Silicon Strip Detector

1
Silicon Strip Detector

• Extended status report
• Lilian Martin
• March 27th, 2006

2
Detailed Items

• Hardware
• Hardware on the cone
• Hardware on the TPC wheels
• Hardware on the pole tips
• Hardware in the WAH and AB
• Hardware on the south platform
• Hardware in the DAQ room
• Hardware in the control room
• Documentation on the hardware
• Documentation on operating the SSD
• Documentation (general)

• Software and data analysis
• Online software
• StSsdDaqMaker raw data converter
• StSsdPointMaker hit finder
• StSvtEvalMaker SSD global alignment
• Pedestal and noise calibration
• Gain calibration
• Simulation chain
• List of contacts

3
Hardware on the cone

• Items
• 20 SSD ladders are installed on the cone. They
  are attached to 4 separated sectors (2 big and 2
  small sectors) . The SSD has been progressively
  installed (1, 10 and finally 20 ladders).
• An electrical shield (aluminized Mylar) separates
  the SSD from the IFC.
• Expertise
• The SSD was designed and built by both the IReS
  and SUBATECH groups.
• The SSD structure, electronics and mechanics are
  described on the web.
• G.Guilloux, S.Bouvier and L.Martin are experts in
  mechanics
• S.Bouvier and L-M.Rigalleau are experts in the
  control and ADC electronic boards
• L.Martin has some knowledge on the task related
  to the shield installation.
• Maintenance
• Failing ladders have been replaced several times.
  HV decoupling capacitances have been replaced on
  all ladders. For this task, the SSD has been
  completely removed from the cone. Maintenance on
  the SSD ladders has been routinely done.
  Dedicated tooling exist to remove and install the
  SSD sector on the cone.
• Failing ladders have been usually replaced and
  shipped back to France for repair. Two spare
  ladders are always available to replace ladders
  featuring new failure (usually on the electronic
  boards).

4
Hardware on the TPC wheels

• Items
• 4 readout boards read and communicate with 10
  half ladders each. Two are installed on each side
  of the TPC and on the top of 2 SVT readout boxes
  (at 6 and 7 oclock). The boards are installed in
  a box mechanically attached to the SVT rdo boxes.
  The backside of the box features a quick
  connector to plug an air hose for the cooling of
  the board.
• The rdo board is a key element since a software
  is uploaded in its FPGA and in the FPGAs on the
  adc and control boards on the ladders.
• Expertise
• The boards were designed by the electronics team
  at Subatech. Christophe Renard designed the
  board.
• Documentations Christophe Renard has posted all
  the information related to his board on the web
  at this URL
• Experts Christophe Renard (Subatech) and
  Stephane Bouvier (Subatech)
• Maintenances
• The board has never needed any hardware
  maintenance during the last few years.
• The weak connection between the optical
  transceiver and its socket on the board has been
  mechanically secured.
• Several generations of rdo boards exist with
  minor modifications which can easily used as
  spares.
• The main maintenance operations were software
  upgrades.

5
Hardware on the TPC wheels

• Items
• 20 flexible air hoses (tygon) emerged from the
  cone ends (10 on each side). They go through the
  air partition apertures and are connected to the
  air manifold.
• 4 air manifolds connect the air hoses to the 4
  vortec and act as a fan in-fan out for the
  induced air flow. They fit inside partly opened
  tubes attached to the SVT water hose shoe located
  at 4 and 8 oclock.
• Expertise
• These elements were designed or selected by
  Subatech.
• A document posted on the web gives some
  indication on how to route the hoses.
• Experts S.Bouvier, G.Guilloux and L.Martin
  (Subatech), B.Soja (BNL)
• Maintenances
• The air manifolds do not need any particular
  maintenance.
• The air hoses are quite soft and were collapsing
  when subject to small bending radius. Their path
  has to be carefully checked to make sure enough
  air flow is going through them especially at the
  location were they emerged from the SVT air
  partition. One has to carefully inspect the
  connection of the hose dedicated to the rdo
  boards since it is located into a very confined
  space
• Some of these hoses have been recently
  encapsulated into more rigid hoses (blue) with an
  inner radius slightly bigger than the air hose
  radius. Some connectors were also added in the
  middle of some hoses to allow quite
  disconnections and prevent the hose to collapse
  at these particular locations.

