CMSSW framework (by Rob)

Almost completed:

1) Validate CMSSW procedure for the basic configuration of the Optical Link • Debug/test the code Check using TIF data

2) Compare CMSSW – Root Analyser3) Compare the performances of different Subsystems (TIB TOB TEC) 4) Study performance and detailed temperature effects (nominal 150C, 100C, 00C, -100C, -150C)

Ongoing:

1. Propose new histograms and variables for link performance (E.g. link noise contribution at different bias values... )

2. Propose a standard procedure for monitoring of link performance during the Tracker operations (Reliability and radiation effects)

Package: /CMSSW_1_4_0/src/DQM/SiStripCommissioning:

/SiStripCommissioningClients/src/OptoScanHistograms.cc (RB) ( /SiStripCommissioningAnalysis/src/OptoScanAnalysis.cc)

Source_run#.root

For each Gain produces:1) Summary Histos_run#.root (not yet implemented!)2) debug_7531.log file

RU file stored on Castor + Cabling from Online DB

/SiStripCommissioningSources/src/OptoScanTask.cc (RB)

GainSummaryPlot.cc (VR)

Summary plots .gif .root

*Actual Gain : 0 *LLD bias setting : 21 *Measured gain [V/V] : 0.69225 *'Zero light' level [adc] : 28.9667 *Link noise [adc] : 0.179505 *Baseline 'lift off' [mA] : 8.56697 *Laser threshold [mA] : 6.69421 *Tick mark height [adc] : 553.8 *Actual Gain : 1 *LLD bias setting : 20 *Measured gain [V/V] : 0.75632 *'Zero light' level [adc] : 30.85 *Link noise [adc] : 0.357071 *Baseline 'lift off' [mA] : 8.34694 *Laser threshold [mA] : 5.29931 *Tick mark height [adc] : 605.056...................... *Optimum LLD gain setting : 1

Analysis Procedure:

W

W/2

Lift Off

Laser Threshold = Lift Off –W/2

Tick Mark hight = W * Base Slope

Measured Gain = Tick Mark * (1/0.8) * (1.024/1024)

Best Gain ~= 0.8

Bias setting = Lift off + 2

Zero Light Level

Laser Bias (I2C)1 count = 0.45 mA

FE

D o

utp

ut

Laser Current = (Laser Thr * 0.45) mA

Ove

rlap

pin

g r

ang

e

50% range

“Tic

k”“B

ase”

Fit to points within the 20% and the 80% of the range

Coded by Rob

RUN # Tnom(0C) Partition

12042 -15 TIBTOBTEC (reduced # of Connection)

11309 -10 TIBTOBTEC

11169 0 TIBTOBTEC

10814 0 TIBTOBTEC

10601 10 TIBTOBTEC

9143 15 TIBTOBTEC

7580 15 TOB PP1

7531 15 TEC

7080 15 TEC

6668 15 TIB PP1

6419 15 TIB

5598 15 TOB

Variables Monitored:

1) Best Gain2) Bias Setting3) Measured Gain4) Base Slope5) Tick Mark hight6) Zero Light Level7) Lift Off8) Laser Thresold

Few Plots

@ different gains

Bias Setting

for different subsystem

Measured Gain

Target Gain = 0.8

Mainly lasers @ Gain = 1

TIB+TOB+TEC @ -100C - Run 11309

Tick Height

Slope vs Tick

Base slope

W = (Slope/Tick)-1

Independent on the subsystems and Gains

(Laser + LLD) (Laser + LLD + APV)

Input tick to link (=APV output tick)

Different behavior TIB, TOB, TEC due to the different low voltage drop (see A. Venturi talk)

TIB+TOB+TEC @ -100C - Run 11309

Baseline lift off (mA)Laser threshold (mA)

Zero Light Level Trim DAQ value

TIB+TOB+TEC @ -100C - Run 11309

This distribution strongly depends on the temperature of each Fed channel.

Gain @ Gain 0

Correlated to Laser th = Lift off – W/2

- Related to the AOH temperature• Double peak in TIB modules• Structured peak in TOB and TEC• TOB always higherTo compare with the distribution from the production

Digression: “Why double peak in laser th distribution?”

