JPS6013472B2 - Magnetic field equalizer - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a magnet device used in a nuclear magnetic resonance apparatus,
In particular, it relates to a magnetic field homogenizing device for correcting non-uniformity existing in the magnetic field produced by the magnet device.

In the nuclear magnetic resonance analysis method, as is well known, the sample to be examined is placed in a main magnetic field Ho of several thousand to tens of thousands of Gauss, and an alternating magnetic field field with a weaker intensity and variable frequency amplitude is applied. The resonant frequency of the substance in the sample can be determined by adding almost perpendicularly to the ratio and making it resonate with the Larmor frequency y of the atomic nucleus.

Therefore, specific elements and isotopes are identified by utilizing the fact that the gyromagnetic ratio of a sample element can be determined by knowing the main magnetic field Ho and the Larmor frequency. In order to make this identification accurately, an extremely uniform polarized magnetic field is required in the space occupied by the sample, and it is widely known that even a non-uniform magnetic field as small as 1×10-Amuro makes accurate identification impossible. It is being Therefore, in order to correct this extremely small non-uniformity of the magnetic field and obtain a uniform magnetic field, a relatively flat conductor commonly called a shim coil for generating a modified Z-positive magnetic field has been inserted into the gap between the magnetic poles that generates the magnetic field &. However, the amount of current applied to this was manually varied via a bridge circuit, etc., and the operation was repeated until a relatively satisfactory uniform magnetic field was obtained.

This operation requires the intuition and perseverance of an expert, and is a troublesome and time-consuming task. Further, the degree of uniformity of the magnetic field Ho is determined by the number of magnetic field gradients to be corrected, and shim coils corresponding to this number are required.

Naturally, as the number of shim coils increases, the gap between the magnetic poles into which the sample is placed will decrease by the volume filled by the shim coils. That is, if an attempt is made to improve the degree of correction of a non-uniform magnetic field, the volume of the magnetic pole gap will be reduced, and if an attempt is made to increase the gap, the number of shim coils will be reduced and the degree of correction of a non-uniform magnetic field will be reduced. Because of this trade-off relationship, efforts have been made to reduce the number of shim coils as much as possible by rotating the sample, and to reduce the volume of the shim coils. In addition, when a current flows through a shim coil, a magnetic field gradient expressed by countless spherical harmonic functions is generated, so in order to obtain the desired harmonic function, the root of the harmonic function is found, and multiple shim coils are used. Techniques have been taken, such as mutually canceling out unnecessary or undesirable harmonic functions depending on the geometrical position of the coil.

For this reason, the design for determining the position and dimensions of the shim coils is complicated, and moreover, the coils have to be arranged in multiple places, and the volume of the shim board on which the shim coils are attached increases, resulting in unfavorable design results. Additionally, when the number of shim coils or equalizing conductors is increased in order to improve uniformity, the thickness of the shim coils or uniform Q conductors may be increased by overlapping in parallel or crossing at right angles in the same place. In addition to increasing the
Adding a single conductor or equalizing conductor was not easy as a practical matter.

In order to improve the apparent uniformity of the magnetic field around the sample, methods such as rotating the sample around its center line are used.

However, a rotating sample requires a different harmonic function that cannot be corrected by the harmonic function corrected for a stationary sample. An object of the present invention is to provide a magnetic field homogenizing device that does not take up a volume large enough to affect a sample in the gap between magnetic poles, and that achieves a high degree of uniformity of the magnetic field.

Another object of the present invention is to modify only a desired harmonic function or magnetic field gradient without affecting other harmonic functions or magnetic field gradients. The object of the present invention is to provide a magnetic field homogenization device that can achieve the desired magnetic field uniformity.

Another object of the invention is to use electrical circuits such as operational amplifiers to direct currents having some kind of correlation to each equalizing conductor, rather than by means such as harmonic cancellation by the geometrical position of the shim coils. The purpose of the present invention is to provide a magnetic field homogenizing device in which the harmonic functions are canceled out or combined in the vicinity of the sample by supplying power to the magnetic field, and the necessary harmonic functions are obtained, thereby simplifying the arrangement of the homogenizing conductor.

