Status FEE-DAQ

1
Status FEE-DAQ

• Walter F.J. Müller, GSI, Darmstadt
• for the CBM Collaboration
• 13th CBM Collaboration MeetingFriday, 13 March
  2009

2
FEE-DAQ forDetector RD... n-XYTER ...
3
n-XYTER profile perspectives

• Profile
• the n-XYTER is currently used for detector RD
• the chip is available, and has for many
  applications (STS, GEM, RICH) at least acceptable
  characteristics
• key asset self-triggered architecture
• Perspectives
• new readout ASICs are planned for all CBM
  subsystems
• some go by the name CBM-XYTER
• CBM-XYTER ! n-XYTER !!!
• The demonstrators planned for 2009 and early 2010
  will use the n-XYTER for proof-of-principle
  studies getting used to self-triggered systems.
• after mid-2010 we'll see a transition to
  demonstrators using the prototypes of the CBM
  readout ASICs.

4
n-XYTER FEB Rev. C
n-XYTER

• In spite of bonding difficulties, managed to
  realize
• 10 Rev. C boards, fully functional
• 4 rev. B boards with 64 channels functional
  (the other 64 may be employed for direct
  silicon det. tests)

We are striving to get the entire system (FEB,
SysCore, software) packaged to allow shipment.
Volker Kleipa is on board at GSI since January.
Slide C.Schmidt
5
Complex Multilayer Data Chain
n-XYTER

• Detector
• n-XYTER-FEB
• n-XYTER
• FEB and Bonding Technology
• Documentation (Manual)
• ADC
• Interconnect
• SysCore
• Firmware
• Embedded software
• Soft configuration
• Ethernet-Interconnect
• PC and DAQ
• KNUT
• GUI
• DAQ System

Realize software design freeze to be
packaged Missing still diagnostic toolbox for
system analysis to make successful deployment in
other labs feasible
Slide C.Schmidt
6
n-XYTER FEBs next steps
n-XYTER

• FEB-Rev D next version to implement all
  learnings
• Relieve PCB specs by tightening bonding
  necessities and challenge
• Targeted 125µm pitch, ILFA insists on 140 µm
  (gap copper shoulders 70 µm
  landing width for bonding)
• Chip-In-Board, two layer bonding and metal in-lay
  cooling contact
• GEM-TPC FEB ongoing dual n-XYTER for PANDA
  GEM-TPC
• n-XYTER Quattro STS-Baby-Sensor double sided
  readout ? towards a beam telescope

Slide C.Schmidt
7
n-XYTER Quatro Double Sided
Silicon Readout
n-XYTER

• Read-Out of double sided Si, 256 x 256 channels
• First setup to occupy complete capacities of one
  SysCore
• Several Modules may be set-up to make a beam
  telescope
• FEE of both sides live on one potential!
  ? Rely upon readout-caps!

Slide C.Schmidt
8
n-XYTER Energy Output Channel (preliminary)
n-XYTER
Pulse height spectra with internal test pulse
scanned, indetermined Cdet so far (5-10 pF)
tests done as is on the table
peak to peak distance 5560 e-, FWHM 1000e, s
425e
2780e (CAL 10)
Pedestal
Slide C.Schmidt
9
Timestamp calibration example
n-XYTER
- random input signals (noise)- delay registers
are set to achive the same probability for all
counter values- example four patterns used for
bit 0 calibration
Slide R.Szczygiel
10
n-XYTER Engineering Run

• How many n-XYTER's are available
• about 200, of those 100 are owned by the CBM
  collaboration
• Planed systems with n-XYTER's
• starter kits ? a lot of small scale detetor RD
• CBM Beam trackers
• CBM Demonstrator 1
• several other projects outside CBM (PANDA GEM
  TPC,.....)
• Not all can be realized with the available
  chips? plan to produce more ? plan for
  engineering run
• cost 110 kEUR, shared by CBM, PANDA, and others
• will give many thousand chips
• opportunity for fixes, e.g. baseline temperature
  coefficient
• Timelines Review end April Submit end
  June