6
Hardware on the TPC wheels

• Items
• 20 LVHV power cables on each side are coming
  from the SSD rack on the south platform and are
  connected to the cables emerging from the cone.
  10 cables on each side carry the HV while the
  others are only providing LV (-2V, 2V and 5V).
  When reaching the STAR magnet, the cables are
  separated and following different paths depending
  on the ladder they are serving. The connectors at
  the end of the cables are AMP connectors and
  feature both male and female pins. Once connected
  to their corresponding cables emerging from the
  cone, the connections are secured using cable
  tights.
• Expertise
• These elements were designed or/and selected by
  Subatech and IReS.
• Documentations posted on the web give some
  indication on how the cables are routed around
  the TPC wheels and their specifications.
• Experts S.Bouvier, L-M.Rigallaud (Subatech)
  and D.Bonnet, F.Littel (IReS)
• Maintenances
• Usually during the shutdowns, the cables are
  disconnected from the cone cables and they are
  removed from the TPC wheels up to the magnet
  radius. During that time the connector may be
  exposed. Few times several connectors had been
  broken, usually the body of the connector made of
  plastic. They are relatively cheap and can be
  easily replaced.
• About 15 spare cables of various length are
  stocked in the STAR stock building. The shortest
  is long enough to reach the farest ladder.

7
Hardware on the TPC wheels

• Items
• 4 slow control cables, 2 trigger cables, 6
  optical fibers and 2 power cables are equally
  installed on both wheels. The trigger cables are
  coming from the trigger rack on the south
  platform. They are equipped at their end on the
  wheel with a flat cable extension featuring two
  connectors and properly terminated. All the other
  cables are coming from the SSD rack. 4 fibers
  (out of 6) are used. The others are available as
  spares and are only routed up to the south
  platform (where a patch panel provides lines
  already reaching the daq room). The two power
  cables are equipped at their end with a short
  Y-shape extension allowing to power the two rdo
  boards installed on each side.
• Expertise
• These elements were designed or selected by
  Subatech and IReS.
• Experts S.Bouvier, L-M.Rigallaud (Subatech)
  and D.Bonnet, F.Littel (IReS)
• Maintenances
• These cables had not needed any maintenance
  operations so far. Two additional optical fibers
  have been installed and are routed up to the
  south platform and stored under the first floor
  near the trigger rack.
• Spare fibers and round to square shape adapter
  can be found in the SSD cabinet.

8
Hardware on the pole tips

• Items
• 4 transvectors (vortec) are installed on the pole
  tips to produce an induced air flow into an hose
  connected to 5 individual ladders. The air flow
  resulted on the Venturi effect produced by 120
  psi compressed air liberated in the vortec
  volume. The flow amplitude is determined by the
  thickness of a shim in the vortec. Before
  reaching the vortec, the compressed air goes
  trough a filter and a manual valve allowing to
  reach a stable pressure of 5 bar). They are both
  very close to the vortec and share a common
  support on the pole tips.
• Expertise
• The system was designed by Subatech and BNL. The
  system has been installed by the STAR Technical
  Operation group.
• Documents the SSD cooling system document is
  available online. The various elements (filter,
  gauge,) are described on the web.
• S.Bouvier and G.Guilloux (Subatech), R.Brown and
  B.Soja (BNL)
• Maintenance
• The vortec have been maintained cleaning and
  shim replacement. It cost is modest and covered
  by the STAR maintenance budget. This maintenance
  has been performed by B.Soja

