ESI: An Echellette Spectrograph
and Imager for the Keck II Telescope

Version 1.0

Principal Investigator: J. Miller
Co-PI: M. Bolte, R. Guhathakurta, D. Zaritsky
Optical Designers: B. Sutin, H. Epps
Project Manager: D. Cowley
Project Scientist: B. Bigelow
January 9, 1996


1 Overview

2 Instrument Science Rationale

3 Optical Design and Performance

4 Overview of Specifications and Operating Modes 5 Functional Descriptions 6 Instrument to Telescope Interfaces 7 Data Processing 8 Project Management 9 Budget

10 Schedule

11 Future Upgrades

Appendix 1. ESI Camera Design

1. Overview

The ESI instrument is a versatile, multi-mode spectrograph and imager. There are two distinct spectroscopic modes: a medium-resolution echellette mode with prism cross dispersion; and a high-throughput mode using prism dispersion only. The spectrograph modes will cover the full wavelength range of the Keck II silvered mirrors (3900 to 11000 Å) in a single exposure. The low-resolution, prism-only mode provides the same spectral range, multi-object capability, and very high through-put. ESI also provides an imaging mode with a field of view of 3.0 arc-minutes. An Epps refracting camera and a single 2K x 4K detector are used for all three modes.

The spectrograph will achieve cost savings through simplicity and the direct use or minor modification of existing Lick optical, mechanical, electronic, and software designs. The data taking and control software will be adapted from existing packages, primarily those developed for the DEIMOS spectrograph. This is made possible by use of the same motors, motor controllers, encoders, etc, as are being used for DEIMOS and existing instruments such as the Lick prime focus camera and MOS.

The ESI collimator will be an off-axis paraboloid, similar to the collimator in LRIS. A single 2K x 4K CCD with 15 micron pixels (1/8 of a DEIMOS array) will be used with a version of the DEIMOS CCD controller. A single, all-refracting camera, also similar to the DEIMOS design, will be used for all three operating modes. There will be two CCD TV cameras; one for acquisition and guiding, and one slit viewer. Two filter-wheels will be provided; one immediately behind the slit, and one in front of the camera for the imaging mode only. A third wheel in the focal plane will carry a set of slit masks. At any one time, a single echellette grating is available, which is fixed in position and in rotation. Changing gratings is a manual operation and normally will not be done during the night. A flat mirror bypasses the grating for switching to the prism-only mode. A second flat mirror bypasses both the prisms and the grating to provide the imaging mode.

Variations on all of the required motion stages and servo-mechanisms have been successfully designed and built in the Lick Observatory labs. The structure will likely be a light, rigid space frame. The only potentially new feature could be active control of collimator tip and tilt, for flexure compensation, if required.

The project is scheduled to be completed in 1.5 - 2 years from inception to commissioning, on a budget of $1.7 million. The PI is Joe Miller, with Co-PI's Michael Bolte, Raja Guhathakurta, and Dennis Zaritsky.

2 Instrument Science Rationale

ESI is an intermediate resolution echellette spectrograph and imager whose capabilities are suited to addressing a wide range of scientific problems. In its primary mode, ESI will be used to carry out single slit echelle spectroscopy. Distant galaxies, both field galaxies and those in clusters, and high redshift quasars are ideal targets. The moderate spectral resolution provided by the baseline grating (FWHM of about 55--70 km/s or 4-5 pixels) and high throughput of the instrument make it suitable for measuring the internal kinematics of faint galaxies. The stars and ionized emission line gas in spiral galaxies and the stars in elliptical galaxies have a typical rotation speed or velocity dispersion of a few 100 km/s so they would be well resolved by ESI. Studies of the internal kinematics of distant galaxies are becoming an increasingly popular tool for deciphering the nature of luminosity evolution in moderate redshift galaxies via their luminosity-linewidth relation (Tully-Fisher, Faber-Jackson, and Fundamental Plane relations).

The wide continuous spectral coverage of ESI (3900-11000 Å) will, at the same time, provide an excellent measure of the spectral energy distribution of these distant galaxies, information that is crucial to understanding their mix of constituent stellar populations. Quasar spectroscopy will benefit greatly from ESI's wide spectral coverage which corresponds to a wide range of redshifts for the intervening hydrogen (damped Lyman , Lyman limit, Lyman forest) and metal-line absorption systems. The slit can be narrowed down to 0.5" (in periods of very good seeing or for bright quasars) to achieve a velocity resolution as high as 40 km/s this is important for deblending lines in the Lyman forest and for studying the kinematics of multicomponent metal-line systems. If funding permits, additional gratings providing significantly lower dispersion (and longer slit lengths per orders) could be made available.

