Minutes of the IR Engineering and Physics Meeting of 22 Mar 96
Quad Testing
John Seeman presented more results from a series of tests on an SSC
permanent magnet quad, and a HER Arc quad, with solenoid and mirror
plates.
SSC P.M. Quad Measurements:
Further Hall probe measurements of this quad show that the vertical
component of the B-field on the x-axis, just outside the magnet, is
symmetric from one end to the other. The peak stray field is 1400 G at
the outer collar, in the middle of the magnet length. This drops down to
725 G at the ends on the outer collar, and also drops off quickly with
increased radial distance from the quad centerline.
Quad with Solenoid:
Measurements presented last week showed that the mirror plate with a
bore shaped like the quad pole tips did not adequately shield the quad
from the external solenoid field. To qualify this, the solenoid axial field
strength was measured with a Hall probe through the solenoid and into
the quad magnet. When the mirror plate was added, the field in the
solenoid increased, as expected, then it dropped off in the entrance to
the quad, as some of the axial field was shunted out to the flux plate.
However, the axial field only dropped by 40%, indicating that 60% of the
axial field entered the quad magnet proper.
In another experiment, a single winding was wrapped around each quad
pole at its root, so a bucking octupole field could be generated. At 30
Amps, this winding produced a 1 x 10^-4 octupole harmonic in the quad.
New, higher precision octupole windings are being made which should
be able to produce ten times the bucking field. These should be able to
buck up to 300 G of solenoid field.
Iron Q2 Update
Fran Younger ran a poisson analysis of a possible octupole winding
configuration for the iron Q2 design. Because of the tight space
constraints from the four septum coils, the octupole windings lie just
inside the main coils, where the coils meet the steel of the pole tips. The
coils contain seven windings run at 5 Amps. They produce a 4 G
octupole field at 4.23 cm (the bore radius of the magnet). This is roughly
the octupole harmonic expected from a 100 G solenoid field.
In essence, this method is using a non-local correction (octupole
windings) to offset a local harmonic problem (octupole on the in-board
end of the magnet). This is the antithesis of the P.M. Q2 harmonic
correction ring, which has been shown by the Lattice Group to be an
acceptable way to cancel harmonics. Since the beta-function is
changing rapidly through the magnet, the concern has been that local
correction for distributed harmonics can not completely negate their effect
on the beam. However, this does not seem to be a problem for the P.M.
Q2 correction scheme. Thus, by inference, the octupole windings should
also produce acceptable correction (although we will check on this,
anyway, with the Lattice Group).
B1, Q1 Harmonic Analysis Update
Mike Sullivan reported on harmonic analysis for B1 and Q1 permanent
magnets. First, harmonic magnitude is more sensitive to the block
magnetization angle error than to the error in magnitude. So far, these
errors have been modeled as square-wave errors, with constant
distribution within the error range. However, recent data from Shin-Etsu
shows that for both magnitude and angularity error, the blocks are spread
in a gaussian-like distribution. Mike added this error distribution to the
MBUILD program, and found that, for the tight sigma shown for the Shin-
Etsu material, the harmonics are reduced on the order of 30-50% from
the flat-distribution simulations. This is significant because it means we
can open up the error tolerance in our procurement specification,
knowing that the gaussian distribution of the as-received material will
produce better harmonics than the spec values would suggest.
MBUILD was then further modified to build variable-radius slices, to
better simulate B1 magnet fabrication. Using standard 0.005"
dimensional tolerances, and +/- 2 deg and 2%, with a remanent field of
10.5 kG, the magnet strength is exactly what is needed (337.5 kG-m).
The 4% margin that had been built into the simplified design of last fall
was "used up" by the tolerances and errors. This simulation includes a
derating factor to account for a relative permeability of 1.05.
For this magnet, the relative harmonic values were 2 x 10^-4 for the
sextupole harmonic, and <10^-4 for all other N. The radius of expansion
for the harmonic calucation was the BSC at the slice being evaluated.
This changes through B1, from a radius of 0.92 cm to 2.0 cm. The intent
of using the varying BSC as the radius of expansion was to find a method
to compare the relative effect on the beams of harmonics from slice to
slice. This method essentially scales the harmonic by the BSC.
There was some discussion as to whether this was the most appropriate
scaling. The BSC changes with (Beta)^0.5, which defines steering
effects of a magnet, but tuning effects vary linearly with (Beta), so this
may be a better scaling factor. Mike will look into this. This method of
scaling harmonics has an impact on Q1 as well. Although the bore of the
magnet is constant, the BSC's are changing rapidly, so the slice
harmonics will scale differently, depending on the location of the slice
along the magnet.
P.M. Quad Update
Dave Humphries gave a quick update on the P.M. Q2 shielding design
progress. Two concepts for solenoid shielding were presented. One
used nested cones in front of Q2 to shunt the stray solenoid field away
from the bore of Q2. An active shield could use a single cone, with
bucking coils to accomplish the same task. Either configuration must fit
within tight space constraints in-board of Q2, where the Forward
shielding plug wraps around the Septum beampipe.
These minutes, and agenda for future meetings, are available on the
Web at:
http://www.slac.stanford.edu/accel/pepii/near-ir/home.html