11
FEE for the Real ThingUMC 018 Process
12
RadHard

• UMC 0.18µm radiation hardness studies
• Update -
• Sven Löchner
• 13th CBM Collaboration Meeting
• GSI Darmstadt
• March 12th, 2009

Slide S.Löchner
13
GRISU test ASIC
RadHard

• GRISU chip
• UMC 0.18µm process
• 1.5 x 1.5 mm²
• 64 pads
• 28 core pads
• 36 pads

Slide S.Löchner
14
Low Energy testing site (SEU Tests)
RadHard

• Installation of a test facility for ASIC
  irradiation with heavy ions at X6 cave at GSI (in
  cooperation with bio physics group)
• Beam monitoring via ionisation chamber
• Dosimetry setup available
• Irradiation of DUT in air
• Easy access

• Disadvantages of setup
• Only one ion source during beam time
• Fixed LET range for ion source

Slide S.Löchner
15
LET testing range
RadHard

• Overview of the LET testing range for the applied
  heavy ions test

Slide S.Löchner
16
Cross-section (Weibull-Fit)
RadHard
Test with Ru beam running
C-12 Ar-40 Ni-58 Ru-96 Xe-132

• ? LETcrit 1.93 MeV cm²/mg
• ? ssat 1.4810-8 cm²/bit

Slide S.Löchner
17
Total Ionizing Dose (TID) tests
RadHard

• TID testing with X-rays
• Irradiation facility at Institute for
  Experimental Nuclear Physics, University of
  Karlsruhe
• 60keV X-ray
• 100 ... 600krad/h
• 9 GRISU chips tested
• Total dose between 800krad and 2500krad(SiO2)
• Operating dose rate between 200krad/h and
  580krad/h
• Two radiation test modes
• single transistor test structure
  measurementsleakage current, threshold shift,
  characteristics
• complete chip measurementstransition times,
  total power consumption

Slide S.Löchner
18
TID tests single transistors
RadHard

• Measurements of the transistor characteristics
  and calculation of
• the threshold voltages for different dose levels
• In total 6 chips are irradiated
• Total dose up to 2.5Mrad
• Dose rate between 200krad/h and 580krad/h

Slide S.Löchner
19
TID summary
RadHard

• UMC process shows good annealing at room
  temperature(at least at high dose rates)
• Simulation models slightly differs from measured
  characteristics (especially between small and
  large Ugs)
• Still to be done
• Second irradiation campaign with low dose rates
• Long term test with a gamma source

Slide S.Löchner
20
FEE for the Real ThingTOF - RPC
21
GSI Event-driven TDC with 4 ChannelsGET4
RPC-GET4

• Harald Deppe, EE-ASIC
• Holger Flemming, EE-ASIC

Slide H.Flemming
22
Requirements for the CBM ToF
RPC-GET4

• Very high time resolution lt 25ps
• Double hit resolution lt 5ns
• Event rate up to 50 kHz per channel
• Capability to measure time over threshold
• Low power consumption with less than 30 mW per
  channel
• Number of Channels 65.000
• Triggerless operation
• Each event combined with a timestamp
• Epoche event on timestamp counter overflow
• Timestamp counters of all chips have to run
  synchronously

Slide H.Flemming
23
GET4 Prototype PCB
RPC-GET4
Bottom side view
Top side view
Slide H.Flemming
24
Summary and Outlook
RPC-GET4

• First full scale TDC Prototype GET4 was submitted
  In Oct. 2008
• Token ring readout and serialiser are fully
  operational
• DNL of TDC core worse than on DANTE test chip
• Time resolution 20 ps 25 ps
• Double hit resolution gt 3.2 ns
• With GET4 a first TDC Prototype for ToF detector
  test is available
• Next Steps
• Investigation on reasons for nonlinearity
• Upgrading of Linearity gt Improve the
  Resolution
• Slow control, etc.

