DEIMOS Quarterly Report
Number 12
March 31, 1997 - June 30, 1997
1. General Items
The CCD Detector Review was held on May 20, 1997, with Dan Fabricant as Chair and Mike Lesser, Gerald Luppino, Phillip MacQueen and Richard Reed on the committee. With assistance from Gerry Luppino, Richard Stover was able to present a very impressive survey of candidate CCD characteristics. It was a good review with a lot of input and opinions from the committee and attendees. A copy of the committee report is attached. Having listened to the committee and thought about our needs and resources, we have decided to participate with the University of Hawaii (UH) and Lincoln Labs on Phase 3 of their CCD development program. A purchase order has been placed by CARA for $316,000. This will likely give us a yield of 16 scientific grade CCD20s, which we would expect to receive at the earliest in January 1998.
We have ordered a complete set of Leach-2 boards from SDSU for both the scientific and FC controllers.
We are in the process of asking SSC permission to build an interim mosaic of detectors using as many of the Phase 1 Lincoln devices or other devices as can be made available to us. We would complete the mosaic with thick Orbit/Stover devices. The strategy would be to have this mosaic together and ready for testing in a dewar by the end of November 1997. The mosaic of science grade devices might be available as early as March 1998, but building an interim mosaic would have the advantage that delay of those devices beyond that date would not hold the project hostage.
The Software CDR was held on June 16, 1997, with Steve Kent as Chair and John Cromer, Jill Knapp, George Jacoby and Al Conrad on the committee. This review also went well. A copy is attached. Highlights are discussed under software.
One major project issue that was uncovered as a result of preparing for the Software Review is that the software for the instrument will likely not be completed before October 1998, making it one of the schedule driving activities. We are taking steps to hire extra personnel and make other adjustments in software.
We have not received the CaF2 for Element 3 from Optovac, nor have we received the replacement for Element 5, which has a fracture. We have been in frequent contact with Optovac and Sandy has visited them personally. Optovac has successfully produced two boules for this element, but both were destroyed in the process of generating them to size for Element 3. Since producing the last successful boule for this element, they determined that their furnaces (in which the boules were produced) were aging and unreliable. They have since replaced their outdated furnaces with three new ones, two of which are currently making boules for Element 3. These boules are due in mid-October and mid-November. It will take approximately an extra month from that time to generate an element and deliver it to Santa Cruz. Given that timetable, the earliest we could expect to have a finished CaF2 lens would be mid-January 1998.
We have not yet determined a schedule for the replacement of Element 5. We did optically figure the fractured CaF2 element we have with the fracture, and it has the correct figure on both surfaces. The fluid couplant wicks into the fracture, which makes it optically of little significance. Assuming the fracture does not grow and destroy the piece, the current element could be used until we replace it with an unfractured one.
We are exploring the possibility of using a different material if Optovac is unable to deliver the CaF2 Element 3. A glass produced by Ohara appears to be a close match to CaF2. We are investigating the effect to the camera performance and the extent to which using this material would change the design of the other elements. We have asked Ohara for a quote on this glass to determine the cost and production schedule.
Element 2 is finished and is the largest optically fabricated piece of CaF2 in the world. George Laopodis, our new optical technician, also completed the field flattener.
We are currently planning a front window in DEIMOS (which was omitted in one of the earlier concepts of the instrument). This window will likely be BK7; David Hilyard is currently getting a price quote for the glass.
We have decided to hire a consultant to help with the design of the camera barrel. The reason driving this decision is that Eric James is needed to work on the ESI Project, due to the departure of their mechanical engineer. Alan Schier (who designed the ESI camera barrel) is available. Eric has completed much of the barrel and lens mount details, and Alan is expected to continue using as much of this design as is possible. The change of engineers offers the opportunity to have a second very qualified engineer review in detail what has been done and perfect the design where appropriate.
The schedule will be delayed slightly, as Alan will not be able to start the project until mid-August. This, coupled with Sandy’s schedule, puts the earliest possible date for the Camera Review in mid-October, and the completion of the lens cells and body in February 1998. This is barely adequate to support our goal of a fully assembled spectrograph in March 1998 (see below).
