Fabrication of Elements 5, 7 & 8 have been completed, although some testing of Element 5 and 8 still continues. David Hilyard has started on Elements 1 and 2. He anticipates completing Element 2, our second CaF₂ about the end of May. Element 1 is the strong aspheric at the front of the camera, and is the last asphere needed. David anticipates completing this Element about mid summer. A new person is now working in the optical lab, George Laopodis. George will likely start work on Element 9 about the end of May.

The ORA optical report indicated a strong temperature dependence of the plate scale. The image will move 0.6 px at the edge of the detector per ° C. ORA is studying passive compensation schemes for the camera barrel.

The structure was delivered from L&F Industries in February. It has initially been aligned on its rollers and is currently having the drive system installed. Some modifications were needed when it was discovered some gussets interfered with the sub structure, and to make allowance for the filter wheel. These minor modifications have been made, and work continues to install studs and brackets etc. The next major addition to the structure will be the nose portion, which is scheduled to be installed about the middle of the quarter.

The slitmask system is currently being fabricated, with plans for completion by the end of the quarter.

Work continues on the grating system, all the basic parts of the system being prototyped. By the end of the quarter we plan to complete the tests and start fabrication of the remaining components.

Jack Osborne continues with the design of the TV system, and at the end of the quarter, Brian Sutin issued a report on the optical design. The TV camera has been received and will be tested during this current quarter.

The end to end conceptual design of the dewar system has been completed and is currently being put into report form for the Detector Review planned for May 20. The spectrograph focus and the cross dispersion direction of the flexure control system are planned to be contained within the vacuum of the dewar system. These systems have been prototyped and tests proved successful. The Carbon-Carbon back plane material for mounting the detector mosaic has been received.

Two thick Lincoln chips have been received, which were packaged primarily to test the packaging concepts and handling procedures. These chips were not made in Lincoln's regular process, as they were part of an experimental procedure that Lincoln was conducting. We did run the devices successfully and were able to verify that we were able to read data from them. However the tests we did indicated that the performance of these particular devices was rather disappointing. Gerry Luppino reports testing a similar device that was processed in the normal way for thinned chips and says the performance is spectacular.

In February, the DEIMOS supervisory computer (a Sun UltraSparc Model 140) was delivered and installed and is now in active use. In early March, we received a beta-test unit of the Lantronix ETS8/P terminal server and began tests.

On March 11, Steve Allen visited NOAO for meetings with Doug Tody to assess progress on the CCD mosaic image display software currently under development there. The NOAO effort is so far proceeding on schedule.

The initial release of the Galil servo motor controller firmware (shared by DEIMOS, ESI, and PFCAM) was successfully tested during the recent Prime Focus Camera (PFCAM) commissioning run at Mt. Hamilton. This included successful testing of the initial release of the dashboard-2 graphical user interface, as well as on-telescope testing of the beta-test Lantronix terminal server.

The software operational procedures for milling slitmasks and the data flows associated with those procedures have been finalized. The scheme adopted requires only the single barcode label that is attached to the raw slitmask stock. Under this scheme, the files containing the slitmask milling instructions will be individually down-loaded to the mill directly from the Unix host computer.

The first release of the software design documentation tool is now operational and will be used in preparation of documentation for the Software CDR. The DEIMOS Software team is preparing for the Software Review now scheduled for June 16th.

During this quarter, the aspheric surfaces on both Element 7 and Element 8 were finished, completing the first two lenses in the camera and two of the three longest lead time surfaces.

Further profilometer tests were done on Element 8, and it looks like the rear asphere is indeed decentered by 0.009 in. The tolerance for this error is only 0.001 in. The optical affects are being studied by ORA.

After much discussion, it was decided to attempt to finish processing the fractured calcium fluoride lens (Element 5). During grinding operations to bring the lens into thickness tolerance, the fracture propagated straight through the crystal grain boundary. It penetrated approximately 24mm further into the interior of the lens. (We had expected any propagation to follow along grain boundaries.) The surface was then polished and figured to specifications without further movement of the fracture. It remains stable at this writing.