To describe the general characteristics of the present invention, the first sub-chapter is that when a uniform magnetic field is obtained by generating a harmonic function by passing current through a uniform conductor or equalizing coil in the magnetic pole gap, (Samurai Kosho 451-381) and Mr. Gorey (
Methods such as Japanese Patent Publication No. 47-18794) have been used.

The feature of these methods is to obtain a magnetic field gradient that is determined from the position, size, shape, etc. of a uniform conductor or coil, and the current flowing through these coils is controlled by using a bridge circuit, etc. They were simply adjusted depending on the degree of non-uniformity of the magnetic field, without being correlated with each other, and a simple electric circuit was used. In other words, the emphasis was on determining the geometric dimensions and shape of the uniform conductor and coil.
The shape and position of the shim coil had to be complicated. In the case of the present invention, the shape of the equalizing conductor may be simply that straight lines intersect symmetrically, and its position and dimensions may be arbitrarily determined according to the designer's convenience. The present invention calculates the correlation between the currents to be passed through each uniform conductor, synthesizes a non-uniform magnetic field for correction using an electronic circuit including an arithmetic circuit such as a summing amplifier, and uses this to correct the non-uniform magnetic field. It is a correction device, and its feature is that it can independently vary a desired harmonic function without changing other harmonic functions in the vicinity of the origin where the sample exists. The second general feature of the present invention is that, as mentioned above, the shape of the uniform conductor is simple, the position is free, and the conductors are only insulated and orthogonal to each other. It is possible to make it thin.

In the case of the embodiment shown in Fig. 13, a polyimide flexible film with a thickness of 50 mm and a steel plate of 35 mm on both sides is used, and the pattern shown in Fig. 14 is created with the Z axis as the center. Since it is realized by simply stacking the X and Y axes rotated 49 times, the total thickness is less than 500 degrees, and the homogenizing soma hardly narrows the magnetic pole gap. BRIEF DESCRIPTION OF THE DRAWINGS The features and advantages of the invention and its effects will become clearer from the description given below in conjunction with the accompanying drawings and from the mathematical description using spherical harmonics and some matrices.

In Fig. 1, it is generally well known that the partial pressure in the Z-axis direction of the magnetic field generated at any point P between the magnetic poles, Hz, satisfies the plasma equation (formula below):
☆☆v2 ratio = 10 sea breams 10 crocodile =.

'1' This property also applies to the correction magnetic field generated when a current is passed through a uniform body.

Furthermore, the magnetic field IH is generally expressed as a scalar potential center as shown in the following equation.

IH = 1gr phantom small ■As is well known, this mind is expanded in terms of the following spherical mirror sum function near the origin of the orthogonal system (the origin in Figure 1).

J=-sa, sheep.

rnP mother (cosa) [A mother cosm ○ 1 B history simm
J] [31 As shown in Figure 2, the needle here is the angle formed by projecting the x axis and y onto the x-Y plane, and dream ≦
There is a relationship of '41 Z=yCOS8.

The component force Hz in the Z direction is expressed by the following equation from equations (2) and '4}. HZi-C.

S pot ten and a half solution '51 formula '3' and formula [H from 51
From the Legendre function number, it can be seen that z is the sum of harmonic functions expressed in Cartesian coordinates and is expressed by the following formula.

HZ=A9 12A8Z 13A Room X+ Mother Fist Y No. 10 A8{z
2-(x2 Y2)
2) 13 eel Y + Shizz {solution-3 (x2 Jushu Jupo A; x confirmation-
(x20Y2)}10 bags B; Y and others-(x20Y2)}19
Mountain crab Z (X2-Y2) 118 poisoned Y1103 book X (X
2-3) 110 dollars *(Matsu 2-Matsu) 10...
...By '6' or more, the order n of the formula [6} is generally
, is a harmonic function of rank m, and the Legendre rank function Pling(co
The constant A corresponding to 3 in so) is (n, m), B
If we express something related to (n, m)' as n>1, m=0n
〉When m>0, it is called a Hojo harmonic* function, and when n=m, it is called an Ogijo harmonic function.If each coefficient is ignored, Hz can be expressed as a composition of the components of each harmonic function.

If this is expressed as (Hz), it will include (Hz) up to order S.
is (HZ)=nmine. Kin,.

