1. U.S. ATLAS Executive Meeting Silicon M&OWBS 3.1 August 3, 2005Toronto, Canada A. Seiden UC Santa Cruz

2. 2 ATLAS Silicon Tracking Silicon pixel detector WBS 1.1.1 Silicon strip detector WBS 1.1.2 Read Out Driver WBS 1.1.3

3. 3 M&O FY05 In FY05 began pre-operations for SCT and RODs. 1.5 FTE for SCT, 1 FTE for RODs in FY05.

4. 4 3.1.1 M&O for Pixels Pre-operations and Operations, start driven by Installation Schedule –Engineering and technical support for pre-operation and first operation of services (cooling, electrical and optical) in-pit. Critical manpower available in case of repair or intervention. –Extensive documentation of “as done” installation, hookup and first operation. Essential for later removal of pixel system (same people not likely to be available). –Contribution to continued technical support for operations, likely to be part of a common pool. Maintenance –Spares, major activity for FY06. –Minor consumables for cooling, services. –B-layer replacement, effort starting in FY07, completion in FY12, assuming replacement starting in 2012.

5. 5 3.1.1 Pixel Manpower and Cost (January ‘05 Plan) Spares Mechanical engineering and technical support for preops, initial operation and documentation Mostly B-layer replacement Technical support pool

6. 6 3.1.1 Pixel Manpower and Cost for FY06 (Aug. 05 Plan) 3.1.1.1Pre-operations postponed by one year. No funding requested for FY06. 3.1.1.3Request for spares is $75k. Developing a very strong CERN based pixel team in FY06. Kevin Einsweiler is now Pixel Project leader. Engineering team (installation group): Neal Hartman, Eric Anderssen, Tom Johnson. Physicist Base: Postdocs Andreas Korn, Marian Zdrazil, Tobias Golling; Graduate Student Michael Leyton, possibly second student.

7. 7 B-Layer Replacement The B-layer is the innermost layer of the pixel detector. It is important to overall tracking and essential for impact-parameter measurements, for example for b-tagging. It is expected to receive a lifetime dose in about three years at the LHC design luminosity. The target year for replacement would be 2012. Removal of the pixel detector and replacement of the B-layer is possible in a half-year or less shutdown. The remainder of the pixel detector would be retained. Overall ATLAS is developing a proposal for a B-layer replacement that is supported by the maintenance part of M&O funding. The U.S. has developed a funding plan focused on specific, critical elements, primarily in IC electronics and sensors. Initial R&D part of Upgrade R&D program.

8. 8 3.1.2 M&O for Strips Responsibilities in 2005: Finish picking chips with one bad channel for potential future use. Contribute to pre-operations and operations effort at CERN. UCSC has one technician (Forest Martinez-McKinney) and one postdoc (Sofia Chouridou) stationed at CERN. They are two members of a team of 19 people stationed at CERN, working on SCT barrels. Plan to add one student (Jessica Metcalfe). Also one engineer (Ned Spencer) responsible for systems issues, stationed in U.S. but makes frequent trips to CERN. In addition, A. Ciocio of LBNL is working on SCT at CERN as of this summer. Forest has been at CERN since January, except for one 6 week period in the U.S., he is now the cooling expert for the ATLAS barrels and is responsible for running the cooling system.

9. 9 3.1.2 Strip Manpower and Cost (January 05 Plan) FY05FY06FY07FY08 EE.5.4.1 Tech1.0.51.51.0 Costs Including Travel: 183k$154k$222k$115k$ In SR Building In ATLAS Detector Believe that it would be advantageous to not reduce presence in FY06. A better plan would be constant funding and manpower at FY05 level through 2007

10. 10 3.1.3 ROD M&O The ROD is a flexible design that is based on FPGAs that use synthesized VHDL for the firmware and DSPs that use compiled C code. This design allows for upgrades in the firmware and software as new needs are understood by use of the SCT and Pixel communities. The SCT/Pixel ROD is one of the most complex ROD designs in ATLAS. Maintenance and upgrades will be required for the life of the ROD, although greatest activity is expected during the first few years of pre operation and early operation. A competent staff and the tools needed to support the ROD must be maintained for the life of the ROD. The activities fall into 6 main areas listed below: 1. Software Support 2. Firmware Support 3. Maintenance 4. Licences 5. Material 6. Support at CERN

11. 11 3.1.3 ROD Software Support Software Support: The software is divided into 3 main areas. The first is the large volume of common infrastructure code that supports the basic operation of the ROD. Some of these basic operations include, configuring the ROD, providing event data to tasks, handling primitives (a primitive supports operations such as read/write access to a specific register, or starting a task, etc), error processing, and handling of communication. In summary, all of the infrastructure required to make the ROD function. The two remaining areas of code are the SCT and Pixel specific code. Examples of subsystem code are histogramming, error counting, and generation of serial module configuration streams. There is an ongoing effort to improve and upgrade the software that will continue to evolve as the SCT and pixel communities use the RODs and find new desires and needs.

12. 12 3.1.3 ROD Firmware and Maintenance Firmware Support: Firmware support is required for the VHDL code that determines how the logic in the FPGAs provides the functionality needed in the ROD. The firmware has been extensively tested in the user evaluation. During these tests problems have been found and corrected. Experience has shown that even when the ROD meets the basic requirements, there are new needs that result in changes. An example of a change could be the algorithm that selects an event to be trapped in the Router and transmitted to the slave DSPs. Maintenance Support: Maintenance is composed of repairing defective RODs. We are projecting failures over a number of years to be 10-20%. Plan includes 10% spares. Will repair boards to maintain spare numbers. Having qualified technicians (to repair RODs) and engineering (covered by firmware effort) is essential for repairing defective RODs, allowing SCT and pixels to perform with a full set of modules.

13. 13 3.1.3 ROD Misc. Costs Licenses support: Licenses are for C and VHDL compilers and hardware simulators. The costs are minimal, 4k per year for the first 3-4 years (until obsolete) and then maintenance of compatible hardware for the life of the experiment. Material: It will be necessary to maintain the supply of spare parts and as necessary make life-time purchases of parts that are becoming obsolete. Services such as replacing a ball grid array FPGA that has become defective, estimated cost is $200 for each replacement. Personal computers (estimate 2) will be needed to support the software in the first 1-4 year. Support at CERN: Support at CERN will be mainly in the commissioning and early running period. We have planned six trips to CERN over a four year period to find complex problems in the ROD system.

14. 14 3.1.3 ROD Manpower and Cost FY 05FY 06FY 07FY 08 EE (Firmware).2.4.4.3 Tech. (Maintenance).2.3.3.3 Computer Prof. (Software).6.6.6.4 Software Support60k$606044 Firmware Support29595945 Maintenance 22393333 Spares61 Misc. Costs 15285828 Total187k$186210150

15. 15 FY06 Summary January 05 Request: August 05 3.1.1316k75k 3.1.2154k183k 3.1.3186k Total:656k444k