Notes and Comments on the OOPS Support Engineering Meeting at Bates, Thursday, February 3, 1994 A number of us gathered in the Bates conference room to discuss the status of various support issues. Present were: Dan Tieger, Larry Bartoszek, Chris Tschalaer, Steve Dolfini, Wooyoung Kim, Stan Sobczynski, Adam Sarty, and yours truly, Steve Williamson (and possibly others that I have forgotten). 1 FLIPPER The first order of business was a discussion of the OOPS module flipper design. This discussion was lead by Larry Bartoszek, who elucidated the, then current, concept for the flipper. Briefly, on each side of the OOPS a trunnion clamp was to tightly hold the OOPS trunnions. Each clamp was fastened to a shaft supported by a pair of spherical bearings. These in turn, were supported above the floor by a vertical beam. The vertical beams, one on each side of the OOPS were to be fastened to a base plate resting on the floor. This design permits the OOPS to be flipped, after it is once clamped using either a motor drive or the overhead crane. A number of points were made concerning the flipper concept: 1. One of the chief costs of the support would be the bearings. 4 bearings in all would be required. No saving would result form use of simpler bearings because gearing would have to be increased to overcome added friction and custom mounts for the bearings would be required. 2. Most felt that there is no particular advantage in being able to flip the OOPS under motor control. Thus the gearbox and motor should be left off and the flipping be done using the overhead crane. 3. Given the relative lack of strength of anchors imbedded in the concrete floor there is no stability advantage to bolting the base to the floor -- it would need to be large anyway -- so it could simply rest on the floor. There was much discussion about the issue of which way the OOPS should be flipped. One approach would be to require the OOPS to flip with the shielding going down. This permits the vertical beams to be shorter and therefore the base can be smaller. Note that if the end cap is installed this method of rotation goes through a stable minimum when the shielding is at its lowest. If the endcap has been removed, then the center of gravity (COG) shifts toward the dipole end of the OOPS and the position with the shielding at its lowest becomes an unstable point of equilibrium. Some attending the meeting felt that this might produce a "jerk" as the OOPS was nudged through vertical unless great care is taken. Others disagreed. A way around this would be to make the support taller and then allow the OOPS to flip, always, with the COG passing below the pivot and through a stable equilibrium. If indeed this philosophy was followed, there would be no need for clamps and expensive bearings. A simple cradle to hold each trunnion would suffice. This new concept would look much like the North Hall support but with the vertical columns farther apart to accommodate the width of the shielding, the overall height greater, and no cross members between the vertical columns to obstruct the flip. This would, according to Larry, be considerably cheaper and simpler, and for that reason this design Page 2 concept was adopted. 2 VERTICAL SPACERS FOR THE SATELLITE A concept for the design of the vertical spacers for the satellites was discussed. Because all satellite out-of-plane positions may now be carried out above the scattering plane (due to changes in the scattering chamber design), one can consider making all out-of-plane "adjustments" by simply inserting out-of-plane spacers between the boat frame and the in-plane spacer. A set of 5 out-of-plane spacers, each twice as thick as the last, would permit angles from 0 to 15.5 degrees in 0.5 deg steps to be achieved with an un-flipped OOPS. By flipping the OOPS, the range from 20.9 to 36.4 degrees could be reached. An additional point at -17.1 would be available by mounting the boom directly on the boat frame. The out-of-plane spacers would range from 1.25" to 20" in thickness. The thinnest spacer would be a flame-cut plate. The next largest 2.5" thick spacer would consist of 2 1.25" spacers. The 5" spacer would be constructed of welded steel bar. The 10" and 20" spacers would be made of beams and plates along the lines of the in-plane spacer. 3 GANTRY TOWER ALIGNMENT Larry had pursued an earlier plan for a 6-strut system to adjust the gantry tower degrees of freedom. The size of the gear boxes and other assemblies associated with this system were found to be somewhat larger than initially expected. To avoid obstructions, caused especially by the 3 vertical struts of the 6-strut system, the designs of the two gantry towers must be quite different. After some discussion, a suggestion was made that the gantry tower be fastened directly to the airpad base. Because this base in turn rests directly on the South Hall floor (no jacks are used), the adjustment of the positions of the OOPS modules must account for misalignment of the tower. This adjustment would then be carried out separately for each OOPS using the vertical and radial motion of the booms holding the OOPSs and transverse, yaw, pitch, and roll at the boom ends with some kind of "XY" table or 6-strut system. Advantages of this approach are: 1. Additional rigidity is achieved with the tower fastened directly to the airpad base. Perhaps the tower members can be made smaller. 2. Some additional space is provided below the tower by the removal of assemblies for its positioning. This can be used to extend down the range of travel of the carts fastened to the booms. This may provide some additional freedom in their design. 3. Fewer obstructions so the second gantry can be more nearly like the first. 4. There must be individual adjustments of roll, pitch, yaw, and transverse translation in any case. 5. Some cost savings. Some disadvantages are: Page 3 1. The degrees of freedom of the individual adjustment of the OOPs become quite entangled and non-orthogonal if the tower is not vertical. 2. In principle all 6 degrees of freedom of both OOPSs on the tower must be adjusted whenever a new alignment is to be done: 12 adjustments in all. 3. The range of adjustment of boom-end degrees of freedom may need to be made larger than is practical. The general feeling was that it is worth pursuing the idea of the gantry tower being fastened directly to the airpad base. The practicalities of this approach should be investigated. 