scintillator hcal read-out andout and calibration hcal read-out andout and calibration felix sefkow...

Scintillator HCAL read-out andread out and calibration

Felix Sefkow

TILC08 at Sendai, Japan March 4, 2008

Outline

• Test beam

• Calibration and monitoring methods

N d t l t i d t t• New read-out electronics under test

Felix Sefkow TILC08, Sendai, March 3-7, 2008 Scintillator HCAL readout & calibration 2

Scintillator HCAL prototypep yp

• Steel scintillator sandwich• 38 layers, 2cm steel absorbers• Scintillator tiles 3x3x0.5cm3

• 7608 SiPMs (MEPhI / Pulsar)7608 SiPMs (MEPhI / Pulsar)

Versatile calibration and monitoring system


Test beam experiencep

• Established the scintillator SiPM technology on large scale (7608 SiPMs)– Robust and stable operation, 95% up-time, 1.6% dead channels (mostly solder)– Noise occupancy 10-3 as expected 0 8 MIP = 25 MeV / hit


– Noise occupancy 10 3 as expected, 0.8 MIP = 25 MeV / hit– Imaging capability nicely demonstrated, millions of events collected

Analysis ongoingy g g

• Electromagnetic showers: Verify detector model and calibration dprocedures

• Hadrons: test simulation models and particle flow performance

Ongoing thesis workN.Wattimena O. Wendt


Next: technical prototypep yp

• Goal: A compact and realistic (i e scaleable) scintillator HCAL

• Integration issues

• Goal: A compact and realistic (i.e. scaleable) scintillator HCAL structure with embedded electronics

– Readout architecture– Ultra-low power ASICs– Calibration system

Til d SiPM i t ti– Tile and SiPM integration– Absorber mechanics with

minimal cracks• See V Zutshi’s talkSee V.Zutshi s talk

• Feed-back from test beam essential


essential– Calibration concept

Calibration

• Pixel photo-diode is non-linear: need p2 reference scales– MIP: muon test beam– Gain: low intensity LED light

• Energy deposition– E[MIP] = A / AMIP * f ( A / Apixel)

• Saturation correction fSaturat on correct on f– From test benchSATURATION CURVE SIPMS WAFER 4-11

Light, MIP

Res

pons

e, p

ixel

s

0

200

400

600

800

1000

1200

1400

0 50 100 150 200 250

G i

SATURATION CURVE SIPMS WAFER 4-11

Light, MIP

Res

pons

e, p

ixel

s

0

200

400

600

800

1000

1200

1400

0 50 100 150 200 250

Gain


SATURATION CURVE SIPMS WAFER 4-11

Light, MIP

Res

pons

e, p

ixel

s

0

200

400

600

800

1000

1200

1400

0 50 100 150 200 250

Temperature and voltagep g

• Signal (charge) depends on gain and Geiger efficiency: A ~ ε * Gg g p g g y• Both depend on overvoltage ΔU = Ubias – Ubreakdown

• Breakdown voltage Ubreakdown depends on temperature

Figures: S.Uozumi• For x = A, G: dx/dT = - dx/dU * dUbreakdown/dT

R ti f lit d d i ffi i t i th f U d T


• Ratio of amplitude and gain coefficient is the same for U and T dependence

Voltage dependenceg p

• Measured for A and G at ITEP test bench, stored in data base• Reproduced in CERN test beam set-up

ITEP

CERN

This and following slides:

N F


N.Feege,A.Kaplan

Temperature dependencep p

• Measured in test-beam only

Felix Sefkow TILC08, Sendai, March 3-7, 2008 Scintillator HCAL readout & calibration 10Red: low ΔU group

Amplitude and gain coeff.sp g

• Check model: dA/dT / dG/dT = dA/dU / dG/dU ?

• Mean (norm) -4.0 -1.7 5.6 -2.6 %/(K or 0.1V) • RMS (norm) 2.8* 0.4 0.8 0.4

mean spread


Correction methods

• Temperature correction: x = x0 [ 1 + dx/xdT (T-T0) ] , x = A, Gp– Instantaneous, but non-local

• Gain correction A = A0 [ 1 + { dA/AdU / dG/GdU } (G-G0) ]0 [ { } ( 0) ]– Local, but not instantaneous

• Combination of both – all to be triedCombination of both all to be tried– Success not guarenteed – may introduce additional errors

V lt dj st t U U ( G G ) / dG/dU• Voltage adjustment U = U0 + ( G – G0) / dG/dU– More radical, but most effective


Bias working pointg p

• Critical for signal / noise optimization, not recoverable offline

Derived from Li h i ld A/GLight yield = A/G

N.D’Ascenzo


Voltage adjustmentg j

• Test beam preparation at DESY vs. CERN

observedTemperature-corrected

Required bias adjustment

Voltage-corrected(prediction)


q j mfrom observed gain shift

(average per layer)

Auto-calibration

• With single photo-electron (pixel) peaks the pixel photo-diode provides its own reference scale

P i i t l f it i t t• Promising tool for monitoring temperature-induced response variations

• Opens possibilities for further simplification• Opens possibilities for further simplification of calorimeter design – No external reference– Small amplitudesp– Loose stability requirements

• Stability of saturation correction


– Under study, so far OK

LED on board

• Attractive option - if auto-Attractive option - if auto-calibration of SiPM sufficient

• Proof-of-principle test borad, check for cross-talkcheck for cross-talk, uniformity, dynamic range,…

First tests:


M.Reinecke, DESYFirst tests:

electrical cross-talk small

Estimate Crosstalk

Scope in averagingdmode

Light in Tile 1Light in Tile 1

Crosstalk to tile 2 and tile 4

≈40mV

≈1mV

Crosstalk ≈ 2 5%

tile 2 and tile 4

Very encouragingOngoing optimization and measurements Crosstalk ≈ 2.5%

Consistent with optical XTg g p

(ligt uniformity, dyn. range,…)

