G. Cavoto 1

Timing Counter Status

G.Cavoto

INFN Roma

Feb 18th 2009

Timing Counter 2008 configurationMonitoring and Calibrations

Timing resolution (TC and MEG)

G. Cavoto 2

Timing Counter (TC)

Two sectors, DownStream and UpStream, each with – 15 scintillator bars readout by PMTs (TICP)

» Trigger » Measure time, and z

– 256 Scint. Fibers readout by APDs (TICZ)» (Meant for) geometrical trigger selection » Measure z

QuickTime™ and aTIFF (LZW) decompressor

are needed to see this picture.

G. Cavoto 3

Hardware upgradesNew EVAL bag for N2Very tight! He concentration at

few times atm. bkg level

Reduced APD noiseLight tight Tedlar foil added

Some maintenance work (fix connectors,cables…)Both US and DS TC remounted in late spring

G. Cavoto 4

TICZ readout tests After installation, test with ‘pass-thru’ cosmics

• All analog channels (8 APD in one ch.) show good signals!

• US TICZ digital readout working!• DS TICZ worsened

by higher environmental noise

• 3 (out of 8) DS TICZ readout sections failed during the 2008 run• “on-board” hardware problems: fixing deferred to 2009

US TICZ

fiber #

G. Cavoto 5

Laser system– Additional timing calibration device

» Providing 532 and 266 nm light at 50 Hz » Optical fibers distribution to TC bars (and XEC)

– Problems in operation:» environment temperature became too high

cavity unstable drift in the feedback signal

electronic saturation, damages of optical components and electronics.

3m cavity 48 MHz, 1064nmDiode pumped Nd:YVO

Acustic-optical pulse selectors

50 Hz2 stages pulse amplifier

50Hz @ 532nm

256 nm-For 2009 run:- Proper box with actively stabilized temperature- Retuning of cavity- New feedback electronics

G. Cavoto 6

PMT gain equalization

• Select Michel positron hitting bar center(magnetic field ON)

• Change HV to havesame average chargefor PMTs of the same bar

Bef

ore

Aft

er

G. Cavoto 7

TC measurements

DRS3 digit izat ion f or Double Threshold Discr iminator NIM pulses t0 and t1 extracted with waveform template fi t to NIM pulses.Checked with cross-correlati on method (opt imal f ilt ering).

t0 T h

veff b0

c0A0

e+

L

hPMT0 PMT1

t1 T L hveff

b1 c1A1.

T : time of positron at the impact point on first hit bar (connected to the positron track from DCH)

h : impact point along bar length (z direction)

effective velocity

amplitude of PMT signal

G. Cavoto 8

TC calibrations Timing with DTD NIM signals

c0

c1

L

2veffb0 b12

b0 b1

Time walk correction for each PMT

Michel nmult=3

Time offset between PMT of the same bar (z offset calibration)

Cosmics

Relative time offset between bars

Boron

Effective velocity for each bar

(z measurement)

Pass-thru cosmics

Absolute time offset between positron

and gamma

_ Dalitz 0

0 e(e )

veff

B(p,)C

G. Cavoto 9

‘Double’ and ‘triple’ hit events

TA

TB

On sample of two hit adjacentbars (‘doubles’) test time walk correction (c0 and c1)

|z|

bar

#

e+

TC

t0C t0B2

t0A 1

2

c0CA0C

c0BA0B

c0AA0A

.

1st bar

On events with three adjacent hit bars (triples)minimize the differences (for all the bars)

T TA TB

ns

G. Cavoto 10

TC time resolution

Upper limit on average time resolution () in 60-90ps range

Includes effect of DRS digitization (~10 ps)

(estimated sending same signal to two DTD inputs)

Estimate of single bar time resolutionAssuming the two bars to have the same intrinsic time resolution

T

2

G. Cavoto 11

TC time resolution stability

• Same TW calibration constants

• Stable over timeno need of different sets of constants

Runs 24xxxRuns 25xxxRuns 26xxxRuns 27xxxRuns 29xxxRuns 30xxxRuns 31xxx

G. Cavoto 12

Inter-PMT offsets

Dedicated cosmics runs taken in different down-time periods– Select event with 1 or 2 hit bar

Assume symmetric distribution along the bars

Most of the bars have this flat distribution

t1 - t0 [ns]

Data taken in different periods very consistent (TICP very stable!!!)

