Background of the Invention
[0001] The present invention relates to passively shimming magnetic resonance magnets to
obtain imaging quality homogeneity in the bore of the magnet.
[0002] To create a highly uniform magnetic field with an electromagnet or array of permanent
magnets, it is necessary to build the magnet to a carefully specified shape, and to
strive to minimize the deviations from the specified shape due to manufacturing variations.
The resulting magnets, however, typically require field corrections to achieve the
desired level of inhomogeneity, due to deviations of the magnet from the design or
due to the presence of ferromagnetic material in the vicinity of the magnet.
[0003] To improve field uniformity, correction coils are typically used. These coils are
capable of creating different field shapes which can be superimposed on an inhomogeneous
main magnetic field to perturb the main magnetic field in a manner which increases
the overall field uniformity. Unfortunately, many sets of such coils are typically
required. A state of the art magnetic resonance (MR) imaging magnet has between ten
and twenty independent sets of correction coils, each with its own power supply to
provide the correct current flow. Naturally, these coils add significantly to the
cost and complexity of the magnet.
[0004] One way of removing the need for correction coils is to shim the magnet passively,
using only pieces of iron to bring an initially inhomogeneous field to within imaging
homogeneity specifications. With the iron placed inside the bore of the magnet a minimal
addition to the size and weight would be required. A passively shimmed magnet would
be cheaper and more reliable than the typical set of correction coils presently used.
[0005] The primary difficulty in implementing such a shimming approach lies in predicting
the locations and sizes of iron pieces required to shim the field. Electromagnetic
coils are generally designed to produce certain terms of a spherical harmonic expansion.
Such a design criteria is difficult to implement with passive shims because the permeability
of iron cannot be reserved, whereas a current reversal through a coil can be used
to obtain a field reversal in a correction coil. Additionally, the size and complexity
of the groups of shim pieces which would be required to produce a single harmonic
would not make this approach feasible. Since magnetic coupling between the shims is
also a complicating factor, shimming with large pieces which inevitably become physically
close to one another increases the difficulty in properly shimming the magnet.
[0006] Passive shimming is currently used to correct large deviations in magnetic fields
than cannot be corrected by the available correction coils alone. The passive shimming
is accomplished by placing a piece of iron in an appropriate place outside the magnet.
The desired level of field uniformity can then be achieved by the correction coils.
[0007] It is an object of the present invention to provide a method of passively shimming
for a magnetic resonance magnet which can achieve a level of field inhomogeneity
required for magnetic resonance imaging, without the use of correction coils.
[0008] It is a further object of the present invention to provide a method of determining
the optimum axial and circumferential location of shims to bring the field homogeneity
to a level consistent with magnetic resonance imaging using only pieces of ferromagnetic
material.
[0009] It is a still further object of the present invention to provide a method of minimizing
the total field inhomogeneity and not selected harmonics.
Summary of the Invention
[0010] In one aspect of the present invention a method of passively shimming a magnet having
a central bore and using shims placed in the bore of the magnet is provided. The method
comprises the steps of measuring the initial field homogeneity in the bore of the
magnet. The magnetic field effect of a shim at each of the predetermined permissible
shim locations in the bore of the magnet is checked independently of one another,
to determine the shim strength needed to improve magnetic field inhomogeneity in the
magnet bore. The locations wherein positive shim strengths were found to be beneficial
are selected and used to determine shim strength at each selected location, considering
all selected locations simultaneously. Locations found to require negative shim strengths
are eliminated and with the new selected locations shim strengths are again determined
until all selected locations remaining require positive shim strengths. The shims
of predicted positive strengths are placed in their selected locations in the bore
of the magnet.
Brief Description of the Drawing
[0011] While the specification concludes with claims particularly pointing out and distinctly
claiming the present invention, objects and advantages of the invention can be more
readily ascertained from the following description of preferred embodiments when
used in conjunction with the accompanying drawing in which:
Figure 1 is an isometric view of a passive shimming assembly;
Figure 2 is a partial sectional end view of the shimming assembly of Figure 1 situated
in the bore of a magnetic resonance magnet;
Figure 3 is an end view of another passive shimming assembly;
Figure 4 is a partial sectional end view of the shimming assembly of Figure 3 situated
in the bore of a magnetic resonance magnet;
Figure 5 is a partial isometric view of the passive shimming assembly of Fig. 3 showing
one of the removable drawer sections being slid into place in the passive shim assembly;
Figure 6 is a flow chart for passively shimming a magnetic resonance magnet in accordance
with the present invention;
Figure 7 is a cutaway isometric view of the bore of a magnetic resonance magnet showing
the locations in which the magnet inhomogeneity is checked; and
Figure 8 is a partial side view with dimensions of Figure 7, showing the locations
in which the magnet inhomogeneity is checked.