After some experimentation with couplant candidates, it appears the best choice for filling the fracture gap (if we decide to use the fractured Element 5) is Dow Corning 200 fluid. The Cargille Laser Liquid is too highly viscous to wick entirely into the gap to make it optically transparent.

A second calcium fluoride lens, Element 2, was fine ground and is presently being polished on the first side.

The third and final aspheric in the camera, Element 1, was being fine ground at the end of the quarter, the spherical side being readied for polishing first before the asphere is plunged on the convex side.

A fused silica window was fabricated for dewar cell test purposes.

The ORA tolerance report indicates a large temperature -dependent effect on the plate scale. The size of the effect amounts to 0.6 px per ° C at the edge of the detector. About two thirds of the effect is due to the three positively powered CaF₂ lenses, whose refractive indices and radii are both highly temperature sensitive. One third of the effect is due to the remaining negative glass elements, whose refractive indices vary oppositely with temperature. A small component, of unknown sign, is due to shrinkage of the Al camera board.

The temperature opening goal for DEIMOS is ± 5° C, and the total error budget for image stability is ± 0.25 px. To meet these will require passive thermal compensation of the plate scale. ORA is studying ways to do this. The favored way at present is to increase the spacing between the triplet and the doublet at lower T, accompanied by a shift in focus. The final design parameters are under study.

Structure

The structure is nearing completion. We have aligned the mechanical components to within tolerances, and all modifications have been completed. We have installed the mounting studs for the camera and grating mounts and machined the mounting holes for the slitmask and TV systems. The drive disk wipers have been fabricated and installed. The remaining tasks are to install the drive motor assembly which is otherwise complete, install the limit system, the encoder, the band brake, the bulkheads and the position angle fiducial.

Slitmask System

The drawings for the slitmask system are complete except for the mounting hardware for the form. This is pending completion of the design for the grating mount to which we will attach the form brackets. Material has been received for the cassette actuator and insertion device bracketry, and fabrication has begun on the cassette actuator. We have received the prisms for the flexure compensation fiberheads and have a conceptual design.

Collimator

The mirror has been aluminized. We have fabricated the alignment clips for the mirror and completed designs for the other alignment fixtures. The cell shield is being fabricated. The cell has been complete for some time as has the counterweight. We need to make and install the slides that attach to the bulkhead in order to install the collimator. The largest task left to do is the design and fabrication of the various handling fixtures.

Grating System

The large 8x12 cell is 90% fabricated and a trial fit was done. Final flexure testing with the large cell, the large dummy grating and the cell mounting hardware will be done soon. Remote locks (also known as clamps) are being fabricated. The flat mirror has been ordered. (This is one of the four units in the grating slide.) The three small 6x8 gratings have been ordered from Spectronics (formerly Milton Roy).

Camera

A study of errors in the athermal design for holding the camera lens elements was carried out. For proper athermalization, the large difference in expansion coefficients of the Al ring and the glass (or CaF₂) must be accurately balanced by the CTE of the RTV gasket. The study showed that the expansion coefficient of the RTV must be known to ± 10% to avoid undesirable stress on the lens elements. Laboring tests were outlined to test the uniformity of the CTE of RTV when constrained between Al and glass in different layer geometry's. Parallel tests were defined to study the chemical action of optical coupling fluids on the RTV. These tests will be carried out next quarter.

