Mosaic-1 Upgrade Project Execution Plan

Transcription

1 NATIONAL OPTICAL ASTRONOMY OBSERVATORY SYSTEM INSTRUMENTATION GROUP 950 N. Cherry Ave. P. O. Box Tucson, Arizona (520) FAX: (520) Mosaic-1 Upgrade Project Execution Plan NOAO Document M1U-AD Revision: 0 Authored by: David Sawyer Please send comments: dsawyer@noao.edu Page 1 of 16

2 Revision History Version Date Approved Sections Affected Remarks 0 01/07/2010 All Original Version Page 2 of 16

3 Table of Contents Revision History...2 Table of Contents...3 List of Figures...4 List of Tables Introduction Project Management ReSTAR Funding Budget Allocation NOAO Base Funding Budget Allocation Project WBS Project Plan and Schedule Project Management Detector System Dewar Assembly Software System Integration and Commissioning Coordination between KPNO Engineering and Systems Instrumentation Contingency Management Configuration Management Systems Engineering Requirements Management Risk Management Error Budget Management Observatory Interfaces Instrument Interface Detector Interface Facility Interface Design Reviews Project Plan Review Final Design Review Safety...16 Page 3 of 16

4 List of Figures Figure 1 - Work Breakdown Structure for the Mosaic-1 Upgrade Project...7 Figure 2 - Schedule for the Project Management Component of the Mosaic Project.8 Figure 3 - Schedule for the Detector System Component of the Mosaic Project...8 Figure 4 - Schedule for the Design Component of the Dewar Assembly...9 Figure 5 - Schedule for the Fabrication and Assembly Component of the Dewar...10 Figure 6 - Schedule for the Software Component of the Mosaic Project...11 Figure 7 - Schedule for the Integration and Commissioning Components of the Mosaic Project...11 Figure 8 - Project vs. Functional Lines of Authority for the Matrix Organization..12 List of Tables Table 1 - Effort levels for the assigned resources and labor skills derived from the project schedule described in section Page 4 of 16

5 1 Introduction The primary objectives of the Mosaic-1 Upgrade are to replace the aging (and obsolete) CCD controllers with modern MONSOON controllers, and to replace the CCD detectors to improve observing efficiency (faster readout times) and performance (quantum efficiency and read noise). Both the controllers and detectors must be replaced to achieve the desired gains in performance and efficiency. Beyond these objectives the only requirements are that the imaging performance must not be degraded, the instrument must use the same mechanical interface and ancillary equipment (filters, corrector, ADC, etc.), and the resulting data files must be compatible with the existing NOAO science data archive. Upgrading the controllers and detectors will provide the potential for several performance and operational enhancements beyond the scope of the requirements and these will be exploited as time and budget allow. There are two options for MONSOON controllers. One is a mature project, called Torrent, in progress to develop a new generation of MONSOON controllers. The Torrent controllers are designed for low power, ease of configuration, and use readily available components. The other option is the Orange controller, a previous generation MONSOON controller that is in use today on many instruments. The Torrent controllers are more suitable for the Mosaic upgrade application and thus are included in the baseline design. However, the Orange controllers will provide a fallback option if the Torrent project is not completed in time. 2 Project Management The responsibility of managing the Mosaic-1 Upgrade project has been assigned to the team of David Sawyer (Project Manager, or PM) and Steve Howell (Project Scientist, or PS). The PM will focus primarily on the administrative aspects of the project (budget, schedule, resource allocation, etc) while the PS focuses on the scientific and operational aspects (developing requirements, engaging the community, etc), although they will work together on most project-related activities. The PM reports to the NOAO System Technology Center (NSTC) Director, David Sprayberry, while the PS reports to the Kitt Peak National Observatory (KPNO) Director, Buell Jannuzi. The project is a joint KPNO/NSTC effort and both centers will provide the direction and oversight for the project. The management team is responsible for defining the requirements and scope of the project, developing a plan for completing the project, conducting reviews as appropriate, and providing suitable documentation. The management team will report on the status of the project at the weekly KPNO projects meeting and provide monthly written progress reports to the KPNO and NSTC Directors. It is expected that a broad group of people from within NOAO will be involved in the Mosaic-1 upgrade during the process of defining the project as well as during the construction, integration and commissioning phases. The management team will engage the NOAO group as appropriate and define work packages as required. Page 5 of 16