) ten・nsa.・Threefold Lord・{kun, m) ten (n, m)′}
:difference. Poison mo three (three mi spicy pickles. *)'ki (n, m)'
}(h,n)+(n,n)']'7'. For example, S=4, that is, up to the fourth order (Hz) is expressed by the following formula. (HZ): (1,0) ten (1,1) ten (1,1!)'
Ten(2,0) Ten(2,1) Ten(2,1)′+(2,2)
ten (2,2)' ten (3,0) ten (3.1) ten (3.1)
'+(3,2) ten(3,2)'ten(3,3) ten(3,3
)′+(4,0)ten(4,1)ten(4,1)′+(4,
2) Ten (4, 2)' + (4, 3) Ten (4, 3)' + (4
．． 4) Ten (4,4)'
・・・・・・'8' Therefore, when modifying the inhomogeneous magnetic field up to order 4 under a stationary sample, each of the modified magnetic fields expressed by the 24 harmonic functions in equation (■) should be approximately It is necessary to generate only the amount necessary for correction independently.

*For simplicity, (4,4) and (4,4)' are ignored, and the harmonic functions of (Hz) up to order 4 and order 3 are summarized in Table 1. Table 1 When the sample is rotated around the Y axis, the magnetic field Hz for any given atomic nucleus is in a plane perpendicular to the Y axis and is averaged along a circle centered on the plane.

In other words, the terms of the harmonic function that include the shapes of ,0),
(2,2), (3,1)', (3,3)', (4,0)
, (4, 2). On the other hand, when the sample is spun at a rotational frequency f, the harmonic function generated when the Y-axis and Z-axis of equation (2) are exchanged for the spinning sideband is 0.
The order term affects the spinning decomposition, and one pair of one
The order term affects the first sideband resonance at f from the center spectrum, the second order of the pair affects the second sideband resonance at 2 from the center spectrum, and the third order of the pair affects the second sideband resonance at f from the center spectrum. It is known that the third sideband resonance occurs at $ from the center spectrum and similarly affects the following.

Following Table 1, replace Y and Z on the formula side and summarize in Omotekawa.

Table □ Column C of Table 0 affects the center line of the spectrum, that is, the spinning resolution, columns f and f' are for the first sideband, and columns 2', 3, and g' are for the second sideband. , is the harmonic function affecting the third sideband.

From the above, it can be seen that the harmonic function for correction as shown in Table 1 for a stationary sample affects the amplitude of the sideband of a spinning sample and cannot suddenly improve the spinning resolution.

I will return to this issue later. Now, if a straight homogenizing conductor located on the XY plane in the magnetic pole gap and parallel to the As is well known, the slopes of Shibumoori, Wani, .・．． ．． ..., Mai in Figure 4, A4 is 4...
..., Souo is Kushi, Wani, Z first 4'Iso' a5 ratio..., and in Figure 6 it is 'Maojitsu, Takayoshihana a5' 66HZ Exile,...

In FIGS. 3 to 6, if the relative magnetic permeability of the magnetic pole is set to r and the mirror image phenomenon is considered, the magnetic field gradient near the origin can be calculated with higher accuracy. Messrs. B. Riegel and J. Kummel published a paper on this problem in the journal of Scientific Instruments (March 1977, pp. 346-349). While this paper calculates each gradient with the equalizing conductor stuck to the magnetic pole surface, the present inventor generalized this and calculated each gradient with the equalizing conductor separated from the magnetic pole surface. The slope was calculated for the straight homogenized conductor of FIGS. 3-6 (FIG. 7). First, we calculate the current l according to the calculation method of B. Riegel and J. Kummel.
The strength of the current of the kth mirror image of n(o)n is 1n(b; CK
1n(0) [91(Here, k=
0, 1, 2,...C white (shir-・)/(rr+・)).

Next, by further developing the above-mentioned calculation method, we can calculate the projection point Z from the center of the magnetic pole gap, that is, the origin, to the Z axis of the uniform conductor.

The ratio of the distance G to the magnetic pole surface and the distance G to the magnetic pole surface is
.