TV Guiding System

The optical layout is finished and checked with the AutoCAD drafting software. The layout is composed of three sub-layouts: 1) the longslit mode, where the single long slitmask is inserted; 2) the multislit mode, where the multi is inserted; 3) the offset guide mode, where there is no mask in place and the TV camera "looks" at a guide mirror placed next to the mask support frame. The TV views this mirror at all times, not just in the direct imaging mode. The hardware is being fabricated to hold the fold mirror, the Canon lens, the filterwheel and the Photometrics TV Camera. These four items are all located about 65 inches out in front of the drive disk. This forms a natural package limit for DEIMOS. The inlet doors (also known as the "hatch") will be located here. The proposed test of the TV system will only check for proper alignment of the components and to verify that there is no shadowing of the TV-CCD. The final test to see that the exit pupil from the actual telescope fits through the lens without vignetting will be done during general spectrograph optical alignment.

2.3 Detectors

Lincoln CCDs

Two Lincoln Labs 2Kx4K CCDs have been tested. These were the first devices packaged by Lincoln. They are thick, front side CCDs and not the thinned, backside devices we hope to use in the instrument. In addition, these two devices were from an experimental lot for which some key aspects of the processing were different than the normal CCD processing. As a result, it isn't possible to make many conclusions regarding the thinned CCDs.

The first device was tested in March. A detailed description of the test results can be found on our web page at http://gardiner.ucolick.org/~ccdev/lincoln/lincoln.html. One output amplifier was very noisy, which made measurements difficult, and the serial register of the CCD had a charge trap near the other amplifier, making measurements difficult through that amplifier, too. Given these limitations we learned what we could. The results were:

• Serial CTE = 0.99998 (accuracy is limited by noise amplifier)
• Vertical CTE is good (no number measured)
• Single-pixel traps: 6 (measured on the half of the CCD read by the "A" amplifier, but still a small number for this sized device)
• A-side amplifier noise: 19e-
• B-side amplifier noise: 4e- (16 microsecond CDS integration of data and 16 microsec. on baseline)
• Dark current : 0.6e-/hour/pixel (measured from 1000s integration)
• Pixel full well: 116K e-
• Additional problems with residual image and spurious charge are detailed in the web document.

The connector epoxied to this device came off during testing. According to Gerry Luppino this was a known problem (weak glue) with this device and it should not happen on any of the thinned CCDs.

The second Lincoln device we tested had severe serial charge transfer problems that made accurate measurements of most parameters impossible. Readout noise appears to be about two electrons, but we can't accurately calibrate the gain, so this number is uncertain.

The web document also shows plots of the surface flatness for the two Lincoln CCDs as well as a charge transfer curve for the second device. Flatness is good to 25 m pv, a little worse than our specification but acceptable.

Gerry Luppino reports that he has a Lincoln 2Kx4K CCD which was fabricated with Lincoln's standard CCD processes and this device has very good performance. We hope to receive this device for testing at Lick before the end of April.

SITe CCDs

We tested a SITe thinned, backside illuminated 2Kx4K CCD. (This is the first one we've seen.) Some test results were:

• Read noise with 16 microsecond integration time on data and 16 microseconds on baseline (16+16) was about 2 electrons. With 1 microsecond integration's, (1+1) the read noise was about 8 electrons.
• QE peaks around 7000A at about 71% at -57C and at about 60% at -130C.
• Fringing was 7% at 8000A, 9% at 9000A, and 8% at 10000A.

We also tested a mechanical sample CCD from SITe which uses their new flat CCD packaging techniques. The device had about 20 m of surface variation. Morley Blouke of SITe says they are doing even better now. If they can make CCDs flat with the quality of the thinned one we tested, they are a potential source of devices for DEIMOS (or ESI).

Lick/Orbit CCDs

We are continuing to make progress on our CCD thinning process. The device thinning seems to be well in hand, but we've been dealing with issues related to the metal pads which are used to make electrical contact to the device. We've made some changes in our processing and are working on several wafers with functional 2Kx4K CCDs. We hope to be testing these devices by the last week of April.