6 2.1 ReSTAR Funding Budget Allocation The total ReSTAR allocation to the Mosaic-1 Upgrade is $520,000. This amount is notionally allocated to capital equipment purchases in the approximate distribution of $500K for replacement detectors and $20K toward computers to drive the replacement controllers. Adjustments to the allocation of the funds may occur as the project progresses. The CCD detector selection was a critical component of the project regarding budget, science and mechanical design decisions and thus the detectors have been ordered from e2v Technologies as of December The cost of the replacement detectors was $520K, or $20K over the ReSTAR allocation, and so the difference will be covered in the base budget. 2.2 NOAO Base Funding Budget Allocation NOAO will supplement the ReSTAR funds with approximately $200K for labor from the base budget, and another $20K to make up the difference in the actual detector cost. These funds will pay for the PM and staff from the NOAO technical pool at the levels shown in Table 1. In addition, the PS will be paid from the KPNO operations budget. Table 1 - Effort levels for the assigned resources and labor skills derived from the project schedule described in section 2.4. Resource Skill Hours FTE Ditsler Tech support vacuum, detectors Daly Software engineer Dryden Electronic technician ET TBD MONSOON electronic technician Hunten Detector engineer NOAO Shop TBD Instrument maker KPNO Instrument Support Trainee KPNO Support Staff Instrument change Moore Detector engineer Rath Mechanical designer/instrument maker Rosin Draftsman Sawyer Project manager/systems engineer Schweiker Instrument specialist DHS Programmer TBD Software engineer Totals Project WBS The work breakdown structure for the Mosaic-1 Upgrade project is shown graphically in Figure 1. There are six major categories that, with the exception of project management, are matched to project deliverables. The detector system, Dewar assembly and software components can be developed in parallel and delivered independently. After that the project becomes more serial in that the system integration effort must follow the delivery of the components and the commissioning effort must follow the completion of system integration. In most cases the WBS is detailed to three levels, and at most four levels. The WBS is designed to provide adequate accounting and tracking for a project of this size while keeping the administrative effort modest. Page 6 of 16

7 Work Breakdown Structure for the Mosaic-1 Upgrade Project Figure Project Plan and Schedule A detailed schedule has been developed that accomplishes the primary objectives of the upgrade effort with realistic estimations of the resources available. The schedule follows the WBS organization and thus this description of the project plan and schedule will be broken down into the major WBS categories Project Management The project management component of the schedule is shown in Figure 2. The general project management responsibilities for the Mosaic upgrade project are projected to require one day a week each for the project manager and the project scientist for the duration of the project. In addition, there are specific project management tasks to development requirements and a project plan (this document), as well as preparing for two design reviews. The detailed design of the Dewar assembly is linked to Review 1 and is shown in Figure 4. The fabrication of the Dewar assembly is linked to Review 2 and is shown in Figure 5. The transition module component of the schedule is shown in Figure 3 and is linked to Review 2 (link not shown) because it is expected that the decision about the controller option will be finalized at that time. Page 7 of 16

8 Schedule for the Project Management Component of the Mosaic Project Figure Detector System The CCD Detector component of the schedule involves the procurement and testing/characterization of the detectors selected for the Mosaic upgrade. The detectors have been ordered from e2v Technologies and are expected to arrive by April 15, In the meantime there is design, fabrication and assembly work scheduled to build a chip mount for a universal test Dewar to allow testing the e2v devices in the lab. The plan is to build a chip mount that will allow the e2v detectors to be installed and tested two at a time. Schedule for the Detector System Component of the Mosaic Project Figure 3 Page 8 of 16