The second mode of ESI is as a prism-only spectrograph characterized by variable resolution of about 3Å/pixel (or about 150 km/s/pixel at 6000 Å), wide spectral range (3900-11000 Å), and very high throughput (~50% including detector). This will be used mainly in multislit mode to study about 20 objects at a time. The prism setup will thus provide an efficient means for obtaining redshifts (accurate to 0.3 pixel or 50 km/s) and spectral energy distributions of galaxies in high redshift clusters (the cores of which are comparable to ESI's 3' field of view) or around a QSO (in order to identify the galaxy resposible for a particular absorption line). It will also go fainter than any other spectrograph on the Keck telescopes.

The third mode of ESI offers direct imaging capability over a moderate sized (3') field. ESI will be unique among Keck instruments in its narrow band imaging capability; this will be useful in studying the spatial distribution of ionized gas (H , [OII], [OIII]) and the nature of star formation in distant galaxies, in searching for emission line objects in some narrow, predetermined redshift range (e.g., searches for galaxies around a QSO or around QSO absorbers), and in imaging searches for Lyman protogalaxies.

Controlled functions in ESI

3 Optical Design and Performance

The primary consideration for the optical design of the ESI instrument was to have an inexpensive (relative to other Keck instruments), high throughput, mid-to-high resolution imaging spectrograph which could be delivered to the Keck II telescope in two years.

The minimum requirements are as follows: a spectrograph mode with dispersion of at least 15 km/sec/pixel from 0.39 to 1.1 microns without gaps, and an imaging mode with a field of view at least 3.0 arcmin in diameter. ESI is designed for the Cassegrain focus, and uses a single 2K x 4K CCD.

3.1 Collimator

Because of the high-throughput consideration, the light loss from a lens collimator was considered prohibitive, and a Cassegrain collimator would not have the required 3-arcmin field of view, so an off-axis parabolic collimator was chosen.

The collimator focal length is 90.0 inches, the maximum length which would fit into the Cassegrain module envelope. A shorter focal length would give a smaller beam size, but the overall performance would be degraded. The final beam size is 6.57 inches if the entire Keck aperture (from tip to tip) is considered.

In order to make enough room for the imaging mirror which may be inserted to direct the collimated beam into the camera, the collimator is off axis by 5.0 degrees. This collimator tilt does not affect the on-axis image, and has little effect in the echellette spectrograph mode. Off axis the spot images at the detector become uniformly illuminated triangular shapes which can become smeared out by the aberrations of the camera (see attached report for spot diagrams).

3.2 Echellette Mode

The overall design was driven by the 'workhorse' mode, where ESI operates as a cross-dispersed grating echellette spectrograph. Figure 1 shows the format of the baseline echellette spectrum on the 2k x 4K CCD. Figure 2 shows the optical layout of the instrument in the echellette mode. The slit width of 20.0 arcsec is the minimum slit length for which the spectral orders never overlap. Slits are shown for the central wavelengths and the ends of the free spectral range.

Figure 9 shows the outlines of the light beams intersecting the camera mouth while the instrument is in the echellette mode. The beams correspond to the center and ends of each order. Some vignetting occurs at the ends of the slit in the two red-most orders, but very little light is lost since the outer parts of the beam are from the wing-tips of the Keck telescope primary.

A major problem with cross-dispersed grating spectrographs is that gratings are most efficient in Littrow, when the incoming and outgoing beams are parallel; however, in this case the camera will block the incoming light. In ESI one prism is used in double-pass right in front of the grating, which bends the incoming and outgoing beams from the grating such that the camera no longer blocks the incoming beam. The grating is operated only a few degrees off Littrow, rather than the usual 20 degrees or more.

To get the appropriate amount of cross dispersion, three 50 degree prisms operating near minimum deviation would be required. Since one prism is operated in double pass, only two prisms are used.

The best material for the prisms was UBK7/BK7Y (blue transmitting BK7), which is also cheap and available in large sizes. Fused silica was a close runner-up, but resulted in significantly larger prism angles.

Optimization of the prism angles and positions for the best format on the chip, and best presentation to the camera mouth while keeping the beams on the grating, gave prism angles very close to 50.0 degrees, so the actual prism angles were both set to this value. The double pass prism is 13.0 inches high and 12.0 inches wide, while the single pass prism is 10.0 inches high and 13.0 inches wide. These are maxima, since precise ray tracing to eliminate unused material has not yet been done.

