KOPIO Closeout BNL, April 22, 2005

1
KOPIO CloseoutBNL, April 22, 2005

• The KOPIO Sub-panel
• Marj Corcoran
• Peter Denes
• Karol Lang
• Dick Loveless
• Leo Piilonen
• Stan Wojcicki
• NSF, NSERC observers

,
2
Preamble

• We thank the KOPIO Collaboration and the RSVP
  Project Office for frank and open discussions and
  presentations.
• We have been impressed with the amount of effort
  put into the planning of the experiment.
• At the conceptual level KOPIO appears to be well
  thought-out and ingenious.
• It is also understood that limited funds
  available so far have led to a yet-incomplete
  detailed design but we have not identified any
  show-stoppers.
• We were unable to thoroughly review the physics
  capabilities
• but have no reason to doubt the claims presented
  in the CDR.
• Findings and recommendations of the recent
  Ritchies Panel have been largely accepted by
  KOPIO.
• We are reminded, however, that the past history
  teaches that upgrades will be necessary to reach
  the design sensitivity
• Below are our findings, comments and
  recommendations.

3
First physics run in 2012
4
KOPIO breakout sessions

• Presentations by
• Mike Marx Status and plans
• Dana Beavis Vacuum (WBS 1.2.1)
• Toshio Numao Preradiator (WBS 1.2.2)
• Nello Nappi KOPIO Trigger (WBS 1.2.7)
• Oleg Mineev Photon veto (WBS 1.2.5)
• Vladimir Issakov Calorimeter (WBS 1.2.3)
• Doug Bryman - Backgrounds
• Discussions with sub-system managers
• Vacuum (WBS 1.2.1) Lang, Piilonen ?? Beavis
  an eng
• Preradiator (WBS 1.2.2) Loveless ?? Numao
• Calorimeter Photon Veto (WBS 1.2.3/1.2.5) -
  Corcoran, Lang ?? Issakov, Mineev
• Charged Particle Veto (WBS 1.2.3) - Piilonen ??
  Frank
• Trigger, DAQ (WBS 1.2.7/1.2.8) Denes ?? Nappi,
  Kettel, Schamberger
• Offline (WBS 1.2.9) Piilonen ?? Poutissou
• No discussions on System Integration and Project
  Services (WBS 1.2.10/1.2.11)

5
1.2.1 Vacuum - Findings

• System consists of
• Upstream Vacuum Decay Vessel ( 1418k)
• Vacuum Transitions (windows membrane) (
  359k)
• D4 Vacuum Box ( 238k)
• Downstream Veto Vacuum Tank ( 393k)
• Vacuum Pumping Station ( 630k)
• Management Activities (includes ¼ engineer for 4
  years) ( 167k)
• Biggest technical challenge
• 12m3, 10-7 Torr, 7 X0
• Investigated many options
• Beryllium, Carbon fiber, Al Honeycomb, Aluminum
• Carbon fiber 1/5 scale under construction in
  Russia
• Spun aluminum seems best vendor quotes and
  fabrication plan
• (SPINCRAFT) a vessel for 381k
• Will order 2 vessels, the first one for tests
• Membrane to separate 10-3 and 10-7 Torr vacuums.
• Charge Particle Veto (CPV) to reside inside 10-3
  volume

6
1.2.1 Vacuum - Findings

• Costs and Schedule
• Probed down to level 5 cost drivers

1.2.1 Vacuum - Comments

• In the past invested heavily in engineering of
  the tank good call!
• More engineering necessary and is budgeted!
• Need to integrate Charge Particle Veto
• mounting
• feedthroughs
• define the membrane
• Costing on D4 and D/S vacuum box agrees with past
  experience
• Need to integrate the Magnet Photon veto system
• Main cost drivers Vacuum tank, pumping station
• Reasonable costs and schedule and the level of
  contingency

1.2.1 Vacuum - Recommendation

• Complete the design CPV engineering integration

7
1.2.2 Preradiator - Findings

• The preradiator is a 3 x 3 m2 sandwich of 256
  (2x2x64) cathode strip drift chambers and 288x27
  6 scintillator planks (8 mm thick) with Pb
  radiators
• Designed to measure gamma ray location with a
  position accuracy of 5mm and an angular accuracy
  of 25mrad.
• The cathode strip chambers consist of 75K
  channels of anode (wire TDCs) and 75K channels
  of cathode (strips ADCs).
• On the outside of the scintillator-drift chamber
  sandwich is an array of 32x36 external photon
  vetos, which are made of Pb-scintillator.
• The Triumf group is responsible for the
  preradiator including the front-end chamber
  electronics.
• They will provide and install all parts of the
  preradiator except the external veto counters.
• 5M from Canada
• The estimated cost of the complete preradiator is
  22.6M with a contingency of 5.4M (31).