Detailed Description of the Drawing
[0012] Referring now to the drawing and particularly Figure 1 thereof, a passive shim assembly
comprising a nonmagnetic thin wall tube 11 is shown. The tube is fabricated of fiberglass
creating a tube with a 1/8" wall thickness. A plurality of longitudinally extending
nonmagnetic channel pieces 13 are equally circumferentially spaced about the interior
of the tube 11. The channel pieces extend the length of the tube and are secured thereto
by screws threadingly engaging the fiberglass tube. The channel pieces each have two
projecting edges 13a extending on either side of the channel. The edges are parallel
to the tube and spaced away therefrom. The edges extend the longitudinal length of
the tube. The channel pieces can be fabricated by extruding aluminum to the desired
shape or if eddy currents are a problem, the channel pieces can be pultruded from
composite material. Pultrusion is a process in which continuous filaments are drawn
through an orifice, which also meters out encapsulating resin such as a thermoplastic.
[0013] Arcuate carrier pieces 15 of nonmagnetic material, such as fiberglass, are slidably
mounted between the two adjacent channels with the projecting edges 13a of two adjacent
channels preventing radial movement of the carrier pieces 15. Ferromagnetic strips
17, which provide the shimming, are stacked to the desired height on the carrier pieces
with the strips having a length less than the length of the arcuate carrier pieces.
The ferromagnetic strips can comprise .010 inch thick low carbon steel, cut to an
axial width of 2 cm. and a circumferential extent at their mean radius of 30 degrees
for a 1 meter bore magnet. The strips are secured the arcuate carrier pieces such
as by threaded fasteners 19.
[0014] Th carrier pieces are anchored in their longitudinal position by clamps 21 which
can be fabricated from aluminum. The clamps, which can be more easily seen in Fig.
2, when tightened by bolts 23 to the arcuate carrier piece, captures a portion of
the edge 13a of channel 13 between the clamps and the arcuate carrier piece fixing
the longitudinal position of the ferromagnetic strips.
[0015] The radial thickness of the entire assembly is kept to a minimum to minimize interference
with precious bore space, which is occupied by gradient and RF coils and the patient
table (which are not shown).
[0016] The axial position of the ferromagnetic strips are infinitely adjustable and very
fine adjustment of the shim strength is available by changing the stack height of
the strips. Thinner steel strips may be used to provide finer strength adjustment.
The channels are placed every 45 degrees around the inner circumference of the bore
allowing eight discrete circumferential locations for the ferromagnetic strips. The
choice of circumferential locations in the present embodiment was selected to shim
for spherical harmonics having degree m=2. Fields with m=2 vary sinusoidally with
2φ, where φ is the circumferential angle. Therefore such fields have peaks or nodes
every 45° circumferentially. Providing capability of placing shims every 45° will
clearly allow shimming of the m=2 harmonics.
[0017] The maximum axial force on a 1 cm. thick shim with the magnet energized is about
20 pounds in a 0.5 T magnet. The shim carrier could be moved when the clamps are loosened
while making adjustment of the axial locations. A handle could readily be made which
would allow easy control of the shim carrier with the clamps loose.
[0018] Referring now to Fig. 3 another embodiment is shown. A passive shim assembly comprises
a nonferromagnetic thin wall tube 31, which in the preferred embodiment is fabricated
of fiberglass material 1/8" thick. A plurality of channel pieces 33 are equally spaced
about the exterior of the tube. The channel pieces extend the length of the tube and
are secured thereto by screws threadingly engaging the fiberglass tube 31. Some of
the screws 35 located at the ends of the channel extend above the channel pieces surface
to position the tube concentrically in the bore of the magnet. This can be more clearly
seen in Fig. 4. The channel pieces have projecting edges 33a on either side of the
channel, extending away from the channel. The edges are parallel to the tube and spaced
away therefrom. The edges extend the longitudinal length of the tube. The channel
pieces can be fabricated by extruding aluminum to the desired shape or if eddy currents
are a problem, the channel can be pultruded from composite material. Increasing the
number of circumferential locations provides greater flexibility in eliminating tesseral
(axiperiodic) harmonics.
