1 Spectrograph Mechanical Design

1.1 Overview

This section describes the complete mechanical design of NIFS. It begins with the overall instrument assembly and works inwards to the smallest sub-assemblies. The order follows that of the optical path from the environmental cover through to the detector output. The discussion also deals with the components which are being duplicated from NIRI, the Gemini-supplied items, and the purpose-built handling equipment which is required to both assemble and test NIFS at RSAA and to routinely service NIFS at the Gemini telescope.

1.2 Changes Since CoDR

The following developments have occurred since the NIFS CoDR.

· The mechanical design has been developed further in all areas.

· The grating wheel has been replaced by a grating turret.

· A flip mirror mechanism has be added to allow object acquisition without moving the grating turret.

· Handling and alignment procedures have been developed.

1.3 Instrument Integration

Figure 43 shows NIFS ready for attachment to one of the side-looking ports of the Gemini Instrument Support Structure (ISS). The overall dimensions and mass of the assembled instrument are the same as for NIRI; length 2400 mm, width 3300 mm, height 1300 mm, and overall mass 2000 kg. The top part of the integration frame will lie just below the primary mirror cell when the instrument is attached to a side-looking port of the ISS. The instrument will be transported to the ISS on a scissor lift air pallet (ICD 1.5.3/1.9, Dwg 89-GP-1000-6100). One of the four handling pads which locate the instrument on the air pallet can be seen on the lower rail of the integration frame in Figure 43.

Figure 43: Complete NIFS assembly ready for attachment to a side-looking port of the ISS.

The large integration frame is at the center of Figure 43. This frame carries the thermal enclosure carrier frame, which in turn carries the two electronics thermal enclosures at its outboard ends. The ISS interface plate is shown at the lower left of Figure 43. The integration frame attaches to this plate and in turn to the ISS. Ballast weights will be attached to the integration frame to bring the instrument to the Gemini specified mass of 2000 kg and moment of 20 kNm (ICD 1.5.3/1.9, p. 4). The integration frame has a single strut at the top to allow the inside doors of the thermal enclosures to be opened without fouling on the frame. The patch panels that connect to the cryostat are in the lower inside faces of the thermal enclosures. These are readily accessible even when the instrument is on the telescope. The hexagonal cryostat is mounted in the integration frame. It can be seen at the center left of Figure 43. The cryostat houses the NIFS spectrograph in its lower part and the On Instrument Wavefront Sensor (OIWFS) in its upper part as seen in Figure 43. One of the two helium cryocoolers can be seen protruding from the cryostat. The SDSU-2 controller for the science detector can be seen just below the helium cryocooler.

NIFS can also be used as a spectropolarimeter in combination with the GPOL polarization modulator. GPOL is located in the bottom of the ISS above the up-looking port. NIFS will have to be mounted on the bottom face (Face 1) of the ISS when used in this way. To do so, the instrument will be rotated into the vertical position using the instrument handling crane in the Gemini dome. Sufficient cable and helium line will be provided to allow NIFS to be mounted on any ISS port.

1.4 Handling the Cryostat

There is limited overhead space in the RSAA workshop so we plan to integrate the instrument in the horizontal position. We believe that with suitable handling equipment this may be the natural way to handle the hexagonal cryostat during routine servicing at the telescope.

Figure 44: Integrating the complete 2000 kg instrument in the horizontal position.

Figure 44 shows the assembly method. The integration frame is carried on an air pallet located under an overhead crane. This air pallet has an open center and outboard air pads in order to leave the area under the cryostat free to the floor for the cryostat trolley. The cryostat trolley is a simple frame with four wheels and the assembled cryostat sits on adjustable jack pads on the frame. These jack pads allow the cryostat to be raised or lowered a few millimeters for leveling and alignment. To install the cryostat in the integration frame nylon slings are attached to four points on the center section. Next, the 600 kg cryostat is lifted 110 mm from the trolley and is passed through the open end of the integration frame with 20 mm vertical clearance. The cryostat will need to be rotated a little as it passes through the frame, and the frame moved sideways a little on the air pads to clear the outboard ends of the helium cryocoolers. The cryostat trolley is now passed under the integration frame, with just a few mm clearance, and the cryostat is lowered back onto the trolley. This places the cryostat on its trolley inside the integration frame and at the right height for attachment to the ISS interface plate. Lastly, the ISS interface plate with attached brackets is lifted into position using the overhead crane. All three heavy components are now brought together and attached. This completes the assembly of the large components.

When assembled, the cryostat window housing protrudes throughout the ISS interface plate. The environmental cover is now attached to the protruding housing. The environmental cover cannot pass through the hole in the ISS plate so it must be the last of the large components to be assembled.

Some servicing can be performed on the OIWFS while the cryostat is in the integration frame in the horizontal position. This is done by releasing the vacuum, removing the vacuum jacket end plate, and removing the two layers of radiation shields below the end plate. The OIWFS optics and detector are then accessible. Similarly, some servicing of the spectrograph can be performed while the cryostat is in the integration frame by using the Gemini instrument handling crane to turn the integration frame completely over. The purpose-built handling equipment shown in Figure 44 will be supplied to Gemini with the instrument.

1.5 The Cryostat

1.5.1 Vacuum Jacket

The vacuum jacket consists of three machined hexagonal forgings and two ribbed end plates. It is a slightly modified copy of the NIRI vacuum jacket. These modifications include moving the O-rings from the vacuum jacket center section to the outer two sections to facilitate assembly, omitting the NIRI detector controller feed-through ports, adding ports and attachment points for the NIFS detector controller, and adding lifting and trunnion attachment points for safe handling of the assembled cryostat. Outwardly the NIFS vacuum jacket will look little different from the NIRI jacket.

1.5.2 Cold Work Surface Plate

NIFS retains the NIRI-designed Cold Work Surface (CWS) plate suspended on titanium trusses near the mid point of the cryostat. Figure 45 shows the view looking in on the CWS plate with the OIWFS-side vacuum jacket components removed. The NIFS CWS plate is a slightly modified copy of the NIRI original on the OIWFS side, but is heavily modified on the spectrograph side to accommodate the different science instrument.

Figure 45: Looking into the open vacuum jacket at cryostat components on the CWS plate.

1.5.2.1 Cold Work Surface Plate Cooling

As noted in §6.6.1 of the NIFS CoDR documentation, the NIRI instrument has experienced some vibration coupling from the helium cryocoolers to the CWS plate. We have modified parts of the original NIRI design to make the cold strap links more compliant, to reduce the number of cold surface interfaces, and to reduce the total number of parts in the cold strap assembly.

Figure 46 shows the cold end of a Coolpower 130 cryocooler. Note that the cold strap assemblies have “S” shaped cold straps, and have copper conductor rings bolted directly to the flanges of the helium cryocoolers. Each cold strap is formed from 24 thin laminations. The laminations will be spaced with paper before bending to shape. When the paper is removed the combined laminated strap will be compliant in all x, y, and z directions.

Figure 46: NIFS helium cryocooler and CWS cold straps.

1.5.3 OIWFS Side of the Cryostat

The OIWFS side of the CWS plate carries the input baffle, the focal plane unit and the OIWFS proper. Figure 47 shows the layout of these components. Only half of the input baffle is shown here to allow a view of the pick-off probe. Only the Optable and Shack-Hartmann optics parts of the OIWFS are shown. The radiation shield and photon shield have also been removed from the OIWFS in this view.

Figure 47: Components on the OIWFS side of the CWS plate. Not shown are the OIWFS gimbal mirror, filter unit, and fold mirror.

1.5.3.1 Input Baffle

The cylindrical baffle shown around the pick-off probe in Figure 47 passes a 3′ diameter field to the OIWFS and prevents scattered light entering the cavity surrounding the OIWFS. NIFS does not use the NIRI beam splitter wheel so without this baffle there would be a large amount of scatter into the OIWFS cavity. The OIWFS radiation shield carries the input baffle. The input baffle does not contact the pick-off probe. The pick-off probe is cantilevered closely from the CWS plate in order that it be as stable as possible with respect to the spectrograph and to ensure that it becomes as cold as possible.