Slide H.Flemming
25
FEE for the Real ThingCBM-XYTER
26
CBM-XYTER
Slide T.Armbruster
27
CBM-XYTER
PreAmp Tests
Slide T.Armbruster
28
CBM-XYTER
PreAmp Tests
Slide T.Armbruster
29
CBM-XYTER
PreAmp Tests
Slide T.Armbruster
30
CBM-XYTER
ADC Tests
Slide T.Armbruster
31
CBM-XYTER
Slide E.Atkin
32
CBM-XYTER
Slide K.Kasinski
33
CBM-XYTER
Time-over-Threshold Proposal
Slide K.Kasinski
34
CBM-XYTER
Time-over-Threshold Proposal
Slide K.Kasinski
35
XYTERFamily Planning,FEE for Tracking
Detector Applications
CBM-XYTER
Christian J. Schmidt, GSI Darmstadt
Slide C.J.Schmidt
13. CBM Collaboration Meeting, GSI, Darmstadt,
March 10 13, 2009
36
Collection of Tracking Detector Specifications
CBM-XYTER
Family Planning
Slide C.J.Schmidt
37
Comparison Tentative Silicon/Gas
XYTER Specs
CBM-XYTER
Family Planning
Slide C.J.Schmidt
38
Silicon-XYTER proposal
CBM-XYTER
Family Planning

• Lowest possible power
• Average signal rate per channel 150kHz
• 128 ch
• Noise 800 enc at 25pF
• Average signal MIP in Si(300) 25ke
• Assume signal degradation to 13ke over det. life
  time
• ADC res 5 bit for MIPs (25ke) gives 4 bit for
  MIP in aged detector 780e
• Dynamic range 4 to 7 MIPs (100 to 169 ke) but
  not necessarily linear over its range
• specials both polarities, switchable bandwidth (
  gm ) optimizes power to needs allow DC
  couppling
• Time resolution 2 ns

TOT solution promises very low power and
architectural simplicity
Slide C.J.Schmidt
39
Gas Detector XYTER Proposal
CBM-XYTER
Family Planning

• Dual shapers for timing and energy resolution,
  energy determination through ADC (similar to
  n-XYTER)
• Distinguish discrimination needs and data
  resolution!
• More advanced signal conditioning desired.
  (baseline restoration etc.)
• Alternate or integrated front-end filtering for
  wire read-out
• Large necessary dynamic range may be addressed
  through a dual realm amplification 0 to 1000ke
  and 1000 to 3000 ke
• Linear interval up to 1000ke needed for TRD
  applications
• PANDA TPC application imposes special needs with
  low mean signal height (25 ke with a dynamic
  range of 30 times the average (750 ke)). This
  should however be congruent with TRD needs.

Slide C.J.Schmidt
40
CBM-XYTER Summary

• Process
• UMC018 very promising.
• Building Blocks
• CSA, ADC, ... in work and converging
• "Silicon-XYTER"
• ToT proposal on the table? needs feedback from
  simulation (resolution, dead time)? clear path
  for system prototype
• "Gas-XYTER"
• requirements have more spread
• what is 'good to have', what has 'physics
  performance impact' ?
• not yet decided whether all can be accommodated
  in one ASIC

41
A Note on FEE

• So far the FEE sessions were essentially
  ASIC Designers meetings
• FEE is more than just read-out ASIC's !
• FEE is the whole chain from detector signal to
  DAQ interface, including the first data
  aggregation stagesand all the cabling up to the
  detector perimeter
• We urgently need a 'strawman' FEE design for all
  CBM sub-systems.

42
DAQ forDemonstrators
43
ROC
Slide U.Kebschull
44
ROC
Slide U.Kebschull
45
ROC
GET4 Readout
Slide S.Manz
46
Demonstrator I
DCB-CBMnet
ROC Readout Controller DCB Data Combiner
Board ABB Active Buffer Board
Thanks to Walter F.J. Mueller for this picture.
Slide F.Lemke
47
Current Status
DCB-CBMnet

• Basic Communication tested between
• DCB DCB
• DCB ROC
• ABB DCB
• Deterministic Latency for MGTs and GTPs proved
  with new designs
• All existing coded parts are directly implemented
  in a clean modular way for reusability in later
  design phases
• First version of Protocol implemented in Verilog
• Test setup for first protocol version is in
  preparation
• gt Next step is the first hierarchical
  communication test with CBM protocol