The baseline grating was chosen by virtue of being close to optimal and readily available from the Spectronic Instruments (formerly Milton Roy) catalogue. This grating is 203 mm square, with 175.555 lines/mm and a blaze angle of 32.3 degrees. The final dispersion is 11.4 km/s/pix (R = 26,300).

Because some of the of the cross dispersion takes place before the light meets the grating, the grating is operated in quasi-Littrow, meaning that the incoming light is at an angle to the plane of dispersion. Operating a grating in quasi-Littrow offers higher efficiency as compared to near-Littrow operation.

The final spacing of the orders on the chip leaves enough room for a slit length of 20.0 arcsec on the sky.

3.3 Low Dispersion Mode

By inserting a mirror in front of the grating, only the prism cross dispersion takes place, and the instrument becomes a multi-object low dispersion spectrograph with a varying dispersion of 50 km/s/pix (R = 6000) at 0.39 microns, down to a dispersion of 300 km/s/pix (R = 1000) at 1.1 microns. Figure 4 shows the instrument in the low dispersion mode. Figure 5 shows dispersion as a function of wavelength in km/sec/pixel in this mode. Figure 6 shows dispersion vs. wavelength in Å/pixel. Figure 8 shows the outlines of the light beams intersecting the camera mouth while the instrument is in the low dispersion mode. The beams correspond to the center and ends of a long 3.0-arcmin slit at various wavelengths from 0.39 to 1.1 microns. No vignetting occurs.

3.4 Imaging Mode

Figure 3 shows the optical layout of the instrument in the imaging mode. By inserting a mirror in front of the first prism, light from the collimator is sent directly into the camera. A curve on the front entrance window near the telescope focal surface serves as a pupil mover to move the exit pupil of the collimator to a point between this imaging mirror and the camera mouth. At the exit pupil, the beam waist is 6.57 inches, where imaging filters or a Fabry-Perot etalon would be placed. If 5.0 percent light loss is acceptable, then a filter or etalon with a diameter of 5.9 inches could be used, which would exclude some light from the outer hexagonal eartips of the Keck primary. Figure 7 shows the light lost due to an undersized Fabry-Perot etalon. The square dots are at the radii which correspond to points on the outer edge of the Keck telescope primary. A 168 mm (6.6 inch) etalon has no vignetting, while a 145 mm (5.7 inch) etalon would vignette 10 percent.

3.5 Camera

The three different camera designs by Harland Epps (see Appendix 1) were originally designed to have a focal length of 12.375 inches. In order to fit the entire spectrum onto the chip, the camera focal length had to be shortened to 12.125 inches. This redesign has not yet been completed because of changes currently being made in the Ohara glass catalogue. Many glasses are being either eliminated or changed for ecological reasons, and information on the availability of the high transmission (i-line) glasses is not yet available. Currently a revised camera design should be available in January 1996.

The new camera focal length gives a final image scale of 0.153 arcsec/pix.

The appendix gives three camera designs, with either zero or one aspheric surface. If the final camera choice includes an aspheric, that surface will be made at Lick. We currently plan to have the other surfaces made at a commercial optical shop.

4 Overview of Specifications and Operating Modes

The following specifications are preliminary, based on a document called "Echelle Spectrograph and Imager (ESI)", dated 6 July 1995. Specifications for the overall design and the three operating modes are described below.

4.1 Instrument Specifications

4.2 Mode 1: Echellette in near-Littrow configuration

The medium resolution mode uses a baseline catalogue echellette grating, and two prism cross-dispersers, one of which is used in double-pass to provide pre-dispersion. Table 1 shows diffraction orders, wavelength ranges, and dispersions for the echellette mode. Figure 2 shows the optical arrangement of the instrument in this configuration.

