Presented below are the TSU responses to the issues raised in the PDR's Technical Report compiled after the 17-18 September, 1998 visit.

The following are replies to specific points in the Gull report:

SCIENTIFIC PROGRAM:

The scientific program for the AST was covered in our presentation to the committee. Since the committee seemed to accept this proposed research program at our review of 17-18 September, we will merely summarize it here. There were a number of specific scientific goals for the telescope, without which the telescope would be of littlt interest to the members of CASS, as well as some general goals. We feel the general goals can be accomplished with a telescope/spectrograph that can satisfy the specific goals. Those specific goals are generally the following:

A) Map the lower chromospheres of cool supergiant stars by observing eclipses of zeta Aurigae binaries in the optical. This requires observations in the violet (3700-4000A ideally; 3800-4000A at least) to detect the eclipsed component and enough lines to make an analysis profitable, a stable spectrograph to reduce the inevitable uncertainties about the velocity scale, S/N>= 200 to detect weaker lines, and a resolution at least R=30,000 but better yet near R=100,000 to begin to resolve the velocity structure of the lines. The stars are all K supergiants near V=4 with fainter B-type components with exposure times near 20 minutes. They will be observed at definite, specified times determined from orbital phases. We will obtain perhaps 50 separate observations per eclipse over timespans of a few weeks, with of the order of five (to ten) observations per night maximum.

B) Determine orbits of a wide range of spectroscopic binaries and triple systems, particularly those containing giant stars to determine masses of stars on the giant branch. This requires radial velocities with errors of a few tenths of km/s, implying resolutions of the order of R=30,000 and a stable wavelength calibration. Stars are generally brighter than V=7 and would require S/N >= 100 for good results. We will obtain these observations over timespans of a few months to ten years, with stars sampled frequently enough to give 50-100 spectra around a star's orbit.

C) Use Doppler images to decode the surface structure of highly spotted stars, distorted binary components, and accretion discs. Stars have magnitudes brighter than V=10. Observations will be primarily in the red (6400 A and H-alpha). Resolution required will be a minimum of R=30,000 with R=90,000 required for resolving surface features in spotted stars better with S/N>=200. For Doppler images of spotted stars, we will obtain of the order of twenty spectra around the star's rotational phase over a timespan of a few days to a few weeks. Exposures are expected to be up to 30 min on the basis of our experience with telescopes at NOAO.

D) Use fluctuations of stellar winds to restrict models for wind acceleration in cool giant stars. These studies require the same kind of coverage as the long-period spectroscopic binaries and require resolutions of order R=30,000 to detect changes in line profiles or to resolve shell lines. The stars are brighter than V=7 and require S/N>=100. Exposures in the red would be 5-20 min.

REQUIREMENTS FOR THE TELESCOPE:

To accomplish the scientific program, the telescope must be able to acquire stars brighter than V=10 and track them for periods up to 20-30 minutes. Because the telescope is essentially built at this point in the 2-m project, there is no point in assessing verious optical prescriptions; the mirror set we are using is a classical Cassegrain giving a final f-ratio f/8 with the focus 18 inches in front of the 81-inch (2-m) primary mirror. The images produced by this optical system must fit into a 200-micron optical fiber, the size of which we have chosen to give minimal f-ratio degradation at the spectrograph. This size subtends 2.5 arcsec at the focus of the telescope, so the telescope should produce appx. 2-arcsec images. We will assume 1-arcsec seeing, which gives a requirement of 1.7 arcsec for the intrinsic images of the telescope itself. Problems that might contribute to the blurring of the telescope's images are (1) environmental seeing, (2) miscolimation of the optical system, (3) warping of the mirrors by misadjustment of their supports, (4) tracking errors that let the star's image wander outside the fiber (entrance aperture) and thus make it effectively larger, and (5) errors in the optical surfaces as manufactured. We can assign various maximum effects for these five factors. Environmental seeing, 0.5 arcsec; miscolimation, 1.0 arcsec; warping, 0.75 arcsec; tracking errors, 0.75 arcsec, and errors in manufacturing, 0.75 arcsec. These give 1.71 arcsec when added quadratically.

We will require that environmental seeing contribute no more than about 0.5 arcsec to the size of the image. In other telescopes environmental effects tend to come from hot air trapped in the dome or from hot air above the primary mirror (or otherwise in the optical path). These effects tend to be either there or not. To assure they are not, we have eliminated most of the sources of heated air in this telescope as part of its design. The telescope is going to be housed in an open structure, which will not trap hot air to flow out a slit. The telescope is going to be evacuated (along with the pump room for the oil bearings) with the air sucked out of the telescope dumped north of the site. This approach should eliminate hot air around the primary mirror. There is, however, a possibly uncompensated source of image degradation in the drive motors for the secondary mirror. The effect of these (as with other local sources of heat) can be measured during our shakedown period by turning them on and off and measuring the size of the telescope's images. If they prove to be a problem, it may be possible to vent them rapidly enough to reduce the amount of heated air in the optical path enough to fit the 0.5-arcsec specification.