Slide F.Lemke
48
ABB
Slide Wenxue Gao
49
ABB
Slide Wenxue Gao
50
FPGA - Framework
Slide U.Kebschull
51
FPGA - Framework
Slide U.Kebschull
52
FPGA - RadTollerance
Slide J.Gebelein
53
FPGA - RadTollerance
Test with Ru beam running
Slide J.Gebelein
54
DABC
Data Acquisition Backbone Core library Jörn
Adamczewski-Musch, Hans G.Essel, Nikolaus Kurz,
Sergey LinevGSI, Experiment Electronics Data
Processing group Release V1.0 Use cases
Work supported by EU RP6 project JRA1 FutureDAQ
RII3-CT-2004-506078
Slide H.Essel
55
DABC release v1.0.01
DABC
Download via dabc.gsi.de
DABC
Controls
Java GUI
Applications
Plugins
Core
bnet-mbs mbs ROC ABB bnet-test core-test net-test
Slim
DIM
bnet-mbs mbs ROC ABB ROOT verbs
Controls No more XDAQ Plugins Implementation of
applications (programmers) Applications Mainly
setup or testing programs (users) external
packages needed
Slide H.Essel
56
DABC design global overview
DABC
DABC
frontend
Linux PCs
datainput
sorting tagging filter analysis
PCIe
frontend
sender
GE
frontend
frontend
IB
Linux PCs
UDP
Linux PCs
TCP
analysis archive
sender
sorting tagging filter analysis
datainput
archive
GE Gigabit Ethernet IB InfiniBand
Slide H.Essel
57
Access layer current view
DABC
online monitor
file output
nXYTER gui
ROClib (KNUT)
Ethernet
Slide S.Linev
58
DABC as common access layer
DABC
TDC gui
nXYTER gui
file output
online monitor
ROC access layer
control API remains the same UDP socket can
be served by DABC directly buffer, thread and
configuration management provided same software
stack for tests and beamtime
PCIe/mprace
Ethernet
optic
Slide S.Linev
59
CBM slow control strategy
DCS/ECS

• 11.03.2009
• CBM collaboration meeting
• DAQ work group

Slide B.Kolb
60
What it should provide
DCS/ECS

• Safe operation of the experiment
• Set up, control and monitoring of detectors,
  DAQ/trigger and environment
• Alarm handling
• Archiving

Slide B.Kolb
61
Strategy (2)
DCS/ECS

• Use
• existing
• well tested
• supported solutions
• Use
• Open Source wherever possible
• Join
• existing development collaborations

Slide B.Kolb
62
Strategy (3)
DCS/ECS

• Choose EPICS!
• Why?
• It works!
• We have in house experience
• It provides solutions for all common hardware
  (HV, LV, gas, )
• Its for free
• Its scalable
• Huge collaboration

Slide B.Kolb
63
Requirements DB
DCS/ECS

• In collaboration with PANDA we have set up a
  database with web access
• http//nuclear.gla.ac.uk/CBM-DCS/
• Register
• Then enter requirements
• Process variables
• Common hardware

Slide B.Kolb
64
The End
Thanks for your attention
65
Backup's
66
n-XYTER
TEMPERATURE SENSITIVITY
Config. for positive input signals NEW
TEMP.COEFF. 0.5 mV/C
Old TEMP.COEFF. 29 mV/C
FAST-SHAPER OUTPUT DC-LEVEL
Config for negative input signals NEW
TEMP.COEFF. 0.7 mV/C
TEMP(C)
Hans K.Soltveit, Physik. Inst. Heidelberg
Slide C.Schmidt
67
Schematic Overview of GET4 Prototype
RPC-GET4
Slide H.Flemming
68
Silicon-XYTER Distillate
CBM-XYTER
Family Planning
Slide C.J.Schmidf
69
Gas-XYTER Distillate
CBM-XYTER
Family Planning
Slide C.J.Schmidf