The extra cost to the project of having Alan Schier work on the design is estimated to be $30,000 has been subtracted from the contingency fund, which stands at approximately $300K (before this commitment).
The structure is now rotating under instrument control. Bob Kibrick is developing a series of tests that Jim Burrous will carry out to verify that we are within tolerance.
The grating and slitmask systems are currently being fabricated. The slitmask system will likely be complete in September and ready for testing, and the grating system in October. The TV system has progressed to the point that we have been able to see an image of the focal plane with the TV lens in place. We have received the tent mirror, and the collimator has been aluminized. We are currently planning to mount these optical components in DEIMOS in October. We should then be able to start initial alignment of a subset of the optical components. The dewar system is in the final design stage with the goal of having a completed mechanical system ready for testing in October 1997.
We undertook a thorough schedule review this quarter. If the Lick shops were fully available to DEIMOS, we estimate that the major pieces – mechanical, camera, detector/dewar, and basic software – could be brought together and installed in the spectrograph in March 1998. Allowing for a further six months of testing puts the delivery date to Mauna Kea in October 1998. This is roughly 6 months beyond the delivery date we have been showing of April 1998. However, since this review it has become clear that there will be a significant further collision between ESI and DEIMOS in the Lick shops during October 1997 to March 1998. Unless significant new resources are made available, we anticipate a substantial further delay in DEIMOS beyond October 1998. A study of this impact and strategies to mitigate it will be conducted in early September.
2. Reports on Specific Areas
2.1 Optics
The calcium fluoride Element 2, the LAK12 Element 9, and the fused silica dewar window were generated, edged, ground, polished and figured to the design specifications. Element 1 was ground, polished and figured on the spherical concave side, leaving the convex side fine ground to the optimum initial sphere ready for aspheric plunge grinding. Element 6 was generated, edged, ground, polished and figured to specification on the convex side, leaving the concave side ground nominally to the design radius and the lens overthick. This allows room to change the radius of curvature or to increase the thickness by 1.5mm if the as-built errors require accommodation.
Element 4 was generated and is presently on hold. An error was made in generating the steep concave side, resulting in a reduced clear aperture from the design. We are presently weighing the options on the best course of action.
We are working on a new approach to plunging the asphere. This is being tested on a scrap piece, and the results are promising. More tests are on order prior to using the method on the real Element 1 aspheric surface. The objective is to obtain better control over the plunge grinding process and better accuracy on centering the aspheric surface on the lens.
The Tent Mirror was received from Zygo and meets specifications.
The three first light gratings have been ordered and received. They are 6" x 8" replicas from Spectronics (AKA Milton Roy) on Zerodur substrates. Rulings are 1200, 900 and 600 lines/mm. The flat mirror was ordered and has been received. It has not yet been aluminized. Note that we no longer plan an 8" x 12" grating for first light, although space for one is being retained. The 8" x 12" 1200-line grating by Spectronics has low reflectivity and costs five times as much as their 6" x 8" gratings. It actually puts through less light.
We set in motion a process to survey the Science Advisory Team on their preference for interference-defined broadband filters and to set thickness and flatness specifications. It is probable that the thicker filters will impact the optical design, which has to be cleared with the designer at ORA.
2.2 Mechanical Design
Structure
The electronics enclosure ring surrounding the cylinder has been designed and built but not yet installed. The rotational limit system has been partially fabricated and installed. The mounting structure for the TV system is built and installed. The drive motor had some surging motion due to incorrect alignment, but this problem has been corrected with some temporary shims. A permanent solution is currently being designed. The position angle fiducial is installed and has been tested and determined to be adequate. The Gurley encoder has been installed. The rotational drive system has been tested and has shown more slippage than specified. We are exploring installing an on-axis Renishaw encoder on the rear bearing.
Grating Cells and Mounts
The 8" by 12" grating cell was designed, built, and installed. It awaits flexure testing. The 6" by 8" cells and the flat mirror cell will be designed next. These cells will all be manually removable.