/G OQ, then the number of mirror images is m, and P=(ah11)/so (
11) Q=(2h〆)/so (12
), and furthermore, in the case of a linear uniform IQ conductor as shown in Figs. 3 to 6, =Y.

/Z. If the parameters (13) are adopted, each gradient of the magnetic field near the origin is calculated as follows. Hitsutomu m stem.

(Dep Uchiga Nanbu (IQ Fortress=
Tuna. {Yoro 2 in Msa・Kimaki} (1
5) Fever-reducing table. Nakajuku etc. CM charge fee
■Wani-Hina-Funeme {Screaming Crocodile Poison 2” Cm2 Imowaniki 2)}
(17) Crocodile - rolled stem. {-empty sheep ticket) 10 Cm song; evil}
(I ■ Hiroshi: - vessel receiver = needle desk. Naka - 3 Misaki rudder origin side 4 - Isori 20QQ} (1902 Juri 2〆
20 riro 4 solution = -a device Z = Tamefuze heart restraint 3) 10 other ■ -ri 3)} (2(tsubo 10 riz 4
20 Rine Tsuba - Kiyoshi Z2 - Kai - Uroko Taku.

Nakakoyaku Sanroku 2-4) Juzu-4 Nori (Spot 4-1 2 2 Juri 4)}
(21) (Bojuri 2 doors Kiaki = Shibuuke = Nozomi Taku.

Ze-4 Kobanawa Wanigi 5 Swords 4) 10Cm-4 NoQ equipment (
22) a5 generation aike thin Z ay5-ay36Z2 = cloth = old☆mi {C leaf I white-no L
r-1 21 tsubo - Mori G2 2) 60 cm across. /
46-1 spots 4ri 213 sleds 4-tachi 2}
(23) (Piri 2〆bath=-Equipment=-Nōichi=First.

{C Kano - Po rudder 2 2 ~ 4 G 2 ten 2℃ tengu - 6. /
Q.

(Pi-gaku・Q2 ri 20 ri 4)}
(2 hearts without 20 ri) 6 formulas (14) to (2
4) is a function of ri, mountain r, m, and is a constant determined each time based on various circumstances at the time of design. The impact will change. When formulas (14) to (24) are applied to m, Hz is m = 20, and the linear slope of Hz is m = 1.
Tortoise The values of each gradient are saturated at m=5 for the quadratic gradient, m=4 for the tertiary gradient, m=2 for the quartic gradient, and m=1 for the 5th gradient. Next, when the linear equalizing conductor shown in FIGS. 3 to 6 is rotated 90 degrees around the Z axis, equations (15) to (2)
4) by exchanging Y and X, and the magnetic field gradient regarding the X component can be obtained.

Further, when rotating these X and Y axes by 45 degrees, if these coordinates are x and Y, the following relationships hold:

For example, in the case of a harmonic function XYZ, if this coordinate transformation is applied,
Z) is rotated 450 times around the Z axis to obtain XYZ.
Table 1 and Table 0 are rearranged using equations (15) to (26) to form table plate and table N.

Table m Table N Table m is a harmonic function for obtaining a uniform magnetic field under a stationary sample, and Table N is a harmonic function for obtaining a uniform magnetic field under a rotating sample.

Next, we will discuss how to generate a magnetic field gradient expressed by a harmonic function.

As shown in Figure 8, 1,,12,13...ln
Z
o and ichiro are arranged parallel to each other on the XY plane.