9
Hardware in the assembly building and the wide
angle hall

• Items
• A setup (compressor, filter, tank) is installed
  in the second floor of the assembly building and
  provide clean and dry compressed air (120 psi).
• A line links the compressor to the vortec mixing
  copper pipes and flexible hoses.
• A solenoid valve and a gauge allows to monitor
  and control the compressed air flow. The are read
  by the SVT slow control system.
• Expertise
• The system was designed by Subatech and BNL. The
  system has been installed by the STAR Technical
  Operation group.
• Documents the SSD cooling system document is
  available online. The various elements used to
  build the compressed air line are described on
  the web.
• S.Bouvier and G.Guilloux (Subatech), R.Brown and
  B.Soja (BNL)
• Maintenance
• The compressor had failed several times over the
  passed few years. A major maintenance has been
  done under the supervision of the STAR Operation
  group.

10
Hardware in a rack on the south platform

• Items
• Two CAEN power supply crates are installed in the
  SSD rack and are filled with LV and HV boards.
  Most of the slots are filled with used boards.
  Three different board models are used to provide
  HV (typically 50V) and LV (-2V, 2V and 5V).
• The boards and the crates are controlled by the
  SSD slow control system.
• Expertise
• The CAEN system was chosen by the IReS group
• The specs of the crate and boards are available
  on the SSD web pages.
• Experts D.Bonnet and F.Littel (IReS)
• Maintenance
• The LV boards have been erratically failing
  during the last runs. During the last shutdown,
  all the boards have been inspected and very poor
  thermal connections between driver components and
  their supports have been identified as the cause
  of the failure. Bad contacts have been secured
  and most of the others have been visually
  inspected.
• The number of spare elements is rather small.
  Only one spare board per model and no spare
  crate. The models used are obsolete and very
  expensive.

11
Hardware in a rack on the south platform

• Items
• Two patch panels are installed in the SSD rack.
  They are used to combine the various power lines
  (-2V, 2V, 5V, HV, senses) coming from the CAEN
  crates together to power individual
  (half-)ladder. One panel is dedicated to the
  N-side ladders (East side) while the other serves
  the P-side ladders (West side). Their are
  installed in between the two CAEN crates. All
  input on the patch panels are equipped with fuse
  boards.
• 40 LV cables and two HV cables. These short
  cables are connected to both the power crates and
  the panels. They are relatively short (1.5 m
  long) and partly dedicated to specific power
  lines. The 2V cables should not be used for an
  other power line. The HV cables are simple flat
  cables.
• Expertise
• The panel and cables were chosen by the IReS
  group and the SUBATECH group.
• The specifications of the cables and panels are
  available on the SSD web pages.
• Experts D.Bonnet and F.Littel (IReS), S.Bouvier
  and L-M.Rigalleau (SUBATECH)
• Maintenance
• The cables and the patch panels never needed any
  particular maintenance operation.
• Few spare cables and fuse boards are stored in
  the SSD cabinet.

12
Hardware in a rack on the south platform

• Items
• A slow control crate installed in the SSD rack.
  The crate is a WEINER VME crate remotely
  controled via its canbus interface. The crate is
  filled with several boards
• A processor board (VME-167) running VxWorks OS,
  an interface board (V288 Caenet) to communicate
  with the CAEN crates and an interface board
  (Corelis CVME 1149) for the JTAG communications.
• A distribution board is the key element for the
  communications with the SSD readout boards.
• Expertise
• The slow control system was designed by the IReS
  group.
• Some documentation on the slow control hardware
  is available on the SSD web pages.
• Experts D.Bonnet and F.Littel (IReS)
• Maintenance
• The slow control hardware has been working stably
  over the past runs.
• A spare of each board (except the crate) is
  available and stored in the SSD cabinets. A spare
  version of the distribution board is currently in
  France.

