Super Belle CDC Shoji Uno (KEK) Dec-12 th, 2008 Basic design Electronics Test of pre-amplifier chips Wire stringing method Schedule
Baseline Design sbelle Belle
Main parameters Present Future Radius of inner boundary (mm) 77 160 Radius of outer boundary (mm) 880 1140 Radius of inner most sense wire (mm) 88 172 Radius of outer most sense wire (mm) 863 1120 Number of layers 50 58 Number of total sense wires 8400 15104 Effective radius of de/dx measurement (mm) 752 978 Gas He-C 2 H 6 He-C 2 H 6 Diameter of sense wire (μm) 30 30
Wire Configuration Present CDC 250 mm 1200 mm New CDC 250 mm
Yesterday CDC meeting One and half hours from 16:00 Main topics Test results of pre-amplifier chips Wire stringing method Several new people joined. One new KEK posdoc candidate 3 persons from KEK electronics group 5 foreigners One Japanese person (non current CDC member) 3 current CDC KEK members 13 in total : more than I expected. We are waiting for more people.
Prototype of readout board by Y. Igarashi ASB + Discriminator ASB + Discriminator ASB + Discriminator ASB + Discriminator Under 20cm FADC FADC SFP Optical Transceiver FPGA (CONTROL, TDC) FPGA (SiTCP) RJ45 RJ45 16ch/board BJT-ASB/Comparator part FADC: over 20MHz / 10bit FPGA : Vertex-5 LXT TDC: 1 nsec counting FADC reading Control FPGA: Spertan3A SiTCP RJ-45 for SiTCP RJ-45 for Belle DAQ timing signals SFP for Belle DAQ data line LEMO input x 3 LEMO output x1 Shielded substrate
Test of Amplifier by N. Taniguchi Three amplifiers Hybrid pre-amplifier (with receiver, gain:10) used in Belle-CDC ASD ( Amp + Shaper + Discriminator ) chip (with receiver, gain:7 ) used in ATLAS- TSC No production, now. ASB ( Amp + Shaper + Buffer ) developed by ASIC group of KEK Can be optimized for Super Belle CDC in near future. Check signal shape using oscilloscope Gas (Ar 90%+CH 4 10%) Fe 55 5.9keV X-ray Amp Receiver Small tube chamber Tungsten wire
Comparison HV=1.7kV ~300mV 100ns Pulse height [mv] pulse height (mv) 10 3 10 2 Belle AMP ATLAS ASD T-ASD ~220mV Belle AMP 40ns ATLAS ASD 1.5 1.6 1.7 1.8 1.9 HV (kv) HV [kv] ~140mV 40ns T-ASD
Comparison Noise level ~5mV Belle AMP ~5mV ATLAS ASD ~3mV T-ASD Saturation HV=1.9kV ATLAS ASD HV=1.9kV T-ASD ~600mV ~600mV distorted saturate
Results on test of amplifiers New ASIC chip is usable after some modifications. We can contact KEK electronics group closely.
Wire stringing Now, I am thinking a vertical stringing, not horizontal. Once, I thought the horizontal stringing. Vertical stringing with outer cylinder. Human can stand inside the chamber and can touch the wire. Inner diameter without the small cell part : ~500mm Inner diameter of present transition cylinder : 580mm Stringing with tension can be done from outer layer. Two-way method ( Belle-CDC ) is not necessary.
Discussion with technical stuffs We just started discussion with KEK technical stuffs. Calculation of deformation and stress Support method etc Need more man power.
