TOP beamtest, electronics and plans

42
TOP beamtest, electronics and plans Gary Varner University of Hawaii for the bPID group Gary Varner , University of Hawaii for the bPID group, BPAC, February 27 th , 2012

Transcript of TOP beamtest, electronics and plans

TOP beamtest, electronics and plans

Gary Varner University of Hawaii for the bPID groupGary Varner, University of Hawaii for the bPID group, BPAC, February 27th, 2012

What tested – detector components

R 5000mml=1200+1310+30mm=2540mm

R=5000mm

(Integrated Front‐end)(Integration atw=450mm

440mm

(Integration at  Nagoya University)

t=20mm

54mm

54mm

2Belle II barrel Particle Identification (p/K separation) Upgrade:  Start running at SuperKEKB in 2015

Goal: Data versus Simulation• Use data to tune/confirm Monte Carlo• Use data to tune/confirm Monte Carlo• Results below were for old PMT type & CAMAC readout

DataPDF (floating norm)

TDC (counts/25ps)TDC (counts/25ps)

ch16~19 ch32~353

Evaluate Highly Integrated ReadoutW fWaveform 

sampling ASIC 

Clock jitterClock jitter cleaners

4

Beam test set‐upp

5x5mm2

Trigger5x5mm2

TriggerVeto counters

Tracker Timing counters

Trigger TriggerTOP Tracker

• 2‐types scintillator trigger counters– Larger counters (50x50mm2) for system checks and beam alignmentLarger counters (50x50mm ) for system checks and beam alignment– Smaller counters (5x5mm2) to realize a collimated beam

• SciFi tracker: 2D, ~3m distance, resolution~1mm• Veto counters: to reduce hadronic shower eventsVeto counters: to reduce hadronic shower events• Timing counters: resolution~22ps

5

Shipping Woes & Repair (Thanks FNAL!)

6

Ready for beam

(not pictured:  Yasuyuki Horii (photographer) [Nagoya], Alan Schwartz    

7

( p y (p g p ) [ g y ],[Cincinnati], Lynn Wood [PNNL],  Marc Rosen & Casey Honniball [Hawaii] )

DAQ system Hawaii Tracker(ASIC SciFi readout)

DC power for pFront‐end Modules

cPCI CPU and fiber

VME/ /

cPCI CPU and fiberCard (custom) = “PC1”; USB readout of CAMACTrig/Timing/

ProgrammingModule

of CAMAC

Dell PowerEdge2970 Server (“PC2”)

Firmware programming (USB to remote JTAG)

8

Trigger/HVCAEN HV

Nagoya Trigger Logic/Timing Modules

Pico‐second Calibration LaserTrigger

CAMAC for 

Trigger

“SuperKEKB RF clock phase measurement, Nagoya timing counters/tracker 9

Some LimitationsO l 20/32 t bOnly 20/32 tubes from Hamamatsu

JT0180 tripped frequently (not used)

82% of electronics h l “ d”channels “good”(bad RF amps, solder connections, etc.)

384 channels @ 2.7 GSa/s “oscilloscope on 

10

pa chip”  (1.5TBy/s)!

Event Sizes/Rates• Read 4 “windows” of 64 samples from each channel (each channel has 512 windows of t )storage); 

• 1 fiber link/module

• Each event ~74kB/module

• Total ~ 230kB/event• Total   230kB/event

• Logging rate obtained was 130‐160 events/spill160 events/spill

(~10MB/s PCI bus + CAMAC USB)

Able to log about 100k beam events per day (~1‐2 M single photons), with comparable number of laser calibration events

11To do better will need to zero‐suppress/online feature extraction (next beam test)

Example single‐photon signals

Clean hit (center of Anode pad) Depending upon amplitude, “cross talk” hit (red) == remove by filteringtalk  hit (red) == remove by filtering

( f f )(16 waveforms from JT0164)

Can (in principle) decouple PMT, i i & d t t lk

12

wiring & readout x‐talk

Start timing/tracking

Veto counter(Sci-Fi)

ScintillatingFiber Tracker

plane (fibe

r  #)

Trigger Trigger7.5x7.5mm2

Trigger

(Sci Fi)

7.5x7.5mm2

Trigger

Y‐p

TriggeriTOP

Trigger

X‐plane (fiber  #)

~30ps Full GEANT4 beamline MC

13

Cherenkov Photon Arrival Patterns (MC)

120°90°

40°

14e‐ e+

Raw (uncalibrated) Timing DistributionsOld, generic MC

15

Rotation: Polar angle: cos() = 0.5

16

Initial Raw Timing Distributions (cos()=0.5)

Data (Exp 10 = 1 shift) GEANT4‐based MC(limited to 50ns window)