[0019] Arcuate shaped drawer pieces 37, as shown in Fig. 5, fit between adjacent edges 33a
and extend the length of the tube. Ferromagnetic strips 41, which serve as the shims,
are stacked to the desired height on the drawers and secured to the drawers selected
predrilled holes 42 the more axial holes provided the finer axial field adjustability.
The sliding axial adjustability of the shims in the Figure 1 embodiment provides infinite
adjustment which might be desirable in some situations. The number of ferromagnetic
strips provides an adjustment of strength. The radial thickness of the shims is kept
to a minimum so that the shims fit in the space provided by the adjustable height
of the extended screws 35. The drawers can slide out of the bore of the magnet to
permit adjusting the axial position and thickness of the ferromatic strips. The drawers
can be removed with the magnet energized and the ferromagnetic strip positions adjusted.
[0020] The position and height of the ferromagnetic strips in the bore of the magnet are
used to create magnetic fields shapes which correct for inhomogeneities in the field
created by the magnet. Flexibility of positioning the steel strips is therefore important
so that all the field shapes needed to counteract fields which could be precludes
imaging quality homogeneity can be obtained. Flexibility is also important since it
is doubtful that the initial prediction of the locations of the ferromagnetic strips
will be perfect, the adjustability of the steel shims is important. Arcuate shims
in specific locations are not intended to eliminate specific harmonics. Rather the
combination of all the shims together are intended to increase the field homogeneity.
In situations where required shim height would interfere with available bore space
wider shims can be used in both embodiments.
[0021] Referring now to Figure 6, a flow chart showing the steps in determining the correct
positions and thicknesses of the shims is shown. The first step in block 45 is to
determine the initial inhomogeneity in the bore of the magnet to be shimmed. The magnetic
field is measured in the energized magnet on a imaginary grid 46. For a 0.5 Tesla
superconducting magnet with a 1 meter diameter bore, a grid having 314 points which
lie on the periphery of thirteen circles and on 2 points 20 cm on either side of the
center of the bore on the z-axis as shown in Figs. 7 and 8 can be used. Five of the
circles have their center point on the z axis and a diameter of 44 cm, six of the
circles have their center point on the z axis and a diameter of 20 cm. The two remaining
circles of the thirteen circles have a diameter of 28 cm. The larger circles are situated
on either side of the center of the bore along the z axis at 7.5 and 10 cm. The smaller
diameter circles are located on either side of the center point of the bore along
the z axis at 10, 15 and 20 cm. The intermediate circles are concentric with the smaller
and larger circles at 10 cm on either side of the center of the bore along the z axis.
The measurements are taken at 24 equally spaced circumferential points along each
of the circles. These circles follow the boundary of the volume of interest, as the
maximum and minimum field values must lie there. The circle locations are chosen to
be close to the ideal magnet field extremities, and the actual inhomogeneity should
be close to that found by sampling their points. Based on a comparison of the measured
field at each of the points, if the difference among the points is over 500 ppm, as
determined in block 47, then the position of large shims for gradient cancellation
is determined in block 51 and the field remeasured with the large shims in place.
If the inhomogeneity is less then 500 ppm than PLAS3D code in block 53 is run.
[0022] The PLAS3D code determines for each permissible shim location the field effect of
an arc shaped steel shim of given axial, radial and circumferential dimensions at
each of the 314 field measurement points. The axial and circumferential positions
of the arc shaped steel shims is a variable in the shimming procedure. For example,
if the allowable domain of arcs is from -90 to 90 cm along the z-axis, the field effects
can be determined with an arc every 10 cm along the z-axis. With higher densities
the algorithm will take longer, but result in more possible shim locations and so
provide generally better homogeneity.
[0023] The magnetic field of magnetized material may be represented as a series of spherical
harmonics expanded about the origin of the magnet coordinate system. The equations
for the magnetic field harmonics are:


where the coefficients A(n,m) represent the volume integral over the shim, a(n,m)
the transducer functions defined by Schenck et al., and P(n,m) the associated Legendre
polynomial. The number of terms required to accurately represent the magnetic field
depends on the size of the volume of interest, for the present shimming purposes,
expansion through order and degree eight is sufficient. The magnetization in the steel
shims may be calculated or assumed.