My Personal Plan for Construction
Backup Slides
Hit rate Hit rate/wire (khz) 100 80 60 40 20 Small cell Inner E41R1128 HER 1.24A LER 1.7A Main 0 0 5 10 15 20 25 30 35 40 45 50 Apr.-5 th,2005 I HER = 1.24A I LER = 1.7A L peak = 1.5x10 34 cm -2 sec -1 I CDC = 1mA 10kHz 20 200kHz Layer of CDC
Hit rate at layer 35 Dec.,2003 410I**2 + 1400*I + 80 740I**2 + 470*I + 80 1600 1400 HER 3000 2500 LER Η it rate (Hz) 1200 1000 800 600 Η it Rate (Hz) 2000 1500 1000 400 200 500 0 0 0.2 0.4 0.6 0.8 1 HER Beam Current(A) 0 0 0.5 1 1.5 2 LER Beam Current(A) I HER = 4.1A Hit rate = 13kHz I LER = 9.4A Hit rate = 70kHz In total 83kHz Dec., 2003 : ~5kHz Now : ~4kHz
Simulation Study for Higher Beam Background by K.Senyo. BELLE BELLE 10 cm MC +BGx1 10 cm MC+BGx20
BG effect on analysis Talk by T. Kawasaki J / ψ ( μμ) K S ( π + π ) D * D * * ( D Dπ,D K 3π ) s B Eff Ratio-1 Nominal 56.8 % 0.0 % 5 BG 56.0 % -1.5 % 20 BG 49.0 % -13.8 % With 40% shorter shaping 20 BG 51.4 % -9.5 % Preliminary B Eff Ratio-1 Nominal 6.48 0.0 % 5 BG 5.69-12.2 % 20 BG 2.28-64.9 % With 40% shorter shaping 20 BG 3.86-40.5 % By H.Ozaki Major loss come from low tracking efficiency on slow particles. Efficiency loss on high multiplicity event is serious. Pulse shape information by FADC readout can save efficiency. SVT standalone tracker will be a great help (not included in this study). Jan24-26, 2008 BNM2008 Atami, Japan 18
Background effect on tracking H. Ozaki BNM2008 D * D * * ( D Dπ,D K 3π ) s Many low momentum tracks, the hardest case for tracking Gain in reconstruction efficiency of B D * D * Excellent with help of SVD 19
Idea for upgrade In order to reduce occupancy, Smaller cell size A new small cell drift chamber was constructed and installed. It has been working, well. Faster drift velocity One candidate : 100% CH 4 Results show worse spatial resolution due to a large Lorentz angle. A beam test was carried out under 1.5T magnetic field. So far, no other good candidate.
Small Cell Drift Chamber normal cell 16.0 13.3 small cell 102 116 128 5.0 5.4 sense wire field wire unit : [mm]
Photo of small cell chamber Just after wire stringing Installation in 2003 summer
XT Curve & Max. Drift Time Normal cell(17.3mm) Small cell(5.4mm) Drift time (nsec) 450 Normal Cell(layer29) 400 350 300 250 200 150 100 50 0-1 -0.8-0.6-0.4-0.2 0 0.2 0.4 0.6 0.8 1 Distance from sense wire (cm) Drift time (nsec) 450 Normal Cell(layer2) 400 350 300 250 200 150 100 50 0-1 -0.8-0.6-0.4-0.2 0 0.2 0.4 0.6 0.8 1 Distance from sense wire (cm)
Chamber Radius Inner radius Physics : Vertexing efficiency using Ks SVD determines the boundary. At present, the boundary is 15cm in radius. Outer radius New barrel PID device determines the outer radius. At present, 115cm is selected, tentatively. The boundary condition is important to start construction. Basically, CDC can manage any radius.
Wire configuration 1 Super-layer structure 6 layers for each super-layer at least 5 layers are required for track reconstruction. Even number is preferred for preamp arrangement on support board to shorten signal cable between feedthrough and preamp. Additional two layers in inner most superlayer and outer super-most layer. Higher hit rate in a few layers near wall. Inner most layer and outer most layer are consider as active guard wire.
Wire configuration 2 9 super-layers : 5 axial + 4 stereo(2u+2v) A 160*8, U 160*6, A 192*6, V 224*6, A 256*6, U 288*6, A 320*6, V 352*6, A 388*8 Number of layers : 58 Number of total sense wires : 15104 Number of total wires : ~60000
Deformation of endplate Number of wires increase by factor 2. Larger deformation of endplate is expected. It may cause troubles in a wire stringing process and other occasions. Number of holes increases, but a chamber radius also enlarges. Cell size is changing as a function of radius to reduce number of wires. The fraction of holes respect to total area is not so different, as comparing with the present CDC. 11.7% for present CDC 12.6% for Super-Belle CDC In order to reduce deformation of endplates, The endplate with a different shape is considered. Wire tension of field wires will be reduced. Anyway, we can arrange the wire configuration and can make a thin aluminum endplate.
Expected performance Occupancy Hit rate : ~100kHz ~5Hz X 20 Maximum drift time : 80-300nsec Occupancy : 1-3% 100kHz X 80-300nsec = 0.01-0.03 Momemtum resolution(svd+cdc) σ Pt /Pt = 0.19Pt 0.30/β[%] : Conservative σ Pt /Pt = 0.11Pt 0.30/β[%] : Possible 0.19*(863/1118) 2 Energy loss measurement 6.9% : Conservative 6.4% : Possible 6.9*(752/869) 1/2
About readout electronics At present, S/QT + multi-hit TDC S/QT : Q to Time conversion FASTBUS TDC was replaced with pipeline COPPER TDC. Three options, High speed FADC(>200MHz) Pipeline TDC + Slow FADC(~20MHz) ASD chip + TMC(or new TDC using FPGA) + slow FADC near detector. ASIC group of KEK Detector Technology Project is developing new ASD chip. New TDC using FPGA is one candidate for TDC near detector.