17

Analysis itemsy• Most of the waveforms look very clean

– Precise timebase/timing calibration needed– Leading edge timing extraction algorithmg g g g

• Photon yield (can do without precise timing)C i i• MC comparisons on‐going

• “direct” light is easy, mirror reflections more g y,complexB lit d t d b k d b• Beam quality good, study backgrounds by overlaying events

18

One complication• At Super‐KEKB the timing of signals should be fixedrelative to trigger (system clock)

p

relative to trigger (system clock)– But it is random with respect to 21 MHz derived (Super‐KEKB RF clock). – thit from waveform must be combined with tFTSW from CAMAC TDC to align hit FTSW

events.– tglobal = thit + tftsw 21 MHz period (~50ns)

thit (ns, uncal.)19

Standard Laser Run ‐ DistributionsDST2‐Based Analysis

Black – laser run dataRed – profile histogram ofRed  profile histogram of same data. Blue – linear fit

No fine calibration applied: assumed 2.7 GSa/s for all samples; 25 ps / count for CAMAC TDC.Time extracted by software fixed threshold discrimination (‐40 ADC counts).

20

Phase Synchronous Triggering @ Beam Test

Signal Generator Laser Control Unit

Softwaretriggered.

Set to trigger on 1‐input only.

Signal Generator (Agilent 33250A)

Laser Control Unit (PiLas EIG1000D) Coincidence UnitMod3

Triggered synchronous w/ 21 MHz front end clock

Mod0

21 MHz front‐end clock.

Laser Head(PiLas PIL???S?S) Gate Generator

SL10s Quartz BarMod1

Mod2

FTSWCAT6 timing/trigger

CAMAC TDC ECL  NIM 

CAT6 timing/trigger

In this mode, the phase of 21 MHz clocks should be fixed 

(25 ps least count) Converterrelative to the global trigger time.

21

Phase Synchronous Triggering @ Beam Test

Signal Generator

Softwaretriggered.

Signal Generator (Agilent 33250A)Mod3

Triggered synchronous w/ 21 MHz front end clock

t = 140 ps

21 MHz front‐end clock.

Input is 5 0 V TTL

Output is 2.5 V pulse.

Input is 5.0 V TTL.

Probably contributed a large amount to the jitter.

Actual FTSW phase measurement jitter may have been less than this measurement.(Triggering for normal beam and laser runs did not have this trigger level issue.) 22

Phase Synchronous Triggering @ Hawaii 

Gate Generator (To21 MHz 

Gate Generator (To Provide Shorter Trigger Pulse)

Gate Generator(Latch Mode)FTSW

clock

Trigger in to FTSW21 MHz synchronous trigger Cable delayed 

f

CAMAC TDC

triggeroutput from gate generator (to match 

f TDC)(Phillips 7186) TDC startTDC stop range of TDC)

23

Phase Synchronous Triggering @ Hawaii

Very fussy – could easily get > 100psWith poor connections/coupling

• Consistent with FTSW clock jitter measured on scope (and previous measurements in Nagoya). 24

Hawaii Test Setup – random phasep p

Signal Generator Pulser (Avtech

Signal generator runs asynchronously

Signal Generator (Agilent 33250A)

Pulser (AvtechAVMP‐2‐C‐EPIA) TTL  NIM

Pulse sent to all ch. of 1 ASIC

Gate GeneratorMod2 Differentiator1  8 splitter

Fanout

FTSWCAT6 timing/trigger

NIM Trigger In

CAMAC TDC ECL  NIM 

21 MHz Latched TriggerTDC Start

TDC 

(25 ps least count) Converter

Stop

25

Measure & Apply Correction

Half of 1/(21 MHz)

For fixed amplitude, well behaved signal… in this case we estimateestimate:core sigma ~ 58 ps (73 ps RMS)

This timing calibration consistent all samples (32k samples, 128 unique timing)

26

Previous result:N b f hit h l ( 0 5)Number of hit channels (cos=0.5)

DataSimulation

• Nhit~7   / 8 MCP‐PMTs (without PMT6)• Good agreement with simulation

(Histogram for simulation is scaled according to number of PMTs.)

27

Extracting number of p.e.Preliminary!! Expectations:  ~1 p.e. per PMT 1‐12

~0.4 p.e. per PMT 14‐24

I iti l

12 p.e. * ~82% electronics ~ 10 p.e. 