9 Before testing the e2v CCDs in the lab it will be necessary to configure the MONSOON controller electronics (Torrent or Orange) for the devices. This configuration is scheduled to begin immediately after the first design review so it can be completed by the time the detectors are delivered in mid-april as shown by the link (items 25 and 18) in Figure 3. The assembly and testing of the MONSOON controller electronics needed for the Mosaic upgrade is being provided by the MONSOON group as a deliverable and is thus not shown in this schedule. Following the Torrent configuration tasks, the Torrent transition module assemblies are scheduled for assembly and testing. The fabrication of the transition modules requires the enclosures to be fabricated and thus the module assembly task (item 29 on the schedule) is linked to item 47 (Figure 5) Dewar Assembly The first two components of the Dewar Assembly are the conceptual design and the detailed design phases as shown in Figure 4. The conceptual design phase primarily involves determining viable solutions for the upgrade components, including a new chip mount for the e2v detectors and the interface and mounting of the MONSOON controllers (either Torrent or Orange). The original Mosaic design was done using AutoCAD and since Solidworks is now the adopted CAD tool used at NOAO, there is a modest effort scheduled to convert the AutoCAD drawing to solid models. At the completion of the conceptual design there will be a design review (Review 1) where the proposed design solutions will be scrutinized. Following the approval of the conceptual designs, the detailed design phase will begin. The product of the detailed design phase will be fabrication-ready drawings for all new components and any modifications to existing components. A final design review is scheduled to occur at this point on March 23, 2010 and after that, upon acceptance of the final design, fabrication will begin. Schedule for the Design Component of the Dewar Assembly Figure 4 Figure 5 shows the fabrication and assembly components of the Dewar Assembly. The machining of the chip mount assembly will be done in the NOAO machine shop using a CNC mill to achieve the necessary machining tolerances for the CCD mounting holes and surface flatness. The machining of the connector pods and Torrent transition modules is scheduled to be done by S. Rath in the model shop, but using the NOAO machine shop as an additional resource is a viable option. During the fabrication phase, the new connector pods and their associated connectors and wiring will also be completed. Page 9 of 16

10 Schedule for the Fabrication and Assembly Component of the Dewar Figure 5 Following the completion of the fabrication phase, the Dewar rework and assembly phase begins with the assembly of the new focal plane. This involves mounting the eight CCD detectors on the chipmount and measuring/verifying the flatness of the detector focal plane. In the meantime, the Mosaic imager will be taken out of service on June 11, 2010, removed from the telescope and transported to the NOAO labs in Tucson. Once the instrument arrives in Tucson, the Dewar assembly will be separated from the filter/shutter mechanisms and the two components will be reworked in parallel. D. Dryden is scheduled to overhaul the filter/shutter mechanism to improve its performance and reliability (This is not actually part of the upgrade project, but is an opportune time to do this necessary servicing), while B. Ditsler is scheduled to complete the Dewar upgrades. The Dewar upgrade starts by removing the existing focal plane assembly and all of its associated wiring and placing it into safe storage. This will preserve the option of returning Mosaic to its current configuration as a technical contingency measure. During the removal of the focal plane careful measurements will be made to ensure that the focus position can be precisely duplicated with the new detectors. After disassembly, the Dewar will receive standard servicing (e.g., cleaning, bake desiccant, etc.) and any modifications that may be needed (e.g., grinding the surface of the radiative cold plate that attaches to the chip mount if better flatness is required.) The new focal plane will then be installed and measured through the Dewar window for focus position and tilt. The shims at the mounting posts will be lapped to adjust any positional errors. Once the focal plane is positioned the new wiring modules will be installed and, after attaching the radiative cold plate, they will be checked for continuity. The remainder of the Dewar will be assembled (LN2 tank, desiccant, and housing), checked for vacuum leaks and put through several pump-flush-pump cycles to maximize the vacuum and hold time. Once the Dewar rework is complete there is a week scheduled to cool the instrument and conduct thermal stability tests, and another week to test the detectors. These are really just functional verification tests but there is a two-week contingency scheduled at this point in case any problems are found or additional tests are required. Page 10 of 16