Felix Sefkow TILC08, Sendai, March 3-7, 2008

Scintillator HCAL readout & calibration

17Mathias Reinecke EUDET annual meeting – Paris 8.-10. Oct. 2007

LEDs: to be continued

• Uniformity not yet sufficient (LED LED i i )(LED to LED variation)– Do not want to –remote-

control or tune each driver– Discussions with manufacturerDiscussions with manufacturer– Could in principle be solved

with pre-selection • Driver uniformity?y

• Dynamic range – the more, the merrier

A Lucaci

• Systematic investigation using CALICE DAQ

A.Lucaci, S.Weber


– ILC-SiPM ASIC and CRC

New ASIC on the test benches

• Auto-triggering and time measurementsgg g• ADC and TDC integrated• Power pulsing, low (continuous) power DAC

50 -100nsGain

selection

Slow Shaper

Slow Depth 16

Low gain Preamplifier

Analog memory1.5pF 0.1pF-1.5pF

Gain

LAL

50-100ns

15ns

HOLD

Shaper

Fast Shaper

Charge measurement

12-bit Wilkinso

nADCDepth 16

High gain Preamplifier

15pF

0.1pF-1.5pF

Conversion

READ

Variable delay

IN

Discri

DESY

4-bit threshold adjustmentDAC

output

Time

Trigger

Depth 168-bit DAC0-5V

80 µsdelayDiscriFlag TDC

Analo


10-bit DACTime

measurementTDC ramp 300ns/5 µs

Common to the 36 channels

g outpu

t

Analogue partg p

pedestal

1,166

Gain uniformity

1,158

1,16

1,162

1,164

DC

vol

tage

(V)

109

110

111

112

113

114

Vout

(mV)

1,152

1,154

1,156

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35

Channel

D

106

107

108

109

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35

Channel

G i d d t l if it d

L.Raux, S.Callier0.15

0.10

Cf=100fF Cf=100fF Cf=200fF Cf=300fF Cf=400fF Cf=500fF Cf=600fF Cf=700fF Cf=800fF

High gain charge output at different Cf Qinj= 1.5pC

• Gain and pedestal uniformity good (as expected from SiGe 0.35)

• Somewhat smaller gain and larger i th i i l ti

0.05

0.00

Vou

t (V

)

Cf=900fF Cf=1pF Cf=1.1pF Cf=1.2F Cf=1.3pF Cf=1.4pF Cf=1.5pF


noise than in simulation500400300200100

Time (x10-9s)

Input DACp

• SiPM bias adjustmentj• Must be extremely low power (~10µW), since always on!• Non-linearity ~2%, just OK

DAC linearity

5

ch0ch1ch2ch30 02

0,03

2,5

3

3,5

4

4,5

5

age

(V)

ch3ch4ch5ch6ch7ch8ch9ch10ch110 01

0

0,01

0,02

0 50 100 150 200 250 300

0

0,5

1

1,5

2

2,5

Volta ch12

ch13ch14ch15ch16ch17ch18ch19

-0,03

-0,02

-0,01


0 50 100 150 200 250 300

DAC

ch20ch21ch22

-0,04

L.Raux, S.Callier

Autotriggergg

• Fast shaper (10ns) and discriminatorC t i 0 5• Can triger on 0.5 p.e.

• Small threshold dispersion• Cross-talk 0.3%• Time walk 6ns

trigger time walk vs injected charge

60

62

64

)

S-curves

80

90

100

Série1

Série2

Série3

Série4

Série5

Série6

Série7

Série8

Sé i 950

52

54

56

58

time

(in

ns)

50

60

70

80

ger e

ffici

ency

Série9

Série10

Série11

Série12

Série13

Série14

Série15

Série16

Série17

0 1000 2000 3000 4000 5000

charge (in fC)

10

20

30

40

trig

g

Série18

Série19

Série20

Série21

Série22

Série23

Série24

Série25

Sé i 26


0388 390 392 394 396 398 400 402 404 406 408

DAC value

Série26

Série27

Série28

Série29

Série30

L.Raux, S.Callier

SPIROC meets SiPM

R.Fabbri, B.Lutz

• Still a lot of tests to be done• Understand response with auto-trigger, control thresholds


• Internal ADC not yet usable

Summaryy

• A nice concept for the 2nd generation HCAL prototype with b dd d l i d lib i i iembedded electronics and calibration system is emerging

• Readout with many new features( )– Auto-trigger (zero-suppression)

– On-chip digitization– Time measurement

Power pulsing– Power pulsing

• Auto-calibration of PPDs is a very powerful tool – but remains a challenge for sensor and electronics designchallenge for sensor and electronics design

• Test beam experience vital and rich source to develop calibration concept


concept

Back up slidesBack-up slides

Felix Sefkow TILC08, Sendai, March 3-7, 2008

Scintillator HCAL readout & calibration 25

Integrated layer designg y g

Reflector Foil100

Polyimide Foil100

HBU Interface500MCSector

wall

100µm 100µm PCB800µm

500µm gapASICTQFP-1001mm high

Top Plate600µm steel

Spacer1.7mm

Top Plate fixing

DESY

SiPMC t ABolt with innerM3 threadwelded to bottom

SiPMTile3mm

Bottom Plate600µm

Component Area:900µm high

HBU height:6.1mm

AbsorberPlates(st l)


plate6. mm(4.9mm without covers => absorber) (steel)

integrated

scintillator hcal read-out andout and calibration hcal read-out andout and calibration felix sefkow...

Documents