Initial scan: charge sharing <~20%

Evidence for “cross talk” double‐hits

Work in progress

28

Work in progress…

T‐1019 Preliminary Summary• Data taking completed less than 2 months ago     (transit back recovery from spending holidays(transit back, recovery from spending holidays working…)

• About 534,000 events at normal incidence, ~300k at cos(theta)=0.5 ( )– Preliminary results show data clean, but two issues found: event time start trigger windowfound:  event time start, trigger window

– Much analysis to be done – a very rich data set

/• Calibration & Debug at KEK/Hawaii• Setting up to extended cosmic running at KEKSetting up to extended cosmic running at KEK

29

Lessons Learned: things to improveg p

• Timebase servo‐loopp– Firmware needs to be re‐writtenPossible hardware change– Possible hardware change

• Better thermal management (85C redline ops)• SCROD v2 (“final” form factor)• HV/packaging SL 10 into module by HPK/Nagoya• HV/packaging SL‐10 into module by HPK/Nagoya• Demonstrate DSP (real time) data reduction• In‐situ (on demand) calibration• Sample pointer dephasing fix (next slide)• Sample pointer dephasing fix (next slide)

30

Sample strobe Dephasing

Ch. 1 <10ps RMSNagoya test with RF pulser

hCh. 7 Ch. 7

31

IRS3 Single Channel 32

• Sampling: 128 (2x 64) separate transfer lanes

Recording in one set 64, transferring other (“ping‐pong”)

separate transfer lanes

• Cleaned up input routing

• Storage: 64 x 512 (512 = 8 * 64)

p p g

Storage: 64 x 512 (512 8 64)

• Wilkinson (64x1): was (32x2)Wilkinson (64x1): was (32x2)• 64 conv/channel

bPID Short‐term Schedule2012 3 4 5 6 7 8 9 10 11 12

‐ Readout and structure confirmation 

Prototype test‐ Cosmic ray‐ Beam test

Set up Data take

Beam testQBB, structure‐ Prototype

T t

DesigningStructure prototype

‐ AssemblyMCP‐PMT‐ B‐field test

Test

MeasurementB field test

Readout‐ Test System check

New version check with PMT Next version‐ PrototypeSoftware‐ Reconstruct

New version check with PMT

Prototype test

Next version

MC + exp. data studyReconstruct‐ Optimization Simmirror test

33

Summaryy

• T 1019 beam test very valuable• T‐1019 beam test very valuable• Analysis  pre‐prod prototypes in 2012

– Verify ASIC– Final board‐stack configurationFinal board stack configuration

• Continued cosmic ray testing at KEK; more i i / i hexperience operating/tuning the system

• Definitive beam test with real‐time feature extraction late 2012/early 2013

34

Back‐up

35

Highly integrated readout

4x4 anode “1 inch”

CH1MCP-PMT (HPK SL-10)

6.4 psσ ~ 38 37RMS

CH2

σ ~ 38.37

36NIM A602 (2009) 438Single p.e. resolution

Trigger/Timing Distribution (FTSW)gg gFrom Nakao‐san’s documentation:

37

Back‐End Data Acquisition

• cPCI crate (above) with custom card (DSP_cPCI, left), that receives data from front end modules.

• This crate collects all data for the cosmic and beam testscosmic and beam tests.– Collects data from front‐end modules by 

Giga‐bit fiber optic link.ll d– Collects CAMAC data via USB.

– Horizontal slice is “FINESSE” back‐end o o ta s ce s SS bac e dprototype for COPPER system

38

Prototype HV/PMT issuesyp

• PMT case at HV• Pins not uniform diameter, location• Difficult to register 8 (32) onto a plane• No detailed information about anode 

fi ti l t i kconfiguration – layout is guesswork (poor cross‐talk performance)• Took 4x iterations before “mostly

HV boardTook 4x iterations before  mostly 

solved” HV sparking/arcing issues (run VERY hot) • Working design does not fit in electronics envelopeW k ith HPK t k b MCP

39

• Work with HPK to package bare MCP‐PMTs into an integrated module

Current Status: Calibrated time resolutionCurrent Status: Calibrated time resolution• Exp. 8 (normal incident), 4000 < event ID < 8000.

• Channel 9 of PMT 16.

• Only odd samples analyzed.

ents

mbe

r of e

ve

Preliminary

NumResolution = (123 ± 14) psec.

Need further investigation.Corrected time [nsec]

Need further investigation.Need extension to all good channels.

40

US DOE “WBS” Milestones/schedule

41

Major (short‐term) milestonesSimplified TOP Electronics Schedule

2011 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012

Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov

Complete beamtestComplete beamtest

Set‐up @ UH, KEK

Cal/T0 studies

Data Reprocess

Beamtest Analysis

C l l iComplete analysis

Intense WBS work

Complete WBS

BPAC Review

CDR Review

Pre CD‐1 Review

CD‐1 Review

Commission CRT run

Fast Feature Extract

TOP CRT run

CRT results (FFE)

IRS3 evaluation

SCROD Rev. B

Carrier Rev. B

Interconnect Rev. B

HV & integrate

Board stack V2

42IRSx Pre‐prod proto

DSP_FINESSE Rev B