prev

next

of 14

View

20Download

1

Tags:

Embed Size (px)

DESCRIPTION

Improving a magnetic shield: what works and what does not. 26 March 2013 Kiril Marinov. Cylindrical shell in an external homogeneous field. A ferromagnetic cylinder in an external homogeneous field B 0 =0.1T. Is there anything else other than using a thicker shielding box?. B in max

Improving a magnetic shield: what works and what does not26 March 2013Kiril Marinov*

- Cylindrical shell in an external homogeneous fieldA ferromagnetic cylinder in an external homogeneous field B0=0.1TBin max
Suggested ideasAdding a second layer of mu-metal besides the steel?Adding steel only where the flux density is higher?Using co-axial cylinders with gaps (zero gauss chambers)? Keep size within reason.

Mu-metalTwo BH curves obtained from different sources, similar but not identical.

Mu-metal vs. 1010: permeability

7 cm steel and 1.8 cm mu-metal vs 7 cm of steel7 cm steel and 1.8 cm of mu-metal7 cm steel onlyImprovement is visible but is the mu-metal layer really working?

7 cm steel and 1.8 cm mu-metal vs 8.8 cm of steel7 cm steel and 1.8 cm of mu-metal8.8 cm steelAll-steel, 8.8 cm-thick shield preforms better

Permeability distribution7 cm steel and 1.8 cm of mu-metalThe mu-metal layer is fully saturated. Bringing mu-metal in contact with or close to strongly magnetized steel results in mu-metal saturation.The permeability of the mu-metal layer is lower than that of the steel layer. This results is poor shielding Introducing gaps between the two materials does not eliminate the problem.

- Boundary conditionsIf 2/1>100 and B1~1.5T is B2>150T? An interface between two magnetic materials: H|| must be continuous across the interfaceThe mu-metal has to saturate. This results in 2
Adding steel where the flux density is higher5 cm-thick can and a second, 5cm layer, 1cm away, covering half of the surface area of the can. Mirror symmetry w.r.t. both X and Y axes, 5 cm-thick can acting alone9-fold reduction of Bmax ; 25% lighter than a 10 cm can. If we get the shield thickness wrong we can still fix this by adding steel at the appropriate places. No need for a good contact between the two layers.

Adding steel where the flux density is higher5 cm-thick can and a second, 5cm layer, 1cm away, covering half of the surface area of the can. Mirror symmetry w.r.t. both X and Y axes, 10 cm-thick can acting aloneIf we get the shield thickness wrong we can still fix this by adding steel at the appropriate place(s). No need to worry about good contact.

Co-axial cylinders with air gaps5 cm-thick can, 1 cm air gap, 4cm steel Solid 9 cm-thick can acting aloneThe 1 cm gap results in lowering the field in the shielded region (by 13%) without increasing the weight of the shield.

Co-axial cylinders with air gaps5 cm-thick can, 2 cm air gap, 4cm steel Solid 9 cm-thick can acting aloneThe 2 cm gap results in lowering the field in the shielded region (by 26 %) without increasing the weight. The gap results in the outer layer carrying higher flux density thus allowing higher permeability and lower flux density in the inner layer.

SummaryThree different strategies for improving shield performance have been considered:Adding a second layer of mu-metal to the steel?Adding steel only where the flux density is higher?Using co-axial cylinders with gaps (quasi-zero-gauss chambers)?Does not work at flux density levels typical for the MICE shielding problem. Works. Allows corrections to be made at a later stageWorks. Could be implemented, if needed.If you can recommend a good Physics article on zero-gauss chambers please, e-mail me. Thanks.