It may be necessary to open of the dewar to effect repairs on the system or to change the filter in the filter wheel. It is important that this is done by somebody who is skilled and knowledgeable about this sort of work. It is not something to be tackle by an innocent astronomert given that it is possible to damage the detector chips with a replacement value of about 400 thousand dollars.
If CIRSI has been assembled for use on either of the telescopes on La Palma, it will have an extended snout on the front plate of the liquid nitrogen dewar. Within the snout there is a black stepped baffle that can only be removed if the front window and its retaining ring alone are first removed. It is not possible to remove the snout or the front plate of the Dewar without removing this baffle unit first. It is easier to remove the bolts that hold from plates onto the dewar while the dewar is still under vacuum. These bolts may be removed before the baffle is removed. Once the interior of the dewar is properly warm (and it is really important to be absolutely certain that this is the case: see pages about emergency warming of the dewar) the vacuum can be released and the front plate removed.
What you see next is the radiation shield. It is held onto the base plate of the assembly by a set of eight M3 socket headed bolts. Note that the cables from the 55 way connectors in the dewar walls to the chips are taped against the side of radiation shield with aluminised mylar tape. This tape should be carefully removed bearing in mind that the wires used for the cable assemblies are relatively fragile as they were chosen to minimize thermal conductivity rather to be particularly mechanically strong. Also note that it is very easy for one of these wires to become trapped under the radiation shield when the unit is being re-assembled. In addition, the chip holder radiation shield is not actually contact with the chip holder and the gap between the two is covered with aluminised mylar tape. This has to the removed before the shield is removed and re-install to make sure that the gap is good and light tight later.
When removing any bolts only assembly it is very important to try and avoid letting any of them slid down between the base plate and then either jam and cause a thermal bridge between cold and warm or to rattle about forever. It is helpful to have a pair of tweezers to remove the bolts from the radiation shield before it is removed from the assembly.
Lift the radiation shield out carefully, avoiding any nasty study fingerprints on the aluminised mylar sheet covers which, apart from looking awful, are likely to worsen the performance of the radiation shield in keeping the system cold.
This exposes the filter wheel assembly. It consists of a base plate, a filter wheel with teeth cut in its age and places for eight cells each of which take four filters which are 32 by 32 mm. There is a round pressure plate which is held against the filter wheel by three springs in the top hat structure on top of the filter wheel cover plate. The filter wheel cover plate has a baffle fitted in its top surface to minimise stray light. This baffle currently also holds a set of blocking filters that are designed to overcome the inadequacies of the preset set of filters from Z to H-long.
The filter wheel itself, the base plate and the pressure plate are all made out of special aluminium plate that has been treated to change its surface into a Teflon impregnated ceramic material. It is very tough and allows the parts to move together even at liquid nitrogen temperatures and under vacuum with very little resistance. The filter wheel itself is held slightly away from the base plate by a thin shim although it is likely that the edge of the filter wheel does actually touch the edge of the filter wheel base plate. The round pressure plate pushes onto the round recess in the middle of the filter wheel and keeps a factor dedicated. The filter will also is located in a brass bush which is centrally located in the filter assembly base plate.
The filter wheel has teeth at its edge which are driven by a gear box mounted on the underside of the filter assembly base plate. The filter wheel is most easily removed from the base unit by installing three bolts into the holes in the middle (recessed) section of the wheel.
The filter wheel itself takes eight filter cells each of which can take four filters 32 by 32 mm square. Each filter cell consists of a base unit and a cover plate. The base unit in some cases is machined out to allow thicker filters to be accepted by the cell. It is important that the filter cell assembly does not, however, cause the base part to protrude underneath the filter wheel itself. This would scrape on the surface of the filter assembly base plate and calls the mechanism to jam. If it is necessary to install thicker filters in this may be done by modifying filter cell cover plate. The undersigned of the filter assembly cover plate has been machined to make sure there is plenty of extra space for thicker filters.