Miscolimation of the optical system could be a serious problem in our optical system because of the fast (f/1.5) primary mirror. We would like to keep the effect of miscolimation to within 1.0 arcsec. The dominant aberration is coma from a shift in the secondary mirror. A 1-mm shift would produce 1.8 arcsec blurring, so the misallignment (decentering of secondary mirror) should be kept to less than 0.5 mm. The secondary support structure is designed to produce gravitational flexures much lower than this. If it does not work as designed, we will program the telescope to reposition the secondary mirror with changes in altitude to compensate. We will collimate the telescope by using the shapes of star images in either the commissioning camera during shakedown or the guide camera during operation.

Warping of the mirrors by the support system will be determined by how well the supports function and how well they are adjusted. We know roughly how much force each actuator is required to support the mirror against gravity from a FEA calculation (see an engineering report on our web site). We may verify the effects of the mirror support actuators two ways. First we may extend the FEA calculation to include non-ideal support, e.g., by eliminating some of the supports and calculating the warping of the mirror. We plan to do these calculations during the commissioning stage for the telescope when the machine is in transit to Arizona or in storage. Second, during commissioning, we can deliberately misadjust the supports and measure the actual effect on the telescope's images. Finally, we intend to run a lab test with one of the axial support actuators to verify that it delivers the designed force (see discussion below).

Tracking errors that let a star drift out of the aperture should be limited to excursions of less than 0.75 arcsec for the <=20 min duration of an observation. This is a very generous allowance for the control system, which we intend to surpass with an actual goal of 0.5-arcsec drift, as corrected by the guide camera. We will verify the tracking with long-duration exposures during our commissioning phase. Problems with tracking will be addressed by additions to the control system that incorporate corrections to the pointing/tracking algorythms.

Errors in manufacturing should be acceptable if they give no more than a 0.75-arcsec blurring of the image, a criterion we have made purposely big enough that we should not have to worry about it. Since the primary mirror was manufactured with a surface error roughly one-wave peak to peak, this criterion should not be a problem (see the discussion of optical quality below).

Beyond the image quality of the telescope, we might define a throughput for the telescope as a system. At this stage of the project, this would be superfluous. In any case, we expect to lose about six percent of the light to blocking and diffraction by the secondary-mirror supports and roughly ten percent to reflection losses at the mirrors.

REQUIREMENTS FOR THE SPECTROGRAPH:

To accomplish our most critical research under CASS, the spectrograph should have a basic resolution of about R=30,000, with R=90,000 (appx) as an alternative mode, if at all possible. The programs we will pursue initially require observations in two wavelength ranges, the blue-violet (3700-4000A) and the red (6300-7000A). Observing wavelengths in between is desirable but not critical. Also, provided certain critical wavelengths are included, we do not absolutely need to observe all of every order of an echelle.

We discussed the preliminary design of the spectrograph at our review in September. It is a white-pupil echelle with an f/7 input from a 200-micron optical fiber colimated to an 8-inch beam. This beam is dispersed by a 52.6 line/mm echelle grating (63.5-deg blaze) from Richardson Grating Lab. operating close to Litrow. At this point, we are looking at designs with various deviations from the Litrow condition, namely theta=2.5 and 0.0 degrees and gamma=1.5 degrees. These angles are defined in Schroeder's book on astronomical optics. Theta measures the pitch of the grating about an axis parallel to its lines and affects the resulution of the spectrograph. Gamma measures the yaw of the grating about an axis perpendicular to the lines and to the input beam; it produces some astigmatism which widens the spectrum. The dispersed beam is reimaged onto a cross-dispersion grating by two concave mirrors and a folding flat. The cross dispersion would be by a large flat grating. There are very few choices for it, there being perhaps two so-called large astronomical gratings from the Richardson Grating Lab with the right combination of spacing and blaze angle. It might be possible to obtain a grating from Russia as an alternative, or to use a very large prism, but our attempts to pursue these directions came to nought. The grating we have chosen has 250 lines/mm with a blaze wavelength near 5800A. The cross-dispersed beam would then be projected onto the detector by a camera having a focal length near 20 inches (400-600 mm) and an aperture near 9 inches. This camera would ideally give images with about 10-micron (20-microradian) resolution over an area 30x55 mm to preserve the high-resolution mode of the spectrograph.

In addition to the basic echelle spectrograph, we are planning a low-dispersion mode for use with fainter objects, both because of our experience with the low-dispersion mode of IUE and at the advice of previous TRC's. This mode will use a separate input fiber that feeds into the second reimaging mirror, which serves as the colimator for the cross-dispersion grating.

