CA2259044C - Flexural pivot bearing - Google Patents

Abstract

A flexural pivot bearing includes a pair of members with opposed closed-spaced faces, which faces are joined by a web of microscopic thickness in a plane intersecting the faces. The members in the web are comprised of an integral body of substantially uniform material and the members are adapted for pivoted mutual flexure about a pivot axis aligned along the web.

Description

FLEXURAL PIVOT BEARING
Field of the Invention 1 This invention relates to a novel flexural pivot bearing which has 2 particular, though certainly not exclusive, application to gravity gradiometry.

3 The gravimeter is widely employed in geological surveying to measure 4 the first derivatives of the earth's gravitational potential function - the gravity field. Because of the difficulty in distinguishing spatial variations of 6 gravity from temporal fluctuations of the accelerations of a moving vehicle, 7 these measurements can be made to sufficient precision for useful 8 exploration only with land-based stationary instruments. This difficulty is in 9 principle avoided by measurement of the second derivatives of the potential -gravity gradients - but only limited success has been met to date in developing 11 a satisfactory gradiometer instrument. Gravity gradiometry is thought 12 especially appropriate to the location of geological structures bearing 13 hydrocarbons, to geological mapping, and to locating high density (e.g., 14 sulphides and iron ore) and low density (e.g., potash) mineral deposits.
Although it is not strictly correct to talk about the gradient of gravity, 16 usage of the term has been universally adopted and will be used herein also.
17 More formally, the second derivatives of the gravitational potential are 18 termed gradients of gravity and constitute the gravity gradient tensor with 19 components gxx, gXy "' gZZ, adopting the convention of taking the z-axis parallel to the local vertical. There are nine such components, only five of 21 which are independent since the tensor is apparently symmetric and the 1 potential is a scalar field obeying Laplace's equation.

3 Background Art 4 The key elements of a gravity gradiometer are a pair of substantially identical spaced masses and the object is to measure differences between the 6 gravitational force on the respective masses. Effectiveness requires 7 measurements of this difference when it approaches only one part in 1012 of 8 normal gravity. Approaches to measuring gravity gradients have thus far 9 fallen into two broad classes. The first of these entails differential modulation of a signal or parameter by the difference between the gravitationally induced 11 accelerations of the two masses. The second technique involves direct 12 measurement of the net gravitational acceleration of one mass relative to the 13 other.
14 British patent publication 2022243 by Standard Oil Company discloses a gravity gradiometer in the first class. An element, described in the patent 16 publication as a mass dipole but more properly termed a mass quadrupole, is 17 mounted coaxially on one end of a photoelastic modulator element 18 positioned in the cavity of a ring laser tube to differentially modulate circular 19 polarization modes in response to application of a torque. In a preferred form, two mass quadrupoles are mounted on opposite ends of the modulator 21 element to balance rotational acceleration noise. A closely related 22 development by the same inventor, Lautzenhiser, described in U.S. patent 23 4255969, employs actual mass dipoles in conjunction with respective 1 photoelastic modulator elements.
2 Another modulation technique involves rotating a platform which is 3 supporting suitable arrangements of mass pairs. Various instruments of this 4 kind have been proposed. One of these consists of electronically matched pairs of accelerometers on a rotating platform. The platform modulates the 6 sum of opposing acceleration signals with a frequency twice its rotational 7 frequency. These modulation systems call for extremely exacting uniformity 8 in the rotation and require the use of bearing, rotational drive and 9 monitoring technology which is not yet of a standard to render the instruments practicably suitable on an appropriate scale for airborne or 11 moving land-based measurements for geophysical resource exploration, as 12 opposed to geodetic surveying. The alternative of directly measuring gravity 13 gradient components necessitates a very high degree of electronic, magnetic, 14 thermal and vibration isolation to achieve the measurement accuracy needed. Machines thus far have had poor spatial resolution and a high noise 16 level.
17 An instrument for measuring the diagonal components gxX, gYY and gZZ
18 of the gravitational gradient tensor is described by van Kann et al in the 19 publication IEEE Trans. Magn. MAG - 21, 610 (1985). This instrument consists of a pair of accelerometers mounted with their sensitive axes in line.
21 The difference in displacement of the accelerometers is proportional to the 22 component of the given tensor gradient and is sensed by the modulated 23 inductance of a proximate superconducting coil. This instrument suffers 24 from the disadvantage that diaphragm springs serve both as mounts for the 1 masses and as gradient sensors. The former of these roles calls for a greater 2 stiffness in the springs while the sensing role necessitates enhancement of the 3 springs' softness. It is also very difficult with the van Kann instrument to 4 achieve axial alignment of the masses and trimming of the spring mountings with the accuracy needed to obtain the common mode acceleration rejection 6 ratios necessary for the accuracy sought.
7 It has been realized that significant advantages can be obtained relative 8 to the van Kann instrument, and direct measurement of gravitational 9 gradients made more easily achievable, by instead measuring off-diagonal components of the gravitational gradient tensor by means of one and 11 preferably two pivoted mass quadrupoles supported by a flexural pivot, and by 12 making the provision of mass support by the relatively stiff tensile spring 13 property distinct from the sensing function provided by the relatively soft 14 bending spring property.
Accordingly, disclosed herein is a gravity gradiometer for measuring 16 off-diagonal components of the gravitational gradient tensor, which includes 17 a housing comprising a pair of electromagnetic shield enclosures arranged 18 one inside the other, and a body including superconducting material 19 mounted within the inner enclosure for fine pivotal flexure as a mass quadrupole about an axis passing substantially through the centre of mass of 21 the aforesaid body. An array of superconducting coils is supported by the 1 outer enclosure and positioned in close proximity to the aforesaid body for 2 diamagnetically applying a rotational force to the body with respect to the axis 3 of flexure and/or for responding by modulation of inductance to pivotal 4 flexure of the body arising from a gravitational gradient across the body.
The array is arranged to apply the rotational force in both rotational directions and 6 to respond to flexure in either rotational direction.