8
1.2.2. Preradiator - comments

• The cathode strip drift chambers are well-suited
  to measuring the position and angle of the ?o
  conversion.
• This is a well-developed technology used in other
  contemporary experiments
• The WBS structure shows good detail and is fairly
  complete
• Contingency of 31 seems small for a project of
  this size and development
• Labor costs for chambers are comparable to
  materials costs (reasonable) but labor costs for
  electronics seems small
• Electronics testing is part of vendor delivery --
  may need additional testing of assembled systems
• Includes funding for 3 years of a system engineer

• .

9
1.2.2 Preradiator findings/comments

• The estimated base cost for the chamber
  electronics (not including contingency) is 6.53M
• 32/channel for both anode and cathode
• Labor for all electronics costed under anode,
  should be redone
• Spares (about 10) should be included
• Scintillator design uses extruded planks with
  holes
• WLS shifting fibers inserted into holes -- nice
  design
• Cathode strip chambers are excellent antennas
• Solid grounds are essential
• May need considerable work/testing during
  installation at Brookhaven
• Triumf group is likely to need additional
  people/support to complete this project
• responsibility for installing/integrating the
  photon vetos not designated

10
1.2.3 Calorimeter - comments

• The design of the lead/scintillator modules is
  well-thought-out and complete. Costing was well
  documented and based on extensive experience.
• The Calorimeter will be a shashlyk structure
  fabricated in Russia.
• The group at Vladimir has an impressive track
  record in producing such devices, including
  Phenix, LHCb, and HERA-B.
• They can produce the modules very
  cost-effectively.

11
1.2.5 Photon Vetoes - Findings

• The photon vetoes consist of four different
  subsystems
• Upstream Veto Barrel Veto Magnet Veto
  Downstream Veto
• Shashlyk log lead/scintillator technology
• These systems need a single-photon veto
  inefficiency on the order of 10-4 for KOPIO to
  succeed.
• Where possible E949 phototubes will be recycled,
  resulting in considerable cost savings.
• Shashlyk technology also used in the Preradiator
  photon veto.

12
1.2.5 Photon Vetoes - comments

• The designs of the lead/scintillator modules
  seemed to be well-thought-out and complete.
  Costing was well documented and based on
  extensive experience.
• Support structures In all cases except the
  barrel veto, the mechanical design of the support
  structure is yet to be designed.
• Responsibility for installing/integrating the
  photon vetos not designated
• Hermeticity is crucial to the success of the
  experiment. Therefore, the design of the support
  structure, especially ensuring it introduces no
  gaps or inert material is of utmost importance.
• Spares For devices which are using 949 PMTs,
  there are some spare phototubes. For detectors
  which need new PMTs, no spares are in the
  baseline. None of the detectors have spare
  front-end electronics. But it is clear that the
  experiment cannot tolerate even one dead channel
  in the photon vetoes

1.2.5 Photon Vetoes - recommendations

• Assign responsibility and provide adequate
  engineering to develop designs for the photon
  veto support structures and integration.

13
1.2.6 Photon Catcher findings/comments

• The photon catcher is the an aerogel-lead device
  sensitive to the Cherenkov radiation produced by
  photons which convert in the lead. The
  sensitivity to neutrons is only 0.3, but even so
  this detector produces about 4-5 dead time.
• This detector is the responsibility of the
  Japanese group, who has assumed all financial
  costs.
• The Photon Catcher needs to achieve an
  inefficiency of 10-2, so the requirements are not
  nearly as stringent as for the other photon
  vetoes.
• The Monte Carlo studies have been validated with
  beam tests at KEK, resulting in excellent
  agreement with Monte Carlo and data.

14
1.2.4 Charged Particle Veto - Findings

• Cost 2.63M 1.27 3.35M
• Schedule detector construction/installation
  done by FY08, electronics delivered by FY09
• Thin segmented scintillator envelope inside
  vacuum tank to detect charged particles with
  gt99.99 efficiency and 100 solid angle scints
  overlap no cracks except for beamline
• Barrel design, construction, assembly by Zurich
  design is advanced
• Downstream design, construction, assembly by BNL
  design is conceptual
• Front-end electronics shares common design with
  other systems, but who constructs, assembles and
  maintains it?
• Each scint viewed by 3 or 2 direct-mount
  phototubes for redundancy this detector is
  inaccessible
• Integration with vacuum vessel is needed at early
  stage of vacuum vessel design feedthroughs at
  vacuum vessel flanges for low voltage and LED
  monitor high/low vacuum barrierincluding
  mounting schemeand its contribution to the
  inert-material budget in front of each scint
  support structure heat-conduction scheme
• Steeper charged particle entrance angle into
  downstream scints ? need thin scint coating
  MgF2? instead of wrapping to meet the
  inert-material budget is this lt20 mg/cm2 or lt8
  mg/cm2?