The direction of the current in the conductor through which the current li is flowing is (i) the third
As shown in the figure, in the case of one direction, odd-numbered orders of Y, Y3, etc.
... is created (ii), and when it is symmetrical to Y and antisymmetrical to the Z axis as shown in Figure 4, the even order Y2, Y
4,... side), in the case of point symmetry as shown in Figure 5, it is the odd-numbered meat of ZY, ZY3,... side, as shown in Figure 6, it is symmetrical to the Z axis and opposite to the Y axis. In the case of symmetry, even order Z of YZ,
ZY2, ZY4. A gradient of magnetic field is created. Now in Figure 8, 1, . If the directions of the currents 12, 13, . If the direction of the current is in the relationship as described in (ii) above, the even-order magnetic field gradients of Y are synthesized near the origin. The same applies to the case of (iii) and G below. Therefore, as an example, cases (i) and (il), that is, the odd and even orders of Y will be taken up and explained.
In FIG. 8, consider the thinness of the i-th uniformizing conductor. From the figure, the position of this is i=Yi/young
(27) A current li flows through this, and this current 1
' is composed of n types of current components, then
li=j≧,lii (28) Also, let the order of the gradient of the magnetic field be P, and use equations (14), (15), (
17), (19), (21), and (23) for ri, and multiplying it by OtsuP+1 is fp(ri), then [18] = [Na2'-2(rij)]- ldiag [enemy,
, [ratio,, is 2, Japan/Germany, ...day 2n-2m]
(35) [18] = [Na2i-・(Rij)]-1d
iag [HI1, Nikyaku&3,...Qn-,,n]=[
8ij] diag [day, . , day 32. Ratio 3,... is n
-M] (3 kalpa type (35) type (
36) 4j-2,j, day 2j-,,j of E2-2,E
If we rewrite it as phantom-, and find elements 18 and 18 of the matrix, we get the following.

18=Q.

・E2J-2 (37)18=
8ii・E2-. (3 mother(ri): Roshu processing arp (290 fp(ri)) and a current applied thereto is expressed as Hp, then Nap(ri)1=Hp (30).

In Figure 8, if we represent all the magnetic field gradients that n sets of homogeneous Q conductors near the origin are divided into even-order and odd-order order into matrix representations, we can see that for the even-order case [Na2i-2 (rij)] [1j,
j] = [day 2, -2, 1] (31) When odd number order [na 2
i-・(rij)][1i,j]=[day 2nd,j](32
).

Under equation (28), the current 1,,12,13,...ln
each i-th current, i.e., lu,1 small 13i
...HPi created by lnj is equal to j-2, j when it is an even number order, 2 days and only j when it is an odd number order, and in other cases HPJ cancels each other out and becomes zero. The currents at that time are expressed as 18 for even orders and lo'ii for odd orders.Since the matrices on the right sides of Equations (31) and (32) are square matrices, respectively. The matrix is zero except when i=j, that is, it is a diagonal matrix. [Na2i-2 (rij)] [1;i] = diag [ratio,, day 2, day 43, ...day 2n-2
, n] (33) (mei2-・kurij)][18]=d
iag [H... daily summary, day 53,...-day 2n-. ，
n] (Solving equation (33) or (34), 1; j
Or, if 18 is found, li is found from equation (28),
In FIG. 8, the amount of power supplied to the uniform conductor of the i-th paper is determined.

Since the two matrices on the left side of equations (33) and (34) have the same rank, 18 and 18' necessarily exist. Also, the determinant is necessarily [Na21-・Curi i)], [Na2,-
, (rii)] exists, which can be expressed as [Qij], [
8;i], then 22, days, . ．． , day 2n・2,n]
2 [QG] diag 18, 18 is originally shown in Figure 8 (Koichi 2). (Gen 1) This is a current component that generates the following correction magnetic field.

Also, E2j-2 is multiplied by each element in the j column of the inverse matrix [Qij] of (Na2 [-2 (sword j)], and 18 is the inverse matrix of [Ri2 (rii)]) 8ij] for each element in column i of
It can be seen that it is multiplied by 2j-,. Also 18,1
8'According to subscript i, current 1,,12,13,...
. . . It is a constituent part of li, ln and is related to i, and is involved in the generation of correction magnetic fields of orders (phantom-2) and (2j-1) near the origin. From the above, [qli]
Each element in the i column of [8, It shows the rate at which the voltage E2- for generating the corrected magnetic field in (a-1) is converted into a current.

Under FIG. 8, the energizing circuit for generating the even-order magnetic field gradient of Y is Eo, according to equation (28) and equation (37).
n potentiometers (variable power supplies) 1 that generate output voltages of E2,...E2n-2, an amplifier 2 whose gain is the value of each element of the matrix [Q, j], and a summing multiplier 3. It can be expressed using, for example, as shown in FIG.

In the same figure, 1 is 1, = q, , Eo 10Q, 2E20...
...Q,nE-2=1,. 11, 20...11M, which satisfies equation (28), and similarly 12,...ln also satisfies equation (28).
It can be seen that 8) is satisfied.