[0024] The determination of the arc field effects need only be done for one circumferential
location at each chosen axial location, and the arc field indexed in fifteen degree
increments to represent the field of any of twenty four circumferential locations.
Typical circumferential arc densities are only 8 to 12 per circle, so that indexing
gives accurate results.
[0025] Once the required data files have been created containing the field effect at each
of the 314 field measurement points for all the predetermined arc shaped shim locations,
the effect of each arc location individually on the chosen grid is evaluated for its
optimum strength, defined as that strength which yields the minimum inhomogeneity
on the imaging volume. Such optimization may be performed using a least squares routine
which solves the following equations:

where Σ is the chosen measure of field inhomogeneity, Bz
m represents the measured field at point m, and C
m is the coefficient representing the field per unit thickness created by a shim at
the location in question at field point m. The equation is written for each arc location
on the grid, and solved to give the shim thicknesses at each location which yields
the minimum field inhomogeneity. This optimization on the thickness assumes that the
field effect of an arc is linearly dependent on its thickness; i.e. that the magnetization
of the arc doesn't change with thickness. This assumption is only strictly true for
arcs which are saturated. If the shim thickness is negative, (not a physically realistic
solution) the location is removed from consideration. The set of locations remaining
is then fed into the SHIMPSV code in block 55 as an initial guess.
[0026] The SHIMPSV algorithm determines where the arcuate shims are to be positioned and
their thicknesses. While the PLAS3D program may come up with many dozens of locations
which require positive strength of shims, only 20-25 of these locations will be required
for shimming. Therefore the algorithm must decide which ones to eliminate. The SHIMPSV
algorithm starts with all the locations needing positive strength shims which is approximately
half the locations initially checked by the PLAS3D algorithm. A linear least squares
optimization is then performed on all the positive shim strengths simultaneously.
The result of the initial run will contain negative strength shims, these locations
are eliminated from consideration. Negative strengths result because the effect of
all the shim locations found to have positive strengths individually is not the same
as the effect of all these shim considered simultaneously. The remaining locations
requiring positive strength shims are then taken, and the least square optimization
run again. The process of eliminating negative strength locations is repeated until
a solution with all positive strength is obtained. The predicted inhomogeneity of
the solution is compared to the desired inhomogeneity in block 57. The field homogeneity
attainable with a given group of positive strength shim locations generally depends
inversely on the number of shim locations being worked with, so the more shims the
better. If a solution with all positive strengths with a predicted inhomogeneity within
specification is not possible, the parameters are altered in block 61 to increase
the number of allowable shim locations tried by the PLAS3D code. It is desirable to
use shims that are longitudinally closer to the center of the bore, since a smaller
shim closer to the center will have a larger effect on the inhomogeneity in the center
of the magnet than a shim positioned longitudinally displaced from the center. If
a solution cannot be reached with a group of locations closer to the center the number
of permissible shim locations can be increased and PLAS3D code run again. A solution
with all positive shim strengths is a physical possibility, and is used for the initial
placement of shims in the bore in block 63. Negative shim strengths would require
a material with a negative permeability. With the shims in place in the energized
magnet the field in the bore is again measured in the 314 locations on the grid in
block 65. If the inhomogeneity deviates from the prediction more than desired, the
SHIMPSV algorithm is run again with arc locations fixed and with the field values
obtained with the shims in place. A least squares routine is then used to adjust the
thicknesses of the arcs in block 67. These thickness changes should be small fractions
of the initial thicknesses, and once implemented should reduce the inhomogeneity
to within the desired range.
[0027] There are several poosible changes in implementing the SHIMPSV algorithm in block
55 which may prove beneficial in certain cases. The locations which are removed from
consideration at any step in the iteration of the SHIMPSV code because of negative
shim thicknesses may be introduced in a latter iteration thus effectively providing
more locations to try. If this results in a solution with more shims, the predicted
inhomogeneity will be generally lower, making it desirable. The least squares determinations
assume a linear relation between changes in shim thickness and the effect on the field.
