Optical fabrication of the camera elements continues. The spherical sides of elements 7 and 8 are complete and David Hilyard is working on the aspheric sides of those elements. He estimates he will complete by the end of December.

One side of element 5 is nearly complete, but work on this piece has been stopped due to a fracture which has developed in it. This element is CaF₂, and is the largest piece to ever be optically figured other than flat. Details of the potential causes of this fracture and the steps planned to avoid the problem in the future are discussed in the Optics report in Section 2. We have ordered a replacement piece that is scheduled to be delivered in early April 1997, at an extra cost of $12,000. At this time the optical fabrication schedule is not adversely effected by the problems with the CaF₂, as David Hilyard is ahead of schedule as compared to the schedule estimates in the fabrication of elements 7 and 8.

At the end of the quarter the CaF₂ and flint test doublet were cooled successfully. No problems were encountered with the cell design, the RTV bonds, or the Dow Corning-200 couplant used between the lens elements.

A recruitment is in progress to replace the optical technician who resigned earlier in the year. We hope to have identified a suitable candidate by the end of the next quarter.

L&F Industries in Los Angeles is currently fabricating the DEIMOS structure including the cylinder, drive disk and substructure. These items are planned to be completed and delivered to the Labs in Santa Cruz by the end of the next quarter. Currently the cylinder is complete, the substructure is waiting inspection by the design engineer, and the drive disk is in process.

Jack Osborne continues the design of the grating system and has prototyped and tested several of the components. The plan is for him to complete this design during the next quarter and order the 600 and 800 line gratings.

Eric James has completed most of the design work on the structure and has shifted to the design of the slit mask handling system. The slit changing mechanism has been simplified by switching to a flat, flexible slit mask frame. This frame would be formed into a cylinder by a curved fixture in the focal plane that will give it the correct contour. Currently frame and fixture tests are ongoing to ensure that the slit masks form to the focal plane within the required tolerance.

Optical design of the TV system is nearing completion, and the plan is to order the ìstandardî Keck II guide camera from Photometrics during the next quarter.

At the end of the quarter, work on the DEIMOS dewar and mosaic was picked up from where it was left at the time of the PDR. A concept that would include both focus and one axis of the flexure control inside the dewar is being developed. A prototype of the CCD packaging that will be used by Lincoln Labs was received from Gerry Luppino, and clips to hold it to an invar mosaic base have been designed and fabricated. Testing of the stability of the mosaicing scheme is planned for the next quarter.

The CCD lab is near the end of their development of a CCD thinning process. The last step that remains to be completed is the etching of the bond pads. Richard Stover is optimistic that they will have a thinned chip in the next couple of months. Lincoln Labs is also nearing the end of their development cycle and plans to have chips to the University of Hawaii by the end of the year.

The DEIMOS software effort focused on issues raised at the software PDR and in establishing the software development environment we will use to build DEIMOS code. As a test of that environment, software development tools built and installed during the previous quarter (e.g., gcc, RCS, CVS) were successfully used to re-organize the existing Keck-I instrument software into a portable, architecture-independent package. An internal review of the DEIMOS database schema was successfully completed at the end of July. De Clarke and Steve Allen both presented papers at the Astronomical Data Analysis and Software Systems (ADASS) conference at NRAO in September. Those papers dealt with the formal designs for the DEIMOS database and for representing and handling the mosaic data in FITS format, and were well received.

A mini review is planned to cover the design of the grating and slit mask systems prior to the start of fabrication of those systems. This review in planned for mid-December 1996.

2.1 Optics

The spherical sides of camera elements 7 and 8 were accepted as finished in July and the aspheric sides were plunge-ground in early August. Work continues on those aspheric surfaces, now in the final figuring stage, and completion is expected on both by the end of November. Initial labor estimates for each aspheric surface were about four months, worked one at a time, since the aspheric figuring process is demanding. However, in the interest of schedule and efficiency, working two aspheres at once was tried, stepping each one along alternately between working and testing. If progress continues as expected, two aspheric surfaces should be finished in the time originally allotted for one.