The disappointing performance of the Spectronics 8" x 12" 1200-line grating caused us to downsize one 8" x 12" slot on the slide (it can still take a 6" x 8" grating). This has substantially reduced the packaging problem for the four positions in the slide. The three encoded positions can now all have Gurley on-axis encoders, and the gratings no longer collide with the 13-position slitmask server.
Fabrication has begun on the final grating support structure. Side B parts have been modified to support lead counterweights to keep DEIMOS balanced. The grating slide consists of a slide with four moving elements, three of which have a drive motor sub-system. Finally, there is a clamping system to precisely hold each slider once it has been driven to the "in-use" position. Previous tests have been done on this clamping system and the slide itself. The next quarter’s design effort is aimed at finishing up the remote operation of the clamping devices and putting the structure on the rotating DEIMOS.
Slitmask System
The slitmask system drawings are complete. The cassette actuator is built, and the brackets to hold it have been built and installed. A 13-mask cassette has been redesigned to take advantage of the extra room that became available due to changes in the grating slide design. This is an increase of three masks.
Collimator
The lifting/handling fixture for the collimator has been designed, built and tested.
Camera
The camera cell design was handed off to Alan Schier. The plan now includes an axial thermal compensator that passively adjusts the axial position of the second doublet. This adjustment maintains constant plate scale under temperature change. Tests to verify the theoretical CTE of the elastomeric lens mounts were begun.
2.3 Detectors
Lincoln CCDs
As described in Section 1, the CCD Detector Review was held on May 20, 1997. The first thinned Lincoln device was tested at Lick the morning of the review. It was a high resistivity CCD thinned to 40 m . As expected, the red quantum efficiency was considerably higher than a standard epi CCD, which is only 20 m thick. The extra thickness of the high resistivity CCD provides greater absorption path length for photons whose wavelength is greater than about 750 nm. This first thinned Lincoln CCD shows a "brick wall" pattern in its response at short wavelengths. This arises from incomplete laser annealing of the back surface of the CCD. The magnitude of the response variations is very wavelength dependent, dropping from 12% at 400 nm to near invisibility at 700nm. More details on this device can be found on the web at http://gardiner.ucolick.org/~ccdev/lincoln/lincoln.html.
Lincoln expected to be able to ship approximately two packaged, thinned devices per week, but since the review Lincoln has been trying to solve some CCD mounting problems and we have seen no further CCDs. Lincoln is also reported to be working on solutions for the laser anneal non-uniformity that produces the brick wall pattern. From learning more about the laser process, we feel the brick wall problem may be difficult to solve and we are thinking about workarounds.
Lick/Orbit CCDs
We have identified enough thick Orbit frontside 2Kx4K CCDs to construct an interim mosaic using these CCDs and whatever devices we acquire from the Lincoln Phase 1 effort. Aluminum nitride substrates and other parts needed to package these devices have been ordered.
We will continue with our thinning effort on the Orbit CCDs, and if we obtain good working 2Kx4K Orbit CCDs these can be included in the interim mosaic as well.
Dewar System
Following the Detector Review, the design of the dewar has moved into the detailed stage. At this point we are nearing the completion of the design of the LN2 can and have a detailed design of the thermal connection between the LN2 can and the detector vessel. We have revised the design of the clips that hold the detectors to the backplane and have designed a thermal connection. The thermal connection and the new clips have been cooled and partially tested in the test dewar and so far to work well.
Testing of the focus drive system continued, and the results indicate that the device will work as designed.
The goal of completing the mechanical parts of the dewar system by mid-October still seems possible, barring interference from other projects.
2.4 Software
The major software emphasis this quarter was preparation for the DEIMOS CCD review held in May and the DEIMOS software CDR held in June. In support of those efforts, this quarter's DEIMOS software effort focussed on five areas:
We consider these in turn.
Initial CCD Controller Hardware & Software Testing
The last of three new 3U-size boards for the second generation SDSU CCD controller was received at the end of March. To support testing of these boards, existing diagnostic software used for the first generation SDSU hardware was upgraded to support the second-generation hardware. The second-generation timing board, clock generation board, and video processing boards were successfully tested and accepted, allowing completion of the DEIMOS CCD controller hardware design in time for the CCD review in May. Richard Stover and Kirk Gilmore of the Detector Laboratory contributed significantly to this effort.
