LumiReview ppt P. Denes p. 1 Luminosity Monitor Review Concept Instrument TAN (TAS), count n Requirements Implications of the requirements on the design of the luminosity monitor New Mechanics Updated design for inert gas ionization chamber Suggestions/Compatibility with solid state detector Electronics Previous test beam results and current test beam needs Bill Turner, who would normally be presenting much of this, can not be here today, so we are filling in (at times perhaps imperfectly)
LumiReview ppt P. Denes p. 2 Luminosity Monitor Concept D1tripletTAS tripletD1 TAN IP 140 m n LL RR Luminosity N MIP from n shower Crossing Angle L + R Instrument TAN Massimo
LumiReview ppt P. Denes p. 3 TAN IP Instrument a Copper Bar
LumiReview ppt P. Denes p. 4 TAN Instrumentation Slot
LumiReview ppt P. Denes p. 5 Detector Constraints - I. Charged particles swept away (Gas) Detector placed after several INT (few ) m ~1 n per 3 pp interactions (in the acceptance) Offset due to ±150 µrad crossing angle Horizontal, vertical or 45° crossing ~ 80 x 80 mm 2 for detector Segment (for position) LHC Project Document No. LHC-B-ES-0004 rev 2.0
LumiReview ppt P. Denes p. 6 Detector Constraints - II. Signal collection time < 25 ns Modest S/N performance: Shower fluctuations ( N MIP )/ N MIP ~ 30% 1 P = 1% needs N ~3000 pp interactions Desired precision Number of events n / pp interaction
LumiReview ppt P. Denes p. 7 Detector Constraints - III. Given large hadronic shower fluctuations, SNR ~ 4 or 5 is sufficient Effect of SNR on N to achieve P
LumiReview ppt P. Denes p. 8 Requirements Update
LumiReview ppt P. Denes p. 9 Requirements Total L Absolute L from experiments L / L ~ 1% Reproducibility ~ 1% Integration time ~ 1s Bunch-by-bunch (most stringent: ) L / L ~ 1% Integration time ~ minutes And bringing beams into collision
LumiReview ppt P. Denes p. 10 Total Luminosity m=0.33 INEL =80 mb SNR=5
LumiReview ppt P. Denes p. 11 Bunch-by-bunch Luminosity m=0.33 INEL =80 mb SNR= bunches
LumiReview ppt P. Denes p. 12 Luminosity Optimization Transverse view at IP Beam 1 ** D beam-beam separation Beam 2 d for ,d << * D = d + know d, measure to get to 0.1 * L / L 0.5%
LumiReview ppt P. Denes p. 13 Current Design Options Ionization Chamber GasSolid Active medium Radiation Hardness Mechanical stability Speed Noise (SNR) Ar + N medium replaceable fixed components low mobility doable CdTe hardness to be shown depends on contacting higher mobility trivial Very high TID - up to ~ MRad/10 yrs Access as infrequently as possible
LumiReview ppt P. Denes p. 14 Argon Ionization Chamber I0I0 = x GAP /v D charge/hadron = Q 0 x x GAP x P [Atm] x N GAP I 0 = 2 Q 0 v D PN GAP = 9.7 e – /MIP/mm x 231 MIP/h x x GAP x P x N GAP = 0.72 nA x v D [µm/ns] x P [Atm] x N GAP V+ x GAP N GAP =2
LumiReview ppt P. Denes p. 15 Electronics Very high radiation levels 50 Cable between detector and preamp CDCD Z 0 (50 ) (Virtual) 50 If sufficient signal, 50 resistor can be real, otherwise 50 resistor has to be “virtual”
LumiReview ppt P. Denes p. 16 Modified Ionization Chamber Design Area constrained: 4 quadrants A ~ 4x4 cm 2 Capacitance per gap C GAP = A/x GAP Gap dimensions Gap topology Number of gaps Gas properties Update of previous mechanical design Goal: simplified construction higher reliability Consider all configurations which fit into Cu bar volume Consider different gases / mixtures (simulation)
LumiReview ppt P. Denes p. 17 Optimizing the layout Coax cable N GAP x GAP “50 ” Ionization Current I0I0 T Time Constant 50 x C DETECTOR
LumiReview ppt P. Denes p. 18 Previous Approach I. Series-Parallel connection x GAP VV3V3V5V5V 2V2V4V4V
LumiReview ppt P. Denes p. 19 Previous Approach II. N GAP = N SER x N PAR Effective gas volume = x GAP x N PAR C DETECTOR = C GAP x N PAR / N SER Parasitics - Hard to achieve C DETECTOR, complex mechanics N GAP = 60 x GAP = 0.5mm N PAR = 10, L = 5mm N SER = 6
LumiReview ppt P. Denes p. 20 Improved Speed Possible Previous Operating Point simulated with MAGBOLTZ
LumiReview ppt P. Denes p. 21 Drift Velocity Velocity(cm/microsec) E(V/cm-atm) Ar+1%N2 Ar+1.5%N2 Ar+2%N2 Ar+3%N2 Simulation Measured Data vs. Simulation Ar (98%) N2 (2%) Ar (97%) N2 (3%) Ar (96%) N2 (4%)
LumiReview ppt P. Denes p. 22 Example: Constant 6 mm Gas Volume Current Waveform into Preamplifier Current [µA] at 1 Atm Ar/N 2 (96::4)
LumiReview ppt P. Denes p. 23 6x1 mm Gaps N has to be even Effective gas length Intrinsic capacitance Drift velocity L x v/C 5 mm 6 mm (10/6) / 0.5 6/1 3.2 cm/µs 4.5 cm/µs 1::1 Previous version This version
LumiReview ppt P. Denes p. 24 Detector Concept One ground “comb” milled from a solid Cu block Four signal “combs” Ceramic insulation/alignment pieces (machineable MACOR) Detector mechanical design: T. Loew, D. Cheng Vessel mechanical design: M. Hoff Fabrication design: N. Salmon, A. Mei
LumiReview ppt P. Denes p. 25 Assembly I. Signal comb Alignment features (explained below)
LumiReview ppt P. Denes p. 26 Assembly II. Ground planes 2 mm Cu / 1 mm gap (i.e. 4 mm between plates) 40 mm depth < 10::1 aspect ratio - OK for machining Solid ground separates all 4 quadrants
LumiReview ppt P. Denes p. 27 Assembly III. Alignment features
LumiReview ppt P. Denes p. 28 Assembly IV. One ceramic face is metallized for bias filter and connections to rad- hard coax cable
LumiReview ppt P. Denes p. 29 Available Space 96 mm 67 mm
LumiReview ppt P. Denes p. 30 Quadrant Dimensions Not to scale 94 Ceramic Stainless Steel Copper
LumiReview ppt P. Denes p. 31 Absorber Bar
LumiReview ppt P. Denes p. 32 Detector Housing (TAN Insert) Detector area Services to detector Direct connect or patch panel Compatible with any detector
LumiReview ppt P. Denes p. 33 Detector Housing Signal+HV Connectors Gas Metal pressure seal Detector Volume Detector Volume Gas
LumiReview ppt P. Denes p. 34 Constraints - I. Thin wall dimension designed so that vessel withstands 15 Atm.