13
Hardware in a rack on the south platform

• Items
• A relay box is installed at the back of the VME
  crate in the SSD rack. This box contains various
  relays and a voltage transformer used for the SSD
  control itself and the STAR interlock
  implementation.
• A power supply (Lambda) dedicated to the SSD
  readout board power is installed at the same
  location. Between the power supply and the rdo
  board power cables a fuse box has been installed
  in order to limit the power consumption.
• Expertise
• The slow control system was designed by the IReS
  group.
• Some documentation on the slow control hardware
  is available on the SSD web pages.
• Experts D.Bonnet and F.Littel (IReS).
• Maintenance
• The relay box and the power supply connections
  have been improved and secured over the past
  years. P.Kuczewski who did some of these repairs
  has some experience on these parts.
• A spare Lambda power supply and some spare parts
  of the relay box are stored in the SSD cabinet.

14
Hardware in the acquisition room

• Items
• A DAQ crate equipped with 4 receiver boards is
  installed in the STAR acquisition room and is
  dedicated to the SSD subsystem. Each board is
  dedicated to a SSD rdo board.
• On the left side of the crate, 6 optical fibers
  are emerging from the DAQ room floor. Half of
  them is directly coming from the TPC wheels while
  the other half is first connected to the patch
  panel located near the middle of the daq crate
  wall.
• Expertise
• The crate and its boards was provided by the DAQ
  group.
• The matching table between the SSD rdo board, SSD
  optical fiber and the DAQ receiver boards is
  posted online.
• Experts the DAQ group.
• Maintenance
• The DAQ group is maintaining the crate and the
  boards.

15
Hardware in the control room

• Items
• A SUN station (ssdsun01.starp.bnl.gov) has been
  installed and is dedicated to the slow control
  and monitoring of the SSD. The epics software has
  been installed on this machine. The VME board is
  booting from this machine. The VME board software
  dedicated to the SSD monitoring is also loaded
  from this machine. A Web server and a channel
  archiver are also running on this machine.
• A Linux box (ssdlinux01.starp.bnl.gov) is
  installed near the sun station and was originally
  dedicated to online monitoring of the SSD data.
  For that purpose, the daq disk are mounted and
  specific software has been maintained for online
  analysis or offline analysis of collected data.
  Recently the slow control software has been
  duplicated from the sun station on this machine.
  The slow control monitoring can thus be done from
  it now.
• Login conditions on these machines are explained
  in the SSD operation guide and/or can be obtained
  from the experts.
• Expertise
• The machines were brought from France and
  installed by the two French groups.
• Experts D.Bonnet, L.Martin, J.Baudot
• Maintenance
• These computers are out of date but are running
  well since few years. Some software from the ITD
  has been installed on ssdsun01 to backup specific
  directories.W.Betts had some knowledge of these
  machines.

16
Documentation on the hardware

• Most of the hardware information is available
  online on the SSD web site
• The  Documents  page contains links to various
  useful document
• The SSD technical proposal and the SSD section of
  the STAR NIM paper
• The status reports presented during the
  collaboration meetings
• The talks/seminars given on the SSD while the SSD
  was under construction.
• A collection of images and pictures (the
   photothèque  section)
• The  Hardware  page contains more technical
  information
• A specific document dedicated to each particular
  aspect (electronics, cooling,) of the SSD
• The mechanical drawings
• Various documents (manual, reports, web pages )
  on specific item.
• The documents describing the tasks performed
  during the shutdowns (past years section)

17
Documentation on operating the SSD

• Most of the useful information to operate the SSD
  is also posted on the web
• The Operations page contains essential
  information to operate the SSD
• Various information important to the shift crew
• The updated SSD operation guide.
• A document on the SSD cooling system operation.
• A page showing typical online SSD QA histograms.
• A list of current problems
• Expert email addresses and phone numbers
• Some useful information for the experts and
  people allowed to play with the SSD hardware and
  software.
• The Online page contains heterogeneous
  information
• Some old information on the slow control
• SSD geometry information
• A link to the page dedicated to the calibrations
  (pedestals, gain, alignment).
• The DATA page contains test results and first
  data taken after a shutdown
• Pedestal and noise measurement
• Calibration results,
• Matching tables between the ladder Id and the daq
  data.