Table 1

Echellette Mode order separation, location, and dispersions
Order Lower (Å) Central (Å) Upper (Å) Range (Å) Height (%) Del (") Width (%) Disp #1 (Å/px) Disp #2 (km/s/px)
6 9366 10146 11068 1703 86 26.6 108 0.39 11.4
7 8117 8697 9366 1249 72 23.6 92 0.33 11.5
8 7162 7609 8117 955 58 23.0 80 0.29 11.5
9 6408 6764 7162 754 43 23.5 71 0.26 11.5
10 5798 6088 6408 610 28 24.7 64 0.23 11.5
11 5294 5534 5798 504 11 26.3 58 0.21 11.5
12 4870 5073 5294 423 -7 28.1 53 0.19 11.5
13 4509 4683 4870 361 -28 30.1 49 0.18 11.5
14 4198 4348 4509 311 -52 32.3 46 0.17 11.4
15 3927 4058 4198 271 -79 34.5 43 0.15 11.4

Order -- order of grating dispersion
Lower -- wavelength at lower end of the free spectral range
Central -- on-blaze wavelength
Upper -- wavelength at upper end of the free spectral range
Range -- the free spectral range (Upper - Lower)
Height -- percent height of order above/below the chip center at the central wavelength of the order
Del -- distance in arc-sec between orders at order center
Width -- percent width of the chip
Disp(1) -- dispersion at the central wavelength in Å/pixel
Disp(2) -- same in Km/s/pixel, R = c/11.4 = 26,000 for a thin slit}

4.3 Mode 2: High Efficiency, low dispersion configuration The high efficiency mode bypasses the grating by inserting a flat mirror behind the pre-disperser prism. Figure 4 shows the instrument in this mode.

4.4 Mode 3: Direct imaging configuration

The direct imaging mode bypasses the grating and the prisms by inserting a flat mirror ahead of the pre-disperser prism, and by retracting the second prism from the optical axis. Figure 3 shows the instrument in the imaging mode.

5 Functional Descriptions

The following section provide descriptions of all the optical, mechanical, electronic, and remote controlled functions in the instrument.

5.1 Optical Systems

The following is a preliminary description of all the optical elements in the instrument, roughly in the order that photons follow to the detector. Descriptions of the mechanisms required for the various functions are discussed in the next section.

In general, refracting optics are anti-reflection coated, with dielectric or sol-gel coatings to be determined later. Reflecting optics (except for the grating) are silver coated to match the Keck II optics. Zerodur is the nominal substrate for the collimator and flat mirrors, but other substrates could be used.

5.1.1 Window

A fused silica window seals the instrument and allows it to be lightly pressurized with dry nitrogen. The window is slightly curved for re-imaging the collimator exit pupil.

5.1.2 Calibration System

A small selection of lamps are provided for wavelength calibration. The lamps are fixed in position ahead of the slit. Some calibration sources may be provided by the telescope. Internal calibration light is reflected from a diffusive coating on the back side of the instrument entrance hatch.

5.1.3 Acquisition and Guide TV

The acquisition and guide (A\&G) TV uses a Lick-style CCD camera, and is mounted in a fixed position in the top of the instrument. The A\&G TV has a field of view (FOV) to be determined later. The A\&G TV detector, pixel size, pixel scale, and lens are all to be determined later. The detector is thermo-electrically (TE) cooled. The TE cooler itself is liquid cooled using the observatory glycol system.

5.1.4 Slit Viewing TV

The slit viewing (SV) TV stares at the slit and monitors the slit during an exposure. The SV TV also uses a Lick CCD camera and is fixed in position in the top of the instrument. The FOV, detector, pixel size, pixel scale, and lens are all to be determined later. The TV detector is cooled as above.

5.1.5 Slit Mask Wheel

There is a wheel with 11 positions for slit masks and one open position for imaging. The slit masks cover a 3 arc-min diameter FOV. Slit masks are tilted relative to the optical axis to feed the SV TV camera.

5.1.6 Slit Filter-wheel

The slit filter-wheel is located immediately behind the slit mask wheel. There are approximately 12 positions for filters.

5.1.7 Collimator

The collimator is approximately 310 mm diameter off-axis paraboloid. Instrument flexure is planned to be compensated by tilting the collimator. The collimator uses a Zerodur substrate with a silver coating.

5.1.8 Imaging Mode Flat Mirror

The imaging mode uses a flat mirror inserted in the collimated beam, ahead of the pre-disperser prism. The mirror substrate material is Zerodur, with a silver coating.

5.1.9 Pre-Disperser Prism

In both spectrograph modes, the first prism pre-disperses the collimated beam. The prism angle is 50.00 degrees. The prism face dimensions are approximately 12 inches by 13 inches. The prism material is UBK7.

5.1.10 Low Dispersion Mode Flat Mirror

The low-dispersion mode uses a flat mirror inserted between the pre-disperser and the grating. The mirror material is Zerodur, with a silver coating.