LumiReview ppt P. Denes p. 35 Constraints - II. Double-insulated, high-pressure SMA feedthroughs Rad-hard (SiO 2 ) cable inch (3.6 mm) ø semi-rigid
LumiReview ppt P. Denes p. 36 Complete Insert Insulated from TAN by 0.5 mm ceramic Rad-hard semi-rigid coax insulated by ceramic beads from housing Compatible with standard lifting mechanism
LumiReview ppt P. Denes p. 37 Detector Mounted in Vessel
LumiReview ppt P. Denes p. 38 Integration
LumiReview ppt P. Denes p. 39 Engineering Solution in Preparation
LumiReview ppt P. Denes p. 40 CdTe
LumiReview ppt P. Denes p. 41 Alternate CdTe Layout Reconstruction with 10-disk geometry is complicated Could be simplified by constructing quadrant detector using x cm 2 CdTe (several sources) 3 x 3 array of x cm 2 CdTe 250 µ between chips
LumiReview ppt P. Denes p. 42 CdTe Assembly Using Spring Contacts
LumiReview ppt P. Denes p. 43 Detector Housing (TAN Insert) Same idea, but more cables (and no gas lines)
LumiReview ppt P. Denes p. 44 Assembly
LumiReview ppt P. Denes p. 45 Current Pulse from 2 x 2 cm 2 CdTe into 50 , 93 pF 7 ke - /MIP, 280 MIP
LumiReview ppt P. Denes p. 46 CdTe vs. Ar+N 2 CdTe - radiation-induced leakage current CdTe - Leakage current ~ T 2 e T CdTe - complicated reconstruction - can be solved with different mechanics CdTe - Faster than Ar+N 2, deconvolution required Ar+N 2 - Active medium “easy to replace” Ar+N 2 - Signal smaller than CdTe (less important - have to average over many pulses due to shower fluctuations) Franco
LumiReview ppt P. Denes p. 47 Simulation 50% of signal per quadrant 6x1 mm gaps 6 ATM Ar (96%) N2 (4%) 4 cm/µs drift velocity I 0 = 1 µA I0I0
LumiReview ppt P. Denes p. 48 Pulse Speed A return to baseline within 25 ns is not necessary if Noise is uncorrelated Averaging over many samples is required in order to smooth out shower fluctuations The pulse shape is linear over the dynamic range Not only linearity at the peak, but also invariance of the shape with amplitude are required In this case, deconvolution is straight-forward
LumiReview ppt P. Denes p. 49 Pulse for 1 pp Interaction
LumiReview ppt P. Denes p. 50 Deconvolution - I.
LumiReview ppt P. Denes p. 51 Deconvolution - II. A 1% error (linearity, mis-termination,...) results in a 20% error on a 1 interaction pulse preceded by a 20 interaction pulse
LumiReview ppt P. Denes p. 52 Pulse Shape Uniformity A variation of pulse shape would mean that a 1 /a 0 is not constant Perfectly Linear
LumiReview ppt P. Denes p. 53 Example - 5% Shape Non-Uniformity a 1 /a 0 differs by 5% in the 2 curves
LumiReview ppt P. Denes p. 54 5% Shape Non-Uniformity Small effect since a 0 is constant. (Similar to saying pulse shape is non-linear, but gain at peak is calibrated)
LumiReview ppt P. Denes p. 55 Timing Error Time window for 1% variation = 2 ns
LumiReview ppt P. Denes p. 56 DAQ FEADCFPGA Delay LHC 40 MHz
LumiReview ppt P. Denes p. 57 Timing Error For a single voltage sample per bunch, some timing error is tolerable. Timing error will, however, influence the de- convolution. Worst case: a pulse preceded by a train of pulses M times bigger. Then, the error on the small pulse is Massimo Alex
LumiReview ppt P. Denes p. 58 Conclusions Gas detector: Much work has been done - 2 test beam campaigns(‘00, ‘01) New mechanical design long-term reliability Ready for engineering prototype of final design 2 technologies (gas, CdTe) - both have promising features, both still need some R&D Plan: May ‘03: 25 ns SPS test beam (gas+CdTe) hadron irradiation of both designs