18
Documentation

• The SSD web site has been migrated from a web
  server at Lyon to the STAR web server at BNL.
• Most of the pages and documents have been
  transferred and should be available online.
• Few documents have been recently posted on the
  STAR Drupal tool accessible from the SSD page
  stored in the sub-system area. (http//online.star
  .bnl.gov/STAR/subsys/ssd/)

19
Software and data analysis

• Online software
• Some online software specific to the SSD is
  located on the ssdlinux01 machine. This software
  is useful to quickly look at online data (besides
  the Panitkin plots) or offline data freshly
  taken.
• The code is run from the /home/star/Online
  directory. It has been developed from the
  template (special.c) provided by the DAQ group.
  Several histograms are produced (and saved in ps
  file) pedestal and rms distributions, ladder
  occupancy, raw signal distributions The code is
  self documented and is the same since several
  years.
• A large fraction of the histograms defined in
  this code has been recycled as specific to the
  SSD in the Panitkin plots.
• In order to analyze data, some daq disks are
  mounted (/evb01_x where xa, b, c or d and
  /starevpa and /starevpb). If not, a simple mount
  command with the directory as an argument will
  mount the disk.

20
Software and data analysis

• StSsdDaqMaker raw data converter
• Goal
• The reconstruction of real data starts with the
  StSsdDaqMaker. This maker reads the raw daq data
  and converts them (with the help of a matrix)
  into a table containing a strip id and the signal
  value in adc count.It reads both physics data and
  pedestal/noise data and produces suited table
• Status
• The code is stable and in cvs
• Know issues
• For some ladders, an offset of one had to be
  introduced in the conversion matrix. Without it,
  the alignment plots do not show meaningful
  pattern. The exact reason for that offset is
  still not fully understood but the ad hoc
  correction works.
• Signals from noisy detectors/ladders are
  converted while they can easily be suppressed in
  that maker. An (existing) configuration can be
  used to ignore specific ladders or wafers.

21
Software and data analysis

• StSsdPointMaker SSD cluster and hit finder.
• Goals
• This maker reads the pedestal tables, convert
  them into a more suitable format and write them
  in a root file on disk with a filename containing
  the associated timestamp.
• This maker reads the real (zero-suppressed)
  physics tables, does the cluster finding, cluster
  matching, hit reconstruction and finally write
  them into a StEvent hit container.
• Several histograms are produced for QA
• Status
• The code is stable and in cvs
• Know issues
• In the cluster matching, the code does not decide
  yet which hit combination is the best. All
  possible combinations are written in StEvent. The
  association with the best (highest) probability
  must (and can easily) be selected.
• The relative gain calibration results must be
  integrated in the code. This is important for
  dE/dx capabilities and for the ambiguous hit
  de-convolution in a high local particle density
  (typically AuAu at full energy)

22
Software and data analysis

• StSvtEvalMaker SSD global alignment.
• Goals
• This maker is used to improve the global SSD
  alignment using track to hit association. It
  reads StEvents containing TPC (SVT) tracks and
  SSD hits, loads a given geometry and does track
  to hit associations based on relative distances.
• Track and hit information are stored in an ntuple
  which is used to identify correlations signing
  possible misalignment.
• Status
• The code is stable and in a private directory
  (/afs/rhic/star/users/lmartin/Aligment2). Some
  documentation exist in the directory
  (README.txt).
• Results for the CuCu at 62 GeV have been posted
  here
• http//www.star.bnl.gov/lmartin/Alignment/CuCu62F
  F/AlignmentCuCu62FF.html
• Known issues
• The code would need some cleaning for a commit.