REQUIREMENTS FOR THE DETECTOR:

The detector must be able to give the required resolution with a acceptable amount of oversampling and have a big enough area to record the whole spectrum without moving it. For resolution, the basic echelle mode (R=30,000) will have input from a 200-micron fiber and produce output pixels of 60-70 microns and 23-26 microns at the higher resolution. For the higher resolution, 2x oversampling suggests we need 12-micron pixels in the CCD. For the lower resolution, these could give enough oversampling that we could bin the data by two pixels in both directions. To fit the whole spectrum (3700-7000 A) onto the chip, without losing the ends of the orders at the longer wavelengths, we would need a CCD approximately 40 mm by 55 mm in the preliminary designs of the optical system of the spectrograph. There are few large-format CCDs to choose from, and the possibilities seem to be (1) a SITe SI424A 2kx2k CCD with 24-micron pixels and (2) the newer SITe 2kx4k ST002A with 15-micron pixels. We have chosen to use the SITe ST002A because it is newer and easier to get (apparently easier to manufacture reliably), because the pixels will allow higher resolution, and because it is physically larger in one dimension, which makes it easier to fit the echelle spectrum onto the CCD, at least with comercially available gratings. Consequently, we bought two of these last year, a grade 2 chip ($63K) for the primary detector and an engineering-grade chip ($13.5K) for a backup. We intend to mount these in dewars with mechanical refrigerators for cooling and get San Diego State to sell us the control systems to run them and integrate them into our observatory's computer system.

OPTICAL QUALITY:

The Gull committee advised us to test the primary and secondary mirrors to make sure their optical quality is sufficient for the project. To verify the optical quality of the primary mirror, we have located the original engineering report for the mirror that gives interferograms made at various stages of its manufacture, including ones from the final polish (See Attachment 1). You may recall that Kent Honeycutt tried to buy this mirror in the early 1980's for use in an astronomical telescope (spectroscopic) and had a consultant assess the mirror and its cell for just such an instrument as we have built. Torus Optics provided interferograms with the secondary mirror showing very good optical quality with repect to their test plate.

ROBUSTNESS OF THE MIRROR-SUPPORT SYSTEM:

The Gull committee advised us to evaluate the robustness of the mirror support system we have built to demonstrate it will support the mirrors adequately. We take this to mean the committee were concerned about the system of passive counterweighted levers we are using to support the primary mirror, whether they will provide the right forces and whether they are strong enough to survive putting the mirror into the telescope. To address this concern, we plan to use a load cell to determine how much force one of our axial support levers would actually exert on the back of the mirror at various orientations and to verify that these are the forces assumed in the design of the levers. We determined the required forces through a finite-element analysis of the mirror (see the engineering report on the telescope design on our web site), but we probably could extend that analysis to determine how sensitive the mirror figure is to errors in the force applied. For the analysis of the support lever, Busby has located a load cell in the TSU civil engineering department.

The other question of robustness we have imagined is whether the weight of the primary mirror would break the supports when it is lowered onto them in the cell. This potentiality concerns us, and we have devised a procedure for putting in the mirror that lowers it gently onto the supports. We are revising this procedure during the mechanical integration of the telescope. All these tests and certifications will be finished before we ship the telescope to Arizona.

REQUIREMENTS FOR THE DATA-HANDLING SYSTEM:

Under current operations, Fairborn Observatory is linked to an internet point of presence in Arizona via a dedicated telephone line. The resulting bandwidth is adequate for the existing photometric telescopes but will not be for the upcoming AIT and AST. As noted in the report of the NASA review on the AST project, we have proposed separately to NASA for funding of a T1 internet link to Fairborn, but, due to cost considerations, the ourcome of this request remains uncertain. Therefore it is clear that we must plan for continued operation of the site without the T1 link.

The AIT and AST will generate much larger volumes of data than can be transmitted back to TSU over the existing link. The only practical alternative to a T1 line is weekly shipment of data tapes from Fairborn to TSU. To support this, we plan a new data handling system for the imaging and spectroscopic telescopes. This will be based around two high-speed computers with large hard drives (~200Gbytes) and tape archiving systems. One system will reside on the local network at Fairborn and will accept the raw data files from the AIT and AST as they are produced. This computer will preprocess the individual frames (dark and bias subtraction, flat fielding, etc.), and two copies of such preprocessed images will them be written to tape at Fairborn. One copy will be archived at Fairborn; the second will be shipped to TSU.

A duplicate computer system will be installed at TSU to handle the pre-processed images received from Fairborn. This system will be connected to our existing network, so the data can be loaded, parsed, and distributed to our partners and collaborators. The duplicate system will also be used for final data reduction and analysis and for maintaining the duplicate archive of the pre-processed image frames.

