Shigemi Sasaki, Elizabeth Moog, Maria Petra

1
Magnetic Design Choices

• Shigemi Sasaki, Elizabeth Moog, Maria Petra

• Magnetic Design
• Evaluation of performance
• Magnetic Material
• Type and reason
• Radiation susceptibility

2
Magnet Pole Dimensions
3
Magnetic field lines
4
Permeability in the pole
5
Demagnetizing field next to pole
6
Demagnetizing field at edge of pole
7
Demagnetization Curves for N39UH
8
LCLS Undulator Model
Model calculation was made by using RADIA. NdFeB
magnets with Br1.24 T, and Vanadium permendur
poles were assumed for calculation.
9
Shin-Etsu NdFeB Grades
10
(No Transcript)
11
Gap dependence of magnetic field NdFeB Shin-Etsu
N39UH
12
Gap dependence of magnetic field
13
Gap dependence of magnetic field
14
Considerations of SmCo vs NdFeB

• SmCo is known to have a greater resistance to
  radiation-induced demagnetization than NdFeB, and
  the Sm2Co17 variety is better than SmCo5.
• Higher coercivity has been found to correlate
  with higher resistance to radiation damage.
• A new grade of NdFeB magnet (HILOP by Hitachi)
  has a higher coercivity than standard NdFeB. We
  estimate that the higher coercivity might make a
  difference of 6 in the radiation dose needed to
  cause a 1 decrease in the field.
• Using SmCo instead makes a bigger difference
  Sm2Co17 gives a damage level of less than 0.2
  out to as high as they exposed the magnets.
  Exposure to cause 1 loss in NdFeB ranged from
  23 to 30 x 1013 electrons (at 2 GeV), whereas
  the dose to SmCo went out to 40, 65, or 175 x
  1013 electrons. So on that scale SmCo wins
  hands down 1.
• 1 T. Bizen, T. Tanaka, Y. Asano, D.E. Kim, J.S.
  Bak, H.S. Lee, H. Kitamura, Nucl Instrum. Meth.
  Phys. Res. A467-468 (2001) 185.

15
Coercivity Dose
16
NdFeB
 
Demagnetization of undulator magnets irradiated
high energy electrons, T. Bizen, T. Tanaka, Y.
Asano, D.E. Kim, J.S. Bak, H.S. Lee, H. Kitamura,
Nucl Instrum. Meth. Phys. Res. A467-468 (2001)
185.
17
Reasons for Our Choice

• Does the extra resistance of as-purchased SmCo
  make a practical difference in the longevity of
  the magnets in LCLS so that it is worth going
  that way?
• Look at how long it might take for damage
• Say it takes 3x1014 electrons at 2 GeV to lose
  1 in NdFeB field.
• Thats 5x10-5 C, or 50 microcoulomb.
• If we can only tolerate 0.01 loss in strength,
  then we can only tolerate 500 nC.
• The beam will be up to 1 nC per pulse, with
  pulses at 120 Hz
• Assuming a loss rate of 10-6 of the beam, how
  long to damage magnets?
• current x time x loss rate 120 nA x time x
  10-6 500 nC gives time 4 x 106 sec
  1000 hrs 1.5 month
• The SmCo might last 15 times longer (0.2 loss
  at 100 x1013 electrons instead of 1 loss at 30
  x 1013 electrons), which would take the time to
  22 months, but thats still not good enough. The
  beam loss rate in the undulators really has to be
  infinitesimal.

18
Continued

• SmCo magnets are slightly weaker than NdFeB, by
  about 10. For a 30-mm-period undulator, that
  would translate into a gap difference of nearly 1
  mm.
• SmCo magnets are more expensive, by about a
  factor of 2.
• We have not yet bought a set of SmCo magnets, so
  we dont know first-hand how uniform the quality
  is, though Shin-Etsu claims they could meet the
  same requirements with SmCo as they do with NdFeB.