Therefore, by changing Eo, the Y0 component can be changed independently, and by changing E2, the Y2 component can be changed independently. If the magnetic field gradients of Y2 and Y4 are to be generated practically, two sets of D conductors are sufficient, and therefore the circuit shown in FIG. 9 can be simplified as shown in FIG. 10. An example of finding the matrix [Qij] at that time is shown below.

For the sake of simplicity, it is assumed that the straight homogenizing conductor is on the magnetic pole face (ie, y = 1) and rr = 10. The position of the equalizing conductor can be selected arbitrarily, so for example, Q2 = 0.0 Q3
=1.2. And as mentioned earlier, crocodile breast = 2 saturation and 0 saturation, so according to equation (17), 2 = 0.
By substituting 3 = 1.2 and m = 2 into equation 21), which spans 6 and m = 5, fp(ri) becomes as shown in the table below.

Table V Table V ignores degree 0, so it is expressed as a matrix,
When this inverse matrix [na2[-2(rij〉]]-1 is found, the following formula is obtained.

The first example of the right-hand side of equation (39) Marshal - Electric current change, 3rd row 'Madomeru Densho' Each figure shows a current conversion rate.

Therefore, from equation (39), the gain of each intensifier 2 in Fig. 10 can be set to the value in parentheses in Fig. 10, and by doing this, the Y2 and Y4 components can be changed independently. be able to. Similarly, an energizing circuit for generating an odd-numbered Y magnetic field gradient is shown in FIG.

In the same figure, 1 is 1, = 8, E3 812E3 0...
...BinE ground -, = 1,, 11, 20 ... 11 · n, satisfying the formula (28), and similarly, ... ln also has the formula (28
) is found to be satisfied. Therefore, by changing E, the YI component changes, and by changing E3, the Y3 component changes.
By changing E5, the Y5 component can be changed independently. If magnetic field gradients of Y1, Y3, and Y are to be generated practically, three sets of equalizing conductors are sufficient, and the circuit shown in FIG. 11 can be simplified as shown in FIG. 12. Even in this case, the matrix [8, J] can be obtained as follows.

As in the previous example, let = 1, 〃r = 10, and let the value of ri be, for example, = 0.414, ri2 = 0.7, ri3 = 1.0
Lock fake chest workplace chest = 15, sea bream chest = 4, crocodile chest: 1 and husband candle heaven.

Susana, equation (15), sword, white 0.414, m=15,
Equation (19) 2=0.7. For m = 4, substitute 3 = 1.0 and m = 1 in equation (23), respectively, and create the inverse matrix [Na 2, -
2(rii)]-1, the following formula is obtained.

According to equation (40), the increase rate 8ij of each intensifier 2 in FIG. 12 may be determined by the value in parentheses in the figure.

The above explanation was about generating magnetic field gradients for even and odd orders of Y, but the direction in which current flows through the equalizing conductor for odd and even orders of Y of ZY is also explained in Figures 5 and 6. By defining as shown in the figure, formulas (16) and (18)
, (20), (22), and (24) in exactly the same manner as above, it is possible to obtain the amplification factor of the amplifier and generate a magnetic field gradient.

Figure 13 shows the odd and even orders of Y up to the fifth order, namely Y, Y2, Y3, Y4, and the odd and even orders of Y up to the fifth order, Z, ZY, ZY
An example for generating 2, katana 3, and kaku 4 will be shown.