Summary When Belle group decides the upgrade plan, we can start construction of the new chamber soon. It takes three years to construct the chamber. Outer radius( and inner radius) should be fixed as soon as possible. Barrel PID determines the schedule. Inner radius should be determined by SVD. Supporting structure should be discussed. One big worry is man power. I hope many people join us when the upgrade plan starts.
Radiation Damage Test Gain degradation Total accumulated charge on sense wire(c/cm) a: 93 Plastic tube d: 94 SUS tube b: 93 Plastic tube + O2 filter e: 94 SUS tube + O 2 filter c: 94 Plastic tube f: 94 Plastic tube
Test chamber and beam test A test chamber with new cell structure was constructed. Part of inner most 20 layer( 8 layers with small cell + 12 layers with normal cell) A beam test was carried out in the beginning of June at π2 beam line of 12GeV PS. We confirmed the simulation for pure CH 4 is correct. Velocity under 1.5T is not faster than the present gas and the drift line is largely distorted due to larger Lorentz angle. Similar performance could be obtained using new S/QT module with less dead time. Many data were taken using 500MHz FADC, which was developed by KEK electronics group. Now, a student is analyzing data. We hope to get information about minimum necessary sampling speed for timing and de/dx measurement.
xt curve for new gas(7mm cell) He/C 2 H 6 = 50/50 Pure CH 4 Drift time (μsec) 100nsec Drift time (μsec) 100nsec Distance from wire (cm) Distance from wire (cm)
Drift Velocity Two candidate gases were tested. CH 4 and He-CF 4 12 10 CH 4 (100 %) He(80 %) : CF 4 He(50 %) : C 2 H 6 In case of He-CF 4, higher electric field is necessary to get fast drift velocity. In case of CH 4, faster drift velocity by factor two or more can be obtained, even in rather lower electric field. Drift Velocity (cm/μs) 8 6 4 2 0 0 0.5 1 1.5 2 2.5 3 3.5 E/P (kv/cm*atm)
de/dx Resolution The pulse heights for electron tracks from 90 Sr were measured for various gases. The resolutions for CH 4 and He(50%)- C 2 H 6 (50%) are same. Number of events 7000 6000 5000 4000 3000 2000 CH 4 (100 %) He(80 %) : CF 4 He(50 %) : C 2 H 6 P10 Pedestal The resolution for He- CF 4 is worse than Arbased gas(p-10). 1000 0-20 0 20 40 60 80 100 120 ADC value
Wire chamber Wire chamber is a good device for the central tracker. Less material Good momentum resolution. Cheap It is easy to cover a large region. Established technology Relatively easier construction. Many layers Provide trigger signals and particle ID information. Wire chamber can survive at Super-KEKB. Our answer does not change after the last WS in 2004. The beam background became smaller even for higher beam current and higher luminosity. We recognize the luminosity term is small, clearly.
CDC Total Current CDC Total current (ma) CDC current vs Beam current 1.4 1.2 1 0.8 0.6 0.4 0.2 Jul-2000 Jul-2001 Jun-2002 May-2003 Nov-2004 Apr-2005 0 0 0.5 1 1.5 2 2.5 3 Beam current (LER+HER, A) Maximum current is still below 1.2mA, even for higher stored current and higher luminosity. Vacuum condition is still improving. Thanks KEKB people for hard work. I hope there is still room to improve vacuum condition further.
Luminosity Dependences Feb, 2004 CDC BG did not change! Inner most Middle Outer
Occupancy No. of Luminosity Readout Q>0 Q>50 10 34 cm -2 sec -1 channel N Total time (μsec) N Hit Belle 1.5 8464 6-300 Babar 0.8 7104 2 ~700 ~350 Occ. = N Hit /N Total (%) Occ./Time (%/μsec) Max. drift time (μsec) Random trigger Occ./Time x Max. Drift time (%) Normalized by Lum.(%) Belle 3.5 0.58 0.4 0.23 0.15 Babar 4.9 2.45 ~0.6 1.47 1.84 x20 Bkgd in Belle CDC ~ x3 Bkgd in Babar DCH
at HL6 in KEK
Curved Endplate Deformation of endplate due to wire tension was calculated at design stage of present Belle CDC (Total tension: 3.5 Ton). Deformation(mm) 35.2 2.03 1.31 Present New
Weight Endplate Al, Thickness : 10mm (12mm) 110kgx2 = 220kg (264kg) Outer Cylinder CRRP, Thickness : 5mm 210kg Electronics Board G10, 48ch/board 0.3kgx315 = 95kg
Present support
Calculation of deformation and etc Deformation of Aluminum endplate Thickness of endplate 10mm 12mm Deformation 1/t 3 1 0.58 Tension of field wire 120g 80g Gravitational sag, Sense : 120μm(80g), Field:300μm(80g) Total tension 4.8ton Stress calculation Thickness of outer cylinder CFRP:mm Transition structure between endplate and outer cylinder Support structure Structure for wiring jig Simpler one as compared with Belle-CDC Etc.