Dewar System

Schematic sketches of the dewar system and its placement in DEIMOS are shown as Figures 1, 2 ,3 and 4. The dewar system is comprised of two main parts, a vacuum vessel containing the detector mosaic, focus mechanism and one axis of the flexure control system. The other component is a vacuum vessel containing the LN2 can. They will be joined by thermal feedthoughs that are connected with copper braid in either a dry nitrogen environment or low vacuum.

The end to end conceptual design is complete, and the major components have had an initial round of testing. Testing of the drive mechanism for the linear feed throughs continues.

By the end of last quarter the principle activity was documentation of the design in preparation for writing the report needed for the May 20th review.

2.4 Software

This quarter's DEIMOS software effort focused on three areas:

• Preliminary end-to-end testing of prototype hardware and software for motion control.
• Finalizing the software operational procedures for milling slitmasks.
• Initial testing of hardware and software components for the CCD data acquisition subsystem.

In addition, design of the DEIMOS keywords and associated data base continues, and several major design issues have been resolved.

Since the CCD dewar/controller review (originally planned for early April) is now scheduled for May 20, the DEIMOS software CDR originally scheduled for April 29 is now scheduled for June 16, so that these two reviews will occur in the planned sequence. We will have the same review board for the Software CDR as we had for the Software PDR.

Motion Control Hardware/Software Testing

During tests conducted in late January, we discovered problems with the Lantronix ETS8/UF terminal servers which we use to connect the Galil servo motor controllers to the network. These problems were not detected during our initial testing of these devices last year since they only appear when multiple serial input channels to the terminal server are operated at very high rates. We confirmed this problem with Lantronix and were shipped a beta-test unit of their newer-model ETS8/P terminal server. We conducted extensive tests of the ETS8/P and have confirmed that it corrects these problems. (A detailed report on these tests was issued in early April.) We are now negotiating with Lantronix to exchange our ETS8/UF units for ETS8/P devices.

The initial release of the Galil DMC-1500 firmware was successfully tested during the Prime Focus Camera (PFCAM) commissioning at the end of March. This firmware provides operation of stages from the manual hand-paddle, as well as operation from a Keck-style keyword library. The initial release of a Keck-II style keyword library that drives the Galil DMC-1500 firmware was also successfully tested during the PFCAM commissioning, as was the initial release of the dashboard-2 user interface builder software. (An earlier release of the dashboard software was also installed at CARA in mid-January.) The PFCAM commissioning also successfully tested both the Galil DMC-1500 and Lantronix ETS8/P hardware under actual observing conditions.

More extensive testing of the dashboard-2 user interface builder software was also conducted by successfully using it to build an alternate graphical user interface for HIRES. This allowed us to exercise all of the features of this tool on a fully operational instrument of similar complexity to DEIMOS.

The DEIMOS supervisory computer (an UltraSparc Model 140) was received and installed and is now being used for DEIMOS software development and testing.

Operational Procedures/Milling Slitmasks

The software operational procedures for milling slitmasks and the data flows associated with those procedures have been finalized. In the process, various options were tested and rejected, including several schemes that involved adding a second barcode onto each mask as it was milled (either by milling the barcode into the mask or adding a second label that was printed as the mask was milled). Tests of milling barcodes into the mask proved unsuccessful on two accounts: 1) the milled barcodes could not be reliably read, and 2) the milling of the barcode took too long and put too much wear on the milling bit. The final scheme we have adopted requires only the single barcode label that is attached to the raw slitmask stock as it is placed into inventory. Under this scheme, the files containing the slitmask milling instructions will be individually down-loaded to the mill directly from a Unix host computer, and the mill will be operated in DNC mode.

Allen has reviewed the manual for the Symbol Technologies barcode scanner and has successfully operated this scanner directly from a serial port on a Sun workstation. This resolves our outstanding concerns about being able to read the slitmask barcodes from the same machine that is down-loading the slitmask milling instructions to the mill. We still need to verify that the barcode scanner can be operated via the Lantronix terminal server.

