Minutes of the IR Engineering and Physics Meeting of 10 May 96 Q2 Magnet After the Q2 Magnet Review, a decision was made to use the iron/copper technology for this magnet. Subsequent to that decision, James Osborn was slated to lead this project at LBL. James has been heavily involved in PEP-II, working on LER magnets, and is already jumping into this new project. Q2/Q4 Region Layout Bob Holmes and Lou Bertolini have updated the layout of the vacuum chambers between Q2 and Q4. This now includes fixed 6.75" flanges on both beamlines, with tapped flanges up close against the Q4 mirror plate. The flanges must be machined and offset on their chambers to get them to fit.. This layout includes the latest on the Q4 mirror plate, which is 0.375" thick, and spaced off the coil by 0.5". Johanna Swan felt that this was as compact an end design for Q4 as practical. For the Q2 magnet, the coils for Q2 are assumed to stick out beyond the iron core by 36 mm. James Osborn thought that this was sufficient. This gives a total space between the flanges and Q2 coils of 128 mm. This space will be filled with: clearance for flange bolt removal; SK1 skew quad; A possible harmonic correction ring for Q2; In/outlets for Q2 chamber cooling; In/outlets of Q2 magnet coil cooilng and power leads. Dave Humphries felt that this was sufficient room for a P.M. SK1 rotating ring, which was split-able. Fran Younger had developed a conceptual design for an iron SK1, back when space was not considered an issue. He will revise that design with the new envelop to see if this is an option. The chambers in the new layout will be completely custom. The LEB chamber is formed Glidcop high-strength copper, with a wire EDM'd transition cone in the Q2/Q4 region. The HEB chamber is stainless steel, and also includes a transition to match the Q4 chamber cross-section. The HEB chamber includes pumping chambers for cartridge NEG pumps, and both chambers are water cooled. Q1 Harmonic Correction by Moving Blocks Mike Sullivan has developed an algorithm to correct any random distribution of harmonics in a Q1 magnet slice. It is based on moving the blocks radially, in a sinusoidal pattern. For radial motions of dr = dr0 * sin(m * theta + phi), a harmonic of n = m+2 is generated, with a phase of phi. For a 32-block magnet, harmonics up to n=15 can be generated, each with independently adjustable normal and skew components. Using this algorithm, and the MBUILD program, a perfect slice can be built, then a "perfect" harmonic overlaid. This also generates a dipole or skew dipole term, plus some higher harmonics which are typically 1% of the target harmonic. This method can cancel any given set of harmonics my superposing block motions to generate the cancelling field. The remnant higher harmonics are cancelled by another iteration of motion. For typical harmonics of 2E-3, at 60 mm radius, maximum block motions of 0.03" are needed to correct. This is just at the range of motion available for the prototype Q1. The available position resolution of block motion limits the quality of correction. For a relative resolution of 0.001", harmonics could be corrected to better than 1E-4. For 0.002", correction was around 1E-4. For 0.005" resolution correction was 2-3E-4. Even for 0.02" resolution, harmonics could still be reduced to the 1E-3 level, showing that just coarse motion alone helps considerably. To investigate the range of adjustment needed, Mike "built" 100 single-slice magnets using MBUILD, and corrected harmonics for each of them.. Lower harmonics were typically in the 2-3E-4 range, and required a maximum block motion of 0.041" to correct up to n = 15. To correct harmonics averaged over 5 slices, a range of 0.057" was needed, but a max motion of 0.030" put almost all the blocks in their theoretically correct location. Results of this study suggest that the needed resolution of the block movement mechanism is 0.001-0.002", and the maximum range of motion should be at least ± 0.03". Andy Ringwall thinks that this is possible, given the design of the prototype magnet and fixturing. He is setting up a fixture test now, which will confirm the resolution of the mover. Q1 Trim Coils and Thermal Shield Stan Ecklund and Martin Nordby presented work on a two-layer Q1 trim coil. This is an 80 cm long coil which relies on a necked-down Q1 vacuum chamber which raclaims room otherwise wasted on vacuum. The two-layer coil allows cancellation of both the n = 6 and 10 harmonic. The design has four inner- and two outer windings, using 0.255" square copper conductor running at 7500 A/in^2. This produces a gradient of 0.894 T/m, which is slightly less than the target of 0.91 T/m (5% of the main Q1 int(GdL)). Harmonics at a reference radius of 4.12 cm are negligible for n = 6 and 10. The n = 14 harmonic is 5E-5 of the main field. Introducing a random misalignment of the coil windings, with an rms of 0.008" produces marginally good harmonics. This effect needs to be studied more. At full field, the coil current is 385 Amps, producing 3840 W total power. Each coil is cooled separately, with water flowing at 0.23 gpm, and 6 ft/sec. The water temperature rise is 6 degC. Between the coils and the inner radius of Q1 there is a Thermal Shield, which serves to stabilize the temperature of the Q1 Magnet. This is a 1.5 mm copper sheet, with brazed-on cooling circuits. With the two-layer Trim Coil covering the first two-thirds of the magnet, the water for the Thermal Shield must be routed from the in-board end of the magnet, where it can snake into the center of each winding. This is the only available space for the Shield cooling. Assuming that 10% of the power from the Trims is dumped into this Shield, the azimuthal temperature variation around the shield is 3 degC. This is slightly higher than the target of 2 degC, but assumes a very conservative heat load. For reference, the heat conducted across the 1.5 mm air gap between the Trims and the Shield, given an average temperature difference of 7.5 degC, is only 46 W. The Thermal Shield is mounted to the Q1 Magnet, with an air gap for alignment and insulation between it and the Trim Coils. The Shield clam-shells over the Trims, then the Q1 Magnet is slid over the Shield. These minutes, and agenda for future meetings, are available on the Web at: http://www.slac.stanford.edu/accel/pepii/near-ir/home.html