23
Software and data analysis

• SSD pedestal and noise calibration.
• Goals
• Pedestal runs are usually taken at the beginning
  of each store during the data taking period ie
  typically few times per day. The pedestal and
  noise values of each strip are used in the
  cluster finder procedure of the StSsdPointMaker.
  They are loaded in the InitRun of the maker from
  the db. The StSsdPointMaker itself is used to
  produce a root file containing the ped/rms values
  before insertion in the db.
• .
• Status
• A complete analysis of the Run V pedestal runs
  has been done and results have been posted here
  http//www.star.bnl.gov/lmartin/Calibration/PedAn
  dRms.html
• This calibration was performed from the directory
  /afs/rhic/star/users/lmartin/Calibration and
  the output files were stored in
  /star/data06/SSD/pedestal_calibration/run5/AfterDa
  q
• Known issues
• The very large amount of data to store in the db
  must be reduced. Typically, only pedestal and/or
  rms values that have changed by x with respect
  to their previous values must be inserted in the
  db. Temporarily fix can be done by using a unique
  cut in the StSsdPointMaker (a default value in
  the StSsdStip object constructor).
• Existing root files are may be in an obsolete
  format. They should be easily reproduced using
  the xml files located in the dedicated directory.

24
Software and data analysis

• SSD gain calibration.
• Goals
• Correction of the gain difference on a chip
  basis. Jerome Baudot had done an extensive study
  of the gain response of the SSD using pulser
  runs. It is assumed that the gain is uniform
  within a given FEE chip. The observed differences
  are of the order of few percents. This gain
  calibration is useful to refine the cluster
  charge matching results and the SSD capabilities
  in terms of dE/dx.
• Status
• Jerome has posted his results here
  http//www.star.bnl.gov/baudot/ssdWork_index.html
  and in the daughter pages (pulser runs and
  calibration).
• Known issues
• The gain corrections are not implemented in the
  StSsdPointMaker. A dedicated table must be
  defined to store the values in the db and to be
  loaded by this maker to correct each adc signal
  on a strip basis. Pedestal and rms values must be
  corrected accordingly.
• Some FEE chip pulsers are not working. They thus
  need an other type of gain calibration. This can
  be easily achieved using real data and by
  selecting non ambiguous hits made of one strip
  clusters and by fitting the resulting charge P vs
  charge N correlations.

25
Software and data analysis

• Simulation chain.
• Goals
• The geant description of the SSD has been
  recently updated.
• The simulation chain is at the moment split in
  three parts
• The StSsdSimulationMaker that takes hits from
  gstar simulated event, generates signals on
  strips and mimick the daq behavior (zero
  suppression, compression)
• The StSsdClusterMaker that takes the simulated
  daq data and does the cluster finding and hit
  reconstruction.
• The StSsdEvalMaker that takes the various tables
  produced by the simulation makers and evaluation
  the performances.
• Status
• The code is in cvs and has been recently cleaned
  by A.Kisiel.
• Known issues
• The StSsdClusterMaker and the StSsdPointMaker are
  doing the same tasks on simulated and real data.
  They should be merged in order to be able to do
  embedding.
• The simulation maker assumes some homogeneous
  defects in the SSD. The best would be to
  simulated data using real data figures (pedestal
  and rms values for instance). Thus the code
  should be modified to read/load pedestal table.

26
Contacts

• People involved during the construction,
  installation and first years of operations who
  are good sources of information
• At Subatech (email firstname.lastname_at_subatech.i
  n2p3.fr)
• S.Bouvier engineer in electronics, broad
  knowledge on all aspects of the SSD, specific
  expertise on the electronics
• C.Renard engineer in electronics, designer of
  the rdo board.
• L-M.Rigalleau technician in electronics,
  designed the ADC board.
• Gerard Guilloux engineer in mechanics, designed
  almost each mechanical piece.
• L.Martin physicist, broad knowledge on all
  aspects of the SSD.
• At IReS (firstname.lastname_at_ires.in2p3.fr)
• D.Bonnet engineer in electronics, specific
  expertise on the module electronics, designer of
  the slow control system.
• F.Littel technician in electronics, specific
  expertise in the cabling and the slow control
  hardware.
• J.Baudot physicist, broad knowledge on all
  aspects of the SSD.