5.1.11 Echellette Grating

The grating is a standard Spectronics (Milton-Roy) catalog item with 175.555 lines/mm, and a blaze angle of 32.3 degrees. The grating is 203 x 204 mm square. There is one baseline (first-light) grating which is fixed in both position and rotation. The grating substrate is Zerodur.

5.1.12 Post-Disperser Prism

The second prism is similar to the first, with slightly different face dimensions. The prism substrate is UBK7.

5.1.13 Camera Filter-wheel

The camera filter-wheel is used only in the imaging mode, and is located between the second prism and the camera. There is a 300 mm diameter open position in the filter-wheel for use in the spectrograph modes. The filter diameter is roughly 180 mm. The filters will be AR coated.

5.1.14 Fabry-Perot

Provision is made for a Fabry-Perot (FP) etalon to be added in the future. There is no FP in the first-light instrument.

5.1.15 Camera

The camera is an Epps design, which is similar in shape and size to the DEIMOS camera. The refracting camera uses all-spherical elements, with a maximum lens diameter of 300 mm (12 inches).

5.1.16 Shutter

There is a shutter between the camera and the detector.

5.1.17 Detector

The detector is a 2K x 4K pixel device, with 15 micron pixels. The detector is likely to be a Lick device, manufactured by R. Stover and the UCO/Lick Detector Group. The detector is housed in a standard Lick-designed dewar.

5.1.18 Stray light

The entire optical systems is protected from stray light by a light-tight, dust-proof enclosure. Individual sub-systems are baffled to minimize stray light problems.

5.2 Mechanical Systems

These sections describe the mechanical functional requirements for all the major components of the instrument. In general:

5.2.1 Entrance Hatch

There is an air powered entrance hatch at the top of the instrument. The hatch seals the instrument when not in use. The hatch is dust- and light-tight. The back side of the hatch has a diffusive coating for reflecting calibration light in to the instrument.

5.2.2 Slit Mask Wheel

The slit mask wheel is a rotation stage with 12 positions. The slit masks are mounted in frames which are removable from the fixed wheel. The rotation stage is servo-powered with both motor and high-resolution encoders, and an optical fiducial.

5.2.3 Slit Filter-wheel

The slit filter-wheel is servo-powered and encoded from the motor. There are 11 positions for filters, and one open position for imaging. The filters are AR coated. The filters are mounted in manually removable frames, and attached to a fixed filter-wheel.

5.2.4 Main Collimator Focus and Flexure Control

There is a single main collimator, kinematically mounted to the focusing stage. The focus stage actuators use roller screws, each of which is driven by a high-resolution motor/encoder. As the collimator is at the end of a long structure which may flex unacceptably under varying gravity loads, the actuators are individually controlled to provide both focus and tip/tilt motions for flexure compensation.

5.2.5 Imaging Mode Flat Mirror

The imaging flat is mounted in a cell and carried on a linear stage. The stage is powered into position with an air cylinder. The mirror cell is positioned by the stage, but kinematically located in its final position.

5.2.6 Low Dispersion Mode Flat Mirror

The low-dispersion mode flat retracts out of the optical path for the echellette mode. The flat mirror is mounted in a cell, and carried on a linear stage. The stage is powered into position with an air cylinder. The mirror cell is positioned by the stage, but kinematically located in its final position.

5.2.7 Echellette Grating

The grating is fixed in position and in rotation. There are no automated adjustments. The grating is mounted in a cell which is kinematically supported. There is an access hatch in the top of the instrument to allow for manual exchange of gratings. The instrument must be removed from the telescope for grating exchange.

5.2.8 Post-Disperser Prism

The post-disperser prism retracts out of the optical path when the instrument is in imaging mode. The prism is mounted in a cell, and carried on a linear stage. The stage is powered into position with an air cylinder. The prism cell is positioned by the stage, but kinematically located in its final position. The prism is AR coated.

5.2.9 Camera Filter-wheel

The camera filter-wheel is located between the camera and the post-disperser prism. The camera filters are used only in the imaging mode, and are roughly 180 mm (7") in diameter. The filters are AR coated. The filter-wheel has positions for roughly 6 filters, and a 300 mm diameter open position for the spectroscopic modes. The filters are mounted in manually removable frames, and attached to a fixed wheel. The rotation stage is servo powered, motor-encoded, and uses an optical fiducial.

5.2.10 Shutter

The shutter is a Lick standard device, identical to the shutter in the Lick prime focus camera (PFC), and similar to the shutter used in DEIMOS. The shutter is air-powered, with limit switches for both the open and closed positions.

