January 31, 2007 1
MICE DAQ
• MICE and ISIS Introduction
• MICE Detector Front End Electronics
• Software and MICE DAQ Architecture
• MICE Triggers
• Status and Schedule
…
Terry Hart, Illinois Institute of Technology,
NFMC Collaboration Meeting, UCLA
January 31, 2007 2
Introduction – MICE Description
• Muon Ionization Cooling Experiment (MICE): international accelerator R&D project designed to provide first demonstration of muon cooling.
• MICE is being built at Rutherford Appleton Laboratory (RAL) using muon beam generated from ISIS.
• One muon cooling lattice cell with detectors measuring muon beam emittance before and after cooling.
Terry Hart, Illinois Institute of Technology,
NFMC Collaboration Meeting, UCLA
January 31, 2007 3
Introduction – ISIS Beam Structure
ISIS proton spills: - 20 ms apart (50 Hz) - About 1 ms wide - Contain ~ 3000 proton bursts - Proton bursts are ~100 ns wide and 324 ns apart
MICE RF cavity duty cycle limits the spill rate for MICE to 1 Hz.
324 ns
100 ns
~ 1 ms = 1,000,000 ns
●●●
Terry Hart, Illinois Institute of Technology,
NFMC Collaboration Meeting, UCLA
January 31, 2007 4
Systems for MICE DAQ• Target System: Titanium target inserted into ISIS proton beam produces
pions which decay into muons.
• RF Cavities: Eight 201 MHz cavities which accelerate muons along length of MICE.
• DAQ: Data from MICE trackers, calorimeter, Cherenkov detectors, and time-of-flight counters will be combined to form MICE events.
…
ISIS proton spill → MICE target → pions,muons → MICE RF Cavities and Detectors
These systems need to be synchronized for MICE data collection.
Terry Hart, Illinois Institute of Technology,
NFMC Collaboration Meeting, UCLA
January 31, 2007 5
Introduction – MICE Detectors
• MICE DAQ acquires data from – Trackers – Calorimeter– Cherenkov detector – Time-of-flight counters
• With 600 kHz muon rate within 1 ms wide spill, detector data will be buffered, read out, and built into an event at end of each spill.
Trackers Calorimeter
Cherenkov detector
Time-of-flight Counters
Terry Hart, Illinois Institute of Technology,
NFMC Collaboration Meeting, UCLA
January 31, 2007 6
Front End Electronics (FEE) for Tracker• Sixteen 512-channel Analog Front End II t (AFE-IIt) boards will read out
data from two trackers.– Tracker data are
• Bitmap of tracker channels with analog charge above preset thresholds• Digitization of amount of charge of each channel• Digitization of time of hit with respect to fixed time
– AFE-IIt boards developed for latest D0 experiment upgrade.• Tracker firmware is being modified to be suitable for 600 kHz muon rate
– Reduce digitization time– Make use of 4-level buffer so that digitization time isn’t dead time
• Expected tracker data rate ~3 Mbytes/spill.
Terry Hart, Illinois Institute of Technology,
NFMC Collaboration Meeting, UCLA
January 31, 2007 7
Tracker FEE FirmwareReducing Digitization Time
• ~ 5724 ns: D0 implementation• ~ 5670 ns: Adjust for shorter ISIS bucket period• 4536 ns: TriP-t pipeline collect data during digitization• 2376 ns: Zero-suppress 30 of 32 TriP-t channels• 1836 ns: Reduce zero-suppression time from 2 to 1 cycle.• 1566 ns: Remove cycles from set-up processes.
– Assumed muon rate is 600 kHz for average time between muons of 1667 ns.– Clock period taken as 18 ns– Assume 2 out of 32 channels of TriP-t have charge data above threshold
We are now starting to reduce the dead time so that it’s comparable to the average time between muons.
Terry Hart, Illinois Institute of Technology,
NFMC Collaboration Meeting, UCLA
January 31, 2007 8
– Clock period taken as 18 ns– Assume 2 out of 32 channels of TriP-t have charge data above threshold
Fraction of Recorded Muons for 600 kHz
No Buffering 1-level Buffer
tdigi = 5670 ns 0.227 0.291tdigi = 2376 ns 0.412 0.601tdigi = 1836 ns 0.476 0.699tdigi = 1566 ns 0.516 0.753
Terry Hart, Illinois Institute of Technology,
NFMC Collaboration Meeting, UCLA
Tracker FEE FirmwareReducing Digitization Time and Using Buffer
January 31, 2007 9
Time-of-Flight Counters FEE
• Will use CAEN V1290 TDC with 25 ps LSB (least significant bit) with constant fraction discriminators
• Considering time-over-threshold discriminator with leading edge and trailing edge measurement allowing more precise time walk correction
• 60 ps resolution on time of flight measurement has been obtained in test beam.
Terry Hart, Illinois Institute of Technology,
NFMC Collaboration Meeting, UCLA
January 31, 2007 10
Calorimeter and Cherenkov Detector FEE
• Charge will be measured with 14 bit, 100 MHz flash ADCs (CAEN V1724) coupled with RC shapers.
• Pulse shape analysis (after shaper) allows time resolution ~500 ps => no additional TDC needed.
• Cosmic ray and beam data taken and analyses underway.
