DEIMOS MEETING MINUTES Friday, February 10, 1995 NS-143 2:30 pm Present: Dave Cowley, Eric James, Jack Osborne, Deanne Lago, Marlene Couture, Sandy Faber, Bob Kibrick, Scott Trager, Garth Illingworth, Bill Shepard. As was decided at the last meeting, a list of topics was published, and a round table discussion of each topic is desired with feedback from all encouraged. Error Budget Jack and Sandy have been working diligently on the error budget. Items will be defined in "datums". An example of a datum would be: the theoretical center of the spectrograph focal plane specified as "datum-A", as this is an idealized point without an exact place, we can approx- imate that with a real thing, which is the tooling ball specified as "datum-A Prime". The error budget table with the skeletal beginnings is almost completed (and is already being used), detailed notes are now being written (approximately 100). The notes define all quantities and explain all reasoning. Slit Mask/Handler/Cutter The slit mask holder is having the biggest impact on the design at this point. The current design is called the "side-wheeler" (drawings attached as an addendum). Drawing #1 is a 3-D drawing which has things hidden so as not to show so many lines. Drawing #2 is the square-on projection/ drafting drawing which shows the two cylinders that store these masks. The side-wheeler has two cylinders of ten masks each located on the rotating instrument, which feed mask frames into the focal plane in a single linear motion. A rotation of the cylinder (or side-wheel) brings different storage locations into position. The advantage of this is that two motors are eliminated from the pre-PDR scheme. This design is also less complicated from a robotic/software standpoint, you always have hold of a mask and never let it go. There is one disadvantage, however; the entire spectrograph must be moved back from the telescope and thus, from the nominal focus. At present, we must move back 4" for the grating slides to have clearance, and an additional 4" for the two side-wheels. This impacts the image quality and therefore, the error budget. (We have not added a category for this yet.) Moving back 8 inches induces an optical aberration called spherical aberration. On axis we can re-tilt the primary mirror to almost take it out completely. We can also remove virtually all the effects of spherical aberration over the whole field of view with the same re-tilt. The procedure for determining the need for re-tilt is simple. A picture is taken in the detector of the telescope segments un-stacked. Then we can use the way the segments spread in the actual spectrograph to deduce a proper stack. The alignment camera is not used at all. At the present time, Jerry Nelson and Terry Mast are writing software to stack the primary for each separate detector that might be used by each instrument. A few test fixtures of the slit mask/skin interface have been fabricated for us to study. We also have made Finite Element Analysis models of skins and have added a category in the error budget to allow for failure of the cylindrical mask to approximate the spherical focal plane. The impact of this is also in the RFQ for the laser-cutter manufacturer, since the material could either be 0.003" thick or 0.005" thick hard stainless steel shim stock (the 0.005" could take longer to cut). The motor list has been re-written. The full spectrograph now has 22 motors, 4 piezos, and 14 pneumatic actuators. The slit mask is now cylindrical instead of spherical. The cylinder can equal the sphere along two lines, at which point the error of the focal surface will be zero. It is a matter of preference as to where in the focal plane those two lines should be. Originally it was thought that one line would be very close to the center of the focal plane and we would accept a larger error at the other end of the focal plane (where the images were worse). However, because the aberration at the edge of the focal plane is astigmatism, we decided to take the balanced approach and have a little bit of error on the inside and a little less error on the outside. Harland is still optimizing the cylindrical plane. A change of scope was circulated for the focal surface changing to a cylinder. Action item: Tune up Eric James' memo regarding the slit mask. A change of scope regarding the 8 inch move will be circulated and given to the SSC for review/ sign off. It will be discussed at their meeting on February 27. It was decided to show Frank Melsheimer our side-wheeler design for the slit mask handler for his review/suggestions. It was suggested to build a prototype of the model after sign-off by Melshe- imer. Plastic material samples for potential slit masks were discussed. The samples are coated with a dye, but this material can be made opaque depending on the coloring. The plastic material is called "cellulose triacetate" (acetate) and is dimensionally stable. (The stainless steel takes quite awhile to cut with the laser cutter, plastic goes much faster.) A sample of the 3-mil annealed steel material with a one arcsec slit was passed around the room. This was cut by FLOROD based on our RFQ. We have also asked for a sample cut of the plastic material and also a sample of a 5-mil annealed steel. It was noticed that most of the slits were perfect except for the extreme top one, which had a slight divot. This divot will make it very hard to flat field and to get sky substraction on that image . It should be investigated exactly what this divot is and if it will be repeated in another cutting. Bill will check this carefully with a microscope to try and determine the problem. We will develop a process to test the samples in the CCD lab. With the hardened material the cost to fabricate one mask is approximately $1. The following is the list of materials still up for consideration: Acetate (cellulose triacetate) Mylar (still obtaining information) Capton (still obtaining information) Steel Non-Annealed (3-mil and 5-mil thicknesses) Indium Tin Oxide (is probably not available in solid form) Material eliminated to date: Aluminum in all forms. It was suggested to circulate the RFQ for the cutter to the vendors ahead of time to enable us to get some non-materials review going. There are handling issues and other general things that need to be considered by possible vendors. After we decide on a suitable material we will circu- late a materials update on the RFQ to the potential vendors. This will facilitate things on both ends. A draft RFQ will be prepared by the end of February. Electronics Budget The Electronics design hours were estimated in the PDR blue book as 4,700. After adjusting for some items that are no longer required, this number will drop down to 3,900 hours. The electron- ics fabrication hours will be adjusted to 2,600, which means an Electronics total of 6,500 hours. The materials are being looked at now. One item that seems to be under estimated is the calibra- tion lamp supplies. They cost approximately $1,500 each (Neon, Helium, Argon, Mercury, Cad- mium), and the Etalon is even more expensive than those. There is one Etalon used in HIRES (in the filter wheel). We should decide exactly what we want/need for DEIMOS and adjust the bud- get appropriately. We had discussed using one on the AO source and another as a calibration source in the instrument. Another item which is not budgeted for at all is lamps that shine off the dome. We will have two UPS systems; one in the vault (where heat will be a concern) and one in the computer room. A comparison of the UPS needs from HIRES to DEIMOS shows that DEIMOS has a much larger number of boards. Attached to the instrument we have the CCD assembly/CCD controller. The VME crates will be housed in the control room not on the Nasmyth deck. How- ever there are more Leach boards that need to be powered. If heating is a concern (for the boards in the vault on Nasmyth), there will be another more difficult task of routing the powering up to the computer room and powering through a UPS up there. Again, the additional load of the CCD controller attached to the instrument would normally be powered through the UPS in the vault (which could be a concern). If that also becomes a problem, we will have to power from the UPS now in the computer room. These items will have to be coordinated early on with CARA. Structural Design Eric will be able to devote more time to DEIMOS by the end of the month. This will put us approximately one month behind where we hoped to be. However since this is a year long activ- ity it's not that serious. We are planning on having Frank Melsheimer here for consultation sometime in mid March, but as stated above, we will send the slit mask handler design to him for review prior to his visit. Camera Optical Design By allowing image size to grow by a modest 25%, we are now able to operate farther to the blue. We have very usable images down to 3900 Å, and DEIMOS will truly cover the whole optical range of the Keck II telescope. Harland is in the process of finalizing the design and getting ready to ask OHARA for an RFQ. We were visited by Rob Sparrow from Optovac, who brought us a 10" piece of scrap CaF2, which the Optical Lab can practice on. Optovac is very happy to generate our lenses and edge them to the right curve. We agreed to place an order for two large and one small boule. One of the new large boules will go on the shelf as a spare at Optovac. The existing small boule, the new small boule, and the other large boule will etched and ground by Optovac. They will buy back this insurance boule if we don't build the second camera and/or don't break anything. Rob Sparrow will also talk to Frank Cook, a wealthy optical designer about a contribution to the DEIMOS project. Harland has increased the vacuum window thickness to 0.50 inches and increased the BFD to 0.25 inches. Simultaneously he replaced two of the four uv-absorbing glasses in the PDR design. It has been decided to adopt a "standard" 6.0-mm filter thickness. Harland is presently trying to "finalize" the preconstruction design to a point where he, Sandy and Garth are all satisfied, so it can be released to the Engineering Lab for comments/concerns regarding the mechanical issues which may remain. Spectrograph Optical Design Review Harland will be going through the entire optical design of the whole spectrograph, not just the camera to check for problems. Optics Lab Equipment Most of the DEIMOS activity in the Optical Lab continues to be focused on "tooling up". The number of machines needed relates directly to the anticipated machine load/man load. In review- ing the planned polishing machine needs from in January, it was felt that the two cameras would now be manufactured serially rather than concurrently. This decision reduces the need for the number of machines previously planned for, subsequently two new machines are being purchased with the "desired" features rather than the planned four used machines with "as is" features. The total cost of these two machines is $14,160, not including tax and shipping. General preparation will continue until the final specifications of the camera are known, at which time the tooling for the individual elements will be sketched and released to the Instrument Lab for manufacture. The Strasbaugh 24" curve generator, partially funded by DEIMOS is expected to be delivered on Monday, February 13, 1995. Harland would like to adjust the timeline to start grinding glass in October, 1995. This date could be adjusted earlier pending delivery of the glass. Software Effort and Software PDR The Software PDR will be moved from May to early July. The written materials will be com- pleted and mailed by the beginning of June. A PO has been written for a large 9 gigabit SCSI disk, a SCSI to ethernet interface that the disk will attach to and a 5 gigabit Exabyte. This equipment is split funded from the DEIMOS Soft- ware Upgrade budget and the General Software Group Equipment budget. This will allow us space to start developing code, to bring over large pieces of other code and give us an area where we can experiment on synthetic DEIMOS images. To date there is not a single software group person who is available to provide 100% time to DEI- MOS. Detector Progress The SITe people were more responsive to the Science Community needs at a recent meeting. Per- haps they are getting worried that they will be left out of the emerging market based on what Lin- coln Labs is doing. The recent meeting was very encouraging. Pperhaps the CCD working group can also pay them a visit. SITe would be a very viable option for receiving CCDs if they can manufacture to our flatness specifications. SITe is working on their field curvature problem (which is our main concern). Dave Dorfan and Mike Edwards from Physics are looking at the design of the on-chip amplifier and subsequent circuit signal chain of CCD detectors. They have contacted someone at Lincoln Labs who likes to design CCDs but is not knowledgable about amplifier design (Dorfan is an expert on amplifier design). The thought is that they can design a new circuit funded by Liver- more (Livermore needs a very active rapid read low noise CCD to do Adaptive Optics). Liver- more has a pool of money for CCD design for that purpose, which is administered by Scot Oliver. However it is not known if this will be accomplished in time to benefit DEIMOS. Schedule The major schedule change is the Software PDR moving to July. Budget A revised budget will be published with Quarterly Report #2. We are short 200 K in our funding, and this amount has been moved from our contingency fund. In addition more funds have been moved out of the contingency for other things, (the optical generator, the glass etc.). There are items we can remove from the project (approximately $250K) by using a much smaller detector, to return to the contengency fund. Quarterly Reports #1 and #2 We are close to publishing both Quarterly Reports. They should go out early next week along with the response to the SSC on the PDR.