To: Distribution 30 Sep 96
From: Martin Nordby
Subject: Minutes of the IR Engineering and Physics Meeting of 27 Sep 96
Hard-Copy Distribution:
Bob Bell | 41 | David Kirkby | 95 |
Lou Bertolini | LLNL L-287 | Jim Krebs | 41 |
Gordon Bowden | 26 | Harvey Lynch | 41 |
Pat Burchat | 95 | Tom Mattison | 17 |
David Coward | 95 | James Osborn | LBL B71J |
Scott Debarger | 17 | Andy Ringwall | 17 |
Hobey DeStaebler | 17 | John Seeman | 17 |
Jonathan Dorfan | 17 | Knut Skarpaas VIII | 18 |
Stan Ecklund | 17 | Mike Sullivan | 17 |
Alex Grillo | 95 | Uli Wienands | 17 |
John Hodgson | 12 | Mike Zisman | LBL B71J |
Hank Hsieh | LBL B71J | ||
David Humphries | LBL 46-161 | Orrin Fackler | LLNL L-291 |
Roy Kerth | LBL 50-340 | Lew Keller | 41 |
Electronic Distribution:
Curt Belser | Rick Iverson | Jeff Richman | Jack Tanabe |
Catherine Carr | Nadine Kurita | Natalie Roe | Rick Wilkins |
David Coupal | Georges London | Ross Schlueter | Fran Younger |
Fred Goozen | Joseph Rasonn | Joe Stieber |
Q2 Shielding Plug Update
In the ongoing saga of the Q2 Shielding Plug design and analysis, Jim Krebs reported on further work on the Plug design. There are essentially three Plug options on the table right now. Each produces a different minimum thickness in the third finger. The options are:
Option | Finger Min Thickness |
Big Bore | 123 mm |
Offset Cutout | 116 mm (LEB side) / 125 mm (HEB side) |
Offset Fingers | 132 mm (LEB side) /102 mm (HEB side) |
The Big Bore configuration produces the largest axisymmetric cutout needed in the third finger to clear all PEP-II stay-clear requirements. It is the only way to make an axisymmetric solution to this inherently 3-D problem.
The Offset Cutout design requires two lathe cuts on different axes. This produces a conical crescent-shaped discontinuity in the transition region between the cuts. However, as Jim's 3-D model showed, the discontinuity does not look too sharp, and may be easily blended out with a subsequent machining or hand-grinding job. The advantage of this design is that it preserves as much axisymmetry as possible, making the analysis and machining easier. Its main disadvantage is that it produces a thin spot at one azimuthal location in the third finger. While Orrin's analysis has shown that the stray field for this design is adequately low, Orrin used the average finger thickness. Reality will introduce some kind of left/right asymmetry to the idealized model.
To attempt to correct for this, we talked last week about offsetting
the entire finger geometry inside the cone. This would presumably
maintain the maximum third finger thickness, will still honoring
the PEP-II stay-clear. Martin Nordby investigated this option.
While this increased the minimum thickness from 116 mm to 132
mm, it introduced more asymmetries. The big source of trouble
was the transition from offset cone around the Q2 Magnet, to on-center
cylinder around the septum chamber. This transition forced the
third finger to taper down unacceptably. Also, by offsetting the
entire finger geometry, this design introduced inherent asymmetry
to the second and first fingers. Not only would the two sides
intercept different amounts of flux, but the path length and area
for channeling this flux would differ significantly. The general
consensus was that this design introduced more asymmetries than
it solved, and was not a viable alternative.
Plug Magnetic Modeling
Orrin Fackler reported on ongoing analysis of the Shielding Plug. For the Big Bore configuration, he got Bmod = 270 G at the front face of Q2 (R = 5 cm). At Z = +2.8 m, Br = 90 G for the same model (R = 4 cm). All model used KEK Venus steel.
Orrin looked at using a "small-finger" bucking coil around the front of Q2. The field this generates in Q2 adequately reproduces (with opposite sign) the stray field distribution from the main solenoid. Orrin also added a larger bucking coil around Q2. Optimizing both coil strengths to minimize int(Br*dz), he got 0.42 G-m integral, with Bmod = 212 G, max. This assumed NI = 618 A-turns on the small-finger coil, and NI = 5900 A-turns on the Q2 coil.