Terry Hart, Illinois Institute of Technology,
NFMC Collaboration Meeting, UCLA
January 31, 2007 11
MICE DAQ Hardware and Software
• MICE DAQ software will be built from DATE framework used by CERN experiment ALICE– ALICE = A Large Ion Collider Experiment– DATE = Data Acquisition and Test Environment
• MICE detectors read out over VME.
• Online data stream will include list of selected physical variables from MICE Control and Monitoring (MCM).
• Data runs will be stopped automatically when a subsystem goes into fault status or in case of connection problem.
Terry Hart, Illinois Institute of Technology,
NFMC Collaboration Meeting, UCLA
January 31, 2007 12
MICE DAQ Architecture
DataDataFlowFlow
Trigger
distribution
Tracker Calorimeter Time of Flight Counters Trigger+ Cherenkov
GigaBit Switch
Event Builder Online Storage Online
Monitoring Run Control
Remote MassStorage
100 MegaBit Switch
VME Crates
Optical links
Linux PCs
MICE Control andMonitoring
ReadoutMCMSubnet
Terry Hart, Illinois Institute of Technology,
NFMC Collaboration Meeting, UCLA
January 31, 2007 13
Synchronization Issue
• 3 subsystems have to be synchronized with each other and with the ISIS Machine Cycle:– Target– RF– DAQ
• An “As Soon As Ready” approach has been proposed– Each system provides a veto signal which is set when the system is not ready to
receive a trigger.– The first ISIS Machine Start (MS) after all vetos are dropped generates the
different triggers.
Terry Hart, Illinois Institute of Technology,
NFMC Collaboration Meeting, UCLA
January 31, 2007 14
MICE Veto
Target Veto
RF Veto
DAQ Veto
MS
Target Request
RF Request
Target Trigger
Protons on target
RF Trigger
RF Power
DT Gate
DAQ Trigger
Target Delay
RF Delay
DT Delay
20 ms Extraction
Terry Hart, Illinois Institute of Technology,
NFMC Collaboration Meeting, UCLA
January 31, 2007 15
Triggers
• DAQ Trigger– Triggers the readout of the FEE output buffer memory– Arrives at the end of spill
– Distributed to all VME crates
• Particle Trigger– Triggers the digitization of the signals arriving at the FEE
(Trackers, Calorimeter, Cherenkov detector, Time-of-Flight counters)– Distributed to all FEE boards– MICE expects about 600 Particle Triggers for 1 DAQ trigger
Terry Hart, Illinois Institute of Technology,
NFMC Collaboration Meeting, UCLA
January 31, 2007 16
DAQ Trigger
• DATE requires dedicated inputs for at least four event types:
Start of Spill Signal sent to DAQ slightly before the beginning of the spill ensuring that all detectors are ready to receive data. This is a preparation signal for Physics events.
Physics Normal data from events occurring in a proton spill. Event is built from data from each detector.
End of Spill Signal sent to DAQ when the readout is finished checking that detectors read out data correctly.
Calibration The following are possible ways to calibrate detectors.- Empty event (pedestal)- Pulser event- Cosmic ray event
Terry Hart, Illinois Institute of Technology,
NFMC Collaboration Meeting, UCLA
January 31, 2007 17
Particle TriggerParticle Trigger Conditions for Normal Events
– Burst: ISIS burst with • Target within the Data Taking (DT)-Gate• FEE not busy
– Beam: Burst TOF0
– Upstream: Beam TOF1
– Traversing: Upstream TOF2
TOF0 TOF1 TOF2
Terry Hart, Illinois Institute of Technology,
NFMC Collaboration Meeting, UCLA
January 31, 2007 18
Particle Trigger
• Simplest Implementation of the Trigger Condition selection using output Register
OReg1
OReg2
OReg3
OReg4
Burst
TOF0
TOF1
TOF2
Particle TriggerRequest
FEEBusy
Particle Trigger
• Setting OReg selects the trigger condition
Terry Hart, Illinois Institute of Technology,
NFMC Collaboration Meeting, UCLA
January 31, 2007 19
September 2007NB: picture does not include target, beam line, hall infrastructure, electronics, DAQ, shifters etc… !
STEP I
November 2007
STEP III: Winter 2008
STEP II
STEP IVFall 2008
STEP V Spring 2009
STEP VI Fall 2009
MICE Status and Schedule
Terry Hart, Illinois Institute of Technology,
NFMC Collaboration Meeting, UCLA
January 31, 2007 20
MICE DAQ Summary• DAQ architecture determined
– DATE framework– Detector FEEs readout out by VME– Control and monitoring established
• Detector FEEs under development– Tracker FEE software being modified from D0– FEE hardware for calorimeter, time-of-flight counters, and
Cherenkov detectors under consideration
• Trigger signals and run modes established
Terry Hart, Illinois Institute of Technology,
NFMC Collaboration Meeting, UCLA
January 31, 2007 21
MICE DAQ Status and Schedule
• February, 2007: DAQ test bench set up at University of Geneva
• March, 2007: Order MICE Stage 1 FEE hardware
• July, 2007: Move MICE DAQ system to RAL
• September, 2007: DAQ working in MICE Stage 1
Terry Hart, Illinois Institute of Technology,
NFMC Collaboration Meeting, UCLA