In the figure, 4 is a polyimide film (thickness: 50 m)
This is a flexible printed circuit board made by the company, and each board is provided with 10 pairs (20 in total) of parallel linear printed patterns of S, - S, o by etching. This linear printed pattern is a uniform conductor. The substrate is arranged so that the center of the magnetic pole surface and the center o of the substrate match on the left and right magnetic pole surfaces of the magnet device, the patterns are exactly facing each other, and the sides A and B have lengths in the x direction. They are installed so that they face each other. Furthermore, the homogenizing conductor is S
,,S4,S7 are shown in FIG. 3, and S2,S,. is shown in Figure 6,
S3 and Ss are connected as shown in Figure 4, and S6 and S8 are connected as shown in Figure 5, so that current flows.
,, S4, S7 to Y, Y3, $ biasing circuit 6, S2,
S, o go to the energizing circuit 7 at ZY, Z, S3, S9 go to Y2
, to the biasing circuit 8 for Y4, Mi, S6, S8 are Z, meat 2, Z
They are respectively connected to the biasing circuit 9 for Y4. Each energizing circuit has exactly the same circuit configuration as that shown in Fig. 10 or Fig. 12, except that the intensification rate of the intensifier 2 is determined by the values of the pupil law and dust of each conductor. The value is as follows. With this configuration, S, , S4, and S7 yield Y
, Y3, because a magnetic field gradient occurs, Y, by operating the potentiometers IY, IY3, and IY5,
It is possible to correct the magnetic field inhomogeneity component of $, and furthermore, by operating other variable power supplies, Z, ZY, ZZ, public 3
, 4, Y2, and Y4 can be corrected independently. Incidentally, if the same pattern as shown in FIG. 13 is provided on the back surface of the substrate 4 by rotating it 900 times around o as shown in FIG. 14, and the same pattern as shown in FIG. Magnetic field inhomogeneous component play in Y direction, ×, X2
, X3, X4, X5, ZX, ZX2, ZX3, and ZX4 can be corrected.

In this way, the harmonic function (1,0), (
1,1), (1,1)', (2,1), (2,1)',
(3,1), (3,1)', (3,3), (3,3)'
, (4,1), (4,1)'. (4,3), (4,3)
That is, (1, 0) and those of orders 1 and 3 of orders 1 to 4 can be generated immediately.

However, the harmonic functions (2,0), (2,2), (2,2)′
, (3,0), (3,2), (3,2)', (4,0)
, (4, 2), (4, 2)', that is, orders 0 and 2 from orders 2 to 4, based on Table m from the magnetic field gradient of each order generated by the apparatus shown in Figure 13. need to be synthesized. For example (2,0), (2,2), (4,0),
(4, 2) can be synthesized using a biasing circuit as shown in FIG. The energizing circuit in FIG. 15 is basically two energizing circuits shown in FIG. 10 arranged side by side, and the energizing circuit 10 is
2, the magnetic field gradient of X4 is applied to the energizing circuit 11' or Y2. Current is supplied to each of the two sets of uniform Q conductors to generate a magnetic field gradient of Y4, respectively. la and lb are (2.0) and (
2, 2), and lc and ld are potentiometers for adjusting the magnetic field gradient expressed by harmonic functions of (4, 2), respectively.
This is a potentiometer for adjusting the magnetic field gradient expressed by a harmonic function of 0) and (4, 2). In the circuit of FIG. 15, unlike the circuit of FIG. 10, each potentiometer IX2, IX4, IY2, IY4 and the energizing circuit
Adding multiplier 3a having three input terminals between the input of
, 3b, 3c, and 3d are inserted, and the output of potentiometer la is sent to summing multipliers 3a and 3c. Therefore, if the potentiometer la is changed, the magnetic field gradient of X2 and the magnetic field gradient of
, 0), that is, a magnetic field gradient of (X2 + Y2) can be generated. Next, the output of the potentiometer lb is directly sent to the summing multiplier 3a, inverted by the inverting circuit 12, and sent to the summing multiplier 3c. Therefore, by using the potentiometer lb, the harmonic function (2, 2), that is, (X2 - Y2
) can be generated. In exactly the same way, variable power supplies lc and ld have harmonic functions (4,0) and (4,2
), that is, they are connected to adder intensifiers 3b and 3d, respectively, to generate magnetic field gradients of (X4 + Y4) and (X4 - Y4). It goes without saying that other harmonic functions can be generated in exactly the same way.

However, since (2, 2)' (3, 2)' and (4, 2)' include X and Y components, the pattern shown in Figure 14 is
This cannot occur in a uniform conductor of
As mentioned in the explanation of 26), a pattern as shown in FIG. 16 is required, including a pattern obtained by rotating the pattern in FIG. 14 by 45° around o. However, as you can see from Table m, the required values are X2, ×4, Z×2, Y2,
Since there are only Y4 and ZY2, there are a total of 5 sets of equalizing conductors, 2 sets for X2 and X4 and 3 sets for ZX2, as shown in Figure 16.
2, 2 sets for Y4 and 3 sets for ZY2, a total of 5 sets are required. Each equalizing conductor is energized by a circuit similar to the example circuit shown in FIG.
), including the harmonic function of (3,2)′, that is, (ZX2−ZY
It is possible to generate magnetic field gradients represented by the respective harmonic functions of 2) and (4,2)'rochi (X4-Y4).