The above method of determining shim locations in the bore of a magnet can be used
with electromagnetic magnets including superconducting magnets and permanent magnets.
[0028] The foregoing has described a method of passively shimming MR magnets assembly which
can achieve a level of field homogeneity required for magnetic resonance imaging,
without the use of correction coils.
[0029] While the invention has been described with respect to preferred embodiments, it
will be apparent that certain modifications and changes can be made without departing
from the spirit and scope of the invention. It is therefore, to be understood that
the appended claims are intended to cover all such modifications and changes as fall
within the true spirit of the invention.
1. A method of passively shimming a magnet having a central bore and using shims placed
in the bore of the magnet, said method comprising the following steps in order:
a) measuring the initial field inhomogeneity in the bore of the magnet;
b) checking the magnetic field effect of a shim at each of the predetermined permissible
shim locations in the bore of the magnet independently of one another, to determine
the shim strength needed to improve magnetic field inhomogeneity in the magnet bore;
c) selecting the locations where positive shim strengths were found beneficial;
d) determining shim strengths needed to improve magnetic field inhomogeneity at each
selected location, considering all selected locations simultaneously;
e) eliminating locations found to require negative shim strength and repeating steps
d and e until all selected locations remaining require positive shim strengths; and
f) placing shims of predicted thickness at their selected locations in the bore of
the magnet.
2. The method of claim 1 further comprising the following steps after step d of claim
1,
predicting the field inhomogeneity with the shims of predicted thickness at their
selected locations in the bore of the magnet;
comparing the predicted inhomogeneity with a desired inhomogeneity;
increasing the number of predetermined permissible locations in step b and repeating
steps b, c, d and e.
3. The method of claim 1 further comprising the steps of:
measuring the field inhomogeneity in the magnet bore with the shims in place;
comparing the desired inhomogeneity with the measured inhomogeneity;
comparing the predicted inhomogeneity with the measured inhomogeneity;
determining incremental shim thicknesses at the specified locations using the field
measured with the shims in place, if the measured field inhomogeneity is too large.
4. The method of claim 1 further comprising the following step:
introducing in step e, a location previously eliminated in an earlier iteration to
help increase the number of selected locations found to require positive shim strengths.
5. A passive shim assembly for a magnet having a central bore, comprising:
a nonmagnetic tube situated coaxially in the bore of the magnet; and
a plurality of predetermined length arcuate strips ferromagnetic of material demountably
secured to said tube, said strips extending in the circumferential direction around
said tube.
6. The apparatus of claim 5 wherein said strips are secured to the exterior of said
tube.
7. The apparatus of claim 5 further comprising means for coaxially situating said
tube in said bore.
8. The apparatus of claim 6 wherein all of said strips have the same length.
9. The apparatus of claim 8 wherein said strips are secured as stacks of strips of
varying heights at different locations about said tube.
10. The apparatus of claim 5 wherein said strips are secured to the interior of said
tube.
11. The apparatus of claim 10 wherein all of said strips have the same length.
12. The apparatus of claim 11 wherein said strips are secured as stacks of strips
of varying heights at different locations about said tube.
13. A passive shim assembly for a magnet having a central bore comprising:
a nonmagnetic tube situated coaxially in the bore;
a plurality of channel members secured to the interior of said tube, said channel
members being equally spaced circumferentially and extending in a longitudinal direction;
a plurality of arcuately shaped carrier pieces slidably mounted between adjacent channel
pieces;
clamp means to secure said carrier to said channels to prevent sliding motion; and
a stack of predetermined length strips of magnetic flux conducting material demountably
secured to each of said carrier pieces, said strips extending in the circumferential
direction.
14. The apparatus in claim 13 wherein said carrier pieces said channel members and
said clamp means each comprise nonmagnetic materials.
15. A passive shim assembly for a magnet having a central bore comprising:
a nonmagnetic tube situated coaxially in the bore;
a plurality of channel member secured to the exterior of said tube, said channel members
equally spaced circumferentially and extending in a longitudinal direction;
a plurality of arcuately shaped drawers slidably mounted between adjacent channel
pieces;
a plurality of stacks of predetermined length arcuate strips of magnetic flux conducting
material demountably secured to each of said drawers, said strips extending in the
circumferential direction.
16. The apparatus of claim 15 further comprising means for coaxially situating said
tube in the bore of the magnet.