Components For The CCD Subsystem

The last of three new 3U-size boards for the second generation SDSU CCD Controller was received at the end of March, and testing of this hardware is now in progress with the goal of completing initial tests in time for the CCD review on May 20.

Initial tests were also completed on the Force 5CE CPU board that will be used in the CCD VME crate. Although the second-generation SDSU fiber interface VME board will not be available until this summer, initial development of the DEIMOS CCD VME crate can proceed using the first-generation fiber interface VME board, although this will limit our bandwidth during CCD readout.

A 100-Mbit/sec fast ethernet hub was acquired and installed in order to establish a private fast ethernet network in our CCD lab. This network will be used to test the link between the DEIMOS CCD VME crate and the instrument computer to confirm that it provides sufficient bandwidth for readout of the DEIMOS CCD mosaic. This 100-Mbit/sec private network now links the DEIMOS test-bed DEC Alpha computer (radec, a machine with 1GB of RAM) with the ESI Ultra-2 instrument computer and Ultra-1 supervisory computer and will be connected to a fast ethernet Sbus board on the Force 5CE.

Cromer at CIT has nearly completed porting of the Keck figdisp image display server to Solaris, using a Solaris machine loaned to CIT by the DEIMOS project. DEIMOS will use this version of figdisp for initial testing and development, until such time as the NOAO image display server becomes available. As noted earlier, Allen visited NOAO in March and determined that their mosaic image display server software development effort was proceeding on schedule, which calls for delivery of the image display server in late fall or winter.

2.5 Electronics

The three boards that reside within the dewar with the CCD were fabricated and assembled. The assembly required "microscope-aided" soldering of .005" pitch connectors that reside on one of the boards. Overall, the three board assembly looked very nice, minimizing the area consumed by associated CCD circuitry but allowing for easy "swapping" of CCDs within the dewar. One of the boards is mounted on the same aluminum nitrite as the CCD itself.

The overall height of the board above the surface of the CCD was found to be .090" to .100". To minimize this height even further, milling of the board will be tried next quarter. Once the CCD connection system has been finalized, several outside board-assembly houses will be contacted for quoting the assembly of these three boards, as they are rather time consuming to assemble in house. Testing of the board/aluminum nitrite assembly in vacuum will be done next quarter.

The new CCD preamplifier has been assembled and will go under extensive testing of input-referred noise and settling time next quarter. Preliminary tests show the amplifier to be operating at gains ranging from 5 to 45.

Progress has been made in the Electronics Shop: 1) All but one of the Galil amplifier panels have been wired up and tested. These units connect the amplifiers making them modular for easier maintenance. 2) The Galil controller for the DEIMOS cradle has been wired up and tested. With this done, we are now able to rotate the instrument. Some refinements are being added to the controller for interlocking the motion, hardware interfacing to other elements of the instrument, and the addition of ìPanicî buttons. Documentation in the form of schematics continues to be improved.

Operational modes and the signal chain for the Flexure Compensation System were defined. Operation will remain in the background, with as little attention from the observer as possible. The three modes are initialize (find lines in current position), guide (maintain current position), and reset (recapture previous setting). Two arc lamps, Ne and Ar, will provide sufficient spectral coverage to guide in all but very bluest settings of the 1200-line grating. This gap could be filled by adding a Hg lamp if desired. All other gratings and direct imaging mode are covered Ne/Ar. The exposure time is a constant 10 s in all modes, and the update time for guide moves is 15 s. The lamps will be on continuously, with exposure control via frame transfer on the CCD. Contamination by scattered light in the spectrograph should be minimal. Plans were also developed for storing FC keywords in the master DEIMOS database, and for testing the FC system before shipping.

Following delivery of the shell and carriage structures to Lick, we measured the runout of the drive disk central hole and the reference shoulder near the rear bearing. Both runouts were smaller than 1 mil, better than expected from the initial measurements made at L&F Industries. These measurements were made by manually rotating the shell, and they will be repeated this coming quarter as soon as the rotation is motor driven.