1.5.3.2 Spectrograph Focal Plane Unit

The spectrograph focal plane unit is an easily removable module containing the Focal Plane Mask Wheel and associated drive and encoding systems, the OIWFS test projector, the cantilevered pick-off probe, and the spectrograph focal converter and cold stop mirrors. This unit is the only part of the NIFS spectrograph located on the OIWFS side of the CWS plate. Figure 48 shows the important parts of the focal plane unit.

Figure 48: Principal components of the spectrograph focal plane unit.

1.5.3.2.1 Pick-Off Probe

The NIFS science field is only about 2 mm square at the telescope focal plane. The pick-off probe supports a pick-off mirror about 94 mm above the telescope focus. The probe picks off the small NIFS science field from the center of the much larger telescope field that continues to the OIWFS. The probe is 10 mm thick and carries two small diamond turned aluminum mirrors at the center of the telescope field. The upper of these mirrors folds the NIFS science field down into a hole drilled inside the probe and ultimately to focus at the Focal Plane Mask Wheel inside the focal plane unit. The feed hole inside the probe is threaded and painted with Aeroglaze black paint to suppress reflections. The closest usable unvignetted guide star to the center of the NIFS field is at a little over 8 mm or 12.7² from the field center. Figure 49 shows the pick-off probe and the f/16.2 cone from a star passing through to the OIWFS. Guide stars will pass through to the OIWFS at a little over 5 mm from the field center, but will be partially vignetted and will not illuminate the two-part Shack-Hartmann prism in the OIWFS properly. The OIWFS then will not deliver accurate focus corrections to the telescope. There may be usable position angles where the prisms are equally illuminated but this would probably require that the object be rotated with respect to the NIFS integral field unit (IFU) and this may not be scientifically desirable.

Figure 49: NIFS pick-off probe showing OIWFS guide star field.

It is essential that deflection of the pick-off probe not cause significant star movement in the NIFS focal plane. Analysis for the 175 g probe showed a deflection of about 1 μm at the probe tip (Appendix D, §13.2). While parallel deflection of the pick-off probe will not cause focal plane shift, rotation of the mirror will. Analysis showed a rotation of 6.3 μrad that would result in an image shift of about 1 μm in the telescope focal plane or about 1 milli arcsecond. This is much smaller than any likely image. The probe is very stiff in shear in both directions; the calculated shear deflections are just a few nanometers in any direction. This analysis was later confirmed by finite element methods.

1.5.3.2.2 OIWFS Projector

A second diamond turned mirror, facing downward from the end of the probe, folds a beam from a test projector built inside the probe to the OIWFS. This projector creates a 0.5″ star image in space at the focal plane of the telescope directly under the probe. This projected star will allow the OIWFS to be tested during daytime or when NIFS is not on the telescope. The projector will also be useful for daytime differential flexure testing of NIFS with ALTAIR when both instruments are on the telescope. The internally generated star image from ALTAIR can be steered past the pick-off probe to form a second spot at the telescope focal plane and both it and the NIFS test projector spot can be simultaneously imaged by the OIWFS detector. Differential flexure will then show as relative movement between the two images as the telescope is moved.

1.5.3.2.3 Focal Plane Mask Wheel

The Focal Plane Mask Wheel region of the focal plane unit is very crowded and has presented considerable design challenges. The twelve position Focal Plane Mask Wheel is located inside the focal plane unit, at the folded telescope focus. This wheel is completely enclosed in a cold aluminum housing that is sealed off from the OIWFS cavity by the pick-off probe carrier plate. The Focal Plane Mask Wheel rotates on lubricant-free stainless steel ball bearings and is driven by a three-stage spur gear drive from a Phytron stepper motor. The stepper motor runs in full step mode and only dissipates heat when moving the wheel. When power is off, the stepper motor sits at one of its magnetic detents. These are at 1.8° intervals and there are an even number of motor full steps between each mask position. The three stage reduction gear system has a ratio of 323.4:1. A motor full step then moves the center of any mask by 0.004″ at the telescope focal plane. There are no detent systems on the wheel but a magnetic drag brake firmly holds the wheel in its driven-to-position,. The magnets apply more force to the disk than the mass of the wheel plus disk, and lock the wheel in both x and y as well as locking it in rotation. The drag brake disk can be seen in Figure 50. This disk is thin hard steel and slides without lubrication on eight small and hard magnets. This system will produce very little swarf and we believe the small quantities that are produced will be retained by the magnets. It is obviously important that swarf does not drift onto the occulting disks. The wheel carries a three level two row Hall effect sensor encoder system that gives absolute encoding for each mask in the wheel. The Gemini-supplied stepper motors are very powerful and would destroy the fine pitch gearing if the wheel should jamb. To avoid this, the gear train contains a slipping clutch and a once per revolution Hall effect sensor on the first gear train for sensing a train jamb. More detail on the gear train drive can be found in Appendix D (§13.3). The grating turret uses a similar drive system. This is investigated more fully in Appendix D (§13.5).

Figure 50: Focal plane unit internal components. The cold stop mirror partly wraps around the Focal Plane Mask Wheel to provide a light seal into the spectrograph.

As well as being simple, the continuous, stop anywhere drive for the Focal Plane Mask Wheel has a number of scientific advantages. It allows occulting disks to be moved left or right across the NIFS field to position the occulting disk on the intensity peak of an asymmetric object. It is also useful for engineering tests as it allows a vertical slit to be scanned across the image slicer to test different regions of the IFU.

The Focal Plane Mask Wheel carries a clear aperture for general integral-field spectroscopy plus an array of masks for special uses and a blocked position which forms an instrument shutter. Apart from the clear mask, which will be an open thin metal frame, all other masks will be metalized sapphire or glass disks. Each disk will be 6 mm in diameter and 0.5 mm thick. Figure 51 shows the Focal Plane Mask Wheel and its masks along with its associated drag brake disk at near full size. Mask patterns will include occulting disks of various sizes, a 60 μm pinhole to simulate a star, and a vertical slit and a Ronchi screen for calibrating the spectrograph. During assembly the wheel will be placed under an alignment microscope to align the slit, occulting disks, and Ronchi screen,. This will be a time consuming operation. All twelve positions of this wheel will not be filled at delivery to Gemini. Extra apertures can be inserted at a future date by opening the cryostat and removing the focal plane unit.

Figure 51: Near full size view of the Focal Plane Mask Wheel.

1.5.3.2.4 Focal Converter and Cold Stop System

The focal converter mirror and cold stop mirror (Figure 52) form the fore optics of the spectrograph proper. The focal converter mirror performs the dual function of re-imaging the telescope pupil onto the cold stop mirror as well as re-imaging the science field and the surrounding focal plane mask onto the IFU image slicer at enlarged scale.

Figure 52: Focal converter and cold stop mirrors.

The focal converter mirror is spherical and is diamond turned into a 114 mm-diameter aluminum disk. A tilted mirror is arranged by cutting the spherical mirror off-center in the diamond turning machine. Only a small area of this mirror is used by the NIFS optical system. The large diaphragm shown in Figure 52 mounts and centers the focal converter mirror and acts as a flexure diaphragm. Three screws in the mirror blank strain against the diaphragm and adjust the mirror in tip-tilt to align the system pupil onto the cold stop mirror. The system also allows a small amount of axial adjustment to focus the entrance aperture onto the image slicer. The tilt adjustment screws are on a large radius and they provide a sensitive and stable adjustment system for the focal converter mirror.

The cold stop mirror is directly mounted to the focal plane unit housing without adjustment. This component is 36 mm in diameter to allow ease of handling and mounting. The mirror itself is just 4 mm in diameter and is diamond machined at 1.5° with respect to the component blank. This will require a wedge jig for the diamond turning machine. This small flat mirror is diamond turned without edge chamfer to provide a sharp edge that cleanly cut off the telescope pupil. A black painted 1 mm wide cavity is formed in the housing around the cold stop to provide a beam dump for rays outside the telescope pupil. Both the focal converter mirror and the cold stop mirror are deeply buried in the aluminium focal plane unit housing to ensure that the cold stop becomes as cold as possible.

Figure 53 shows the region between the focal converter and cold stop mirrors. The f/16.2 beam from the telescope passes through focus at the Focal Plane Mask Wheel and the beam expands to the focal converter mirror. This mirror converts the beam to f/256 and forms an image of the telescope secondary mirror on the cold stop mirror near the center of the figure. The cold stop then returns the f/256 beam and it passes out through a hole in the focal converter mirror and continues downwards to the spectrograph.. A stack of thin plate baffles between the mirrors prevent the spectrograph from bypassing the cold stop and seeing directly to the sky, this stack is enclosed in a cylinder that has only one input hole. Baffling near the imager slicer augments the cold stop baffling. This efficiently baffles the NIFS field.