Installation Removing and installation can be done using similar small bar in horizontal direction. Cathode installation in vertical direction CDC installation in horizontal direction
Signal Shape Each signal shapes are not same. Rise time : ~10sec Pulse width : ~200nsec. Maximum drift time : ~300nsec
Timing resolution Good timing resolution for the drift time measurement is key item. 250MHz sampling is not good enough. Present leading edge measurement shows 130μm resolution. Spatial resolution (μm) 700 600 500 400 300 200 100 0 FADC test Before correction After correction 0 10 20 30 40 Sampling width (nsec) 1nsec resolution is required.
Sampling rate for energy loss measurement Slow sampling rate is good enough for energy loss measurement. 20MHz is OK. de/dx Resolution measurement (%) 分解能 ( % ) 12 10 8 6 4 2 0 ns 0 100 200 300 Sampling サンプリングレート width (1/T) (nsec)
Purposes of the Readout board Prototype A study of the CDC readout scheme Charge measurements by FADC Drift time measurements by FPGA base TDC A evaluation of ASB for CDC readout A study of the noise diffusion from the readout board to CDC We hope to study about CDAQ/Front-end data transport.
ASB part diagram ASB Amp. Shaper Buffer 4ch/chip Gain : -360mV/pC ~ -1400mV/pC (4 step variable) Power consumption : ~18mA
LOGIC part diagram ASB Discriminator ASB Discriminator ASB Discriminator ASB Discriminator CONTROL 3x4 TIMING LVDS 4x4 pairs LVDS 8 pairs FADC Ti ADS5287 FADC Ti ADS5287 5 LVDS CLKs LVDS 8 pairs 5 LVDS CLKs 16 16 TDC (with FIFO) De-serializer FIFO CONTROL RocketIO GFP RocketIO GFP Vertex-5 LXT (XC5VLX50T, IO:360pin) 48 CLK For GFP LVDS 2 pairs LVDS 2 pairs LVDS 4 pairs Spertan3A SiTCP 4 8 16 PUSH SW CLK 50MHz 3 8 32 SFP (Optical connector) 15 LED DIP-SW TEST PIN SFP (Optical connector) 100base PHY TEST PIN DIP-SW RJ-45 RJ-45 LEMO LEMO LEMO LEMO DAC (Vth) 8 Sampling CLK 20~40 MHz CLK 125MHz CLK 42.33MHz
Schedule plan 2008/11 Specification design 2008/11,12 design and drawing the circuit ASD part (T.Taniguchi-san) Digital part (M.Saito-san) 2008/12 end order the substrate 2009/2 Check the mask pattern M.Ikeno-san etc 2009/3 Start the practical study
Introduction ATLAS-ASD Pre gain 0.8 V/pC, main x 7 Fe 55 ASD buffer +HV ASD oscilloscope T-ASD (Taniguchi-san, ASIC group) ~ 7V/pC Fe 55 +HV ASD oscilloscope
Conclusion Gain : T-ASD is smaller (0.6 x ATLAS ASD, half of Belle AMP) Noise level: T-ASD is smaller Resolution : T-ASD seems to be better than ATLAS ASD Rise time is ~ 20 ns for ATLAS ASD and T-ASD Saturation : ~ 1.9kV Beam test using test chamber and new ASIC chip ~ Apr, 2009
ATLAS ASD HV=1.6kV HV=1.7kV ~80mV ~220mV 40ns HV=1.8kV HV=1.9kV ~500mV ~600mV distorted
T-ASD HV=1.6kV HV=1.7kV ~50mV ~140mV 40ns HV=1.8kV HV=1.9kV ~300m V ~600mV saturate
Belle AMP (HV=1.8kV)
900 800 700 600 500 400 300 200 100 0 1.5 1.6 1.7 1.8 1.9