11 2.4.4 Software The user interface and user software must be changed as part of the upgrade project. Currently Mosaic-1 is run with an IRAF ARCON package as well as a few independent GUIs that control specific features of the instrument/telescope combination. Given the success of NEWFIRM s user interface and control system (NOCS), we are proposing this option to use for the new Mosaic-1 user interface. The advantages of choosing NOCS are 1) it has had two years of use and tuning at the Mayall telescope, 2) has complete and plentiful documentation, 3) a number of 4-m users are already experienced with its use, 4) the software will be going to Chile with NEWFIRM and so manuals already exist to teach the new system to the CTIO staff, 5) it is likely that NOCS will be used by KOSMOS at the Mayall, and 6) 90% of the functionality required to operate the new Mosaic-1 upgrade is already written and operational. The NOCS interface is already integrated with a Data Handling System (DHS) for NEWFIRM and only minor changes will be needed to modify the existing DHS for Mosaic. There is a modest software effort required to adapt the NOCS interface and DHS system for the Mosaic instrument. Figure 6 shows the tasks with rough estimates of time to complete them. These are only rough estimates of time, but there is plenty of schedule slack in this effort (item 74 to 75, almost two months) to absorb any increase in effort that may result from better estimates. Schedule for the Software Component of the Mosaic Project Figure System Integration and Commissioning The system integration phase of the project will involve the final assembly of all the subsystems in the lab where they will be tested as an end-to-end system. This component of the schedule requires that the Detector System, Dewar Assembly, and Software subsystems have all been completed, and thus the start date for this effort is driven by those links. The lab testing will involve functional verification, troubleshooting if required, and conducting the acceptance tests. An acceptance test plan will be prepared in advance and will ensure that the project requirements have been met. KPNO operations staff will be engaged during this system integration phase to provide a training opportunity. Once the integration is completed and the instrument has been accepted, it will be send to KPNO to start the commissioning phase. There is a two-week period scheduled for test and engineering in the telescope environment. Although the schedule shows the instrument being installed on the telescope at the start of the T&E period, it will probably occur a few days later to allow the functionality to be verified at the telescope before installation on the telescope. Most of the T&E testing will be done during the daytime and should not impact nighttime operations at the telescope, but a few T&E nights will be scheduled as Page 11 of 16

12 well. After the T&E period, there will be a few nights scheduled for conducting science programs to tune the operational efficiency and verify the scientific performance of the instrument. Schedule for the Integration and Commissioning Components of the Mosaic Project Figure 7 During the commissioning phase it is expected that KPNO staff will be involved to receive training and accept hand-over of the instrument. The user documentation will be completed and delivered as well. 2.5 Coordination between KPNO Engineering and Systems Instrumentation The PM will work with the appropriate managers within KPNO Engineering (T. Abraham) and NSTC/Electronics (M. Hunten) to identify and utilize their resources to the extent possible. The PM will utilize these resources in a matrix fashion by providing the project-related assignment and direction to the individuals, while leaving the functional obligations to their respective managers. Resource conflicts or concerns will be raised with the responsible manager and if a resolution cannot be reached the appropriate Director(s) will be notified. The PM will report the project status to both the KPNO Director (B. Jannuzi) and NSTC Director (D. Sprayberry) on a regular basis and, as part of the report, will provide a summary of the resource commitment and schedule status. Project vs. Functional Lines of Authority for the Matrix Organization Figure 8 Page 12 of 16