G. Cavoto 13

MEG physics runs TC hitmap

t1-t0 [ns] t1-t0 [ns]

Distribution nicely “justifies”

Need veff to have the correct z scale

Before After

DS

US

G. Cavoto 14

Effective velocity with TICZPass-thru cosmics, 2 hit fibers (clusters) expected

Single-cluster inefficiency: 27%

due to cosmics geometrical inefficiency + dead/hot fibers (5%)

zbar-zfiber (cm)

Bar16

z fiber 1

2veff (t1 t0)Using

v eff (c

m/n

s)

Syst err.1.5%

G. Cavoto 15

Inter-bar offset extraction

• Gaussian fits.– Cosmic background evaluated in runs with no-beam and

same trigger requirements (TC & XEC)

• Syst error on calibration constants at 50ps level

T[ns]

Bar17-signal-bkg (rescaled)

Bar14-signal-bkg (rescaled)

T[ns]

Boron events 4.4MeV (XEC) and 11.7MeV(TC)

T TXEC LXEC

c

t0 t12

LTC

c

G. Cavoto 16

Inter-bar offsets monitoring

Sep08/1013/1020/1025/1010/1124/1101/12

Tmean [ns]

bar#

- No clear trend vs time- Relative offsets stable

Bar15: change in DRS board

tim

e

relative to bar 17

Weeklymonitoring(periodic DB updates)

G. Cavoto 17

DCH-TC match hitmap

Extrapolate tracks from DCH to TC bars

Given a track and a TC bar hit matching efficiency is 91%Some data/MC discrepancy

zzDCH zbar [cm]

G. Cavoto 18

MEG absolute time offset Dalitz 0 events

– Same topology as signal !

– Gamma/positron energy range (can be chosen) same as for signal

– Worse resolution due to LH2 target Comparison with signal is not exact

0 e(e )

ns

Te TXEC LXEC

c

T

LeTC

c

Centre of signal window

Control sample (flight length correction)

MC validation

Bar17μ = 26.06±0.01ns = 267±10ps

G. Cavoto 19

Timing systematics check

Boron (E=4.4MeV)

bar#

T

(e

) [n

s]

Dalitz (E>45MeV)

Residual time walk effecton XEC timing

Dalitz before calibration

Dalitz after calibrationwith Boron

bar#Residual inter-bar differences(different target for 0 Dalitz)

T

(e

)-T

(

e)(b

ar1

7) [n

s]

Relative to bar 17

G. Cavoto 20

Summary

• TICP very stable over the 2008 run.• TICZ: 13 out of 16 sectors working. • Calibration strategy developed and

successfully applied to data.• Data monitored during data-taking.• Some fine-tuning still needed

– Residual bar differences – Detailed data-MC comparison

G. Cavoto 21

Outlook for 2009

– Detector activity• New APD electronics

– Shorter shaping time to fit trigger timing– Currently prototyping new el. boards

• LASER• Upgrade of TC slow control

– On track for a 2009 data-taking period

– Data analysis• Calibrated TC data and TC simulation available for MEG

analysis• MEG timing measurements well underway

G. Cavoto 22

Back up

G. Cavoto 23

New APD electronics

• Requirements: – fast signal ( 3-5 times faster) 15ns achievable.– better EMI immunity– Implement redundant schemes for the command transmission (I2C)– Improve over all settings capability ( on board test /calibration signal,…)- Bread-boarding is started, final design of the prototype board will be

sent for the production by Dec 08.

Test pulse response

Move Amplifier tothe APD board and/or differential input

To insert APD into trigger