Beam Secondary Shower Acquisition System: Cable Conclusions & Possibilities Student Meeting Jose Luis Sirvent PhD. Student 27/05/2013

1.Cable Impact on pCVD signal 3 Studies carried out for the moment (With 2 reports available): – pCVD Signal estimations from Wire Scanners Dynamic range first approximations for SPS – SPS max Charge per bunch (All BWS)_Totem: 120pC – SPS max Charge per bucnh (Prototype Location 51740)_Nominal: 60pC Dynamic range first approximations for LHC – LHC max Charge per bunch E=2e-6m 7TeV (Beam_Sigma=181um)_Nominal: 226pC – LHC max Charge per bunch E=2e-6m 450GeV (Beam_Sigma=716um)_Nominal: 32pC Proposal of Dynamic range: 2fC (~MIP) to 400pC (~2*Qmax) – Impact of Long Cables (250m of CK50) in the pCVD Signal (Report) 250m Cable Bandwidth: ~ 10MHz (To avoid 25ns pulse overlapping needed > 40MHz) Amplitude loss of 5dB Observed offset due pulse overlapping 5% Settling time ~ 66ns (first 3 bunches) – BWS Scan Simulation Though Long CK50 Cables (Report) pCVD Signal generation from BWS and Beam characteristics (Script) Matlab processing and simulation of cable as an approximate filter (Script) Sigma measurement not affected by cable Very low temporal resolution at bunch level (too long decays) Not constant offset. This is good, bad…?

1.Cable Impact on pCVD signal BWS Scan Simulation Though Long CK50 Cables (Report) – A) Why with Matlab? Problems with Pspice when custom input signals are “too heavy” Simulations could not be completed with Pspice, only 10% segments … Matlab has tools enough to approximate the response of the cable.

1.Cable Impact on pCVD signal BWS Scan Simulation Though Long CK50 Cables (Report) – B) Let’s see the response in “Time domain” Approximations with Butterworth & Chevychev 1 st order filters is quite similar Approximation with FIR2 & FIR1 filters not at all satisfactory See Matlab (left) VS Pspice (Right)

1.Cable Impact on pCVD signal BWS Scan Simulation Though Long CK50 Cables (Report) – C) Now we can generate the pCVD spected signal and introduce it in cable model. Modulated train of bunches with a Gaussian envelope All calculations performed from BWS and Beam characteristics for pCVD at 2m

1.Cable Impact on pCVD signal All this leads us to some interesting conclusions: – 1. Cable is a limiting factor for BbB (Bunch by Bunch) acquisition – 2. Long cables can lead to a poor SNR in Gaussian tails (possible induced noise) Could be possible to avoid with a “smart” amplification scheme. Long CK50 cable noise to be studied in practice. – 3. It’s required a BW >> for 250m link!! (Undesired signal shaping) A) Better Cables? B) Study different techniques?

2. Looking for inspiration Experiments using Diamonds or useful technology for us: – ATLAS Beam Conditions Monitors – CMS Beam Conditions Monitors – CMS BHM (Background at High Radius Detector) – LHC Beam Loss Monitor – CMS Upgrade of the Hadron Calorimeter – Dessy & Fermilab Many synergies, shared technology, collaboration & knowledge – Very nice developments already available (ASICS) that could fit well all our specs – This is full of very intelligent people!

2. Looking for inspiration ATLAS BCM (pCVD module): – Double Decked detector: SNR + 50% – 45º Tilted: Signal increasement factor sqrt(2) – BW limit ~ 300MHz: SNR + 20% – 40 dB Amplification in 2 stages (Algient MGA BW500MHz & MiniCircuits Gali52 BW2GHz) – Microstript technology 50ohm. – Package very well shielded and isolated – Diamonds very well characterised and tested. – Connections: HV, LV & Signal – Signal cable Length to ADC: 15m – SNR 11:1 for MIP HV Signal *V.Cindro et al. “The ATLAS Beam Conditions Monitor”. 2008, Journal of Instrumentation Vol3, Feb2008. Double Decked Detector (Simplified Scheme)