We are still awaiting receipt of the second-generation VME interface board, but this is not expected to be delivered until the end of August. Its delay has no schedule impact as we can use a spare.
Software Development, Design, and Documentation Tools
Significant effort went into refining various software design and documentation tools used to complete the software design presented at the CDR. The "etcha" tool for software system design was significantly revised in response to CDR needs. Drawings produced with this tool and related software were used in the CDR. A few improvements remain to be made; the first production version should be released in early September.
The "documentation generator" suite of tools was used to produce both the Keyword Dictionary and the Database Dictionary, which accompanied the main CDR document. These lengthy reference texts were entirely machine-generated. The documentation generator suite is now in production and will be used for ESI, PFCAM, and DEIMOS. The generators can be run online (dynamically) via Web interfaces, as well as producing static (paper) documentation.
Figdisp Enhancements
In late April, Cromer at CIT 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. Steve Allen has subsequently ported John Cromer's Solaris version to several other architectures/systems, including Intel/Linux and DEC Alpha/Digital Unix. In addition, Allen has identified and corrected several areas of inefficiency, resulting in improved performance for large images.
Entry of Keywords Into Database
Tucker has completed entry of all of the currently known DEIMOS motion control keywords into the keyword database, thus making these keywords available for GUI prototyping. This database was used to automatically generate the DEIMOS keyword document that was distributed at the software CDR.
Document Completion - Software CDR
The DEIMOS CDR document and the DEIMOS keyword and database document (Lick Observatory Technical Report #80) were completed and distributed at the software CDR. On line copies of these documents are available at http://www.ucolick.org/~deimos/swcdrtc.html. On line demos of the keyword and database documentation generators are available via http://www.ucolick.org/cgi-bin/Tcl/memes.cgi.
Motion Control Hardware/Software Testing
Testing of the replacement ETS8/P Lantronix terminal servers was completed in April and a report issued at that time. That report detailed problems encountered with the existing ETS8/UF units and demonstrated that these had been resolved by the newer ETS8/P model. Negotiations with Lantronix resulted in a favorable exchange agreement, and the DEIMOS ETS8/UF terminal servers were exchanged for the newer ETS8/P model in June.
The initial release of the Galil DMC-1500 firmware along with the initial release of the Keck-II style keyword library that interfaces to the Galil firmware was successfully demonstrated using the Prime Focus Camera (PFCAM) instrument during the DEIMOS software CDR in June. The PFCAM instrument was operated both using the manual hand-paddle and via keywords, providing interoperability between the two modes. The communication between the Galil controller and the instrument computer used for this demo was via an ETS8/P terminal server.
Software Operational Procedures - Milling Slitmasks
Work commenced on testing some of the DEIMOS slit mask-milling procedures in terms of using the new CNC mill to generate LRIS slit masks. We have started to explore the feasibility of translating slit mask designs into CNC mill instructions on a Unix host in order to allow us to embed identification information into the file of mill instructions. This work will continue into the next quarter.
Data Reduction Software/Calibration Plan
The coordinate systems for all stages of the instrument and the detector system were fully defined, including the mosaic. Key coordinate transformations for astrometry and wavelength calibration were specified. A start was made on a plan to automatically calibrate DEIMOS in both direct imaging and spectroscopic modes, and the time and database requirements were estimated. This work will continue in the winter 1998 quarter.
2.5 Electronics
Both the new pre-amplifier design and the CCD signal electronics were accepted at the Detector review. Updated noise figures for the new pre-amplifier, stemming from comments made at the review, were incorporated into a final overall noise analysis spreadsheet for the entire DEIMOS CCD signal chain. Following the review, work has now begun on the design of the printed circuit CCD interconnect boards and their associated circuitry and cabling. These are needed to implement the CCD mosaic connection to the new SDSU-II CCD controller and are being designed flexibly to accommodate both MITLL and Orbit CCDs in a variety of amplifier readout configurations. All electronics and printed circuit boards for this arrangement are expected to be completed by next (fall) quarter.