13 2.6 Contingency Management The contingency for budget and schedule are being managed through the scope of the upgrade effort. The schedule presented in section 2.4 was developed to meet the budget and schedule constraints of the project. There was a strong desire (goal) to build a new Dewar instead of retrofitting the existing one to minimize science downtime and risk associated with the upgrade project. However, this goal was not possible within the budget and timeline available and was not necessary to achieve the project requirements and thus was eliminated from the Project Plan. Additional project goals relating to user tools and operational modes may or may not be completed as time and budget allow. A source of schedule contingency will be available through additional labor (i.e., an increase in the base budget). The project will utilize staff from the KPNO and NSTC groups and, provided resources are available within those groups, they could be reassigned to the project to increase the work force. There is a modest cash contingency of $50K from the KPNO operations budget. Some of this contingency may be used to fabricate new Dewar parts if they have to be done by outside machine shops. This provides a contingency for the NOAO instrument shop that is notionally available for the Mosaic work but is currently involved in a large project (ODI) that could impact their availability. One of the technical risks of the project is the completion of the Torrent controllers. Torrent is well along toward meeting a production readiness state in March 2010 but is not complete and has not demonstrated performance at this time. A contingency plan to mitigate this risk will be to use the existing MONSOON Orange controllers instead of Torrents. The choice of controller will have minimal impact on the design phase of the project and thus this decision will be postponed until the Mosaic final design review (March 23, 2010) to give the Torrent project more time to progress. 2.7 Configuration Management This project will implement modest configuration management techniques to maintain revision control of drawings and documents. The PM and PS will have overall authority for acceptance of released documents and approval of changes thereafter. PDMWorks Enterprise software will be used as the document database for version control. The mechanical solid models will be developed in an initiated state where all project staff has access to them. Once the drawings are released by the PM or lead engineer the shop personnel will have access to them and no changes will be allowed unless the PM or lead engineer approve the changes and modify the drawing state to in revision. This process has been used effectively for several recent projects at NOAO. Documents will be entered into the PDMWorks database upon their initial release. There will be group access levels to the database that are appropriate for the types of documents. For instance, only a management group will have full access to documents with sensitive financial information, whereas the technical group can have full access to the technical documentation. In most cases any general user will have read-only access to project documents. 3 Systems Engineering The PM, David Sawyer, will also act as the Systems Engineer (SE) for the Mosaic-1 Upgrade project. The SE will ensure that the technical requirements and engineering decisions flow down from the science requirements as appropriate. Since this is a project to upgrade certain components of an existing instrument, only those components require validation. As such, this project will define a limited set of Page 13 of 16

14 requirements documents including the Science Requirements (SRD) in the context of the CCD detectors, and an Operations Concept Document (OCD) to define the user interface, observing modes and diagnostic tools. In addition to the requirements documents, the SE will be responsible for developing the Acceptance Test Plan (ATP) as well as any interface control documents (ICD) that are required for the mechanical, electrical and software interfaces. 3.1 Requirements Management The PS will lead the effort to define the science requirements. Once that is complete the operations and technical requirements can be defined as they flow down from the science requirements. The PS will lead the effort to define the operations requirements and the SE will lead the development of any technical requirements. For each operations and technical requirement that is defined, it will include statements about why the requirement is justified and how the requirements will be verified. The justification may include direct flow down from science requirements, derived to meet a science requirement, or allocated as a design choice. The verification methods may include demonstration of an actual component, controlled test of an experiment, or analysis through calculations or modeling. The verification statement will be used as the basis for developing the acceptance tests. 3.2 Risk Management There is a significant schedule risk since this is a fast-track project occurring simultaneously with other priority projects for KPNO. There is a fixed pool of resources that are distributed among project and operations responsibilities, and thus there is a limited group available for this project, especially in the area of detector technicians. As such, the availability of resources may be out of the project s control and could impact the schedule. The fast-track plan of this project is designed to avert some of the risk by utilizing key resources before they are needed for other scheduled projects. There is moderate Base (labor) funding budget risk since this project plan has minimized the scope of work as much as possible to match the budget allocation. Since the labor resources will be provided by the KPNO and NSTC groups it should be possible to supplement the Base budget by reassigning resources within those groups. There is a moderate technical risk with the Torrent controllers since Torrent is currently a development project and is not yet at a point where the performance can be demonstrated. The design of Torrent reuses much of the MONSOON Orange technology to mitigate the technical risk, but it also incorporates some new technologies. A contingency plan to mitigate this risk will be to use the existing MONSOON Orange controllers instead of Torrents. This contingency plan will require a modest additional effort to develop an electrical interface for the Orange controllers, and a mechanical package. This decision will not have any impact on the Dewar design and thus does not have to be considered until the Mosaic Final Design Review in late March Error Budget Management Error budget analyses will not be needed for this project since the components being upgraded will meet or exceed the performance of the components they are replacing. Page 14 of 16

