To: Distribution 13 Nov 95 From: Martin Nordby Subject: Minutes of Near IR Engineering Meeting of 10 Nov 95 Beam Pick-up Concerns of D.C. Dave Nelson reported on preliminary work he has done on beam pick-up in the Drift Chamber, and RF shielding requirements. Dave summarized the experiences of other Drift Chambers. SLD sees no RF pick-up from SLC, but has a lot of aluminum shielding. Mark II on PEP-I saw some pick-up, as did CLEO at CESR (although this may be cable cross-talk, per CLEO people). To minimize possible pick-up from the BPM at 90 cm, the BPM cables will be solid shield coax cable. Dave built time-domain models for the PEP-II machine, with and without the gap in the bunch train. The gap introduces a 136 kHz frequency, but also many other harmonics, since it is essentially a square-wave gap. Dave then looked at the FFT of the response of various components. His initial results show that, with the Beryllium I.P. vacuum chamber, the transmission out of the chamber is nearly 10 times that of other comparable machines, and copper plating the chamber does not help. The next step is to look at how RF shielding of the SVT and Support Tube help attentuate this transmitted signal, and at how the D.C. ground wires affect this. Q1 Magnetic Analysis Update Mike Sullivan reported on his ongoing analysis of the B1 and Q1 permanent magnet. He has found that a 32 block hybrid quad/dipole Q1 design (the current working design), with "reasonable" tolerances provided by Martin Nordby, produces field qualities which are two times those assumed by the Lattice group (this was encouraging news). Furthermore, most of this error is dominated by tolerances of the block remnant field magnitude and orientation (assumed to be +/-1% and +/-1 degree, respectively). The harmonics arrived at for the computer-generated magnets assume that there is no sorting of blocks, and no correction attempted. One large harmonic did appear at n = 34, which typically is not included in lattice simulations. This raised a discussion on the possible impact of such a high-order harmonic. However, since the harmonic magnitude varies with 1/radius^(n-1), this dies off quickly. Mike has now turned to how best to sort the blocks to further reduce harmonic content. He has started with Gordon Bowden's sorting algorithm of grouping blocks for a given slice which have similar remnant field error vectors (magnitude and direction). This reduces the odd harmonics, but not the even. Mike, with Gordon's consulting, is trying to understand this phenomenon, and to further optimize the sorting method. Further work will be directed towards finalizing the preferred sorting method, including sorting of blocks and slices. Also, Mike will look at a correction scheme for the quad blocks, so that harmonics can be reduced by moving blocks radially. This was done successfully for the CESR P.M. quads, and is being planned as the primary method for harmonic correction for Q1. As reported a few weeks ago, the dipole ring for Q1 does not need any sorting or adjusting. An unadjusted dipole slice produces harmonics which are only 20% of the allowable value. Q1 Assembly Method Andy Ringwall updated the status of the Q1 assembly planning, including the sequencing and fixturing. During asembly, the fixturing must hold both quad and dipole blocks rigidly, despite large (up to 100 lbs) and non-uniform forces exerted on each of the 48 blocks in a slice. Furthermore, as mentioned above, the quad blocks must be adjustable, so that undesirable harmonics can be minimized. Finally, if undesirable harmonics are still present after assembly, there is still provision to magnetically shim each slice. Finally, since there are over 2100 blocks needed for just the Q1 magnet, the entire assembly and testing process must be efficient (i.e.: low-cost). The assembly process proceeds as follows: 1) Measure each block with Helmholtz coils to determine remnant field strength and orientation. Zach Wolf felt that this could be done to an accuracy of less than 0.2%. 2) Sort blocks for quad (and possibly dipole) slices.. 3) Assemble dipole ring by positioning blocks with respect to pins on the support collar. Measure harmonics using a rotating coil. Pot dipole ring in epoxy, and re-measure. 4) Assemble quad blocks in their fixture (in pre-assembled dipole ring), with radial adjustment possible for each block. 5) Insert rotating coil and measure quad/dipole field harmonics. Radially adjust quad blocks as needed, and re-measure harmonics. Find offset quad center. Pot quad blocks in epoxy, and re-measure. 6) Machine away support mounts which are epoxied to the end face of each quad block. 7) Re-map slice and add magnetic shims, if needed. Possibly pair slices to cancel sextupole harmonic, and measure the pairs. 8) Assemble slices into partial, then complete Q1 assembly. Slices are compressed together, pinned to ensure good alignment, held with temporary external clamps, then bolted together with threaded rods running the length of the magnet. Quad Block Assembly Fixture The assembly fixture for the quad blocks is built off of a precise datum surface to minimize tolerance stack-up. Each block has a temporary support epoxied to one of its trapezoidal faces, which mounts onto an individual sliding mechanism which allows radial motion. A slice is assembled flat, with blocks lowered into place vertically by a cross- slide with a support arm mounted to it, which indexes to the correct location, but provides adjustment to ensure accurate azimuthal positioning. The assembly fixture is removed after all 32 blocks are positioned, so the measurement coil can be inserted down the bore of the magnet. The coil then remains for the duration of the testing and block adjustment. Further work includes refining the assembly and post-assembly machining methods, better understanding the asembly magnetic forces, and how they differ from the as-installed block forces, and developing a plan for stabilizing the magnetic behavior of the SmCo material. A review of the Q1 design and assembly is planned for Thursday, 30 Nov, 1- 3:00 pm in the SLAC Yellow Room, with TV link to LBL. Q2 Design Update Dave Humphries gave a quick update on the design of the permanet magnet Q2 option. There is some question as to the required tune range for Q2, and the resulting trim capability of the magnet. This is partially a function of the desired energy range over which Q2 (and Q1, for that matter) can fully function. Also, changes in this requirement will affect the mechanical and magnetic design of the trim slices for Q2, which are counter-rotating Halbach quad arrays. Stan Ecklund and John Seeman will nail down the specifications for Q2 within the week, so Dave can proceed with the design. Dave will provide Mike Sullivan with preliminary values for quad gradient versus position for this tapered quad design. Mike can then get this to the Lattice group to simulate the result of effectively moving the center of Q2 out- board by tapering it. As the interface with the Q2 magnetic shielding in the Detector, and supports for the beamline components become more difficult problems, the decision to commit (or not) to a permanent magnet Q2 design is becoming more critical. We will shoot to have a Conceptual Design Review for Q2 in mid-December, at which time a decision will be made. The schedule for this decision impacts the design of the Detector door steel, which is on their critical path, and is due to be released for bid by early January. These minutes, and agendas for future meetings, are available on the Web at: http://www.slac.stanford.edu/accel/pepii/near-ir/home.html