2. Looking for inspiration CMS BCM1: – sCVD Diamond sensors used – Signal transmission by Analog-Optohybrid driver: Linear optical driver (optical analog signal) – Digitalization on Counting room – The designs planed to be updated on LS1 (BCM1 & BHM) “CMS Online Beam Background Monitor Upgrade Plans (Marina Giunta) 13 th May 2013” Planning System as CMS-HF Architecture (Qie10-based) Quicker electronics, Peack time <10ns, 25ns BbB… *A.Bell et al. “Fast Beam Conditions Monitor BCM1F for the CMS Experiment”. Nuclear Instruments & Methods in Physics Research. Dec 18,2009

2. Looking for inspiration CMS Upgrade of the Hadron Calorimeter: 3 Sensors: HB (Barrel),HE(Endcap) & HF (Forward) Common scheme, Electronics and architecture (Quicker development) QIE10-Based Acquisition. 3 Main hardware systems: – 1. Detector module – 2. Front-End Electronics (Tunnel) – 3. Back-End Electronics (Surface) Long Link transmission: Optic SMF 1310 FE-Cards – QIE10: Charge Integration and Encoder V10 – FE-FPGA: RH Data Alignment & Formatting – GBT: RH Serializer & Optic link 4.8Gb/s – VTTX: Versatile Link technology (RH PD/PD…) Jake Anderson et al. “Progress on the Upgrade of the CMS Hadron Calorimeter Front-End Electronics”. TIPP2011- Technology and Instrumentation in Particle Physics J. Anderson et al. “CMS Technical Design Report for the Phase 1 Upgrade of the Hadron Calorimeter”. CMS-TDR September 2012.

2. Looking for inspiration QIE10 Characteristics and functionality(*): Rad-Hard Charge-Integrating ASIC (25ns) Fast, Wide dynamic range, Dead-Timeless ADC (Latency: only 4x25ns) Very High dynamic range: 3.2fC  340pC (Fits well our initial estimations!!) – LSB 3.2fC (Almost MIP for pCVD) – Non linear charge digitalization scheme: 6 bit FACD mantissa + 2 Exp (4 Ranges) – TDC capability: Produces TDC info based on the Rising/falling edge of pulse (2 configurable 8bits thresshold levels) Inputs: – Reset (CLK Alignment) – Charge signal (from pCVD) – CLK – Programable stuff ( Thressholds, Pedestrials…) Outputs: – Q : Charge Integral – T1: Arrival time (500ps resolution) – T2: Falling time (500ps resolution) QIE10p4 Already available! QIE10p5 soon (maybe also available) * “CMS Specifications Document for the QIE10 ASIC. 2010”

2. Looking for inspiration What is nice form CMS HE/HB FEE ? – QIE BOARD – 12 Channels per board – Integrated RH ProASIC3 FPGA, GBT & VL (4.8Gb/s GBT link) – Provides already digitalized and formatted information in an Optical link – It’s possible place near the BWS (RadHard components) – Needed a ngCCM Control Board (Clock & Control Module): » Deliver clock (Good quality) » Orbit signal for data marking and synchro » QIE10 RESET » Calibration signal » I2C com for config of GTBX,QIE10 and FE-FPGA (Remote programming) » Monitoring of Voltages & Temperatures

3. Just an initial architecture proposal CMS HE/HB FEE-Inspired (Very simplified & Very preliminar) pCVD Diamond Detector Bias Circuit Line amplifiers (Matching Impedance) Detector Module (ATLAS BCM) HV LV Signal (<15m of CK50) Front-End Electronics (CMS HE/HB) QIE10 RH FPGA (ProASIC3) GBTVTTX ngCCM HVLV SMF or MMF 4.8Gb/s optical link (RX: CLK + Control) (TX: D_Signal + Diag.) 250m Back-End Electronics (PFGA Development Board) GBTVTTX FPGA (Control) DDR2 Memory Ethernet Link Diagnostics System BST (Beam Synchronous Timing) Counting-Room Data Bases Ethernet Link Expert Application Data processing Systems integration Tunnel Surface

4. Final statements 1. Signal cable as limiting factor (Experiments in general use optical links for high speed signal transmission) 2. QIE10 is a great candidate for the digitalization of the pCVD signal (Reaches specs in terms of Timing & Dynamic range) 3. By Using the Know-how of our section & potential collaboration with CMS our system could be built (Relatively quickly). 4. By following LEGO ® mentality we can use a big amount of CERN Developed ASICS. 5. What now? Present research plan to “Universidad de Barcelona” for PhD. Approval