15 3.4 Observatory Interfaces Instrument Interface The mechanical mounting of the Mosaic-1 imager and its ancillary equipment will not be changed as a result of the upgrade and thus a telescope interface control document (ICD) is not required. However, there will be changes to the CCD controllers that are mounted to the Dewar and so the SE will ensure that the new MONSOON controllers are mounted in such a way that they do not interfere mechanically or block access to any other equipment on or around the instrument. In addition, the SE will ensure the mounting of the controllers is optimal for servicing. For electrical interfaces the new controllers will only require 115V AC power and two data fibers. These services are already available at the instrument on the telescope and thus no changes are required. If the MONSOON Orange controller is used, there will have to be provisions for mounting the power supply near the instrument. The SE will work with KPNO personnel to determine the optimal location for mounting the power supply. Interface requirements for shutter and focal plane temperature control will have to be defined by the SE. Currently these functions are controlled within the CCD controllers that are being replaced. The Torrent controllers will have the capability of providing the shutter and temperature control but the Orange controllers do not. In either case, the requirements will have to be defined so that a new control interface can be designed and implemented. The software interface requirements will be described as part of the Operations Concept Document (OCD). The OCD will describe the interfaces required for the user GUI, the data handling system and the data pipeline Detector Interface A detector engineer will provide a detector interface document that is customized for the e2v CCD44-82 detectors that have been chosen. The document will describe controller configuration requirements, such as bias and clock voltage levels, clock timing, power sequencing, etc. The document will also detail the signal interconnection scheme from the CCD device, through the vacuum barrier, to the MONSOON controller Facility Interface The Mosaic-1 upgrade will have minimal impact on the telescope facility and is not expected to require any modifications or additions to the existing facility interfaces. 4 Design Reviews Since the Mosaic-1 upgrade project is a relatively minor project in terms of budget and scope, the review process will be limited to two reviews during the project. There will be a project plan review before the detailed design phase, and then a final design review before the fabrication phase. 4.1 Project Plan Review The project plan review will be held after the planning phase is complete and the project has been defined. The review will include the science requirements, operations concept and project plan including Page 15 of 16

16 the schedule, budget, risk assessment and contingency measures. The review will also cover the preliminary designs for the Dewar, detector and software systems. 4.2 Final Design Review The final design review will be held once the detailed mechanical design is complete. This review will be an internal review and is intended to identify any technical issues or concerns with the design before the major fabrication phase begins. Components of the project that are considered low technical risk may be fabricated before the final design review if there are cost/schedule benefits to doing so. 5 Safety The Mosaic-1 Upgrade project will be conducted in a way the meets the safety policies of KPNO and NOAO. The Mosaic-1 project will include handling heavy equipment, handling liquid cryogens, running machines, and working in observatory domes with overhead cranes. The personnel that will be required to conduct the work are all trained for proper safety practices. Working with liquid nitrogen will only be done in areas that are designated for such work (i.e., they have adequate ventilation or an oxygen warning sensor). Page 16 of 16