Using the small-finger coil only, Bmod = 91 G, but int(Br*dz) = 1.98 G-m (coil current = 618 A-turns). This shows that the two criteria: Bmod, and int(Br*dz) do not respond the same to changes in stray field distribution. Stan Ecklund recommended adding a third criteria: Bz at the entrance to the quad. Since int(Br*dz) is sensitive to the distance over which the integration occurs, and Bmod is more a global figure-of-merit value, neither predict the expected octupole in Q2. However, Stan showed last week that Bz at the entrance to the Q2 bore is essential entirely converted to octupole in the magnet (at the 70% level). Thus, optimizing to minimize Bz should, really, produce the lowest possible octupole in Q2, for a given field distribution.
John Seeman opined that the real Q2 is offset with respect to
the opening in the Shielding Plug. He thought that this offset
in the entrance angle of the solenoid stray field should produce
odd multi-poles in the magnet. He thought that a horizontal dipole
winding could help buck this out. James Osborn said that the magnet
already has two such windings to correct for the inherent asymmetry
of the magnet, and that they, or offsetting the magnet with respect
to its ideal position could do the job.
Q2 Position and Raft Design
Scott Debarger reported on the status of the Raft design. He has put in the new Q2 Magnet cross-section, in its new position (per Mike Sullivan's $88k MAGBENDS run). Because of these changes, the magnet is not centered inside the U-trough of the Raft any more. Fortunately, to center the support under the magnet, the support must move back towards the HEB, making the support, and cutout in the Plug more symmetric. Scott will move the U-trough to find the exact number, but he thought the move would be 1-1.5 cm.
Before Scott finalizes the Raft, James will nail down the Q2 Magnet cross-section, and send it to Scott.
Also, the current U-trough design leaves no room in front of the
Q2 Magnet for Orrin's "small-finger" bucking coil. However,
if the U-trough flares slightly larger, there should be room enough
for it inside the support. This should improve its ability to
buck the stray field, and only require a small additional cut-out
in the third finger. Scott and Orrin will work up the placement
for this coil.
Q1 Prototype Status
Andy Ringwall and Stan Ecklund reported on the status of the Q1 prototype slice assembly and testing. All P.M. blocks are here and have been tested in the new Helmholtz coil. The dipole ring is assembled and tested in the new rotating coil (with dipole and quad bucking coils). The RFP for the production order went out today.
Results from block measurements show that the Shin-Etsu SmCo material had a tight distribution of Br for all blocks. The distribution was <1% spread for 100 blocks. Angularity tolerance was not quite as good. 90% of blocks were within the +/-2° specification, but some were outside this.
To fully saturate the material during pulse-magnetization, a 55 kOe field is needed. However, the coil used was under-strength, and aluminum fixturing was used, so the initial magnetization produced under-strength blocks. When this was corrected, the blocks were heated to correct for the anemic coil strength, and wooden fixtures were used (don't worry, no nails, either!).
Prototype assembly has proceeded well. Holding the blocks by aluminum
tabs works well, as does epoxying the dipole blocks to the outer
support ring. The insertion fixturing provides four degrees-of-freedom,
which are all needed during assembly. The fixturing worked well,
but the clearances were too tight, so assembly was tricky. Andy
thought that gaps could be opened up, with little loss of accuracy,
so assembly would be easier.
Stan showed that the calculated field strength for the dipole was 0.2352 T, based on Br = 1.05 T, and mu = 1.05. Average values for the 100 prototype blocks are: Br = 1.067 T, and mu = 1.064,
Stan folded the magnetic actual values for the specific blocks used in slice #1 into his model. This produced an expected filed of 0.2392 T, which agreed within the error bars of the measurements with a stretched-wire test.
The harmonics of the slice are all within 1 G-m (when integrated over the 23 blocks of Q1), and agree in distribution very well with the predicted trends. Magnitudes of most harmonics were within a factor of two with that predicted for the magnet. For n = 3 and 4, measured harmonics were 5 x 10^-4, while all others were <10^-4.
This is an extremely good start, since the values are both very reasonable, and also agree well with predictions. These harmonics will be corrected slice-by-slice using the quad block radial motion, but their small magnitude will mean less radial motion is needed.
Andy and Stan will also look at external stray fields around the
slice.
These minutes, and agenda for future meetings, are available on the Web at:
http://www.slac.stanford.edu/accel/pepii/near-ir/home.html