Should you need to install new filters into a new cell it is important that the edges of the filters are masked so that no stray light passes through the edges, or around the edge of the cell if the filters only fit loosely. A black paper mask should be fitted on either side of each filter. It should be cut to be just under 32 x 32 mm and have a clear area cut out that is at least 28 x 28 mm.
When a new filter is being installed take great care to record which filter cell is where. Unless you have very good reasons to do otherwise it is important to leave the metallic plate in place on one of the positions as this is the only reliable guide to dark current in the system.
The position of the filter wheel is encoded two micro-switches that run along a single track on the underside of the filter wheel. When position 0 (normally occupied by the metal plate) is directly over the detector s both micro-switches are closed. One micro-switch is also closed when each of the other filter positions is directly over the detectors.
The drive shaft from outside the dewar must only be turned clockwise so that the filter wheel itself is also only able to turn clockwise. This is important partly because of the design of the micro-switches chosen and partly because the slots into which each locate to indicate the position of the filters are designed with a sharp drop on entry and a smooth ramp on exit. This gives a more accurate indication of the precise position having been reached. It takes just over three full turns clockwise of the drive shaft (actually 3.1 turns) to move from one filter position to the next.
If it is intended to remove the filter assembly from the dewar body it is important to look at the drive mechanism between the feed-through (Ferro-Fluidics type) and the filter wheel body gear box. The filter wheel assembly is located fairly sloppily within the dewar and it is important that you take care to reposition it as carefully and centrally as you find it.
When you remove the filter wheel assembly from the dewar take care because there are wires going to it (to the micro-switches) and the thermocouple.
The filter wheel gear box is now exposed. It consists of two spur gears that convert rotation on the drive shaft into a rotation of the filter wheel itself. The structure is relatively sloppy at room temperature since it has to cope with a very substantial change in temperature without jamming. We have no idea how well this mechanism will survive with usage. Replacements for gears are provided with the system. One is machined down and it is important that each is fitted in the correct way. It is also essential that the loading spring is installed in the gear box if it is not to jam.
Once the filter wheel assembly is removed the chip box with its solid copper heat strap is exposed. When working with the chips it is very important to make sure that at all times they are electrically shorted in order to give them full anti-static protection. When the chips are normally installed in the dewar each of the 55 way connectors has a metal cap over it which is lined with conducting foam. This ensures that all the pins are connected together. When the chip box is removed it is necessary to remove the two blocks through which electrical contact is made in order to pass through the cold heat shield. It is helpful to mark carefully on the base plate the exact location of these connector blocks before they are unbolted so that they can the replaced accurately later. The connectors that go to this box from the 55 way connectors should be removed and replace with sub-miniature D-connector shorting plug. These plugs travel with the CIRSI system. All four connectors should be installed before the chip box is removed.
The copper heat strap is bolted onto to copper blocks which go through to the liquid nitrogen cylinder and also onto the chip box. This box is, in fact, insulated from the base plate with a single sheet of aluminised mylar. It does not contact either the filter wheel assembly or heat shield and is cooled directly through the copper heat strap. The heat strap should be removed and the bolts holding the chip box down should also be removed. Once they are all removed the chip box with its cables may be removed from the dewar and the cover of the chip assembly removed to expose the chips.
We are now entering slightly more scary chapter in which there is a serious chance of doing irreparable damage to the chips. Increasingly it is important that the work is done by some of who knows they are doing.
To remove one of the chip holders first take the cover off the chip box exposing the four chips with the connectors and cables that go through the wall of the assembly. The first thing is to remove the connector from the mating part that is bolted to the side of the chip box you wish to remove. It should be replaced with a shorting plug from the other end of the cable you have just taken out. The chip box is removed by undoing three bolts in the base of the chip holder.