TIB Double Sided TIB Single Sided

Fiber 1

Fiber 2

Fiber 3

Ccu26

ccu28

ccu30

Ccu27

ccu29

ccu31

CERN, 9th October 2007 V. Radicci 12

Laser temperature “Tracker Map”: (fiber #3 @ gain 3 @ T=150C)

• “CMS Tracker Map Visualization” tool used here to map the laser temperatures:

- TOB is hotter- two different temperatures for TIB modules (green and yellow in the map) double peak in the TIB distribution

Laser Threshold distribution

Color scale from Red = hot to Blue = cool

TIB Layers TOB Layers

TID Disks TEC Disks

- Powerful tool also for mapping the tick height, base slope…- Interesting results will come from the correlation between Tick and Laser Temperature

CERN, 9th October 2007 V. Radicci 12

CMSSW vs RootAnalyser:

CMSSWRA

(T=-100C)

Difference channel by channel between Best Gain(CMSSW) – Best Gain (RA)

3.4% with diff>0

1) Comparison of Best Gain chosen for each laser

CMSSWRA

Difference channel by channel between Bias (CMSSW) – Bias (RA)

16% with diff>0 mainly ±1

1) Comparison of Bias setting chosen for each laser

Due to the different gain

Good agreementRobust performance of new code

Measured Gain @ Selected Gain comparison with the simulation

T=150C T=100C

T=00C

T=-150C

T=-100C

Range 0.62 V/V – 0.96 V/Vindependent from temperature in Agreement with simulation!

G0 G1 G2 G3

T=150C 3.8% 84.4% 10.4% 1.2%

T=100C 7.3% 84% 7.8% 0.7%

T=00C 18.9% 76.4% 4.1% 0.5%

T=-100C 34% 63.4% 2% 0.3%

T=-150C 40.1% 58.3% 1.3% 0.3%

Gain spread looks reasonable.Usual effect of increased gain with higher temperature. Values to be compared with simulation

Temperature dependence

– all distributions are fitted with a Gauss function, the mean and the error are plotted as a function of Temperature

mA

I

T

TTthref

3.5ln

CT

KT

ref

025

70/65/60/55/50/40

Nominal cooling

temperatures !!

The parameter T have to be determined by comparing with other measurements: PLC, DCU

- Laser threshold current is temperature dependent.- Different values with gain and subsystems at the same nominal temperature!- The current @ the laser threshold (Ith) can be used to evaluate the actual laser temperature T according its known exponential behavior:

TWarm – T+10TWarm – T0

TWarm – T-10 TWarm – T-15

In order to compare the AOH temperature and1) the temperature of the cooler or of the air (PLC measurements),2) the temperature of the sensor and of the hybrid (DCU readings) we can considering the temperature variation ΔT=(T1-T2), measure on the same element (laser, hybrid…), for

different experimental condition.

ΔT evaluated from Ith for different values of T

Mpv from the fit

Warm : +100C 3.4 4.2 4.7 5.1 5.5 6

Warm : 00C 10.7 13.4 14.8 16.1 17.2 18.8Warm : -100C 17.2 21.5 23.6 25.8 27.5 30.1Warm : -150C 20.8 26.0 28.6 31.2 33.3 36.4-100C : -150C 3.4 4.2 4.7 5.1 5.1 6

50T40T 55T 60T 65T 70T DCU-Hybrid

DCU-Sensor

- -

- -

23.2 22.8

28.1 27.6

4.9 4.9

PLC

Warm : +100C ~4.9

Warm : 00C ~15.1

Warm : -100C ~23.5

Warm : -150C ~27.7

-100C : -150C

To be checked!!!!

A. Venturi

KT 55 Now “the laser is a good thermometer” !!!

Note large ΔT over nominal

Effect of T on link gain and tick mark height

(Laser + LLD)

Tick height at FED (Laser + LLD + APV)

Input tick to link (=APV output tick)

Laser alone

W = Tick Height / Base Slope

Measured gain prop to Tick

Fit slope: G0: -0.24 G1: -0.26 G2: -0.27 G3: -0.27Decrease = (0.55±0.02) % / 10C

Fit slope: G0: -2.4 G1: -3.1 G2: -4.4 G3: -5.8Decrease = (0.49±0.03) % / 10C

• Gain change seems dominated by laser– Individual lasers vary more and do not

necessarily follow average trace• APV tick is constant!

Fit slope: G0: 0.080 G1: 0.084 G2: 0.086 G3: 0.086Decrease = (1.0±0.1) % / 10C

Fit slope: G0: 0.18 G1: 0.19 G2: 0.18 G3: 0.18Decrease = (1.0±0.1) % / 10C

Particular behavior of the Zero light level an explanation could be considering the fed temperature channel by channel

Conclusion

– Links commissioning software/analysis in use on TIF data

• CMSSW code stable– New parameters included

» Laser threshold, noise, tick input height– Able to check out detailed effects

– Move development effort now onto details of performance (e.g. noise) and monitoring (Ana talk)