5.2.11 Detector and Dewar

The detector is mounted in a fixed, Lick-standard dewar, similar in design to the Lick Prime Focus Camera dewar. The half-full hold time is at least 18 hours. The dewar must be refilled manually with liquid nitrogen. There are no automated adjustments for the dewar. An auto-fill mechanism will be considered.

5.2.12 Instrument Structure

Conceptually, the structure is a welded and bolted steel space frame, with threaded mounting points for all the major subsystems. The structure is kinematically attached to the Cassegrain instrument rotator, and must provide sufficient rigidity for the rotator bearing. There are lifting and handling points on the structure for installation and storage of the instrument.

5.2.13 Thermal Enclosure

The instrument is housed in a thermal enclosure which protects it against temperature variations, dust, etc. The enclosure is similar in concept to the HIRES and DEIMOS enclosures, and made of modular panels which are attached to the instrument structure. There are removable panels in the enclosure for access to the various instrument subsystems.

5.2.14 Cooling System

All major sources of heat in the instrument are actively cooled, using the telescope glycol system. These include:

The cooling system is similar in concept to HIRES and DEIMOS. Heat sources are mounted in enclosures with air-to-liquid heat exchangers. Temperature is sensed by the instrument computer, which cycles fans to control temperature.

5.2.15 Telescope Simulator

The instrument must operate at all rotation and elevation angles encountered at the Cassegrain focus of the telescope. Hence the instrument may see complete reversal of the gravity vector at a variety of elevation angles. A telescope Cassegrain focus simulator will be designed and built for testing the instrument during construction. The simulator will remain at Lick. The simulator is servo-powered in both rotation and elevation angle. The instrument will be thoroughly tested on the simulator prior to shipment to confirm performance under all operating orientations. The simulator will be able to operate the instrument horizontally, to test for possible operation at the Nasmyth focus.

Mounting of the instrument on the Lick 3m for further testing will be considered.

5.3 Electronic Systems

The instrument will make the fullest possible use of existing electronic components and software which have been used previously at Lick. Examples include Galil motion control servo-systems and motion control software. The electronic systems are mounted together in a temperature-controlled enclosure. The enclosure is attached to the instrument structure, and possibly within the main instrument enclosure.

5.3.1 Motion Stages

Commercial motion stages will be used whenever possible. Linear stages will be fitted with both software and hard (power off) limit switches. Rotational stages with greater than 360 degrees of travel use encoders and fiducials only. Travel will be encoded by motor encoder and/or high-resolution linear encoder if required. Optical fiducials are used on all stages.

Linear stages use the following hardware:

All motors are as follows: High-resolution rotating stages (slit mask wheel) also use the following: There is also "times 4" interpolation inherent in the Galil controller, yielding a total of 4000 pulses/revolution for motor encoders and 900,000 pulses/revolution for high-resolution encoders.

5.3.2 Flexure Control System

Flexure control will be provided by the collimator focus actuators. Position feedback will be provided by the motor-encoders and/or high-resolution encoders if required. Flexure will be open-loop controlled, based on a look-up table of flexure at various elevation and Cassegrain rotator angles. There is no provision for closed-loop control of flexure in the initial instrument configuration.

5.3.3 Air-Powered Stages

Air-powered stages will use integral switches whenever possible. Air-cylinders are operated by latching solenoids for minimum heat dissipation.

5.3.4 Environmental Sensing

Instrument temperature, air pressure, humidity, coolant temperature in/out, coolant pressure in/out, and coolant flow will be electronically sensed and logged by the instrument control system. Temperatures inside of thermal enclosures (electronics, calibration sources) will be monitored for active temperature control.

5.3.5 Connectors

In general, connectors will be consistent with DEIMOS design. Wherever possible, individual motion stages will be attached with local connectors to allow quick removal and replacement. Air and coolant connections will be made with no-leak quick-disconnect type fittings. Instrument to telescope connectors will also be consistent with DEIMOS design.

5.3.6 Thermal Control

Again, thermal control will be consistent with DEIMOS design. Large heat sources (calibration, electronics, CCD controller) will be actively cooled using temperature sensors, flow controls, fans, and the telescope glycol system.

5.3.7 Detector System

The instrument uses a single, thinned, back-side illuminated detector, which is one-eighth of the DEIMOS detector array. Detector electronics and control will be derived from the DEIMOS system. Detector electronics will be located as required by the detector, and outside of the main instrument enclosure if possible.

5.4 Control Functions

The following lists the control functions requiring servo-motor systems, high-resolution encoders, and/or air-powered actuators:

  1. Slit Mask Wheel:
  2. Slit Filter Wheel:
  3. Collimator Focus and Flexure Control:
  4. Imaging Mode Flat:
  5. Low-Dispersion Mode Flat:
  6. Post-Disperser Prism
  7. Camera Filter-Wheel
  8. Total Motor count: 6
  9. Total High-Resolution encoder count: 1
  10. Total Actuator count: 6

6 Instrument to Telescope Interfaces

ESI is a Cassegrain focus instrument, and attaches to the telescope Cassegrain rotator. The following sections detail the various interfaces to the telescope.

6.1 Optical

The optical interface to the telescope is a converging beam coming from the f/13.7 secondary mirror. The optical design assumes that the primary and secondary are aligned and spaced to give a nominal 1.38 arc-sec/mm plate scale at the Cassegrain focal plane.

6.2 Mechanical

The mechanical interface to the telescope is through the Cassegrain module instrument mounting flange. This flange is attached to the Cassegrain rotator bearing. Conceptually, the instrument will provide a stiff mounting ring to connect to the mounting flange. The instrument will be kinematically attached to the mounting ring, to avoid passing moment loads between the rotator bearing and the instrument. The kinematic mount will allow for collimation and alignment of the instrument with the telescope. The instrument mounting flange is metric, and ESI will generally follow this precedent (unlike HIRES and DEIMOS whose designs are based on English units).

6.3 Air

Compressed air consumption is estimated to be comparable to DEIMOS and HIRES; 10 CFM at 100 PSI.

6.4 Power

Power consumption for ESI has not yet been calculated in detail. However, based on the smaller number of motion stages and associated electronics as compared to DEIMOS and HIRES, ESI is estimated to consume somewhat less power than those instruments.

6.5 Cooling

The main thermal load areas in ESI are similar to those in DEIMOS and HIRES, consisting of the CCD electronics, instrument electronics, calibration sources, and the two CCD TV systems. All heat sources will be cooled to the ambient temperature, except for the electronics enclosure. Following DEIMOS and HIRES, ESI will use commercial grade electronics which are not rated for use below 5 C. In maintaining temperature in the instrument, ESI will utilize the full 1500 Watts of cooling allowed for instruments, at a nominal flow rate of 2.6 gallons per minute (GPM) and 80 PSI. The cooling connections will be made using Parker series 60 quick-disconnects.

6.6 Control

Hardware and software for instrument control are changing rapidly; any detailed description of the instrument control system and its interface to the telescope will change before the instrument is completed. Additionally, DEIMOS is faced with many of the same requirements, and eight times as many detectors per side. Consequently, the control interface is still to be determined, but will likely be a simplified version of the system developed for DEIMOS.

6.7 Data Handling

As with the control interface, ESI will follow the approach adopted by DEIMOS.

6.8 Instrument Handling Equipment

The main handling fixture will be a carriage for installing and storing the instrument. Storage boxes will be provided for elements such as the collimator, which will requiring periodic service or re-coating. Optical alignment tools will be provided as required.

6.9 Support Equipment and Facilities

No special support equipment or facilities are required. The telescope coating facility is sufficient for ESI reflective optics. Liquid nitrogen is required for the CCD dewar as usual. Vacuum and leak detection systems for the dewar will be required, should already be available for DEIMOS and HIRES. Video recording and playback facilities will be used to document installation and maintenance procedures.

6.10 Documentation

A full set of standard documentation will be provided with the instrument, including:

7 Data Processing

This section will evolve over time to provide a general description of the data flow for images and instrument control, in terms of software and hardware.

7.1 Data Flow

Specifications for the hardware and software required for acquiring and storing images are still to be determined, but will be based on one-eighth of the requirements for each DEIMOS detector array. The areas to be addressed include:

7.2 Data Acquisition

Specifications for the hardware and software required for acquiring data will also be based on the systems developed for DEIMOS. The areas to be addressed include:

7.3 Data Reduction

Specifications for the hardware and software required for reducing and storing data will be based on the systems developed for DEIMOS. The areas to be addressed include:

7.4 Instrument Control

Instrument control system will be based on recent Lick system designs, and areas to be addressed include:

Motion control and I/O will be provided with Galil systems, which are later versions of the components used in HIRES. The new components are listed with part numbers:

Motion stage and I/O control components
Quantity Part Number Description
1 DMC-1580-72 controls first 8 independent stages and offers 72 analog input/output bits for environmental monitoring
2 Galil AMP-1140 amplifier module for 4 independent stages
1 Lambda LR5-56 24 Volt, 23 Amp power supply for 8 stages

For control and minimum heat dissipation, air cylinder actuators are run by latching solenoids. These will be Skinner Magnelatch Valve: p/n H935RBM2150-120A. These are 110 volts AC. The air cylinders are dual acting (e.g., air pressure opens and air pressure closes them).