The small measured runout allows simplification of some alignment fixtures. The fixtures required for aligning the collimator are designed, and the components are in house. They will be fabricated and tested with the installed collimator in the coming weeks.

Initial tests with one of the alignment telescopes strongly suggested the need for adding a camera for video display and/or image processing in order to achieve the required ease and accuracy. An in-house ST4 camera will be coupled to the alignment telescope. The telescope-to-camera relay system is designed and will be fabricated in the coming weeks. Tests will be made over the distance required for aligning the full spectrograph.

Work continues on defining the details of the overall alignment plan.

Assembly & Test

In the coming weeks the collimator mirror will be installed in its cell, and the mirror cell assembly will be installed and aligned in the shell. We will begin to test and document the DEIMOS assembly procedures.

Preliminary designs and quotations were received for narrow-band mirrors from Barr Associates. These can be placed in the grating slots to allow narrow-band imaging with DEIMOS. Because they are located at the pupil, they service the whole FOV, but they have to be used in conjunction with a wider-band regular filter. Two grades are available: A very durable type with off-band reflectance of ~ 3%, and a less durable type with off-band reflectance of ~ 5%. Both have FWHM of ³ 150 A . The former is approximately $10,000 and the latter is about $6,000 in 7 x 8-inch sizes, including substrate. It was originally hoped that the bandpass could be turned over ~ 10% in wavelength by tilting the filter using the grating drive. Owing to the unique design of these filters, this proves impossible, unlike normal interference filters used in transmission. The positive side is that the bandpass is quite stable (to 50 A at 5000 A ) over the DEIMOS FOV.

Since tilting is of little advantage, these filters might work just as well in LRIS as in DEIMOS, where they could be put into the spare grating slots.

The contact at Barr is Sue Fox; her address and phone are 2 Lyberby Way, Westford, MA 01886 (508-692-7513).

As of the end of the quarter, we had expended $2,559,075 against project funds, or about 51% of the budget. Of the expended amount, $1,295,100 was for labor and $1,263,975 was for materials and supplies. During the quarter we spent $216,844 on labor and $159,000 for materials, as shown in Table 1. Details of the budget are shown in Table 2.

The major expenditures were $13,000 for a replacement CaF₂ element, $23,000 for the tent mirror, $15,000 for gratings, $12,000 for CCD controller cards, $30,000 for structural fabrication, $8,000 for dewar parts and equipment, $15,000 for grating system parts, $9,000 for a Sun-Ultra computer and software, $4,500 on travel including people traveling to L&F Industries and Lick for reviews, and $4,000 for optical fabrication equipment, including a replacement camera for the Zygo and a micropolishing machine.

Tables 3 and 4 graph these expenditures.

We have now completed spending the NSF grant funds and have requested funds from CARA.

Figure 5 shows a summary of the current schedule. The projected date of first star light remains approximately the same as last quarter, spring of 1998.

The critical path of the project is shown on Figure 6. This quarter marks a change in the critical path. Last quarter it was the fabrication of the camera elements and subsequent integration of the camera into DEIMOS. This quarter it is the detector system, including the dewar. A couple of factors have caused this change. One is that we now have two opticians in the optical lab, and their primary priority is to work on DEIMOS camera optics. Another is that, although we have tested thick chips, we have not received thinned ones for testing, which is delaying our decision on which CCDs will be used in DEIMOS. Also the dewar design is going a little slower than planned. Additional help has been hired to help with the dewar drawings, which should ensure we are able to keep the current schedule and not slip further.

Engineering in general remains close to the critical path, with an expected shift to the Instrument Lab towards the end of the quarter. We are intending to increase Instrument Lab manpower by one (temporary) person and to contract some fabrication to handle the expected level of work.

The following is a list of milestones for Quarter 11, together with the progress made on them: