Minutes of the IR Engineering and Physics Meeting of 26 April 96 Forward End Q2/4/5 Raft Design Scott Debarger has been modifying the cantilever section of the Forward End Q2/4/5 Raft to fit through the tight hole of the Q2 Shielding Plug. The new cantilever section is a cone which fits tightly around the septum chamber, and then transitions to either a larger cone or discrete struts which fit around the in-board end of the Q2 magnet. This design requires using the steel Q2 magnet design. Using a full cone around the septum chamber, the deflection of the cone is small: for a 10 mm thick cone, deflection is 0.0015 in., for example. Orrin Fackler's magnetic analysis currently allowed for 15 mm of radial space around the stay-clear region needed for the septum chamber, so a 10 mm cone thickness is consistent with Orrin's analysis. Scott is looking at further maximizing the stiffness of the cone by using extra space below the beamlines, which is inside the stay-clear region, but not being used by the septum chamber. Also, he is investigating using a U-shaped member, or split cone, which would allow the septum chamber to be removed without disassembling the entire support structure. However, this would significantly reduce its stiffness. Inside this cone region, there is 740 cm^2 of cross-sectional area which is not being used by the chamber itself. This appears to be sufficient room to route the 100 cm^2 of machine services coming out the Support Tube through the cone, so they do not have to exit out the face of the detector barrel. However, space in this cone does not yet include joints and water connections for the septum masks. Backward End Q2/4/5 Raft Design Martin Nordby has started to look at the Q2/4/5 Raft on the Backward End. This Raft reaches into the DIRC S.O.B. and strong support tube to hold the backward end of the Support Tube and Q2 magnet. Because of the tight radial space in this region, and the need to remove a large fraction of the backward end Q2 Shielding Plug, the Raft cross-section needs to be minimized. This has forced two design constraints on the Raft. First, the Raft must be supported at its in-board end off the DIRC strong support tube. Since this support establishes the position of the Raft and Support Tube components, the magnetic and earthquake motion of the strong support tube must be well understood (and minimal). Martin, Jim Krebs, and Les Dittert are investigating this. (For the Raft to cantilever all the way into the Support Tube, it would need a 18 in. x 18 in. cross-section, which is far too big for the available space). The second constraint on the Raft design is that the truss system planned-for on the Forward End blocks removal of the Plug. This means that a simple box beam-type girder under the beamline must be used. Martin showed a 16 in. wide by 8.5 in. deep girder which was stiff enough to take gravity loading. Because the detector is on seismic isolators, the Rat pivots inside the S.O.B. during an earthquake. This relative motion between the Raft and the S.O.B. must not result in the two colliding, or severe damage to the Drift Chamber or Support Tube could result (since clearances are so tight). Thus, a +/- 5.4 in. (max.) stay-clear must be maintained on either side of the Backward End Raft to allow for this motion. This stay-clear region tapers down towards the in-board end of the Raft. Two other components which also need to exist inside the S.O.B. inner diameter are rails for removing the Q2 Plug, and cableways for the Drift Chamber cables. Martin showed a possible design for Plug removal rails which are based on a concept from Jim Krebs. The rails are built into a fixed piece of the S.O.B. magnetic shield cylinder, and attempt to maximize the amount of Plug material being removed. Unfortunately, the resulting cross-section left very little room for any Drift Chamber cables at all. This clearly needs more work, including investigating how much of the cables could fit inside the Q2/4/5 Raft. One final issue identified is that there does not appear to be enough axial space between the back of the S.O.B. and the front of the Q5 magnet to remove the Q2 Plug in one piece. This needs further investigation. Backward End Space Issues Jim Krebs has also been investigating the space on the Backward End. His design concept for the Q2 Plug removal rails includes room for Drift Chamber cables. However, no room was left for any type of Q2/4/5 Raft. Jim felt that the Q2 Plug may be able to be further reduced in size, possibly even shortened, to avoid axial space problems. Jim and Georges London will investigate that. Another Plug issue that arose is that routing all the cables for the D.C., S.V.T., and PEP-II I.R. out below beamline requires a fairly large, non-axisymmetric hole in the Plug. This has two potential problems. First, this most likely allows too much field leakage out the hole. Second (and more of an issue, based on Forward End experience) is that to model this requires a 3-D magnetic analysis. Jim, Martin, and David Coward will look at other possible routing schemes, especially since there is not enough room below beamline, anyway. Drift Chamber Cable Requirements David Coward gave an impromptu encore of his talk at the Electronic Integration discussion at the Collaboration Meeting. To summarize, there are three types of D.C. cables: Data acquisition, high voltage, and low voltage. Based on pending decisions regarding needed reliability of the D.C. electronics, the cable volume for each type could change dramatically. Data acquisition: the plan is to multiplex 16 each 60-megabit cables onto a single 1 gigabit fiber optic link. If this occurs at the D.C., only the fiber optics will need to penetrate the magnetic containment. However, if the multiplexing happens outside the detector, then the full complement of cables must be routed out the S.O.B. This decision changes the total D.C. cable cross-section by a factor of two. High voltage: here there could be anywhere from 1 to 160 HV cables running into the D.C. The issue here is what fraction of the D.C. are you willing to leave dead if you lose HV. On the order of 60 parallel HV circuits is expected, but not yet decided. Low voltage: here, the issue is power. Total power lost in these cables is 2500 watts. Assuming multiplexing at the D.C., about 75 in^2 of cable cross-section is expected. However, this assumes no packing fraction, and no provision for NEC regulations regarding packaging of signal, HV, and LV cables. Another issue involving the cable routing is how they cross the end of the D.C. volume. The electronics cards for the D.C. are mounted into boxes which mount to water-cooled radial ribs on the D.C. end (for cooling and structural support). The cards plug into readout cards which are mounted directly to the pins coming out of the endplate. Thus, to remove a card box requires pulling it axially off the readout card. If cables are water- falled down to below beamline, they must be flexible enough, or have service loops, such that they can be moved out of the way. This minimizes the access problems to any individual card box or readout card. To-Do List Much new information was presented, but it showed that plenty of work lies ahead. Here is a summary of "action items" which came out of the meeting: --Finish conceptual design of Forward End Raft cantilever, so re- optimizing of Q2 Shielding Plug fingers can begin (Debarger). --Work on shortening/modifying the Back End Q2 Shielding Plug to make it easier to remove (Krebs, London). --Finish conceptual design for Backward Q2/4/5 Raft, especially at the inboard end, where it must fit through/around the Q2 Plug and D.C. cables (Nordby). --Develop alternate Q2 Plug removal concepts that work around the Q2/4/5 Raft design. Can these tracks mount to the S.O.B., for instance (Krebs). --Finish D.C. cabling Technical Note, and make decisions on cabling options (Coward, et al). --Develop alternate routing plans for the D.C. cables (Coward, Krebs, Nordby). --Investigate DIRC horse collar/S.O.B. magnetic and seismic motion (Krebs, Dittert). We will update this at the IR Meeting in a few weeks. These minutes, and agenda for future meetings, are available on the Web at: http://www.slac.stanford.edu/accel/pepii/near-ir/home.html