15
1.2.4 Charged Particle Veto - comments

• The Charged Particle Veto is well designed. Some
• issues vis a vis integration with the vacuum
  vessel
• (feedthroughs, support, heat transfer, high/low
  vacuum
• barrier) need to be resolved.
• The costs of this subsystem, probed down to WBS
• level 7 for the cost-drivers, are
  well-documented with
• price quotations for materials/equipment and
  reasonable
• estimates for fabrication/installation. No
  spares.
• Fairly complete WBS. All aspects of the necessary
  work, including material procurement, design,
  labour, tooling, integration, and installation
  are included. Not clear who is responsible for
  the readout electronics, though.

16
Electronics
Many different components in many different WBS
elements
CLOCK

• Chamber readout owned by preradiator
• Photon readout is common
• Photodetector owned by subdetector
• Rest is owned by ? (Virtual) electronics group
  or subdet.
• Integration owned by subdetector
• Clock owned by L1 Trig
• L3 trigger 400 node processor farm
• L2 trigger TBD

Anode
TDC
Cathode
ADC
40 MHz
Logic/Pattern Boards
ADC
PMT
Base
ADC
APD
Super- visor
250-500 MHz
DAQ
Front end and digitizers
Collectors
L1 Trig.
17
Electronics - comments

• Certain elements well developed and prototyped
  others in (pre-) conceptual design
• Significant engineering needed
• For prototype to production board design
• For trigger electronics design
• For FPGA coding
• Work estimates reasonable for an iteration
  perhaps short for final production
• Testing and integration manpower needed
• Schedule is tight for certain items
• L1 trigger has 6 months float ? need
  manpower/engineering by then
• 10x10 calo prototype assumed to use final
  electronics ? 18 months to finalize calo
  electronics

18
Electronics - comments

• The group appropriately exercised the contingency
  methodology proposed by the project.
• Relative to charge recognizing, importantly,
  that the construction project will include
  significant engineering design and development
  activity KOPIO presents no significant
  electronics challenges

Electronics - recommendations

• Add people (and use that as a metric)
• Consider forming an electronics group (at least
  for everything other than the chamber front-end)

19
1.2.9 Offline - Findings

• Cost 0.78M 1.18 0.92M
• Schedule hardware purchased as late as possible,
  software development schedule to be fleshed out
  in Fall 2005 with work done on collaborators
  existing equipment
• Hardware purchase commodity components  tape
  silo (4 tape drives, 600 tapes, 1 TB/tape) for
  raw data disk farm (100 TB) for skimmed data 84
  dual-node processor farm for real-time
  processing 15 workstations network switches
  racks
• System manager (0.5 FTE) is shared with L3
  trigger.
• Software simulations, reconstruction,
  generation of calibration constants, quality
  assurance monitor, data analysis, skimming, data
  management
• One professional programmer and many physicists
  to design and manage analysis framework, data
  format, data management system, detector
  description language, and documentation
• Full-time software managers (physicists) needed
  for simulation, event reconstruction, calibration
  analysis, physics analysis, quality assurance
  monitor. Additional manpower (physicists) needed
  to do the work.

20
1.2.9 Offline - comments

• A plan needs to be formulated this year for the
  software development effort.
• The hardware costs of this subsystem, probed down
  to WBS level 6 for the cost-drivers, are
  well-documented with price quotations for
  equipment (based on recent TRIUMF purchases) and
  reasonable estimates for performance improvements
  in 4 years.
• Software costs are dominated by salary of one
  computer professional. However, additional
  professionals will be needed.
• All aspects of the necessary work, including
  equipment procurement, design, labour, tooling,
  integration, and installation, are included in
  the WBS. Timetables are to be solidified for
  software work.
• Additional manpower needs are likely.
• L3 and offline should consider adopting a common
  processor-node spec.

21
Contingency

• Currently, bottoms up Lockheed formula.
• This appears to be artificial and not adequate.
• Top-down contingency proposed by the RSVP
  management is 45. It is more reasonable but we
  are unable to assess if it is correct.
• This would give a total cost of KOPIO to be
• 53M 1.45 ? 77M spares
• We recommend that KOPIO refines contingency
  analysis to improve credibility of the cost
  estimation.

22
Response to the charge

• Technical approach and feasibility
• Conduct early beam tests as feasible
• Refine simulations to improve credibility of
  background calculations
• Completeness of the plan (WBS)
• Reasonably complete
• NO spares!
• Readiness to proceed to construction
• Significant engineering required
• Expansion of the Collaboration necessary
• Likely duration of the experiment
• Past experience for this type of experiments
  teaches that upgrades and improvements will be
  necessary to reach the design sensitivity
• (Technical) metrics of progress
• Adding personnel is the most critical need
• Costs and schedule
• Refine contingency analysis to improve
  credibility of cost estimation