The calcium fluoride lens blank for element 5 fractured during the final figuring run on the first side. After careful inspection of the blank at Optovac by their fracture mechanics expert, it was thought that a micro-fracture at the intersection of the bevel and the optical surface was the starting point, with the fracture terminating about 1.25" in from the edge at the intersection with crystal grain boundary. Inspection of the blank in their strain viewer showed that there was little stress left in this area of the blank. They recommended that the bevel be reground and polished before proceeding, since there were many other micro-fractures present at the bevel-optical surface intersection. The bevel was carefully ground and polished and the blank stored. After about a week, it was noticed that the fracture had propagated slightly (0.05-in) at the grain boundary intersection. Because of this instability, we do not think it is safe to continue with this blank, and a new lens blank was ordered for $12,000, with delivery expected in April 1997.

In the future a method sheet will be written for each CaF₂ lens, containing the details of the operational sequence process, including helpful insights from experts at Optovac on working with this material. We are, in effect, the test ground for making large CaF₂ lenses with this kind of thickness-to-diameter aspect ratio and center-to-edge geometry. Optovac is very anxious for us to succeed and have been very helpful and forthcoming with information. They closely reviewed our fabrication procedures on the first blank and suggested only that, we constantly monitor the bevel for microfractures and re-bevel as necessary. Other than that our procedures were deemed fully appropriate.

The next two lenses that will likely be worked are element 2, another CaF₂ lens, and element 1, the third asphere in the camera.

The final design and tolerance report from ORA was received, and Faber is studying it. We will report on its contents next quarter.

Mast, Hilyard, and Faber formally accepted the figure on the collimator mirror, and the central plug will be removed and various dimensions on the mirror measured to aid later collimation of DEIMOS. The large-scale slope error on the mirror (per beam footprint) is 0.21 arcsec (1-D rms), which corresponds to an rms image distortion in the focal plane of 0.05 px (0.006 arcsec on the sky). This is better than the stated tolerance for large-scale slope errors, which is 0.28 arcsec. The measured small-scale 1-D rms slope error (on scales smaller than a beam footprint) is 0.40 arcsec. This corresponds to a 1-D image blur of 0.10 px rms (0.03 arcsec on the sky FWHM) using geometrical optics, and less than half this using an approximate physical optics argument. This error is also better than the tolerance, which is 0.47 arcsec for small-scale slope errors. In summary, the collimator mirror surface quality is very good.

The cylinder has been fabricated and is now awaiting completion of the other structural items being made at L&F Industries. The hatch covers have also been ordered from L&F. The various interior bulkheads have been made and are in our shop awaiting the arrival of the structure.

The drive disk is currently in the fabrication stage at L&F Industries. It should be completed in late November. Several features have been added to the drive disk to aid in installing and aligning the various systems that mount on the disk, including a series of tooling balls mounted at precise locations. The front surface will be made flat and perpendicular to the rotational axis at L&F. The edge will be precision ground to be cylindrical and concentric with the center hole.

The carriage is nearing completion at L&F. It also has a precision machined reference surface on the rear bearing to aid in assembly and alignment. The roller assemblies and the rear bearing assembly are now complete.

A band brake has been designed to fit over the drive disk. This brake will have several functions. It serves as a locking device to prevent the rotation of DEIMOS during servicing. It also resists upward loads of the cylinder during earthquakes, and it provides thermal insulation and dust protection for the edge of the disk.

An encoder housing and mount have been designed for positional feedback for cylinder rotation angle. The drive system for driving DEIMOS along the Nasmyth rail systems has also been designed and resembles the systems that move the handling fixtures for the Cassegrain instruments.

We intend to send a crew to L&F in Huntington Park to assemble and test the entire structural system when the various parts have been completed. Any major alterations or adjustments would be better done at their facility where they have the machinery and tooling to do it. This visit is scheduled for mid-December.