Figure 53: Ray path between the cold stop and focal converter mirrors. The baffle stack prevents the spectrograph from bypassing the cold stop and looking out directly to the sky.

1.5.3.3 On-Instrument Wavefront Sensor

The NIFS OIWFS is a near copy of the NIRI original. Changes to the original design are detailed in the following sections. These changes are not expected to alter the OIWFS performance. The principal changes to the NIRI design are shown in Figure 47.

1.5.3.3.1 Optable

Since NIFS does not use the large NIRI focal plane mask wheel, we have retained extra metal in the Optable component that was originally removed to clear the NIRI focal plane mask wheel. This stiffens this Optable structure which will in turn stiffen the CWS plate and hence aid in stiffening the NIFS spectrograph. The Optable can be seen in Figure 47.

1.5.3.3.2 Shack-Hartmann Optical System

NIFS will not use the NIRI OIWFS focusing mechanism. We have replaced the NIRI focus mechanism and Shack-Hartmann lens housing with a new fixed module that carries the lens system and detector directly from the Optable.

This module includes manual only focusing for the OIWFS via slotted holes when the cryostat is open. Figure 54 shows the Shack-Hartmann optics housing.. A fiberglass plate and insulator washers provide electrical isolation from the OIWFS Optable for the entire Shack-Hartmann assembly.

Figure 54: Shack-Hartmann optical system housing and manual focus slide.

1.5.3.3.2.1 Shack-Hartmann Prism

RSAA is currently quoting for the construction of two-facet Shack-Hartmann prisms for NIFS, GNIRS, and NIRI. Figure 55 shows our preferred prism and prism cell option. The prism is 19 mm square and 3 mm thick. Each of the two faces slope outwards from the central roof edge at an angle of 0.366°. The taper from the roof edge to outer edge of the prism is 61 μm. The prism is placed in a well-collimated beam and at the system pupil. It divides the beam into two collimated sections. These rays partly cross over, enter a common camera, and form side-by-side images of the same object on the OIWFS HAWAII-1 detector. The principal difficulty with manufacturing a two-facet prism is polishing the shallow roof edge to a sharp corner. During final polishing at RSAA, an edge runner technique would be used to polish the prism surfaces. Optical shop trials are currently underway to determine surface figure errors adjacent to the apex of the roof edge. For this design, scatter and final image performance are entirely dependent on finish and figure near the roof edge. There should be no vignetting from a good quality two-facet prism.

Figure 55: Two-facet Shack-Hartmann prism and prism cell.

1.5.4 Spectrograph Side of the Cryostat

In Figure 56, the NIFS cryostat has been removed from the integration frame, rotated 180° from the position shown in Figure 45 using an overhead crane and returned to the cryostat service trolley. Then the vacuum jacket end plate and bottom section have been removed to reveal the NIFS spectrograph. Note that the spectrograph occupies a little over half of the space available inside the radiation shield. The flat panels in the radiation shield frame are not shown in this view. The radiation shield can be removed in one piece from the CWS plate by reaching down to the lower frame of the shield with long tools. The spectrograph vent is intended to quickly vent the spectrograph cavity in the event of sudden loss of vacuum as the cavity will only vent slowly through the pick-off probe, which is the only other air entry point.

Figure 56: NIFS cryostat spectrograph side up on the service trolley. The spectrograph is inside the radiation shield frame.

Figure 57 zooms in on the cryostat. The radiation shield, spectrograph cover, and skirt have been removed to reveal the spectrograph components on the science side of the CWS plate. If necessary the spectrograph cover and skirt can be removed from inside the radiation shield by reaching down with long tools.

Figure 57: View from above showing the spectrograph modules and towers distributed around the cold work surface plate.

When assembled, the spectrograph towers, skirt, and cover are bolted to the CWS plate and the tops of the towers are bolted to the cover. This forms a very stiff low profile structure.

1.5.4.1 Spectrograph Modules and Towers

Figure 58 zooms in closer to the spectrograph. In this view the sheet metal baffling has been removed to clearly show the spectrograph modules and towers attached to the CWS plate. The spectrograph optical beam path passes upwards from the OIWFS side of the CWS plate and enters the filter-fold tower. It first passes through an order blocking filter in the Filter Wheel immediately above the CWS plate. Fold mirrors inside the filter-fold tower then fold the beam out to the left in Figure 58 and 90 mm above the CWS plate. The beam then passes rightwards to the tri-fold tower, leftwards to the image slicer, rightwards again to the tri-fold tower where it passes through the IFU mirror arrays and fold mirror, then leftwards to the Bouwers collimator mirror tower, rightwards to the tri-fold tower again, leftwards through the Bouwers collimator corrector tower to one of six gratings, then rightwards through the spectrograph camera to the science detector enclosed in the detector housing. Analog signal from the science detector passes down through ribbon cables to the CWS plate then rightwards under the skirt and radiation shield to hermetic D connectors in the vacuum jacket. Individual cabling now carries the signal to the SDSU-2 detector controller attached to the lower right exterior of the vacuum jacket.

Figure 58: Spectrograph modules and towers spread over the CWS plate. The ray path between the towers begins near the right edge of the filter-fold tower and zigzags down to the detector.

All three stepper motors inside the spectrograph cavity operate in full step mode. The motors will only dissipate heat for the time it takes to change a filter, grating, or to place the flip mirror. They will be powered off during detector integrations.

All of the parts inside the skirt, including the skirt cover and baffles, will be painted with Aeroglaze IR black paint. When the blocked position in the Filter Wheel is inserted in the beam, the NIFS spectrograph cavity will have a very low instrumental background as this aperture is the only entry hole into this chamber.

1.5.4.2 Filter-Fold Tower

The filter-fold tower carries the spectrograph filter wheel, the first and second fold mirrors, and some baffling. The Filter Wheel is located as low in the tower as is possible to maximize the f/256 beam footprint on the filter and to minimize the total beam envelope on the filter. This minimizes the effect of pinholes and non-uniformities in the filters.

1.5.4.3 Filter Wheel

Figure 59 shows the component outline of the filter mechanism. The Filter Wheel will accommodate a total of eight filters. This may include a metal blank to act as a cavity shutter. The filters shown in Figure 59 are 25 mm in diameter. The actual used area of the filters is small. The f/256 beam footprint on the filter is less than 4 mm and the total ray envelope is less than 6 mm in diameter. The beam from the cold stop is not perpendicular to the CWS plate so the filters are tilted by 3° in the Filter Wheel such that the filter surfaces is normal to the chief ray. This ensures that the filters are working at their design wavelength. Mounting the filters square to the beam may produce small filter ghosts but we believe that operating the filters at their design wavelengths is a higher priority.

Figure 59: View with parts of the filter-fold tower and some of the baffling removed to show the Filter Wheel, filter drive, and the first and second fold mirrors.

The Filter Wheel is driven by a three stage gear drive from a Phytron stepper motor (Appendix C, §13.4). Most of the drive components are copied from the Focal Plane Mask Wheel including the magnetic drag brake, slipping clutch, and the Hall effect sensor encoding system. This will again provide a simple, direct, stop-anywhere drive system.

1.5.4.3.1 First and Second Fold Mirrors

Two small flat diamond turned mirrors are mounted inside the filter-fold tower. The first of these mirrors folds the f/256 beam from the cold stop out parallel to and 90 mm above the CWS plate. The second mirror aims the beam towards the image slicer via the first facet on the tri-fold mirror and takes up path length so that the image slicer fits inside the cryostat. Both of these mirrors are bolted firmly to a common carrier plate and both can be rotated to steer the f/256 beam. This adjustment allows the 2 mm square f/16.2 focal plane aperture mask to be aligned onto the 29 mm square f/256 image slicer.

Figure 60 shows both of these mirrors set for some FEA analysis used to determine if bolt clamping forces will distort the active area of the mirrors. The axis mirror is about 20 mm high and the take up mirror about 30 mm high. The active area on the diamond turned mirrors is small and well away from the mounting surfaces. Analysis showed that the deformation was small. Both mirrors pivot about a pin near the optical surface and both blocks are extended and have an adjustment point at their extremities to give fine control of the mirror adjustment.