The chip holder consists of two parts. The base units is machined to be very flat. The chip itself is mounted in a socket on a printed circuit board with a short cable to a sub-miniature D-connector bolted to the outside of the chip holder base plate. The top part of the chip holder, the chip cover plate is held in position with a dowel pin and bolted at three points. A nylon bolt at one edge provides downward pressure onto the top surface of the chip, at the edge of the chip to hold it against the flat surface of the metal base plate. To disassemble the chip holder start by removing this nylon bolt and then remove the two socket headed bolts that hold the chip cover plate onto the chip base. The dowel pin can make it difficult to lift the cover plate,. Some leverage can be brought against the two bolts that hold the printed circuit board onto the base plate. The design of the machining of the base plate and cover plate ensures that it is only possible to lift this cover plate vertically and therefore avoid damaging the chip once the cover plate is removed the chip carrier printed circuit board may be seen to be held in position by three bolts. The disassembly is quite straightforward.
The principal reason you've got this far into the system is probably because one of the quadrants or one of the chips has ceased working. If a whole chip has ceased working you have probably done the wrong thing in getting this far since it is very unlikely to be a problem with the printed circuit board or the chip socket. It is much more likely to be a problem in the electronics of the multiplexer or the cabling onto the chip box.
The chip maybe removed from the printed circuit board by undoing the three bolts that locate the PCB, removing the pillars on which the board is mounted, and gently pressing down on the socket so that it is the central part of the base plate which presses against the underside of the chip. Store the chips safely in one of the anti-static boxes that travel with the CIRSI system.
The chip socket is a complete nightmare for the HAWAII chips. As supplied by the manufacturer the contact pins on the side of the chip socket do not make good contact with chip. These need to be pulled out away from the side of the socket so that they make contact with chip itself. It is very difficult to position each of the contacts so that they press firmly against the side of the chip rather than against either the top or the bottom edge. They must also be accurately in alignment with the contacts and not displaced slightly sideways stopping them from making a good contact. Once the chip has been inserted it must be checked visually for accurate alignment on all four edges. It is also important to visually to make sure the contacts appear to be good. The chip may be connected to the multiplexer box using the simple adapter cable that travels with the CIRSI system for testing purposes.
You should also check before inserting the chip that there is a yellow shim spacer at the bottom of the socket which is intended to make chip sit at the correct level within the socket.
The re-assembly of the chip holder is relatively straightforward. It is important to make sure that the chip is held with a light but firm pressure against the reference plate surface. This is done by slightly tensioning the chip against the base plate using its intrinsic springiness. Before attaching the cover plate remember to install the thin plastic shim which sits on the top face of the edge of the chip so that the cover plate presses against this and not against the chip itself. Adjust the nylon bolt so that there is no residual springiness is in the cover plate.
After this the assembly procedure is a fairly straightforward reversal of everything you have done to get this far. Things to remember as you re-assemble the units are as follows.
1. Make sure you remember to read connect all the electrical connections to the chips.
2. Make sure that the copper strapped is well bolted down at both ends. The quality of these contacts will have a lot of impact on the final temperature of the detector chips.
3. Make sure that the thermocouples are all back in their correct places. These are valuable diagnostics of system performance and give the first warnings of degrading vacuum.
4. Make sure that the filter wheel assembly is properly position and checked that it takes smoothly throughout at least one revolution without their being any sticky parts.
5. Make sure that the heat shield is firmly bolted in place. The heat shield is very thin and therefore it takes a long time to cool down and heat up if this contact is poor. Check you have not trapped in the wires underneath the heat shield. Make sure again that the gap between the heat shield and the chip box is properly covered with the aluminised mylar. Make sure that the cable bundles are attached to the side of the heat shield with a piece of aluminised mylar tape about one-third of the distance from the connector that goes through the shield.
6. Make sure that there is no dirt or other contamination on the O-ring that seals from plate onto the dewar. Simply rub a clean finger around the O-ring and also around the mating surface on the front plate before it is put in place.
7. Tighten up to bolts around the rim of the from plate gradually, so that there is a fairly uniform pressure over the whole plate.
8. Try to get the dewar back on the vacuum pump as soon as you can after it is sealed. Always minimize the time that it is open to the atmosphere because this will minimize the amount of contamination vacuum pump has to manage.
Prepared by cdm@ast.cam.ac.uk Please contact with comments, etc. Last updated 22 May 2001