4 GANTRY RAILS Stan pointed out that each of the vertically moving carts carrying the horizontal booms that support the OOPSs are constrained by 8 THK linear bearings riding on 4 sets of rails (2 front and 2 back). Some worry was expressed that this over-determined situation might lead 1) to an excessive effort required to align the rails and/or 2) to the front or the back set of rails getting the entire load rather than sharing the load which in turn would lead to possible excessive wear, jamming, or failure of the rails or bearing blocks. The rail alignment precision will require careful machining of the surface on which the rails are mounted. Deformation of this surface will probably require shimming of the rails using optical techniques to check their positions. It was noted that the manufacturer recommends the arrangement of bearings that is currently planned. In order to understand the specifications on the bearings in this case, the assumption must be made that the structure supporting the rails is perfectly rigid and that the sharing of the load is somehow produced when the rails themselves deform. The THK manufacturer must be consulted to determine if indeed this is the operative assumption. If so, the flexibility of the structure supporting the rails can only improve the load sharing. Another possibility is to use only a pair of rails. Because we are already using the largest THK rails available, these would probably be custom built rails perhaps employing Hillman rollers. The experienced with Hillman rollers on OHIPS has been that their motion is non-reproducible. However, a better design might eliminate this problem. Another point is that such rails would necessarily be larger and might generate interference. The first step in attacking this problem is to understand how the THK bearings will share the load. This will require some discussion with the manufacturer and possibly FEA calculations. 5 STATUS OF THE SEXTANT SUPPORT CONCEPT Chris Tschalaer presented an elaboration of the sextant support concept that he introduced at the last engineering meeting. This support employs a vertical arc to constrain the vertical motion of the OOPS, much as the tether constrains its horizontal motion. The arc, supported by a tethered base is situated at a radius forward of the OOPS quad. Rollers fastened to the dipole support ride in a track machined into the arc. The base floats on airpads. The bulk of the Page 4 weight of the OOPS module is not carried by the arc, but rather by cables extending from each module up to a support gantry. This gantry consists of a tower resting on the base behind the OOPSs with a cantilevered arm that extends over the modules. The overhead crane is employed to adjust the length of cable from the cantilevered arm down to an OOPS module. This in turn adjusts the out-of-plane angle because of the arc constraint. The cable is then locked to the cantilevered arm. Larger drift distances are accommodated by moving the dipole relative to the roller assembly that rides in the arc track. Chris presented calculations that indicate that under static loads and with reasonable dimensions for the cross section of the vertical arc, the deformation of the arc will not be an issue. A number of objections were raised, however, concerning this support concept: 1. As an OOPS is lowered while being constrained by the vertical arc, the vertical arc must compensate for the tendency of the OOPS to drop down vertically with an upward vertical force. The arc must be sized for the resulting load, not just the static load when the cable is carrying the weight. 2. Likewise, considerable forces will be put on the dipole support. Quite clearly the existing stainless steel dipole clamp and interface are not sufficient. They must be replaced. This in turn will necessitate repeating the OOPS internal alignment. More seriously, there was concern about the effects of loading the joint between the dipole and quad. The experiments require that the optical properties of the spectrometer not change significantly with the spectrometer position. 3. There was considerable concern about the lack of fine adjustments. The sextant support relies on the precision and alignment of the vertical arc to constrain the pointing and location of the OOPS modules. Overall leveling of the arc could be accomplished with jacks on the tethered base Possibly overall yaw of the arc could also be managed. But the individual spectrometer roll, pitch, yaw, and transverse and radial position can not be corrected. This "dead reckoning" approach violates experience with Murphy's law and its alignment corollary: "the degree of freedom that you can not adjust is the one that will be out of alignment". 4. The sextant support, with no individual yaw adjustments and the vertical arc at relatively small radius can not provide the double-yaw (yaw-yaw) positioning mode that permits one to place spectrometers nearly above and below the beam line. 5. Perhaps the most important issue is cost. It is likely that the quantity of steel required for the sextant support is not much different than that needed for the "minimal" support. The complexity of the rail system of the minimal support is traded for complexity in the vertical arc rail and the need to re-design the dipole clamps for the sextant support. Both systems employ air pads fastened to a tethered base and a rigid vertical tower. At first glance then, the cost of the two systems will likely turn out to be of the same order. If that is the case, then the other considerations mentioned above lead us to prefer the minimal support. Only if substantial cost savings, like a factor of two, can be realized, should we pursue the sextant support. This point was made at the previous engineering meeting. Initial engineering, sufficient to provide a realistic cost for the Page 5 sextant support, has already been approved by the lab director. Both support designs will therefore be carried out in parallel until the next engineering meeting, in roughly one month. At that time with cost estimates in hand, a decision will be made as to which support will be built.