7.5 Computing Facilities

The computing requirements for the various instrument functions in terms of processing power, RAM, disc space, system displays, UPS, storage (tapes, optical discs), etc. are to be determined, and again, will be similar to the systems developed for DEIMOS.

The areas to be addressed include:

8.8 Facilities

The UCO/Lick technical facilities include several well-equipped laboratories which have produced a substantial number of complicated instruments for Lick, Keck, and other observatories and institutions. In general, no major additions to the Lick facilities are anticipated for the design and construction of the ESI instrument. However, some new test equipment will be required, such as the telescope simulator, and temporarily, a Keck Cassegrain instrument module.

8.9 Project Implementation and Control

Reports for the Keck Science Steering Committee (SSC) will be prepared quarterly by the project manager, with start dates TBD. Reports will summarize the activities of the previous quarter, compare progress against goals and schedule, review the status of risk items, set goals and milestones for the next quarter, and update the budget compared to the budget from the previous quarter and to the original budget. Reports should be completed within 30 days of the end of the quarter.

8.10 Design Reviews

Due to the relatively conventional size and design of the instrument (as compared to DEIMOS and HIRES), the ESI project will follow a more informal design review process. Additionally, the ESI project will be operating on an extremely short schedule, which will preclude formal review of the complete optical and mechanical design prior to the beginning of construction.

8.10.1 Preliminary Design Review

Instead, the ESI project team will hold a Preliminary Design Review (PDR) early in the project schedule, which will be presented to a select committee drawn internally from the Keck, CARA, and Lick communities. This review will present the conceptual optical, mechanical, electronic, and software design of the instrument, and will draw heavily on results and designs from concurrent UCO/Lick instrument projects such as the Lick PFC, MOS, the Lick/Keck HIRES image rotator, and the Lick/Keck DEIMOS. Tolerance and sensitivity analyses of the optical systems, preliminary structural designs, and optical and mechanical error budgets will be presented.

8.10.2 Progress Reviews

Instead, a series of progress reviews (PR) will be held to assess the progress of the design and fabrication of the instrument. These reviews will be small and informal. The reviews will address the progress of the optical, mechanical, electronic, and software development efforts. The reviews may concern one, several, or all of the various technical programs.

8.10.3 Pre-Shipment Review

A final, pre-shipment review (PSR) will be held following completion of testing of the instrument in Santa Cruz. This review will present the tested performance of the instrument, and the packing, shipping, and commissioning plans.

8.11 Integration and Test Plan

Mechanisms and components will be tested as they are completed. These units will be combined as sub-assemblies and tested as soon as all the parts are available. The sub-assemblies will be installed and tested using the instrument structure and telescope simulator in the UCO/Lick instrument lab. All features of the completed instrument will be thoroughly tested in the labs in Santa Cruz.

8.12 Critical Components Plan

At this time, the two critical components appear to be the optics and the instrument structure.

The final optical design of the instrument awaits publication of a new glass catalog by Ohara, and the availability of certain glasses may affect the final design. For the extremely short schedule of ESI, lead times for procuring the optics dominate the project critical path. If the Ohara glass catalog is not available in time, the optical designers will pursue alternative glass sources.

The stability and rigidity of the structure will determine the need for active flexure compensation, and the method of its control. If excessive structural flexure is elastic, the compensation system can run in an open-loop mode, based on elevation and rotation corrections from a look-up table. If the structural flexure displays significant hysteresis, a closed-loop control system with a source and a feedback sensor will be required. These options will only be followed if preliminary structural design and finite element analysis demonstrate that the optical and mechanical error budgets for flexure cannot be met by passive means.

9 Future Upgrades

The ESI wish list goes here...

Appendix 1. ESI Camera Design


Wilson, R. W., Jenkins, C. R., ``Theoretical point spread functions for modal adaptive optics'', Proc. SPIE, {\bf 2201}, 117-128, March 1994