SYSTEM DESIGN NOTE

SDN0002.03 - Filter Wheel Design Requirements

Prepared by	Date	Approved by	Date	Rev.	Rev Date
Jay Elias	3/31/99	N. Gaughan	4/1/99	a	4/7/99

1. Introduction

The instrument nominally contains two filter wheels. Although it is possible that the functionality can be reproduced with a single filter wheel, it should be clear that somewhat less flexibility is available, as noted for specific cases.

The filter wheels are located in the instrument preceding the slit/decker combination and after the Offner relay system. The reason why they precede the slit is to ensure that any reflections off the filters will be stopped by the slit and will not propagate through the spectrograph to produce “ghost” spectra.

The filters are intended to restrict the wavelengths of light entering the spectrograph. This is to ensure that different wavelengths of light do not arrive at the same point on the detector when in spectroscopic mode, and that only the range of wavelengths best suited to locating an object arrive when in acquisition (imaging) mode.

Because the filters serve somewhat different purposes in the two modes, it may often be necessary to change filters when changing between acquisition and spectroscopy. Filter changes cannot therefore defeat the purpose of acquisition – the object must remain centered on the slit, and in focus. This has implications for filter properties alignments, and for mechanism positioning. In addition, the need to change filters frequently means that the time to change filters must be relatively rapid, so that telescope overhead is minimized.

2. Basic Requirements

2.1 Number of Positions

The filter wheels must contain at least enough positions for the following:

• 7 order-sorting filters, which cover most of the wavelength range between 0.91 and 5.5 µm. (Note that these will not probably all be included in the “as-commissioned” instrument.)
• 1 cross-dispersion blocker, covering the approximate wavelength region 0.9-2.5 µm.
• 1 lens for the pupil-viewing system. Note that this should reside on a different filter wheel than the main order-sorting filters, in order to view the pupil through different passbands. If there is a single filter wheel, a filter must be provided in conjunction with the lens, preferably for the 2-µm window. The filter wheel on which the lens is located is determined by the pupil-viewer design (distance from the slit required).
• 2 masks for determining detector focus. These mask off opposite halves of the filter aperture (parallel to the slit). As with the pupil-viewer lens, these should reside on a different filter than the main order-sorting filters, for similar reasons. Preferably, the masks are located on the filter wheel furthest from the slit (“front” wheel). Note that it may be possible to use a single mask which is positioned to two locations; in this case it would have to be slightly pie-shaped. This approach would also mean that the software would have to position the mask in a non-standard way.
• 1 or more positions for “custom” filters in addition to the 11 listed above. These comprise both filters that users may want to use for specific observing programs, and which therefore will not reside in the instrument permanently, and filters that are not included in the list above, but which will see more general use. The latter might include a K’ filter for easier acquisition in the 2-µm window, a blocker straddling the K and L windows or narrowband filters for acquisition of emission line regions or planetary “hot spots”.

While some filter changing is inevitable – additional filters will not be purchased all at once, and users may want to reclaim personal filters – it is highly desirable to avoid frequent changes; the number of custom positions should therefore be >1 if at all possible. Filters are not cheap, so it is also reasonable to assume that they will not proliferate in vast numbers. A reasonable configuration would be one that places the 8 baseline filters in one wheel, and the lens and masks in the second, with 5 “custom” positions. Note that each wheel must contain an “open” position to allow filters in the other wheel to be used, so there are effectively 9 positions/wheel. An increase to 10 positions/wheel would permit another filter which could be accurately focused; this allows use of 9 filters in the rear wheel (closest to slit) and 6 in the front wheel. This should be the baseline design (20 holes total; 15 filters).

2.2 Filter Optical Properties

The filters must have the same optical thickness; this is currently to be set by the vendor, but is likely to be equivalent to roughly 3 mm of BK7 glass. This requirement is set by fact that there is no provision for adjusting telescope focus via the wavefront sensor when filters are changed.

The filters are dimensioned to as not to vignette the full length of the slit (100 arcsec) or the full width of the acquisition field (20” width, not cenetered [+15” – 5”]).

The filters are also tilted at an angle of approximately 2.7°; the reason for this is avoid ghost images due to reflections from the filter or from double reflections inside the filter.

The filters must operate at cryogenic temperatures; they can be no warmer than approximately 70K in order to avoid long-wavelength (>5 µm) emission onto the detector. The filters will be designed for use at such temperatures.

The filters are located relatively close to the slit, so that any inhomogeneities will show in somewhat out of focus at the slit. The overall uniformity of the filters is not a concern, but point-like defects – pinholes, scratches, hairs and the like – will complicate flat-fielding and constrain overall repeatability of observations. For the most part, these problems can be avoided by care in handling, though manufacturing defects can be a problem.

2.3 Light Baffling

The baffling of filters in the wheel, as well as around the wheel, is a concern. In particular, it should be noted that filters are normally not coated out to the edges of the substrates, and it is therefore important to mask off uncoated edges (on both sides if necessary).

2.4 Thermal Issues

As indicated above, the filters cannot be warmer than 70K. The energy incident on the filters is relatively modest, under a mW¹, so they do not need massive cooling. On the other hand, they do need to cool promptly when the instrument is cooled initially – and they must warm up promptly when the instrument is warmed.

The same masks used to block light leaks around the edges of the filters may provide a suitable conduction path.

Most of the mass of the filter wheels is, of course, in the wheels themselves rather than in the individual filters. These must be adequately coupled to the cold structure or cooling system.

2.5 Repeatability

The repeatability of the filter wheel positions, as well as the alignment of the filters, is covered in SDN0002. This information is reproduced here for convenience, but in case of any discrepancy the information in SDN0002 is considered to have priority.

2.6 Speed of Operation

The selection of one filter to replace another should take no longer than one minute. The two wheels may be moved in parallel. Specifically, it is permissible for the two “open” positions (or any other two positions) to coincide during the course of filter wheel motions. This is part of a requirement that reconfiguration between acquisition and spectroscopy should take 1 minute or less.

2.7 Lifetime

The baseline observing scenario is described in SDN0005. Under the assumptions used in the scenario, the number of filter wheel operations (for each wheel) during the operational life of the instrument will be somewhat less than 10⁵.
 

¹  The energy incident on the filter is the light reflected off the pick-off mirror and passed by the Offner cold stop. Most of this is at longer wavelengths, so an emissivity of unity is a reasonable approximation. If we take the area of the pick-off mirror (with supports) as ~7 cm², the blackbody power into an f/16 beam is ~0.35 mW.
 

---

---

If you have any questions or suggestions regarding this website, please contact Melissa Bowersock.

NOAO Intranet Services

NOAO Copyright  Statement