The computers will probably be Pentium II machines running the Linux operating system. Currently, only SCSI technology is available off the shelf to build arrays of disks to the capacity needed for this project. In the near future, this should become possible in Linux with cheaper IDE hard drives. Archiving on these machines will initially be on magnetic tape, as optical storage technology for these data volumes is not yet economically available. Funding for these systems should become available shortly through a separate grant to our partner at Western Kentucky University.

Data from our existing APTs are reviewed each morning at TSU in order to identify quickly any problems with the APT operation. This will not be possible with the AIT and AST in the absence of a T1 link. Therefore, software must be put in place on the Fairborn computer to allow Fairborn personnel to review the AIT and AST operations each day. This can be a subset of the complete data-processing systems to be installed on the TSU computer. Critical quality-control frames can also be compressed and transmitted to TSU over the existing communications link for daily review.

If a T1 link eventually becomes available at Fairborn, the pre-processed frames will then be transmitted directly each morning from the Fairborn computer to the TSU data reduction and archiving computer. This will have several advantages: (1) Complete quality-control analyses can be performed each day by TSU scientists to detect any problems with the telescopes, instruments, and detectors to minimize downtime on the telescopes. This has proven to be very effective in getting the best performance from the APTs. (2) The progress of the scientific programs can be more effectively reviewed daily to enable timely revisions of the observing requests and maximize the scientific return. We will initially do this moderately well with nightly summaries of the images taken, which, of course, can be transmitted over our existing internet connection; but these files will not have the full range of information on image quality that daily reduction gives. (3) Real-time control of the telescopes can be exercised to facilitate engineering tests, trouble shooting, and target-of-opportunity observations. (4) Observations requested by our students, partners, and collaborators can be available the next day for analysis.

PERSONNEL FOR THE 2-M AST PROJECT:

The Gull committee had several specific thoughts about further personnel to work in various aspects of the 2-m project. These may be summarized as (1) a research associate to help with spectrograph assembly and integration and to get the CCD detector working, (2) an optical consultant with experience in designing spectrographs and their cameras to design a camera for the spectrograph, (3) an engineer (or technician) with optical/mechanical experience to help assemble the telescope and test it, and (4) unspecified extra administrative personnel to oversee the project schedule. We have considered these recommendations carefully, decided to adopt most of them in verious ways, and made progress on several of them.

The Center of Excellence in Information Systems (residence of CASS) is a small organization consisting entirely of key personnel holding well defined roles, as is our operating contractor Fairborn Observatory. For the past ten years, the Center's policy has been not to expand by adding permanent employees unless there is a compelling reason to do so. This policy has been necessary because of the Center's limited permanent budget. Instead, we maintain a small cadre of key personnel and try to fill temporary needs for expertise or too much work by using consultants. We are thus in the mold of the Space Astronomy Lab at Wisconsin which produced such NASA successes as OAO-2 and WUPPE. The Gull committee have pointed out areas in which we could use more people to augment the AST project, specifically, help with the mechanical assembly of the telescope, help with the optical design of the spectrograph, help with the integration of the spectrograph and shakedown of the CCD detectors, and help with the data handling system.

The Gull committee were most concerned with the design and fabrication of the spectrograph and advised us to add 1) and optical consultant experienced in camera design and 2) a postdoc with experience in designing spectrographs and characterizing CCD's who might assist with the spectrograph. We have been looking for persons to fill both these roles, and have located a potential optical consultant.

We have begun looking for an optical consultant to design the spectrograph and its camera and have located a potential candidate at the Aerospace Corp. He is David Warren, who has experience designing infrared spectrographs for large astronomical telescopes. Eaton will be going out to California to talk with him about the project in mid February, 1999. We thought the Gull Committee were going to contact Harlan Epps for us to see if he might be interested in this project.

We have decided to try to hire a research associate in instrumentation for the Center who will aid in integrating the spectrograph into the 2-m telescope system. To find such a person, we have been asking advice of instrumentalists about possible candidates for this position with limited success. We have collected names of a few potential candidates and will soon begin contacting some of them.

We are using the following mechanical/optical technicians for help in assembling the telescope in Nashville and testing it in Arizona: In Nashville, primarily Andre Hedrick, whom the Gull Committee met, occasionally for advice about mechanical questions and for extensive help with integrating the electrical controls of the telescope into the mechanical structure. We have also retained Mark Wells of Huntsville, Ala., to assist in such mechanical tasks as assessing leaks in the oil-pumping system and assembling the secondary-mirror cell and designing and building absolute-encoder mounts for the axes of the telescope. In Arizona, we are using primarily Lou Boyd, the director of Fairborn Observatory, our host. He has worked with us for many years. Furthermore, he is a brilliant instrumentalist. We can expect him to concentrate on the 2-m project during those times one of us is physically at the site for tests.