Fabrication drawings for the kinematic mounts on the Nasmyth platform are complete, and the material is on order. These will be made early so that DEIMOS can be mounted on them in the Lick Instrument Shop. A special mounting scheme has been designed for mounting DEIMOS in the shop so that we can access to the underside.

A series of teleconferences with Keck, the DEIMOS team, and the NIRSPEC team has settled on a final concept for the Nasmyth rail system. Keck will design and build the system for the instruments.

The thermal tests of the test doublet for the camera are underway with no noticeable adverse results. They were preceded by rough calculations of the differential rates of cooling of the multiple lens elements versus the aluminum mounting rings and camera body. These rates suggested that with reasonable cool there should be no danger of thermal breakage of the camera elements in shipping to Mauna Kea. The thermal cycling of the test doublet (containing CaF₂ and flint glass elements) confirmed this, no couplant leaks were detected at any temperature, and crossed polaroids did not reveal new stress concentrations.

We are now proceeding with the concept that the flexure compensation in the x direction (perpendicular to the spetrum) will be done inside the detector housing rather than by moving the camera. This will simplify the design of the camera mount. Further design work on the camera cell and the camera mount will begin again in February.

The collimator cell and several of its parts have been painted or otherwise coated. The interior mounting parts have not been fabricated pending final measurements of the mirror.

There has been no work done on the tent mirror design since the last report. We are in the process of getting quotations for the fabrication of the mirror itself.

The design of the slitmasks is now being done by Eric James and will ramp up during the next quarter. A major shift is away from rigid to flexible (or even no) outer frames. The new design is much flatter and is deformed into a cylinder on insertion into the focal plane. This permits storage in a linear cassette similar to LRIS and elimination of the complicated caterpillar mechanism. Ten slots are still available.

The test instrument rotator was made smaller to stow out of the way when the real rotator arrives. It is now a permanent shop facility. The kinematic mounts for the grating rotator are still undergoing modification and tests. The Mark II design for the grating mounts nearly met specifications (9 arcsec rotation under PA change) and a Mark III improvement is underway. All four necessary electronics boxes were mounted onto the rotator for flexure tests. This includes a new network server. A cable-take-up system was designed and installed onto the new grating slide mechanism. This mechanism is being studied, and preliminary drive designs are started. No progress was made on the automatic grating kinematic lock devices.

The full-size model of DEIMOS was disassembled and discarded to make room in the shop for other projects and equipment.

A large 8 by 12 grating cell is currently being designed. A purchase order is being prepared for two 6 by 8 gratings from Milton Roy. The Science Advisory Team is being consulted on the choice of these gratings. We purchased a Canon 200mm lens for the TV guider system and have taken some preliminary field images using 35mm film. Since this system will also be used by ESI, we have committed to packaging it early in conjunction with the ESI engineers. ESI has ordered a Photometrics camera system, and we will share mechanical information about it before purchasing one.

We are working with Terry Mast in writing a proposed alignment plan. Towards that end we purchased several lengths of graphite/epoxy tubes to be used to locate the collimator mirror from the slit area datums.

2.3 Detectors

We have received and tested one of the two remaining DEIMOS design CCD wafer runs from Orbit Semiconductor. Based on wafer probe testing there is still a slight problem with charge trapping, but it appears this can be overcome by adjusting voltage levels. With this caveat, there were two flawless devices plus four devices with one or two column defects that would also be judged instrument grade, for an overall yield of about 18%. During this quarter we have also tested two other runs of the same devices made by Orbit for other customers. Those runs yielded 44% and 69% on the same grading scale.

The Lick thinning effort has completed all steps necessary to produce thinned CCDs except for the final step of removing the silicon from the bond pads. This final step has been giving us considerable difficulty, and we are testing a new processing method which should avoid most of the troubles we have encountered.

Lincoln Labs continues to make progress on their CCD development effort, although they too have been slowed a bit by equipment problems and a switch in the processing line from 100mm diameter silicon wafers to 150mm diameter silicon wafers. We should learn more about the Lincoln Labs results at the ESO conference on CCDs to be held October 8-10 in Garching, Germany.

We plan to mount the Lick thinned CCDs in the same style of mechanical package used by Lincoln Labs. Gerry Luppino has been producing the Lincoln packages, and we have obtained the package drawings from him. Fabrication and testing of the Lick packages will begin next quarter.

2.4 Software

Two areas comprised the major focus of this quarterÆs DEIMOS software effort: code management and portability, and DEIMOS keyword/process flow management. With respect to the former, Clarke acquired and tested the CVS code management system. Clarke also collaborated with Allen and Tucker on a revision of the Keck Tasking Library (KTL) and Lick instrument software source trees, intended to simplify future multi-site development, maintenance, and installation. As a proof of concept, existing Keck-II instrument software (specifically, the HIRES instrument and CCD control software and the Keck Tasking Library) were re-organized into the model we intend to use for DEIMOS. This model maintains the source code under CVS (as was recommended by the PDR review board) and provides the ability to generate executable code for a variety of hardware architectures (e.g., Sun/SPARC, Digital/Alpha) and operating systems (e.g., SunOS, Solaris, Digital Unix) from a common set of source code and makefiles. In the second area, Clarke and Allen have developed database schema and applications for the automatic generation of source code from the central database of DEIMOS keywords, software agent specifications, and other instrument information. The DEIMOS developments in both of these areas will be presented at the Keck II Software Coordinating Committee (SCC) meeting to be held in Santa Cruz on November 19.

Allen has developed an ideal mathematic model of DEIMOS which defines the required characteristics of the mapping between pixels and spectra, as well as a method for representing this model in terms of FITS tables and related header information. This model describes where the light from each wavelength in each slitlet will fall, and will allow DEIMOS software to convert directly from pixel coordinates on a DEIMOS spectral image into the corresponding object and wavelength information. The FITS tables and world coordinate system (WCS) issues involved in representing such a model for mosaic-based instruments was the focus of the paper that Allen presented at ADASS and will likely impact the development of standards for the archival storage of images from such instruments.

Clarke attended the TCL developerÆs conference in July and brought back several promising public-domain TCL-based packages which will likely be used in the development of DEIMOS user interfaces. These are now being evaluated.

Work continued on assembling tools for prototyping graphical user interfaces for DEIMOS. Clarke conducted further tests of LuptonÆs ktcl (a version of TCL which provides access to the Keck Tasking Library), including the use of the monitor functions.

In response to recommendations from the software PDR review board, in early July a questionnaire to assess observer preferences regarding the use of IRAF or IDL for DEIMOS quick-look data reduction was assembled and distributed to approximately 150 prospective DEIMOS users at UC, CIT, UH, and NASA. Responses were received from roughly 30 persons and tabulated in August. Somewhat more than half the respondents favored use of IRAF, and less than half favored IDL. There was no consensus that we should switch our present development plan from IRAF to IDL.

An internal review of the DEIMOS database schema, keyword database, and software design tools was held at the end of July. A schema was presented for describing keywords, agents, tables, and headers. Software design tools were presented that included applications to generate documentation, source code, and sample headers from the keyword database and to generate code and dataflow diagrams from the schema. Clarke presented several proof-of-concept demonstrations to validate the functionality of these tools, and there was a general consensus that the schema designs presented were sound. The designs presented at this review were more formally developed and presented in ClarkeÆs paper at ADASS. The material from this paper (and AllenÆs) has been posted to the Web.

Work continued on collecting and investigating alternative image display tools, including an image display server from LLNL and an extremely promising graphical display environment (PAD) developed by an academic consortium. At the ADASS conference in September, Doug Tody of NOAO strongly encouraged Steve Allen and De Clarke to consider the use of the NOAO mosaic image display server (currently under development) for the DEIMOS project. Apparently several other institutions that are developing instruments containing CCD mosaics have already agreed to adopt the NOAO display server. The DEIMOS project is not yet prepared to make such a commitment, but we have invited Tody to Santa Cruz for further discussions. TodyÆs visit is tentatively planned for the first week in December.

The second-generation SDSU/Leach CCD controller boards that we were expecting to receive by mid-September have not yet arrived, and are now not expected until early December. The DEIMOS test dewar was also not completed during this quarter, so testing of multi-chip readout with first-generation SDSU boards was not possible. In case the delivery of the second-generation SDSU boards slips beyond December, we are considering other options, such as the CCD controller system developed by Chris Stubbs of the University of Washington as a fall-back. A visit to the University of Washington to see the Stubbs controller is currently being arranged by Stover.

An IBM laptop computer (running Linux) was purchased with partial funding from DEIMOS and used extensively during the ADASS conference. We plan to use it as yet another development platform to insure portability of DEIMOS code and to evaluate its use for local control of DEIMOS on the Nasmyth platform during instrument commissioning.

As expected, the availability of Kibrick and Tucker to the DEIMOS software effort was severely constrained by the installation and commissioning of the HIRES image rotator, which proved to be considerably more difficult than planned. However, lessons learned during that commissioning should simplify our software development and testing efforts for the DEIMOS instrument rotation task. Nonetheless, the prolonged commissioning effort has delayed TuckerÆs start on the DEIMOS motor control software.

As a result of those delays and delays on the CCD controller front, and the possible impacts of adopting the NOAO mosaic image display server on our development plans, we have decided to defer our software CDR (originally scheduled for the end of the year) until March 1997.

2.5 Electronics

The new CCD pre-amplifier has been designed and will be prototyped in November. It reduces the noise on the CCD input by a factor of 10, increases the output drive current, and can operate up to a 10MHz clock rate - plenty of margin for DEIMOSÆ 500KHz CCD clock rate requirement.

The electronics lab has received many DEIMOSÆ electronics components that have been ordered over the last three or four months. At this point nearly all of what will be needed to complete the electronics for DEIMOS is in house. Many boards have been fabricated, built, and tested and await installation on the various stages of the instrument. A modest amount of work has been done on other projects that should be directly applicable to DEIMOS, such as the redesign of the Power Monitor Card for the CCD controller chassis.

A small amount of work has been done updating the schematic drawings of the instrumentÆs electronic architecture.

We are currently preparing a form to be filled out for each motion stage of DEIMOS (and all other instruments) that are currently being designed. This form will record the specifics of each motion, including the encoding scheme, active travel, resolution, accuracy, etc. The forms are completed at each stage of the development of the system so that we record not only the performance we expected at the time of design but also actual values achieved when the system was assembled and integrated. Among other things this should help ensure that all the relevant values are passed on to the programmers and reduce surprises. This information will also be valuable as part of the instrument documentation for trouble shooting .

No progress was made on flexure compensation during this quarter.

All PI activity this quarter is included under the preceding sections.

To the end of September 1996, $1,797,930 had been expended on the project, including $910,956 on labor and $886,974 on materials and supplies. Table 1 summarizes the budget and expenditures, with details shown in Table 2.

During the quarter $136,764 was expended on labor, the major areas being mechanical design, optical fabrication and software. $114,052 was expended on materials and supplies. The major expenditures were for the mechanical structure and drive system ($60,000), the grating handling system ($14,000), the motor control system ($14,000), and the slitmask system ($5,500).

Figure 1 shows a summary of the project schedule. We are still anticipating commissioning the instrument in the first quarter of 1998. The critical path is shown as Figure 2. Fabrication of the camera elements remains on the critical path. However David Hilyard is making good progress against the estimated fabrication times, and the problems that have occurred with the fabrication of the CaF₂ element 5 are not causing a change to the anticipated commissioning date. Should they be repeated with the re-fabrication of element 5 or the other CaF₂ elements, there will be a schedule impact.

Mechanical engineering and the detector system are also close to the critical path, having four weeks or less of float. At this time no major problems are slowing down the engineering. The design of the dewar system has be re-started and is on schedule.

Figure 3 shows the activities planned for the next year in more detail.

The following is a list of milestones for the quarter from the last Quarterly Report, together with the progress made on them:

1. The DEIMOS cylindrical structure and substructure are completed but awaiting inspection. Fabrication of the drive disk is in progress and is scheduled to be completed by the end of November 1996.

2. Design of the slit mask handling system has begun again and is continuing.

3. The spherical surfaces of elements 7 and 8 were completed, and work started on the aspheric surfaces and is expected to be complete by December 1996.

4. Due to the problems with the CaF₂ element 5, work on element 2 was delayed. Work on element 2 is currently scheduled to begin late in the current quarter.

5. Fabrication of the test dewar began but is not expected to be complete until late in the current quarter. No 2Kx4K chips are likely to be available much before the end of the current quarter.

6. The report on optical tolerance analysis was received from ORA and seems very complete. Review is still in progress.

7. The optical design of the camera was entered into Zemax by the engineering group, which allowed several design and fabrication questions to be explored.

8. The optical quality of the collimator was reviewed and accepted.

9. Terry Mast is heading the effort to produce an alignment plan and is making good progress.

10. Considerable progress was made on the grating mount prototype, but testing has not finished.

11. The design of the grating slide mechanism is nearing completion.

12. Thermal testing of the RTV/fluid coupled test cell was started and will continue into the next quarter.

13. Progress was made on the TV guider design, but it is not yet final. We expect to finalize it and purchase the Photometrics TV next quarter.

14. A response to the Software PDR is being prepared but is not yet complete.

15. Other commitments have prevented the preparation of a revised Software budget. A list of priorities has been prepared.

16. Bob Leach has not completed the analog board but expects to before the end of the year.

17. Internal review of global data design plan was successfully completed at the end of July.

18. Instrument GUI prototyping is continuing, incorporating new tools obtained at the TCL developer's conference in July.

19. Simulated mosaic images have not yet been constructed pending the installation of more memory in our development machine, although simulated FITS headers have been automatically generated. A memory upgrade is in progress.

20. The start of motor control software development was delayed by the protracted commissioning of the HIRES image rotator. TuckerÆs portion of that commissioning effort has now been completed.

21. The advantages of IDL vs. IRAF were evaluated, and it was decided in the absence of convincing evidence to the contrary to stay with our current choice of IRAF.

Milestones for the next quarter:

1. Receive the DEIMOS structure from L&F Industries and install it in the Instrument Lab.

2. Complete the design of the slit mask handling system.

3. Complete the design and testing of the grating tilt drive and handling systems.

4. Hold a ìmini-review" of the slit mask and grating systems prior to the start of fabrication.

5. Purchase the tent mirror and 600 and 800 line gratings.

6. Complete the conceptual design of the dewar system including the focus and FC systems that may be contained within the dewar.

7. Carry out metrology on the completed collimator mirror.

8. Complete the initial alignment plan.

9. Complete optical fabrication of elements 7 and 8 and begin work on CaF₂ element 2.

10. Complete stage description sheets for each of the DEIMOS motions requiring software control.

11. Complete the CCD test dewar and install a 2Kx4K test CCD.

12. Decide on the choice of CCD controller.

13. Finalize the TV guider design and order the TV camera system.

14. Complete the Software budget revision.

15. Begin a fabrication error budget for the camera.

16. Present DEIMOS proposals for source code management, directory structures and makefiles at Keck II SCC meeting on November 19, in pursuit of acceptance/adoption by CARA.

17. Confer with Doug Tody in early December regarding NOAO mosaic image display server and decide if DEIMOS should commit to its use. As a fallback, continue investigating other image display servers.

18. Begin motor-control software.

19. Begin tests of second-generation SDSU/Leach CCD controller as soon as it arrives.

20. Visit the University of Washington to evaluate capabilities and software requirements for StubbsÆ CCD controller to determine if it is a viable fall-back to the SDSU CCD controller.

21. Complete memory upgrade of development machine and begin generating and testing the manipulation of full-sized DEIMOS images.

22. Begin building instrument hardware simulators sufficient to support continued GUI prototyping and testing.