[Brief explanation of drawings]

Figures 1 and 2 are diagrams for explaining the partial pressure Hz of the magnetic field generated at any point P between the magnetic poles, and Figures 3 to 6 are diagrams for explaining the arrangement of the equalizing conductor between the magnetic poles. Figure 7 is a diagram to explain the effect of mirror image, Figure 9 is a diagram to explain 4 sets and n sets of equalizing conductors arranged between magnetic poles, Figure 9 is a diagram to explain the effect of mirror image. - Figure 12 is a diagram showing a circuit for energizing the equalizing conductor, Figure 13 is an example diagram showing the connection between the equalizing conductor pattern and the energizing circuit, and Figures 14 and 16.
The figure shows another pattern of the equalizing conductor, and FIG. 15 is a diagram showing another example of the energizing circuit. 1: Potentiometer, 2: Multiplier, 3: Addition multiplier,
4: Flexible printed circuit board, 6, 7, 8, 9, 10
, 11: energizing circuit, 12: inverting circuit. Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 10 Figure 7 Figure 9 Figure 11 Figure 12 Figure 13 Figure 14 Figure 15 Figure 16

Claims (1)

[Scope of Claims] 1. Two pairs of linear equalizing conductors that are parallel to each other and have equal distances from the central axis of the magnetic poles are arranged symmetrically within the magnetic pole gap in a direction parallel to the magnetic pole surface, and linear equalizing conductors are connected in series to form one set, multiple sets of equalizing conductors are provided at different distances from the central axis of the magnetic pole, and the current values supplied to each of these multiple sets of equalizing conductors are set at a predetermined ratio. 1. A magnetic field equalizing device characterized by comprising a power supply means for maintaining and changing the magnetic field. 2. Two pairs of linear equalizing conductors having equal distances from the magnetic pole center axis and parallel to each other are arranged symmetrically within the magnetic pole gap in a plane direction parallel to the magnetic pole surface, and the two pairs of linear equalizing conductors are connected in series to form one set, multiple sets of equalizing conductors with different distances from the central axis of the magnetic pole are provided to form a equalizing conductor group, multiple groups of equalizing conductors are provided, and each equalizing conductor group is Magnetic field homogenization, characterized in that a number of power supply means corresponding to the number of the homogenization conductor groups is provided for changing the current value supplied to each of the plurality of homogenization conductor groups while maintaining a predetermined ratio. Device. 3. The magnetic field homogenizing device according to claim 2, wherein two groups of the homogenizing conductors are provided, and the respective homogenizing conductor groups are arranged at an angle of 90 degrees. 4. The magnetic field homogenizing device according to claim 2, wherein three groups of the homogenizing conductors are provided, and each of the homogenizing conductor groups is arranged at an angle of 45° with respect to each other. 5. Two pairs of linear equalizing conductors having equal distances from the magnetic pole center axis and parallel to each other are arranged symmetrically within the magnetic pole gap in a plane direction parallel to the magnetic pole surface, and the two pairs of linear equalizing conductors are connected in series to form one set, a plurality of sets of equalizing conductors having different distances from the central axis of the magnetic pole are provided to form a equalizing conductor group, a plurality of equalizing conductor groups are provided, and the plurality of equalizing conductor groups are provided. 1. A magnetic field equalizing device comprising power supply means for changing the current value supplied to each of a plurality of sets of equalizing conductors existing over the area while maintaining a predetermined ratio. 6. The magnetic field homogenizing device according to claim 5, wherein two groups of the homogenizing conductors are provided, and the respective homogenizing conductor groups are arranged at an angle of 90°. 7. The magnetic field homogenizing device according to claim 5, wherein three groups of the equalizing conductors are provided, and each of the equalizing conductor groups is arranged at an angle of 45° with respect to each other.