The Gull committee also stressed the need for help for Eaton in assembling the telescope at the State Hangar for the purpose of safety. We are addressing this concern by being sure there is always a second person from our program at the site (in addition to personnel from the Tennessee DoT) when we are moving any large parts of the telescope around with the crane. To do so we found a student, Kenneth McDavis, to work one morning or two a week at the hangar to help with telescope assembly, we used Andre Hedrick extensively for the purpose during several weeks in November, and we have used Busby and Henry occasionally as well.

The Gull Committee questioned potential conflicts between the AIT and AST schedules. Since Boyd is building the AIT for us, this should be a good point to address that concern. We think there should be a minimum of conflict between these two projects. The AIT is being built entirely by Fairborn Observatory through a contract with TSU. TSU's contact with Fairborn is Greg Henry, who is not involved in the AST project. It is true that the principals in Fairborn are themselves involved in the AST project, but they are also involved in other projects that conflict with both the AST and AIT. Consequently, recognizing this, we have limited their roles in the AST project to what we think are either manageable or absolutely necessary and have given them roles which, although in many ways critical, can be accomplished during slack in what may be seen as critical paths of the project. Consequently, Hedrick, not Epand, is the primary programming consultant for the basic motion controls for the telescope, and for the same reason we decided years ago not to get Boyd to design our telescope mount.

Below is a listing of some personnel other than Eaton and Busby involved in the 2-m AST project:

QUESTIONS ABOUT COMMISSIONING AND OPERATIONAL PHASES:

The Committee questioned plans for staffing at the FO site in Arizona during commissioning and operations. We have formulated a rough breakdown between the tasks to be done in Nashville and those in Arizona, and this is reflected somewhat in the attached schedule. In Nashville, we plan to make all of the mechanical tests that we anticipate will require physical modifications to the telescope structure. Of course, there will be other, unanticipated problems, but our strategy should minimize them and reduce our vulnerability to having to get Boyd to make mechanical changes. We will also verify the ability of the drive motors and control electronics to move the telescope through its full range of motions and verify our ability to use several motor encoders simultaneously to determine the telescope's position. In Arizona, we will conduct tests of the integration of the telescope into its observatory and those that require observation of stars.

We anticipate spending up to one week a month at the site in Arizona for acceptance and performance testing during the time May-October 1999. Eaton and most likely, one other person, possibly Hedrick, would make these trips. They should cost about $1400 per head for customary travel expenses. We have the money for this in our travel budget, and we have located a motel in Patagonia, Ariz., that would serve as a headquarters for these expeditions.

Once the spectrograph is finished and put together in Nashville, we will take further trips to Arizona to integrate it into the observatory and assess its performance.

During routine operations, the telescope and its instruments will be maintained by Lou Boyd and any extra staff he may decide to hire. We will not routinely have TSU employees on site for the 2-m telescope, or, for that matter, for any of our other telescopes.

Development of the data transmission and reduction schemes are being handled by Willard Smith and Francis Fekel of TSU with help from Jeff Hall and Lou Boyd.

SCHEDULE:

The schedule in late 1998 looks like this for the various subcomponents:

I.  Telescope (mechanical integration of major systems):  To be 
    finished by end of February 1999.  

   A. Finish base skirt and air circulation system (to Feb 7, 1999)
      1. Rivet base flanges to motor covers (1st day)
      2. Fit drive and motor covers into main shroud (2nd day)
      3. Mark and cut bottom edge of main skirt (2nd day)
      4. Add holders for air filters to bottoms of drive shrouds (3rd day)
      5. Connect motor covers to fork with 3-inch line (4th day)
      6. Paint drive and motor shrouds.  Add weather stripping to 
         seal gaps as needed (2 days)
   B. Integrate mirror-support system with surrogate mirror (to Jan 28)
      1. Finish putting lateral support levers into cell. 
      2. Finish simulated attachments for lateral supports and attach 
         hard points.  
      3. Adjust support levers and hard points.
      4. Add covers to back of mirror cell.
   C. Machine and annodize tilt-drive sector? (all of Feb. 1999)
      1. Decide whether and how sector should be processed (Feb 1-7)
      2. Procure tools or find machine shop to do it (Feb 8-14)
      3. Find shop to anodize sector (Feb 8-14)
      4. Send out for anodizing (in March or April)
   D. Put actual mirror into cell and perfect attachments for
      the mirror supports (Feb 1-14 or Feb 20-Mar 7)
      1. Pull out surrogate mirror (1st day)
      2. Lift mirror out of box, put on stands, and put lifting fixture 
         into center (1st day)
      3. Finish replacement lateral puck and glue onto mirror (1 week)
      4. Design and construct pull attachments for lateral supports (1 week)
      5. Lift and put mirror into cell (2nd day)
      6. Attach lateral supports and lateral hard points (2nd day)
      7. Glue pucks back onto mirror (2-3 days)
      8. Add real axial hard points (1 day)
   E. Fix leaks in the oil bearings (Feb. 14-20)
      1. Get instrument for inspecting seals in situ
      2. Test seals to find leaks
      3. Devise strategy for perfecting primary seals
      4. Finish constructing secondary oil-return system (1 week+)
   F. Assemble drive tractors and add them to the telescope (Jan 20-31)
      1. Press drive rollers into housings (1st day)
      2. Shrink fit flanges onto ends of rollers (1st&2nd days)
      3. Add motors to drives (3rd day)
      4. Add motor-holding flexures and cover plates (3rd day)
      5. Place azimuth drives into telescope and adjust clamps (4th day)
      6. Modify shrouds to accomodate azimuth drives (5th day)
      7. Place altitude drive into telescope and adjust clamps (5th day)
   G. Modifications to top end (quadrapod) (Jan 15-Feb 14)
      1. Set up centerline with plumb bob
      2. Measure location of baffle housing and decide on modifications 
         to support struts to make it symmetric to within mechanical 
         tolerances (2 days)
      3. Modify struts (3 days)
      4. Place quadrapod onto telescope and remeasure centering (1 day)
      5. Modify feet as necessary and pin to mirror cell (2 days)
      6. Integrate evacuation lines into quadrapod (1 day)
   H. Painting and insulation (Feb 15-March 15; beyond integration)
      1. Touch up paint on mirror cell and fork (1 day)
      2. Add reflective insulation to selected parts of telescope (3 days)
   I. Tests of telescope evacuation system (March 1-7; beyond integration)
      1. Hook up evacuation lines (1st day)
      2. Hook up temporary air outlet to fan and run test [Busby] (2nd day)
      3. Observe air flow through telescope and design flow restrictors 
         [sheet metal plates] to manage it. [Busby and Eaton] (2nd&3rd days) 
   J. Miscellaneous jobs and modifications
      1. Enlarge pockets for lateral-support counterweights in mirror 
         cell (1 day--during packing for transport)
      2. Increase clearance for secondary mirror cell in quadrapod
         (1 day--post integration)
      3. Test perpendicularity of tilt axis to azimuth axis and shim as
         required (run test during Jan/Feb--shim in Arizona?)
      4. Decide on bearing retainers for drive tractors; design and build
         if deemed necessary (decide in Feb-Mar, build in Mar-April)
      5. Many others that cannot really be anticipated.


II.  Telescope (mechanical integration of minor systems):  

   A. Assemble secondary mirror cell and verify its operation [by
      Mark Wells, consultant] (Feb.-March 1999)
      1. Glue secondary mirror onto its mounting stud. (1 day)
      2. Modify secondary mirror cell and associated parts to make them
         fit together.
      3. Finish designing spring-loaded hold-downs and flexures.
      4. Make hold-downs and flexures.
      5. Design limit switches for mirror cell.
      6. Assemble mirror cell and add linear drive motors.
      7. Test mirror positioning with motors
   B. Design, build and integrate mounts for coarse absolute encoders 
      for the two axes [Mark Wells] (February-March 1999]
      1. Finish design of the azimuth-encoder adapters (60% done)
      2. Design tilt-encoder adapter
      3. Machine adapters
      4. Order encoders
   C. Add small systems to telescope, likely in Arizona. (April-May 1999)
      [the attachment holes for the azimuth-encoder adapter and secondary 
      mirror cell are already in place]



III. Telescope (transportation to Arizona):
 
   A. Construct pallets and boxes for various subsystems of the 
      telescope (Jan 1-Feb 15, 1999)
   B. Arrange transport to Arizona [Busby] (February-March 1999)
      1. Make firm estimate of sizes and weights of loads
      2. Contact shipping companies for advice about how to ship
      3. Establish contract for shipping telescope
   C. Arrange transport to the site at Fairborn Observatory
      1. Establish schedule for shipping parts from warehouse in 
         Tucson to site, if applicable
   D. Assembly of telescope at the site (7 days+5 contingency)
      1. Put base and oil pump in place, level base, hook up oil 
         lines, and test oil-pumping system. Temporarily hook up
         electrical connection for interior of base (1st day)
      2. Place fork on base, hook up central pivot, and test oil
         bearings.  Attach main electrical cables into fork. (2nd day)
      3. Assemble 30 of 36 axial counterweights into mirror cell, 
         place mirror cell into the fork, and verify that the tilt axis 
         is perpendicular to azimuth axis. Adjust tilt axis, if necessary.
         (3rd day)
      4. Pull mirror cell upright with special winch, secure with auxillary 
         stay bar, and add tilt drive sector.  Verify runout of tilt drive
         sector.  Add remaining six axial support levers, tilt cell 
         horizontal, and secure with stay bar.  Add the six lateral support 
         levers. (4th day)
      5. Place primary mirror in cell and align mechanically, i.e., with 
         shop gauges.  Assemble quadrapod [top end of telescope] and place 
         on telescope with dummy secondary mirror cell.  Verify balance 
         of telescope about tilt axis. Stow for closing enclosure with 
         85-degree stay bar. (5th day)
      6. Add azimuth drive tractors to telescope and verify their tracking 
         around the slewing ring with temporary electrical setup.  Add tilt 
         drive tractor and verify tracking on drive sector.  (6th day)
      7. Add electrical enclosures [for computers, drive electronics, etc]
         to telescope. Verify working of limit switches, zero-point detectors,
         etc., and add base skirt.  Hook up primary electrical connections. 
         (7th day)



IV. Telescope (control system):

   A. Add drives/control electronics to telescope in Nashville (Feb 15-March 1)
      1. Wire 120-volt line into temporary umbilical
      2. Design and build enclosures [perhaps temporary wooden versions to 
         be replaced with plastic on the way to Arizona] to hold drive 
         amplifiers, computer, and auxiliary electronics on the telescope.
      3. Wire dirve motors up to their drive amplifiers.
      4. Start wiring limit switches and zero-point detectors in to 
         control board
   B. Perform mechanical/electrical tests of telescope motions 
      (servo systems) in Nashville (March 1-March 15)
      1. Use hand-held controller to move individual motors and to move 
         groups of motors as units; test this procedure first in the lab
         on the control simulator.
      2. Hook the drives up to computer and run increasingly complex motion
         profiles
   C. Finish device drivers for control of telescope [Andre Hedrick]
      (February-March 1999)
   D. Integrate device drivers with motion-calculating software 
      and perform tracking tests (March-April 1999)



V. Telescope (control/tracking testing in Arizona):

   A. Build temporary CCD camera for testing telescope (Feb-April 1999)
      1. Design temporary instrument for use in tracking tests [using spare 
         instrument head already bought for that purpose] (Feb 1-Feb 31 1999)
      2. Purchase CCD camera for tracking tests (Feb-March 1999)
      3. Design attachments and shrouds to integrate camera into instrument
         head (March 1999)
      4. Get mounting built physically [Krebs, Wells, and/or Eaton] (April 1999)
      5. Get camera to work with its frame grabber [Hedrick] (April 1999)
   B. Integrate camera into basic control system for telescope motions
      (April-May 1999)
   C. Test telescope with one-week trips to Arizona, as needed (May-October 1999)



VI. Telescope (instrument head):

   A. Assemble instrument head mechanically [Wells and Eaton] (March-April 1999)
      1. Bolt existing pieces onto platform
      2. Decide how to build the pickoff mirror
      3. Fabricate pickoff mirror and spares
      4. Decide what further parts are needed; design and make them.
   B. Get guide camera to work (May-Aug 1999)
      1. Get a student to get camera working under Windows
      2. Get camera working with Italian device driver under linux
   C. Build shrouds to integrate instrument head into telescope (appx 1 week)
   D. Buy fiber optic cable (April-June 1999)
      1. Get quote from Polymicro
      2. Pressure Busby to set up contract
   E. Make up fiber optic cables for telescope (Sept-Nov 1999)
      1. Decide on lengths of cables with test in Arizona
      2. Decide on terminations and jacketing
      3. Decide who makes up cables (Polymicro, Barden, or TSU)
      4. Get cables made up and tested


VII. Enclosure (mechanical):

   A. Get air compressor and UPS through Fairborn Observatory. (Feb-Apr 1999)
      1. Estimate maximum power consumption for telescope drives
      2. Estimate Power consumption for auxiliary instruments
      3. Get Boyd to run air consumption tests for moving roof
      4. Negotiate with Boud to get stuff purchased
   B. Design and build mechanical controls for motions of building.
      [Fairborn Observatory] (March-May 1999)
   C. Get brush seals to close gaps in building. (March-May 1999)
      1. Visit Fairborn Observatory and measure clearances to seal (Feb ?, 1999)
      2. Decide on seal and order it (2 days)



VIII. Enclosure (controls):

   A. Design building controls (for opening and closing the front door 
      and rolling the top section on its rails) [Lou Boyd] (Feb-May 1999)
   B. Decide on capacity of UPS and get Boyd to buy it. (Feb-May 1999)
   C. Wire up and test this control scheme. [Boyd] (April-May 1999)
   D. Integrate air conditioner(s) into enclosure. (summer 1999)



IX. Spectrograph (optical design):

   A. Retain and optical consultant for the spectrograph (and especially
      camera) design (February-March 1999)
   B. Hire a postdoctoral associate/instrumental specialist to assist in
      the spectrograph design, spectrograph integration, CCD integration, 
      and management of the AIT.  (February-April 1999)
   C. Finish optical design of the spectrograph and camera. (March-May 1999)



X.  Spectrograph (mechanical design):

   A. Design mounts for the optical elements of the spectrograph. 
      (1 month; late 1999-early 2000)
   B. Contract with machine shops to build and deliver these mounts. (2 months)



XI. Spectrograph (optical fabrication):

   A. Bid out camera design for fabrication (June-July 1999)
   B. Monitor contract and accept parts (1 year?)
   C. Assemble and test spectrograph camera (2 months)
   D. Integrate camera into spectrograph (1 month)



XII. Spectrograph (calibration sources):

   A. Finish preliminary design for integrating calibration lamps
      into the fiber-optics system
   B. Decide on calibration lamps to buy
   C. Finish designing the mounts for lamps and fiber-optic coupling
   D. Buy or construct the mechanical parts
   E. Assemble and test calibration bench


XII. CCD detector:

   A. Decide on dewer design (February 1999) 
   B. Get dewar built (2 months)
   B. Contract to have CCD controller(s) built. (March 1999)
   C. Integrate CCD controller into control system for the telescope/
      instrument (1 month after receiving detector/controllers)



XIII. System integration:

   A. Put spectrograph on site (post testing)
   B. Start running the whole instrument


XIV. Wastes of time:

   A. Mechanical test of axial support lever for mirror support (Jan 20-
      Feb 20, 1999)

                    AST Cost Estimates (US$)

Component              Year1      Year2      Year3      Year4      Year5

Telescope
  Optics             520,000       2,75                10,515
  Mount                         316,359     365,691       859
  Consultants                    86,526       6,000     6,500
  Misc/Contengency                                                10,000

Site Preparation                 47,558

Enclosure                        28,859     119,450    11,536

Control System
  Components          23,205      3,739                 5,000*     5,000*
  Consultants                    30,000      18,000    20,000
                                                       20,000*

Detectors                                    76,600    73,000*

Spectrograph
  Components          48,000     25,500      31,507    70,000*   205,200*
  Consultants                                          50,000*    50,000*

Data Handling                                                     80,000*

Mgt fees, Ins, Trans             21,609      10,732    10,500
                                                       15,000*

* Indicates components or services to be purchased.

            Details of Estimated Remaining Project Costs (US$)

     Enclosure

UPS               $15,000
Controls            5,000
Seec Enc            2,000
Contingency         2,200
                    -----
Total:             24,200


     Telescope

Misc/Contingency   10,000
                   ------
Total:             10,000


    Control System

Computers          10,000
Jeff Hall          20,000
                   ------
Total:             30,000


       Detectors

Dewars             22,000
Controller         36,000
Integration        15,000
                   ------
Total:             73,000
                 

      Spectrograph

Camera            200,000
Fiber Opt           7,000
Stands             12,000
Control Parts       1,200
Misc Parts          5,000
Consultant*        50,000
Post Doc*          50,000
Contingency        50,000
                   ------
Total:            375,200

     Data Handling

Data Stor*         80,000
                   ------
Total:             80,000


  Mgt Fees, Insurance, Transport

Insurance           5,000
Shipping           10,000
                   ------
Total:             15,000



*  Expenditures not included in original costs estimates.  
These estimated expenditures will be accounted for by a 
budget realignment for Year 5 in our Renewal Proposal 
submission.

                  SUMMARY OF ESTIMATED PROJECT COSTS (US$)

    Component          Cost        Projected Cost   Tot Projected Cost     Orig Projected Cost
                     (2/1/99)        (Remaining)     

 2m Telescope        1,314,525         10,000           1,324,525              1,400,000            
 Site Preparation       47,558            0                47,558                 77,000

 Enclosure             159,845         24,200             184,045                250,000

 Spectrograph          105,007        375,200             480,207                164,100

 Detectors              76,700         73,000             149,600                140,000

 Control System         94,944         30,000             124,944                110,000

 Mgmt Fees              42,841         15,000              57,841                 75,000

 Data Handling                         80,000              80,000                    0
                     ---------        -------           ---------              ---------

  TOTAL:             1,841,320        607,400           2,448,720              2,216,100

Excerpt from original engineering report on the 2-m mirror showing interferograms with respect to null lens for the final polish. (Forwarded to J.A. Eaton by Dr. Kent Honeycutt)