To: Distribution 2 Oct 95 From: Martin Nordby Subject: Minutes of Near IR Engineering Meeting of 29 Sep 95 Support Tube supports At last week's BaBar Collaboration Meeting, a working decision was made to support the Support Tube off the detector. On the forward end (upstream LEB) this means using a "Flagpole" design. This has a hook in it to allow the EC Calorimeter to be removed as a whole, before it is split into halves. The flagpole would not need to be removed for EC Calorimeter removal. Rough analysis by Martin Nordby shows that it is stiff enough to tolerate gravity and EQ loads. The next step is to refine the design by more detailed analysis, and look at the effect of the Flagpole on the Q2 shielding. Support Tube Mini-Review This review covered the design and analysis to-date of the Support Tube. It's intent was to show working designs for the Tube and the assembly sequence, and to identify areas requiring further work. The chief design criteria for the Support Tube are that it be 0.5" thick, with a maximum radiation length of 0.9% X0. Deflection under load should be minimized to reduce the possiblity of hitting the Drift Chamber during an earthquake. The carbnon tube must be able to take a 150,000 in-lb bending moment during an earthquake without trouble. John Hodgson presented two designs for the center carbon tube, both of which meet the above criteria: 1) Single-wall: 0.070" thick carbon-fiber composite cylinder, with fiber orientations of +/- 10¡ and 90¡. Overall deflection of this tube, when joined to 0.5" thick stainless steel tubes is 1.5 mm. Peak fiber stress is 6461 psi (well below failure for carbon fiber). During EQ loading, the single wall design has a safety factor of at least 2.5, compared to the theoretical critical buckling stress. However, this safety factor is dependent on the assumption of how the ends are restrained. The SF can be as high as 3.5 if the ends are assumed to not rotate around the local azimuthal axis. Buckling is initiated near the ends of the carbon tube. 2) Sandwich construction: 0.5" thick sandwich, with 0.015" carbon fiber skins suroounding a 0.47" thick core of Rohacell. Fiber orientations in the skins is +/- 15¡ and 90¡. Total deflection is 2.4 mm under gravity, with a max fiber stress of 10,400 psi. The sandwich design has a safety factor of at least 2.8. This is set by the shear modulus of the core material. Buckling analysis was done using ANSYS finite-element models, large- deflection analysis. A few isuues were raised which warrant more investigation. --What happens at the close-outs at each end of the tube? What is the shear stress in the epoxy, and are the carbon fibers any weaker? This could affect the strength of the cylinder. Also, how stiff is the close-out? Since buckling is initiated in this region, if it tends to be less stiff, buckling may be initiated at a lower-than-expected load. --What affect does an elliptical tube have on critical buckling load? How sensitive is the critical buckling load to eccentricities or non-uniformities of fabrication? This goes for both designs, but especially for the thin-walled option. --When a transverse load is applied to the single-wall design, it buckles at a much lower applied moment (in the model). Is this really true? If so, what criteria for lateral loading should be set? Theoretically, the Support Tube will not see any transverse loads, but what if someone jumps on top of the Tube when it is being installed? --The sandwich design is not quite stiff enough. How could it be stiffened without increasing the radiation length? There may be other options for the core design which allow more carbon fiber to be used. Some of these issues are questions concerning fabrication. The next step in this design is to talk in more detail with manufacturers about both designs to better understand the fabrication process. The bolted joint connecting the stainless steel tubes to the carbon-fiber center tube was also analyzed. 24 high-strength stainless steel bolts are needed for this joint. Since the joint must be only 0.5" thick, the bolt stresses are high, and the joint becomes the weak link in the Support Tube design. Assembly Sequence Martin Nordby presented a preliminary assembly sequence for mounting the three Support Tube pieces over the assembled components. Each piece is mounted individually, with the center carbon tube being the trickiest. Rolling supports register in grooves in the Q1 magnet collars to allow the component assembly to be well-supported on either side of the tube being installed, while ensuring that they are positioned accurately, to avoid damaging the SVT. To enable the carbon tube to be installed without needing holes through the tube for temporary supports, the entire Q1/B1/SVT package must be cantilevered off the back of the Q1 magnet. This loading is the worst-case load for the Q1 magnet collar design, and sets the collar and bolt sizes. To summarize, the Support Tube design and draft assembly sequence show that there are work-able designs for the Tube and asembly fixturing which fit in the space constraints. No big red flags were raised, and the consensus was that future detailing of the Tube design and assembly fixturing should not un- earth big problems.