* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download 20090826_FlatPanel_status
Survey
Document related concepts
Public address system wikipedia , lookup
Time-to-digital converter wikipedia , lookup
Flip-flop (electronics) wikipedia , lookup
Signal-flow graph wikipedia , lookup
Negative feedback wikipedia , lookup
Switched-mode power supply wikipedia , lookup
Resistive opto-isolator wikipedia , lookup
Pulse-width modulation wikipedia , lookup
Schmitt trigger wikipedia , lookup
Oscilloscope types wikipedia , lookup
Oscilloscope wikipedia , lookup
Dynamic range compression wikipedia , lookup
Wien bridge oscillator wikipedia , lookup
Regenerative circuit wikipedia , lookup
Analog-to-digital converter wikipedia , lookup
Transcript
Flat-Panel PMT Status of Readout at Edinburgh Development of amplifier for FP-PMT signals First FP-PMT signal traces (with a 2-channel readout…) Conclusions RICH Upgrade meeting, 26.08.2009 Stephan Eisenhardt University of Edinburgh Task: Build a Reference Readout . Adapt signal output of FP-PMT to charge sensitive ADCs need 64x amp: gain ~100 PMT signal: 0.06 – 0.24pC (0.375 – 1.5 Me-) ADC Range: 0 – 400pC 256 channels Resolution: 0.10pC RICH Upgrade meeting, 26.08.2009 Stephan Eisenhardt 2x32 channels 2 How to … Commercial fast amps too expensive for 64 channels: – – – – channel gain: 10 e.g. 16 channels for ~ £ 3000.OK for 8 channels, as done for MaPMT too expensive for 64 channels ( = £ 24000.-) Alternative : build your own… – – – – large gain low noise linear in charge amplification fast (base of pulse shape <15ns) RICH Upgrade meeting, 26.08.2009 Stephan Eisenhardt 3 Our Solution Base: op-amp LMH 6702 – – – – – – – non-differential, wideband op-amp ultra low distortion bandwidth at gain 2: 1.7GHz (output signal <0.5Vpp) bandwidth at gain 10: 370MHz operated in inverting mode fast slew rate: 3100V/ms low noise: 1.83nV/ Hz Our layout: – two-stage amplifier – 50W input and output termination – each stage with gain 10 operation outside recommended range… – strong noise filtering – signal GND and power GND separated RICH Upgrade meeting, 26.08.2009 Stephan Eisenhardt 4 Biasing Restriction by the input circuit of the V792: may not exceed +15mV – risk of damage to the input circuit – i.e. positive overshoots may kill the expensive devices… Solution: – – – – adjustable bias circuit in the output capacitive decoupling (100nF) bias adjustable (0…-5V) use low noise trimmer (500W, 25 turn) Operation: – mandatory study of output signals before connecting to V792 – set biases for all channels individually to a safe value… RICH Upgrade meeting, 26.08.2009 Stephan Eisenhardt 5 The Prototype Built and studies by Iain Longstaff (summer student): – using two evaluation PCBs – signal source: MaPMT@1000V gain ~1.2Me(centre channel) – use dark counts = single photoelectrons – signal split to: - amp input - oscilloscope – readout: oscilloscope 50W input RICH Upgrade meeting, 26.08.2009 Stephan Eisenhardt 6 MaPMT Signal Trace MaPMT pulses 10mV/div amp pulses time scale 500mV/div 4ns/div >2000 pulses RICH Upgrade meeting, 26.08.2009 Stephan Eisenhardt 7 Prototype Results Voltage gain: LMH6702 ConfigB 2 Amps + BiasA: Voltage In vs Voltage Out – saturates: at 4V - bias – due to output current limit of 80mA & 50W termination 3 Output Voltage (V) 3.5 Charge gain: – effective gain: expected: 142, found: 132±3 – linear over the whole spectrum, if integrated over 16ns 2.5 2 1.5 1 0.5 0 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 Input Voltage (V) LMH6702 ConfigB 2 Amps + BiasA: Charge In vs Charge Out Noise: – measured standard deviation of pedestal (HV of MaPMT off) – std dev @ input: 0.58mV – std dev @ output: 19mV 0 -6E-12 -4E-12 -3E-12 -2E-12 -1E-12 0 -2E-10 – ratio: 33 << gain – noise at input limited by measurement, not by amplifier RICH Upgrade meeting, 26.08.2009 -5E-12 -1E-10 Output Charge (C) Stephan Eisenhardt -3E-10 400pC -4E-10 -5E-10 1pC -6E-10 -7E-10 Input Charge (C) 8 Amplifier Status Single Channel amplifier PCB: – PCB designed – 130 PCBs produced – first PCB populated and under test Motherboard: – – – – – – – amp circuit currently in PCB design 64 amplifier PCBs plug vertically in alternating from front and back side same input track length (capacity!) impedance match where possible 64 ch power bus bars input 64 biasing circuits GND planes 4 output connectors a 16 ch connector – complication: high density input connector • 2x40 pins, 0.75 mm pitch • connecting to both sides of the PCB • connecting to GND on the edge RICH Upgrade meeting, 26.08.2009 Stephan Eisenhardt room for bias circuits 9 Two-Channel Readout How to read out single FP-PMT channels? (without a proper setup…) – need proper grounding of other channels… Recipee: – use 4 PCB mount connectors – roll copper foil – squeeze between solder pins (shorts 16 GND to 64 signal pins) – leave gap to feed out 2 channels… RICH Upgrade meeting, 26.08.2009 Stephan Eisenhardt 10 Two-Channel Readout – feed lines directly into LEMO cables… – and ground well (aehm, sort off…) – prop up FP-PMT in dark box – and look for dark counts – caveat: no cross-talk study possible yet (not neighbouring channels) RICH Upgrade meeting, 26.08.2009 Stephan Eisenhardt 11 FP-PMT Signal Trace FP-PMT pulse 10mV/div trise~0.8ns amp pulse time scale 500mV/div 4ns/div 1 pulse RICH Upgrade meeting, 26.08.2009 Stephan Eisenhardt 12 MaPMT Signal Trace MaPMT pulse known reflection from Y-split 10mV/div amp pulse time scale 500mV/div 4ns/div trise~1.2ns 1 pulse RICH Upgrade meeting, 26.08.2009 Stephan Eisenhardt 13 FP-PMT Signal Trace FP-PMT pulses 10mV/div amp pulses time scale 500mV/div 4ns/div >1000 pulses RICH Upgrade meeting, 26.08.2009 Stephan Eisenhardt 14 FP-PMT Signal Trace FP-PMT pulses 10mV/div amp pulses time scale 500mV/div 10ns/div >1000 pulses RICH Upgrade meeting, 26.08.2009 Stephan Eisenhardt 15 FP-PMT Signal Trace FP-PMT pulses 10mV/div amp pulses time scale 500mV/div 40ns/div >1000 pulses RICH Upgrade meeting, 26.08.2009 Stephan Eisenhardt 16 Conclusions Amplifier development took longer than expected (not aided by me being ill for >1 month…) Prototype amplifier works very well – linear in charge amplification – large gain – low noise PCB status – amplifier PCB just back from production – motherboard PCB on the drawing board FP-PMT signal trace – fast signal indeed: trise ~0.8ns as in data sheet – ringing on the trace: not yet understood RICH Upgrade meeting, 16.04.2009 Stephan Eisenhardt 17 Spare Slides RICH Upgrade meeting, 16.04.2009 Stephan Eisenhardt 18