The instrument rotation drive system was investigated and characterized using the Galil drive system and the Gurley Precision Inc. high resolution encoder. After some mechanical rework, the drive system is now in good working order. The encoder system however seems to be less then ideal due to possible slippage in the friction drive system encoder. We will be adding an extra Renishaw on-axis encoder to the rear bearing, which should cure this problem. The impact on the electronics is minimal.
Work started on assembling the final CCD controllers and chassis. We are starting first with the flexure compensation controller, which is needed early for lab testing. Power supplies and chassis frames were assembled. Work started on the first CCD interconnect boards (see above).
The design for the 24-channel analog input board was completed and the board sent out for manufacture.
2.6 Flexure Compensation
System design work was largely completed on the Flexure Compensation System, and portions were written up for the Software CDR. The major features of the system were described in earlier reports. We decided to add a third, Hg lamp to the existing Ne and Ar lamps to make sure that all blue 1200-line grating tilts will have an available spectral line on the FC detector. We also decided to use the FC signal chain and associated electronics to lab test CCDs before the main mosaic system is ready. Fabrication of this system has started. An in-house review of the FC system is planned for February 1998.
2.7 Alignment Plan
Work on the alignment plan was postponed during this quarter in order to devote resources to optics fabrication, camera design, and detector mosaic design and assembly.
The goals for the alignment plan in the coming quarter include the design of a camera alignment and test facility. This facility will be located in the metrology tunnel of the optics laboratory and will be used for the ESI and DEIMOS cameras. We would like this facility to provide for the use of auto-collimating telescopes, a phase-measuring interferometer, a Hartmann test, and the visual inspection of images.
Additional goals for the coming quarter include continued detailing of the spectrograph alignment plan and initial alignment of the collimator mirror in the shell.
A camera and associated coupling optics were ordered for the auto-collimating telescopes.
2.8 Assembly and Test
Installation of the collimator was postponed this quarter. In the coming quarter 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.
3. Report from the PI’s
The PI of DEIMOS has officially been transferred to Sandra Faber from Garth Illingworth, who is now a Co-I.
The PI would like to take this opportunity here to state her opinion that the schedule conflict and resultant delay between ESI and DEIMOS is likely to be serious and deserves careful study to minimize the impact on both instruments. Such a study is planned. Schedules presented in this report presume no impact and are therefore likely to be optimistic.
4. Budget
As of the end of the quarter we had expended $2,968,000 against project funds, or about 59% of the budget. Of the amount expended, $1,579,400 was for labor and $1,388,600 was for materials and supplies, including travel. During the quarter we spent $284,600 on labor and $196,000 on materials and supplies as shown in Table 1. The details of the budget are shown on Table 2.
Major expenditures during the quarter were for the CCD controllers ($124,000), mechanical hardware including the electronics enclosure, parts for the dewar and grating systems ($38,000), expenses for reviews ($18,000), and software ($6,000).
The cumulative expenses are graphed as Tables 3 and 4.
We have now started to receive funds from CARA. In our request to CARA we asked that they retain $327,000 so that if we decided to participate in the Phase 3 of the CCD development program with University of Hawaii and MITLL (which we have subsequently decided to do), the contract could be written directly by CARA. This follows the same path as previous development contracts were written to this consortium by CARA.
5. Schedule
Figure 1 shows a summary of the current schedule. The major change in the schedule this quarter is the delay of the projected delivery of the instrument to Hawaii to October 1998. As discussed previously, this is principally due to the detector development schedule, the late delivery of the CaF2 Element 3, and the software schedule.
The critical path of the project is shown on Figure 2. The CCD detectors remain on the critical path, particularly development of the mosaic. This path could be changed if we decided that we wanted to deliver the instrument with the interim mosaic. However, in reality with the delay in camera development and the software schedule, we would not be able to deliver the instrument to Hawaii earlier than the 4th quarter of 1998. Although not strictly on the critical path, completion of the grating design has gone slower than anticipated and used up its entire schedule float.
6. Milestones
The following is a list of milestones for Quarter 12, together with the progress made on them:
Milestones for the next quarter: