Документ взят из кэша поисковой машины. Адрес оригинального документа : http://www.apo.nmsu.edu/Telescopes/SDSS/eng.papers/19891013/19891013.html
Дата изменения: Tue May 19 04:52:16 1998
Дата индексирования: Sun Apr 10 05:35:42 2016
Кодировка:

Поисковые слова: п п п п п п п п п п п п п п п п п п п п
Comments on Packaging the CCD Camera Array

Comments on Packaging the CCD Camera Array

2.5 m Survey Telescope Technical Note 19891013

 

Walter Siegmund and Charles Hull

The following thoughts have occurred to us recently regarding the 6x5 ccd camera array for the Cosmological Survey Telescope. Current experience is applicable to most aspects of the design. Unique problems include avoiding catastrophic vacuum failure and space constraints. The solution to the former problem suggested by Jim Gunn is to place the ccds in individual vacuum chambers so that a vacuum failure at most affects one chip. This seems to us to be the best solution as well, but it does acerbate the second problem. This must be solved while maintaining easy access to the chips, buffer electronics, and seals. For reasons of economy and maintenance, the chips should be mounted in identical interchangeable modules. It is desirable to avoid a tangle of LN2 plumbing.

We propose that the ccds be mounted in individual vacuum containers which do not contain LN2. Heat is conducted away from the ccd via a copper cold finger to a cold plate which is shared by all 30 ccd modules. This cold plate is part of a single LN2 dewar which can be large enough to provide a conveniently long interval between refill times.

The drawing shows the details. The left half of the section is taken through the center of the ccd module. The right half is just behind the near surface of the ccd module.

Starting at the bottom is the LN2 reservoir and the cold plate. Located at the ccd locations on the cold plate are spring loaded platforms which are thermally coupled to the cold plate with copper braid (not shown). When assembled, the ccd cold finger (a copper tube) depresses the platform and the spring provides enough pressure at the joint to insure good thermal coupling.

Above the cold plate is the outer wall of the dewar. This surface contains holes at the location of each ccd to allow penetration of the cold finger. Welded to the surface at each hole are standoffs (thick washers). The upper surfaces of the standoffs are machined tangent to a sphere centered at the center of curvature of the focal plane. Surrounding the cold finger is a stainless steel thin walled tube. This provides a vacuum jacket for the cold finger and part of the LN2 vacuum seal. The balance of the vacuum seal consists of a stainless steel bellows and the dewar seal plate which holds an o-ring which is pressed by a spring against the machined surface on the standoff on the outer wall of the LN2 dewar.

The ccd is inside the ccd can, a cubical container, which is open on the top. The ccd can, the window, the filter assembly, and the ccd assembly is called the ccd module. The 30 ccd modules and spare modules are identical. Each module is evacuated and cold pumping material can be thermally coupled to the cold block. The ccd assembly consists of the ccd, the temperature regulated copper block which it is thermally coupled to, the cold pc board, the ccd connectors, the ccd spring clips, the cold finger and its associated stainless steel vacuum jacket, bellows and dewar seal plate, the ccd module lower seal plate and its associated warm pc board. This ccd assembly can be removed from the ccd can through the top. This allows access to the ccd and the cold and warm pc boards for troubleshooting. As the ccd assembly is removed, it is unplugged electrically at the bottom of the ccd can.

In front of the ccd is a quartz window which makes a vacuum seal via a o-ring to the top of the ccd can. Above this is a square retaining frame for the window. The filter assembly consists of the filter and the filter frame. The filter is mated with index matching fluid to the window. Captive machine screws are used to attach the filter assembly to the ccd can.

To remove the filter assembly, a cover plate with handles is attached to the the filter assembly with machine screws. Once the screws which attach the the filter assembly to the ccd module are removed, the filter assembly can be lifted off. (If the ccd can screws are removed, the filter assembly and ccd module can be removed as a unit.)

To remove a ccd module, a similar procedure is followed. The cover plate is attached to the ccd can, the ccd can screws are removed and the ccd module is lifted out. In both cases, the cover plate protects equipment during handling.

The ccd modules mount in rails which have the u-shaped cross-section shown. Each of five rails hold six modules. The modules plug into a mother board which route the common clock signals and individual video signals to connectors at the ends of the rail where external cables can be connected. If both ends of the rail are used, the mother board need only carry in parallel enough conductors to serve three ccds. A viable alternative to a mother board may be ribbon cable.

Electrical signals are transmitted through the bottom of the ccd can via vacuum feedthroughs. These provide pins on both sides which mate with standard connectors. Not shown in the drawing are mechanical guides for the mating of the two connectors and the mechanical support of the connectors mounted on the warm pc board. Depending on the number of signals required and the friction between the connectors and the feedthroughs, levers or screws may be required to allow safe extraction of the ccd assembly or ccd module.

The ccd module mounting surface of the rails is machined so that the ccd surfaces are tangent to the arc of a 11 m radius circle. The plane of this circle is normal to the drawing.

The ends of the rails are bolted to the instrument frame which bolts in turn to the telescope. The LN2 dewar bolts to the instrument frame, also. The rail mounting surface of the instrument frame is machined so that the ccds are tangent to a 11 m radius circle which is parallel to the plane of the drawing.

The rows and columns of the ccd are not accurately aligned with the ccd package. To align the ccd in the ccd module, the ccd module is placed in a special fixture which is accurately oriented with respect to a fiducial pattern which is projected onto the ccd. The actual orientation of the ccd is determined from the image and the fixture allows precise correction of the ccd angle by rotating the cold finger once the ccd has warmed up, the vacuum has been released and screws loosened. The details of providing adequate compliance in the mounting of the warm pc board have not been worked out but should not present any difficulty.

The ccd modules are not actively pumped. The ccd is cooled to only about -100 °C so the vacuum does not have to be as good as for a LN2 dewar. The LN2 dewar, on the other hand, can be continuously pumped and this can be accomplished with a single vacuum line.

The ccd is supported from the bottom of the ccd can via a thin wall stainless steel tube. For 150 mm wall tubing, heat conduction is 0.36 watts and transverse ccd motion as gravity goes from 0 to 1 g is 5 mm. This seems acceptable to us, but if not, the tube diameter could be increased without significantly increasing the heat load on the dewar. Simple calculations show that the copper cold finger can easily extract several watts from the cold block if it is kept at -100 °C.

The temperature of the cold block is regulated in a manner similar to existing designs. The drawing shows one way that this could be done. The temperature is sensed at the center of the cold block by a temperature transducer. This signal is used to control a heater in the cold finger to regulate the about of heat transported from the cold block to the dewar cold plate. In this scheme, heat from the regulating heater does not flow through the cold block and associated temperature gradients do not develop.