Figure 60: NIFS first and second fold mirrors under FEA analysis. The beam footprint on each mirror is near 9 mm in diameter.

1.5.4.4 Tri-Fold Tower

The tri-fold tower carries the tri-fold mirror and the IFU pupil and field mirror arrays. The tower components will be assembled and aligned on an optical bench. The tower is bolted firmly to the CWS plate without adjustment. Figure 61 shows a partly dismantled tri-fold tower and the crowded beam path around the pupil and field mirror arrays.

Figure 61: View with parts of the tri-fold tower removed to show the tri-fold mirror and the pupil and field mirror arrays.

1.5.4.4.1 Tri-Fold Mirror

The tri-fold mirror is a 90 mm square aluminum block with three flat mirrors diamond machined into one side of the part. These three mirrors fold the beam from the filter-fold tower out to the image slicer, from the field mirror array out to the Bouwers collimator mirror, and from the collimator mirror out to the grating. The tri-fold mirror is aligned to the tri-fold tower on an optical bench and firmly bolted to the tower without adjustment.

1.5.4.4.1.1 Tri-Fold Mirror Machining

We envisage machining the tri-fold mirror by fly cutting the three surfaces in a diamond turning lathe. The angular tolerances between the mirror facets needs to be held with 0.01°. This can be achieved with a good rotary table in the diamond turning fly cutting machine. We can use the tri-fold tower to support the mirror for machining in a diamond turning lathe. Figure 62 shows the mirror reversed on the tower ready for fly cutting. The angles between the facets of the mirrors are near 7° and 16°.

Figure 62: Diamond fly cutting the tri-fold mirror using the tri-fold tower as a support in the diamond turning machine.

1.5.4.5 Pupil and Field Mirror Arrays

Perhaps the biggest challenge for the entire NIFS project is the construction of the IFU, and perhaps the most difficult part of the IFU to construct are the pupil and field mirror arrays. Each of the arrays has 29 toroidal mirrors diamond machined into an aluminum bar 4.5 mm thick. Diamond machining these mirrors requires an x,y,z axis diamond turning machine and there are few of these machines in existence. The University of Bremen in Germany has a Moore 500FG freeform diamond turning machine. They are currently testing methods for cutting the mirror arrays on this machine.

Figure 63 shows the mirror arrays and their support structure removed from the tri-fold tower. The mirror elements are arrayed on curves along the aluminum bars such that the arrays are concentric about the center of the image slicer. The resulting arrays form a concave set on the pupil mirror array, and a convex set on the field mirror array. The individual mirrors in the arrays are all concave. The arrays face each other and are about 27 mm apart in their support structure. We plan to machine and pin the arrays to the mount such that they are pre-aligned, but we have included some adjustment in the support for each array in case pre-alignment cannot be achieved.

Figure 63: Pupil and field mirror arrays in their support structure. The mirror arrays face each other and are separated by about 27 mm.

The pupil mirrors convert the f/256 beam from the image slicer elements to f/16.2 and form a staircase slit on the surfaces of the field mirrors. The field mirrors then form a common pupil at the grating. The mirror array blanks have a natural frequency in excess of 1000 Hz (Appendix D, §13.7) so we do not expect any cryocooler-induced oscillation to affect the arrays.

1.5.4.5.1 Mirror Array Machining Method

As noted in §5.5.4.5, the mirror arrays will be cut on an x,y,z freeform diamond turning machine. Figure 64 & Figure 65 show approximately how this will be done. The mirror blank is held in a jig on the bed of the freeform diamond turning machine. A small boring tool is placed in the spindle of the machine and the machine proceeds along the Computer Aided Machining (CAM) path and bores both the alignment pin holes. The machine now stops for a tool change to the diamond fly cutter then proceeds along the CAM path to form all 29 mirrors. The vertical strokes in the CAM path in front of each mirror curve towards the blank with a radius that is slightly different to the fly cutter radius. The combined action of the CAM path and fly cutter forms the desired toroidal mirrors.

Figure 64: Pupil mirror machining. This figure is a transparent wire frame view showing the mirror facets generated on the hidden edge of the blank.

Figure 64 shows the pupil mirror array being generated and Figure 65 shows the field mirror being generated. The length of the mirror arrays is about 60 mm. Note the differences between these figures: The pupil array is machined with a fly cutter of shorter radius then the field mirrors. Not also that the total pupil mirror array is concave when viewed from the fly cutter, while the total field mirror array is convex when viewed from the fly cutter. Lastly note that the looping CAM paths that the fly cutter follows are of very different lengths.

Figure 65: Field mirror machining. This figure is a transparent wire frame view showing the mirror facets generated on the hidden edge of the blank.

1.5.4.6 Image Slicer Tower

The image slicer tower carries the f/256 image slicer. Figure 66 shows the tower assembly. This tower is a vital part of the NIFS IFU. The slicer stack itself consists of 29 small spherical mirrors sandwiched together. The tower consists of the slicer stack plus top and bottom mounting blocks. Initially the top and bottom blocks are manufactured thicker than required and are assembled with the image slicer stack on an optical table. Now the center height and tilt of the center spherical slice is measured using an alignment telescope. Next both blocks are individually machined to bring the center height of the central slice to 90 mm above the optical table, to correct the tilt of the central slice, and to set the total height of the tower to 180 mm. The entire tower can be rotated on a central pin in the bottom block for rotation alignment when attached to the CWS plate.

Figure 66: Image slicer tower assembly with slicer baffle removed. The tower is 180 mm high.

1.5.4.6.1 Image Slicer Stack Construction

The image slicer stack is constructed from two 12 mm thick aluminum anvils. These sandwich the 29 aluminum slices. Each of the 29 slices has a mirror formed in one edge during diamond machining. Each mirror measures about 1×29 mm. Two sets of 10 mm diameter dowel holes on 100 mm centers pass through the stack. One set aligns the stack for diamond machining the 600 mm radius spherical surface to form the mirrors. The second set aligns the stack with a progressive rotation of about 0.1° between adjacent mirrors to form the fanned stack.

Figure 67 shows the fanned stack with the top slice removed. Bolts that compress the stack have been omitted for clarity. To the right of the figure is the top slice turned into plan view. This plan view shows the machining pin hole pairs at the top of the slice and fanned pin hole pairs at the bottom of the slice. The lines from the lower pair of holes are horizontal and show the angle of the extreme slice, this angle is near 1.8°. All four pin holes in each slice are jig bored for each individual position before the stack is assembled. When assembled the slices fan about a vertical line called the fanning axis that is square to the anvils and passes through the face center of the central slice.

Figure 67: Image slicer stack construction. Slice +14, the top slice, is shown removed from the stack.

1.5.4.6.2 Image Slicer Stack Machining.

As the optical center of spherical image slicer mirrors is offset with respect to the image slicer stack module, two of the stacks can be made in one setup on a diamond turning lathe. We intend to manufacture two of these assemblies to ensure that NIFS has a spare of this critical optical component. Figure 68 shows two image slicer stacks on a common plate and shows the separation of the modules during machining. The carrier plate in the figure is about 200 mm in diameter. Note that the diamond machined spherical surface partly continues into the left and right anvils.

Figure 68: Diamond machining two image slicer stack assemblies in one operation.

1.5.4.7 Bouwers Collimator Tower

The Bouwers collimator tower (Figure 69) carries the largest and perhaps the simplest optical component in the NIFS spectrograph. The tower consists of an aluminum block measuring 52×142×60 mm into which the spherical collimator mirror is diamond machined and the top and bottom mounting blocks. Initially the top and bottom blocks are manufactured thicker than required and are assembled with the collimator mirror on an optical table. The center height of the spherical mirror is now measured using an alignment telescope. Next, both blocks are individually machined to bring the center height of the sphere to 90 mm above the optical table and the total height of the tower to 180 mm. The entire tower can be rotated and moved on a central pin in the bottom block for focus and rotation alignment when attached to the CWS plate.

Figure 69: Bouwers collimator tower with baffle removed. The tower is 180 mm high.

1.5.4.8 Bouwers Corrector Tower

The Bouwers corrector tower (Figure 70) is the simplest of the spectrograph towers. It carries the collimator corrector lens, lens clips, and lens baffle. The lens is a circular Calcium Fluoride disk with flats on opposite sides to clear the collimated beam and the camera body. The lens is retained and aligned by two semi circular plates. Coil springs in the body of the tower force the lens forward to seat onto the semi circular plates. The entire tower can be rotated on a central pin in the bottom block for rotation alignment when attached to the CWS plate.

Figure 70: Bouwers corrector tower. The tower is 180 mm high.

1.5.4.9 Grating System

The design of the NIFS grating system has presented the biggest mechanical design challenge of all of the NIFS mechanisms. Only an overview of the system design will be presented here. More detail on the system including a discussion of positioning accuracy and stability, which is near 300 nm, is presented in Appendix D (§13.5). Figure 71 shows the grating system assembly. The grating turret rotates inside the stationary cylindrical grating cover. The turret carries the gratings and is in turn carried on a vertical 40 mm diameter shaft that spans from the CWS plate to the spectrograph cover. The turret rotates on the stationary shaft on a pair of preloaded angular contact bearings. This forms a very stiff structure and allows precise rotation control for the grating turret and in turn the gratings.

Figure 71: Grating module assembly. Six gratings are housed inside the cylindrical cover, only the active grating is visible through the grating aperture.

In Figure 71, one of the gratings can be seen through the active grating aperture in the cover. This cover protects the gratings during alignment and baffles the grating pupil to a rectangular beam that captures the diffracted beam from the image slicer.

1.5.4.9.1 Gratings and Grating Adjustment

The grating turret carries six gratings that are replicated on aluminum substrates. These gratings are longer than both the geometric and diffracted pupils in order to allow some rotation of the turret without the pupil moving off the end of the grating. Each grating is attached to the turret via tilt and face rotation adjustment mechanisms. The third adjustment required to fully adjust the grating position is supplied by the continuous rotation around the turret support shaft. Figure 72 shows the grating adjustment system. Face rotation and out-of-plane tip are achieved by flexing aluminum bars that connect the grating blank to the turret and act as stiff springs. These bars also act as thermal shunts to cool the gratings. The grating blank surface shown in Figure 72 is shown as horizontal but will in fact be individually cut to suit the particular grating blaze angle. This surface angle will vary by ±4° from the horizontal. The extremes of blaze angles are still accommodated inside the grating cover.

Figure 72: Grating flexure mounting and adjustment system.

1.5.4.9.2 Grating Drive System

The grating drive system uses a three stage spur gear train to rotate the grating turret and a friction drag brake directly on the turret to hold the turret in its driven-to-position (Figure 73). This “Drive it to any angle and leave it” system has a big advantage over a grating turret with discrete detented positions. This system will allow the gratings to be driven to their zero order position for set up purposes, allow spectral regions of interest to be driven to preferred positions on the detector, and allow low dispersion gratings to be used at other than their intended wavelength regions in higher orders.

Figure 73: Grating module internal components. The turret drive begins at the stepper pinion and proceeds rightwards to the turret. The friction drag brake is just visible under the turret.

The grating drive system uses a 1008:1 three-stage spur gear drive system to give one pixel shift per motor full step at the detector. Figure 73 shows the gear train and slipping clutch systems inside the drive gear housing. This housing slides in a tongue and groove on the CWS plate to adjust the final gear mesh to the turret drive gear. The large diameter hard metal-to-metal friction drag brake effectively holds the turret in position and helps to cool the turret and in turn the gratings.

1.5.4.9.3 Flip Mirror Drive

The flip mirror system places a flat mirror in front of the grating. This mirror reflects the collimated beam directly into the camera to re-image the undispersed IFU staircase directly onto the detector. The detector image of the staircase is then collapsed in software to form an image of the objects in the telescope focal plane. This allows NIFS to directly image the astronomical object of interest, to align the object in the aperture, and to align occulting disks to the object. Figure 74 shows the flip mirror assembly with the mirror in the down or “out-of-beam” position. More detail on the flip mirror design can be found in the Appendix D (§13.6).

Figure 74: NIFS flip mirror module. This figure shows the mirror in the out position with the diamond turned face of the blade facing down to the CWS plate.

In Figure 74, a stepper motor drives the mirror shaft via a 4:1 reduction gearing and a torsion spring. With the motor power off, the mirror is then spring loaded against hard stops in both the inserted and retracted positions. The stop for the inserted position is adjustable to control the angular position of the mirror and hence the position of the staircase on the detector. By driving the motor 10° past the nominal (90°) positions in either direction sufficient preload is developed to counteract the changing effects of gravity as the telescope is rotated. This spring loading is adequately reacted by the power off motor detent torque. Hard stops are provided to prevent over-winding the torsion spring. These limit the driven gear rotation to 110°.

1.5.4.10 Baffling System

Figure 75 shows the principal components of the NIFS baffle system. A large part of the baffle system consists of conforming apertures cut in thin sheet metal. This system approaches as closely as possible an infinite cavity system where scraper baffles surround the beam. Any scatter from the baffles enters the nearby infinite cavity and does not return to the optical system. The main baffle stack consists of U-shaped pieces of sheet metal with turned out bottom edges that attach directly to the CWS plate. The baffles attach only to the CWS plate and not to the spectrograph cover. Struts between the tops of the metal sheets stiffen the assembled structure so that it does not vibrate. All three stepper motors are under sheet metal covers to prevent radiation from the windings scattering into the spectrograph cavity.

Figure 75: Principal components of the NIFS baffle system. The system approximates an infinite cavity baffle.

A zero order trap collects and dumps the bright zero order light from the high dispersion gratings. There are open apertures above and below the tri-fold mirror. The tri-fold mirror baffle prevents scattered radiation that passes the tri-fold mirror from reaching the detector area. The spectrograph camera moves with respect to the detector to focus the spectrograph. There is a small labyrinth baffle at the join between the camera and detector housing to prevent scattered radiation in the cavity from bypassing the camera and directly reaching the detector.

Figure 76 is one of many section views used to visually evaluate the baffling system. In the figure, part of the first tri-fold mirror facet can be seen through the baffle aperture. Just to the right of this mirror is the facetted face of the pupil mirror array and the rear edge of the field mirror array.

Figure 76: Baffle evaluation. Looking through the image slicer baffle aperture to the pupil mirror array.

Figure 77 looks through the Bouwers collimator mirror baffle onto the tri-fold mirror. Two of the mirror facets can be seen through the baffle. The Bouwers collimator is the largest optical element in NIFS and has the tallest baffle aperture. Baffle apertures to the right of the collimator baffle are progressively less tall. The minimum aperture height is reached at the grating pupil just in front of the spectrograph camera.

Figure 77: Baffle evaluation. Looking through the Bouwers collimator mirror baffle.

1.5.4.11 Spectrograph Camera

The NIFS spectrograph camera is a five element in-line refractive system. These five lens elements are housed in a three part camera body. The first two parts of the housing are made from aluminum and carry the first four lens elements. These two parts of the housing have integrated towers that reach down to the CWS plate and up to the spectrograph cover. Figure 78 shows the camera assembly. A very stiff structure is formed for the refractive optical system when these two parts of the housing are bolted together and to the CWS plate and cover.

Figure 78: View of spectrograph camera. The integrated towers are 180 mm high. The camera 30 K cavity is cantilevered from the right hand end of the barrel.

A purpose-built wave washer and a tangent contact washer seat against the lens surfaces to hold each of the camera lenses in place. All of the lenses are loose radially in the barrel at ambient temperature but have zero diameter clearance at the 65 K operating temperature. During the cooling process, the lens temperature will lag behind the housing temperature so this mounting system has the potential to damage lenses through differential contraction. Detailed thermal analysis shows that this is not a problem, however, as explained in Appendix D (§13.9). Additional analysis of the lens retention system is presented in Appendix D (§13.8).

1.5.4.11.1 Camera 30 K Cavity

Reaching rightwards from the camera barrel in Figure 79 is the camera 30 K cavity. This 30 K cavity begins as a thin wall stainless steel tube that carries the silica camera field flattener lens at its rightmost end. A copper ring at the end of the thin stainless cylinder thermally connects to the 20 K tie point via a cold strap. This tie point then connects through a hole in the CWS plate to the helium cryocooler second stage. This same tie point also connects to and cools the detector housing. An aluminum detector baffle cylinder with internal baffles reaches rightwards from the copper ring almost to, but does not touch, the detector housing. This baffle cylinder and the field flattener lens form the 30 K cavity. Figure 79 shows the lenses, 30 K cavity, detector, and detector mounting board.

Figure 79: Camera cross section showing 30 K cavity and detector.

The 30 K cavity has a dual purpose; it cools the silica field flattener lens to a temperature where it acts as a long wavelength cut-off filter and it presents the detector with a cold cavity that minimally contributes to the instrumental background as seen by a 5 μm cutoff detector. The aluminum baffle fits to a labyrinth baffle on the detector housing to seal the camera cavity to the detector housing but allows for focus movement between the two modules.

1.5.4.11.2 Camera Focus Adjustment

Each of the camera towers has a tongue protruding from the base and these mate to a slot machined in the CWS plate. This guides the camera during focus adjustment and aligns the stationary detector housing to the camera. When the towers are released from the CWS plate, a fine focus adjustment system on the CWS plate moves the camera barrel left-right in Figure 78 to set the focus with the necessary 60 μm resolution. Once in position, the camera is firmly bolted to the CWS plate and in turn to the skirt and spectrograph cover to form a very stiff structure.

1.5.4.12 Spectrograph Detector System

The NIFS spectrograph detector system consists of the detector support block, the detector housing, and the detector, detector mounting board, and associated detector cabling. The detector support block attaches to the CWS plate and the spectrograph cover to form a stiff support structure. Other detector components are mounted from this structure.

1.5.4.12.1 Detector 60 K Housing

The NIFS detector housing is a closed aluminum box that carries the HAWAII-2 detector, the detector mounting board, external JFET output amplifiers, and two temperature control systems. Figure 80 shows an exploded view of the detector housing. The entire detector housing is electrically and thermally isolated from the detector support block on fiberglass strips. The housing is cooled via a brass cold strap that connects from the housing down to the 20 K tie point near the CWS plate. This cold strap is electrically isolated from the detector housing at the housing end via Silpad strips and fiberglass washers to eliminate ground loops. A heater and temperature sensor system located near the cold strap attachment point controls the detector housing temperature at about 60 K and holds this temperature to better than 20 mK (§5.5.4.12.3).

Figure 80: Exploded view of the detector housing showing the detector and associated components.

1.5.4.12.2 Detector, Detector Mounting Board, and 65 K System

The detector mounting board (Figure 81) is a hybrid of rigid and flex circuits. It is carried on a rotating ring inside the detector housing via fiberglass standoffs. This ring allows the detector and mounting board assembly to be rotated a little to align the detector columns with the dispersed spectra. Flex circuits, that allow board rotation, connect tracks on the rigid part of the mounting board to connectors on the stationary detector housing. A tab in the lower left corner of the detector mounting board mates with two screws in the detector housing to allow fine rotation adjustment. Once adjusted, the ring is screwed rigidly to the detector housing.

The detector has a separate temperature control system on the rear side of its mounting board. The fiberglass standoffs on the rotating ring electrically and thermally isolate the detector mounting board from the detector housing. An aluminum strip cold strap about 0.6 mm thick connects the central copper block on the detector mounting board to the inside of the detector housing. Silpad strips and fiberglass washers are used at the detector housing end to electrically isolate this copper block.

The center of the detector mounting board is cut away (see Figure 80) so that the outer active rows of pins from the Yamaichi socket can be soldered to the board while the inner group of pins pass through the board and are soldered into a copper block that controls the detector temperature to 1 mK.

Figure 81: Rear side of the detector mounting board showing the JFET amplifiers and the detector temperature control system.

Figure 81 shows the assembly of the detector mounting board from the rear. At the center rear of the board is the detector temperature control system and the JFET amplifiers. A temperature sensor and heater maintain the 150 g copper block at about 65 K with a RMS temperature variation of about 1 mK. Note that the detector is a little warmer than the surrounding detector housing. The copper block itself is supported from the central 220 pins of the Yamaichi zero insertion force (ZIF) socket, to which the block is soldered. The block cools the detector through these 220 passive pins in the center of the HAWAII-2 ceramic chip carrier. The copper block does not touch the circuit card, and the block is supported through the Yamaichi socket. The socket, detector, and copper block are supported by the outer 216 active pins soldered into the mounting board. Some of these pins can be seen in Figure 81. Eight JFET amplifiers are clamped to the copper block and are cooled and temperature controlled along with the detector to about 1 mK. Leads from the JFETs pass down about 7 mm to holes in the circuit card. These will be flexible enough to allow for the 0.15 mm temperature-induced differential strain between the copper and fibreglass parts.

1.5.4.12.3 Detector Temperature Control

Figure 82 schematically shows heat flow paths around the HAWAII-2 detector. At the center of the figure is the HAWAII-2 detector. To the right of the detector is the detector circuit card and to the right of this is the 150g detector copper block. These are enclosed in the detector housing which is coupled to the detector copper block and the 20 K tie point of the cryocooler by separate cold straps. Heaters are attached to both the detector copper block and the detector housing. The detector copper block is mounted via fiberglass isolators to the inside of the detector housing. The detector housing to mounted via fiberglass isolators to the detector support structure. Flex circuits from the detector mounting board couple to the detector housing and then pass to thermal shunts on the CWS plate and ultimately to the cryostat wall.

Figure 82: Detector heat flow schematic.

The dominant heat flow is from the detector copper block via an aluminum cold strap to the detector housing and then via a brass cold strap to the 20 K tie point. The detector copper block is heated to 65 K and controlled to ±1 mK by a coupled resistor that continuously dissipates ~ 100 mW. This value is chosen to be large compared to the 2 mW that is dissipated when the detector is read out. Of the ~ 100 mW, ~ 81 mW flows directly from the copper block via the cold strap to the wall of the detector housing. A further ~ 16 mW flows along the copper tracks of the flex circuits to connectors attached to the detector housing. If these connectors are efficient thermal shunts, this heat will be conducted to the detector housing. However, the connectors may act more as thermal isolators, in which case this heat will be conducted to the shunts on the CWS plate. The CWS plate shunts operate at close to the detector copper block temperature so the heat flow will be small. Any heat flow through the flex circuits must first pass from the detector copper block into the detector via the central 220 passive pins of the ZIF socket, then out of the detector through the surrounding 216 active pins to the detector mounting board. Radiative transfer is insignificant as there is only 5 K temperature difference between the detector copper block and the detector housing.

The total heat flow along the brass cold strap to the 20 K tie point is ~ 160mW.

The 150 g mass of the detector copper block is chosen such that the 2 mW dissipated for 10 s during detector readout is sufficient to raise the temperature of the copper block by no more than 1 mK.

The 100 mW detector copper block heater is sufficient to raise the detector copper block temperature from 60 K to 65 K in ~ 20 min.

1.5.4.12.4 Detector Cabling

Two separate flexible circuits, one for clocks and one for biases and signals, lead from the rear of the detector housing (Figure 83). These flex circuits pass down to the CWS plate, under the skirt, under the radiation shield, through the floating shield, and then to two hermetic D connectors in the vacuum jacket wall. Both of the flex circuits are clamped separately between curved aluminum blocks where they pass under the skirt and radiation shield bridges. These curved blocks act to shunt heat flowing in along the flex circuits from the detector vacuum sub-plate to the 65 K CWS plate. The curved block system also provides light seals for both the radiation shield and spectrograph cavities.

Figure 83: View on the rear side of the detector housing showing the flex circuits and vacuum sub plate. The circuits will be twisted by 18° where the micro D connectors mate.

Figure 83 is a view of the rear side of the detector housing. The flex circuits are manufactured from black-filled Teflon fabric and have narrow copper tracks and a mesh ground plane. These flex circuits are 0.3 mm thick and 20 mm wide where they begin at the detector housing and expand to 40 mm wide at the detector vacuum sub-plate. Both the radiation shield and spectrograph skirt have V-shaped apertures that mate over the thermal shunt bridges to form an effective light seal that requires leaking light to pass around four corners to enter the cavities.

Figure 84: Exploded view of the flex circuit thermal shunt system. Shunt load springs compress the shunt blocks and sandwich the flex circuits in each bridge.

Figure 84 shows an exploded view of the thermal shunts. The shunt blocks are 12 mm wide and have a 10 mm radius of curvature. This curve provides a light seal that is about six times the 0.3 mm thickness of the circuits. During servicing a sheet metal bracket clips to the top of the skirt and tabs on the bracket clamp the free ends of the flexible circuits back to the wall of the skirt. This frees space to manoeuvre the detector housing and ensures that the fragile flex circuits are not damaged when removing or replacing the detector housing.

Cabling from the outside of the vacuum sub-plate to the SDSU-2 detector controller will be a mixture of individual wires and co-axial cables.

1.5.4.12.5 Detector and Detector Housing Safe Handling

It is vital the HAWAII 2 detector be safely and easily placed in the detector housing and the housing be safely assembled in the spectrograph. Detector housing assembly begins in a clean room where the detector mounting board and rotator ring are assembled in the detector housing and the flex circuits are connected through the rear wall of the housing. Shorting plugs and ribbon cable are now installed on the rear side of the housing to safely short the detector card. Next the housing is placed rear side down on a table. With full anti-static precautions, the detector is loaded into the Yamaichi ZIF socket. Lastly the detector housing front plate is mated to the housing with a cover over the detector hole and attached with many small screws. The detector housing handling tool (Figure 85) is now attached to the detector housing. The cover blade is slid over the naked detector and the assembly is stood up on the runners at the bottom of the housing. The housing is now ready for installation in the spectrograph.

Figure 85: Safe handling of the detector housing.

Figure 85 shows the detector housing, handling tool, and the assembly being mated to the spectrograph. To the right of the figure is the detector housing and the handling tool. The left panel of the figure shows the housing being lowered into position in the spectrograph. At assembly, the housing is lowered until the housing runners contact the CWS plate. The housing is then maneuvered as close as possible to the detector support structure. Now the detector cover blade is lifted to expose the detector to the camera. The assembly is lifted 1 mm. The runners now clear the CWS plate, and the assembly is pushed towards the camera to mate the detector housing attachment pins to the detector housing attachment lugs on the detector support structure. Four screws in the lugs are now tightened to firmly attach the detector housing to the support structure. Then the cold strap is attached to the 20 K tie point. Now with full anti-static precautions, the shorting straps are removed, the detector flex circuits are connected, and the handling tool is removed from the detector housing. The pin and lug attachment system allows the detector housing to be removed and replaced with a positioning accuracy of better than 20 μm at the pins. This should allow the detector to be removed for servicing and returned without re-alignment of the spectrograph.

This detector handling system will allow the detector to be serviced in an integrated instrument. Figure 86 shows NIFS turned completely over using the Gemini instrument service crane so that the spectrograph is in the upper part of the cryostat. The vacuum jacket end plate, radiation shield panel, and spectrograph cover are then removed. The detector housing is now accessible and can be removed for service and the camera can be re-focused, if necessary. Much of the spectrograph can be serviced in this position including all of the mechanisms. The only parts that cannot be serviced here are the optical components that would require re alignment of the spectrograph.

Figure 86: Servicing the detector in an integrated instrument.

1.6 Mechanical Assembly and Optical Alignment

NIFS can be assembled and aligned piece-by-piece or opened and re-aligned with the minimum of disassembly. The text here will assume both initial assembly and alignment and complete assembly and re-alignment. The figures in this section show complete assembly of the spectrograph with only minimal parts removed to allow alignment telescope access as this conveys a better understanding of the instrument. A dummy CWS plate will be built and installed on an optical table in the optics laboratory. This will be used for pre-alignment and testing prior to assembly. Most of the sheet metal baffling can be removed and installed to provide access at any time during the assembly and alignment procedure. Full electrical control of the mechanisms will be required during alignment. Consequently, the component controller thermal enclosure will need to be in the clean room with the cryostat.

Preparation for assembly will involve washing and cleaning all parts to high vacuum standards and the grouping of all these parts in a clean room or other clean assembly area. Assembly should begin with the placement of the floating shields and CWS plate in the vacuum jacket center section. The titanium tines need to be sprung inwards at assembly in order to place spacers under the outboard end of the tines. The CWS plate needs to be made accurately parallel to the vacuum jacket center section bolt flange. This may best be done with a large surface plate and gauge block stacks. Next, the external support components for the cryocoolers are attached to the center section and the cryocoolers are mounted on these. Inside the vacuum jacket, the cold ends of the cryocoolers are attached to the CWS plate via cold straps, the getter and sorption trap are installed, and the CWS plate heaters and temperature sensors are installed. Lastly the science end of the vacuum jacket is temporarily assembled such that the cryostat will stand on its service trolley as in Figure 45.

1.6.1 OIWFS Side of the CWS Plate

1.6.1.1 OIWFS

The OIWFS Optable is separately assembled in the optical laboratory. This contains the field lens, both flat fold mirrors, the combination lens housing, gimbal mirror unit, filter unit with apertures but minus the infrared band pass filters, Shack-Hartmann unit, and the OIWFS detector. The OIWFS Optable is now lowered onto the CWS plate using an overhead hoist and attached to the CWS plate. The gimbal mirror and filter units are now wired and tested. A Taylor-Hobson alignment telescope is now attached to the window port of the vacuum jacket center section (see Figure 87) and used to sight right through to the detector. Alternatively the detector can be temporarily removed, replaced with a target, and the target can be sighted through the entire optical system. With suitable illumination, the field defining apertures in the filter wheel will be simultaneously visible while viewing the detector so the alignment and focus position of these can be checked for all filter positions. The full optical system is designed to perform at 65 K but will still provide adequate imaging performance for alignment purposes at room temperature. Image quality and positioning accuracy of the apertures in the filter unit can be assessed and the filter unit slid around on the Optable to align and focus the apertures. Lastly, with full anti static precautions, the OIWFS detector can be wired and a first-estimate position of the focus position of the Shack-Hartmann optics can be set on the focus slide using warm-to-cold focus shift estimates from Zemax optical ray traces.

It is not possible to run the HAWAII-1 detector at room temperature so final focus adjustments for both the detector and the field-defining apertures in the filter wheel will be made after temperature cycling the instrument. It will be necessary to place both images from the Shack-Hartmann prism as near as possible to one central corner of one of the detector quadrants. The detector is adjustable in x,y position to achieve this detector positioning. Initial adjustment will be made at room temperature using an alignment mask mounted in the detector socket.

1.6.1.2 Focal Plane Unit

The NIFS focal plane unit is now installed on the CWS plate, connected to wiring, and tested.. There are no NIFS alignments on the OIWFS side of the CWS plate other than the OIWFS test projector in the focal plane unit. This alignment is preset in the optical laboratory before installation in NIFS. The NIFS input baffle, OIWFS photon shield, radiation shield, all of the wiring, and the vacuum jacket front and end parts are now assembled.

The cryostat is now turned over with aid from the overhead hoist and placed back on the service trolley OIWFS-end down.

1.6.2 Spectrograph Side of the CWS Plate

1.6.2.1 Optics and Towers from the Mask Wheel to the Image Slicer

The cryostat is now opened on the spectrograph side down to the vacuum jacket center section level. The Taylor-Hobson alignment telescope is mounted on the entrance window port and adjusted to be square to the vacuum jacket. A spot light is now pointed backwards down through the hole in the focal converter mirror to illuminate the cold stop mirror. The alignment telescope is now focused onto the cold stop. This will require that the telescope be focused beyond infinity to about -18 m where the telescope secondary mirror would be. The Taylor-Hobson telescope will reach this focus. Now the focal converter mirror is adjusted on its three-point mount to center the cold stop. Figure 87 shows the required setup. The focal converter mirror is flush with the CWS plate in this figure.

Figure 87: Adjusting the focal converter mirror to align the pupil onto the cold stop mirror.

After alignment of the cold stop, the filter-fold tower is installed on the CWS plate. This covers the focal converter mirror and its adjustment system. The filter drive system is installed, wired, and tested. The tri-fold tower is now installed. The IFU pupil and field mirror arrays that it carries have been machined to provide alignment and this alignment has been pre-tested in the optical laboratory.

The Taylor-Hobson alignment telescope is now mounted on a bracket on the side of the vacuum jacket and arranged to view the f/256 image slicer focus (Figure 88). A lamp source spot projector is placed on the entrance port of the vacuum jacket center section to illuminate the NIFS Focal Plane Mask Wheel via the pick-off probe. The image quality at the image slicer focus is now visually assessed using the 60 μm pinhole in the Focal Plane Mask Wheel as an input object. This setup essentially tests the surface figure of the focal converter mirror, cold stop mirror, the two fold mirrors in the filter-fold tower and the first facet of the tri-fold mirror. Now a glass target marked with centerlines and the edges of the image slicer is placed at the image slicer focus. The Focal Plane Mask Wheel is rotated to place the standard field mask at the mechanical center of the instrument. Now both of the fold mirrors in the filter-fold tower are adjusted to place the image of the field mask in the Focal Plane Mask Wheel onto the temporary image slicer target.

If all of the optical system through to the detector are in place, the remainder of the optical system through to the detector can be inspected at this time by placing the flip mirror in the beam or setting one of the gratings to zero order.

The image slicer is now installed on the CWS plate and the alignment telescope removed in preparation for the next phase of alignment.

Figure 88: Aligning the f/256 beam at the image slicer. Inspecting image quality at the image slicer focus.

1.6.2.2 Optics and Towers from the Image Slicer to the Grating Pupil

After installed the image slicer, the collimator mirror tower and the collimator corrector tower are now installed on the CWS plate. The optical system is now complete to the grating pupil. The grating turret and grating cover are not installed, this allows the collimated beam to pass by the grating turret shaft and out of the instrument about 20 mm above the edge of the vacuum jacket center section.

The vertical slit in the Focal Plane Mask Wheel is rotated into position and the lamp house spot projector over the cryostat window provides illumination to the optics through this slit. A Hilger-Watts collimator is now arranged on a support bracket such that it intercepts the emerging collimated beam. Figure 89 shows this alignment set-up. The Hilger-Watts collimator is permanently focussed at infinity, it has a long focal length, and it is very sensitive to wavefront aberrations. This optical tool is now used to focus the spectrograph collimator. The collimator mirror tower is slid back and forth and rotated on the CWS plate to bring the vertical slit into focus as seen through the Hilger-Watts collimator. It may be necessary to rotate the spectrograph collimator corrector tower on the CWS plate to fine tune this focus. By placing a focus screen at the grating pupil, and in a darkened room, the pupil position can be checked and corrected by rotating the spectrograph collimator mirror on the CWS plate.

Figure 89: Aligning and testing the collimated beam with a Hilger-Watts collimator.

Once the spectrograph collimator mirror is properly focussed the intermediate foci at the image slicer and field mirror array can be checked. If a suitable amount of light is scattered over the image slicer it should be possible to simultaneously see the illuminated Focal Plane Mask Wheel slit as a series of spots and the illuminated slices themselves. If a suitable amount of light is scattered over the field mirror array it should be possible to see the illuminated slit as a series of spots and some elements of the fully illuminated mirror array. The image slicer can now be rotated on the CWS plate to place the staircase at the center of the central field mirror in the horizontal direction.

By tilting the Hilger-Watts collimator up or down on its mount it should be possible to see the staircase on all of the mirrors below the central mirror and at least some of the mirrors above the central mirror. By this method the staircase can be traced along the field mirror array. The beam from the extreme mirrors above the central mirror will be vignetted by the edge of the vacuum jacket and will not reach the Hilger-Watts collimator. If the slit in the Focal Plane Mask Wheel is now replaced with the square aperture, the full staircase can be seen more easily. If the image slicer is misaligned in tilt with respect to the pupil mirror array then the staircase will not form properly on the field mirror array. If this occurs, parts of the staircase will be seen to be missing near the cusps between adjacent mirrors. In this event, the tilt on the image slicer will need to be adjusted by re-machining the image slicer top and bottom mounting blocks.

1.6.2.3 Optics from the Grating Pupil to the Detector

The grating turret with all six gratings attached is now assembled on the turret shaft, the grating cylindrical cover is fitted, and the turret bearings are pre-loaded. If there are fewer than six gratings fitted then grating blanks will be added to balance the turret around the shaft. Next the grating drive is installed, wired, and tested. Then the flip mirror assembly is installed, wired, and tested. The pre-tested spectrograph camera is mated to the tongue and groove focus adjustment on the CWS plate and its focus adjustment is fitted. Lastly, the detector support structure is fitted and the camera 30 K cavity cold strap is connected to the 20 K tie point.

A glass sheet marked out with center lines and the outer edge of the HAWAII-2 detector is supported at the spectrograph camera focus. Now the Taylor-Hobson alignment telescope is supported on a cantilevered bracket such that it can view the camera focus. Figure 90 shows this set up. Next the 60 μm pinhole in the Focal Plane Mask Wheel is driven into position and the flip mirror is inserted in front of the grating. The lamp house spot projector provides illumination to the optics through this hole. The spectrograph camera is designed to work at near-infrared wavelengths but will transmit longwards of about 600 nm so optical images will be visible at the focus of the alignment telescope. The image quality at the camera focus can be determined visually now. The image quality will be assessed against data derived from warm Zemax ray traces.

Figure 90: Inspecting the camera image quality. Pre-alignment of the flip mirror and all six of the gratings.

The Focal Plane Mask Wheel is now rotated to place the square aperture in the beam. A vertical staircase slit image then forms at the camera focus and can be seen in the Taylor-Hobson telescope. The tilt on the flip mirror is adjusted to place all of the staircase onto the glass target in the vertical direction. The Taylor-Hobson telescope will need to be lifted, lowered, and tilted to see the top and bottom of the staircase. The flip mirror is removed and each of the gratings are now driven to a position where their respective zero orders pass into the camera. Each of these zero orders form a staircase slit and each of the gratings is adjusted in tilt to place this staircase on the detector outline in the vertical direction. There are access holes to the grating adjustments in the far side of the grating cover so each grating will need to be driven to this position for adjustment, then back to the pupil position for testing. Following this pre-alignment procedure, the glass target and Taylor-Hobson telescope are removed to allow attachment of the detector.

The detector housing is now fitted with a HAWAII-2 bare silicon multiplexer which will allow imaging at room temperature. The detector housing is placed in the spectrograph and wired to the SDSU-2 detector controller. Further test and alignment will proceed using computer images. The staircase will be visible on the detector output so the spectrograph camera can be moved along its focus groove to focus the staircase onto the detector. Rotation adjustment screws at the back of the detector housing are accessible in this configuration. The detector is rotated inside the detector housing to align the detector with the staircase. If the detector surface is not parallel to the focal plane, the tilt in x and y is noted and the detector housing is dismantled. The tilt can now be corrected by moving the detector mounting board with respect to the rotation ring on the screw clearances between the ring and the fibreglass isolators.

The white light lamp in the lamp house spot projector is now replaced with a suitable spectral lamp. Suitable spectral lines can be now found between 600 nm and 1100 nm. Each of the gratings are now finally aligned in face rotation and vertical tilt to place spectra onto the detector. Spectra can be positioned on the detector in the horizontal direction via the continuous grating drive. Spectral line curvature will curve the staircase slit so Zemax arc maps for 600 nm to 1000 nm will be used to help align the gratings.

Once all gratings are aligned, the spectrograph skirt and cover are assembled over the spectrograph and all bolts are tightened down to full operating torque. Further images are taken with the detector during this process to ensure that assembly of the skirt and cover do not change the alignment of the spectrograph.

After complete grating alignment, the spectrograph cover is removed and the detector housing is removed using the detector handling tool. Any remaining sheet metal baffling is now installed and final checks are conducted on the three mechanisms inside the spectrograph cavity. A functional HAWAII-2 detector, engineering or science grade, is fitted to the detector housing and the housing is returned to the spectrograph. The camera housing is now focused by an amount determined from Zemax warm and cold ray traces. The spectrograph cover is returned and the radiation shield and vacuum jacket are assembled. NIFS is now ready for vacuum pumping and cold testing.

1.7 Mechanical Risks

1.7.1 Lens Mounting Method

The camera design provides accurate lens alignment by containing them within close-fitting housing bores. The fits between lenses and bores are arranged to be near zero at the cryogenic operating temperature. The differential thermal strain characteristics of the lens and housing materials ensure that clearance will exist at higher steady-state temperatures. During cool-down, however, it is possible that temperature lag in the lenses will cause an interference condition to develop, resulting in lens damage.

This possibility is thought to be unlikely, but should it prove to be a problem, the mount design will have to be modified. Tests will be undertaken using a calcium fluoride blank.