7 The two enclosures are conveniently close fitting oblong boxes and the 8 quadrupole body is preferably a matching solid body of a complementary 9 shape. In a preferred embodiment, there is a superconducting coil on opposite sides of each arm of the quadrupole body to either side of the flexure 11 pivot axis. There may also be further coils to either side of the body at the axis 12 and at each end of the body, for monitoring translational movement of the 13 body.
14 The term "superconducting" is used herein, according to the normal convention, to denote a material which at least is superconducting below a 16 characteristic critical temperature. A suitable such material is niobium, 17 which has a critical temperature of about 9K.
18 The aforedescribed gravity gradiometer is of course preferably 19 supported in a system which is shielded electrically, magnetically, thermally and vibrationally in a manner similar to that described in the 21 aforementioned End-of-Grant Report.
22 There are preferably a pair of gradiometers coupled together in a single 23 instrument with the axes of flexure of their respective quadrupole bodies 24 parallel and preferably coincident, but with the quadrupole bodies aligned " '~ CA 02259044 1999-11-29 1 mutually orthogonally and orthogonal to the axes of flexure.
2 The pivotal flexural mounting for the mass quadrupole body may 3 comprise a flexure bearing such as the commercially available Bendix'~
pivot.
4 It has been found, however, that this bearing is less than wholly satisfactory as it is constructed of several different metals secured together and this creates 6 significant problems due to different thermal expansion coefficients and other 7 parameter variations which become critical at the kind of accuracy desired in 8 the present context.
9 Summary of the Invention The invention provides a flexural pivot bearing which comprises a 11 pair of members with opposed close-spaced faces. These faces are joined by a 12 web, of microscopic thickness, in a plane intersecting the faces. The members 13 and the web are comprised of an integral body of substantially uniform 14 material, and the members are adapted for pivoted mutual flexure about a pivot axis aligned along said web.
16 A particularly useful application of the invention is the mounting for 17 the mass quadrupole body of the aforedescribed gravity gradiometer according 18 to the first aspect of the invention. For this purpose the bearing is preferably 19 cut from a single mass of a superconducting material such as niobium.
The two members of the bearing may be a generally annular body and a 21 second body within the annular body.
22 Also disclosed herein is a gravity gradiometer having broad application 23 and not requiring vigorous sensitivity needing superconducting componetry 24 comprising:

_7_ 1 a housing;
2 a body mounted within the housing for fine pivotal flexure as a mass 3 quadrupole about an axis passing substantially through the centre of mass of 4 said body; and an array of transducer devices supported within said housing and 6 positioned in close proximity to said body for applying a rotational force to 7 said body with respect to said axis of flexure and/or for responding by 8 modulation of inductance or capacitance to pivotal flexure of the body arising 9 from a gravitational gradient across said body, wherein the array is arranged to apply said force in both rotational directions and to respond to flexure in 11 either rotational direction. Preferably, in this case, said body is mounted by 12 means of a flexural pivot bearing comprising a pair of members with opposed 13 close-spaced faces, which faces are joined by a web, of microscopic thickness, 14 in a plane intersecting the faces, wherein said members and said web are comprised of an integral body of substantially uniform material, and said axis 16 of flexure is aligned along said web.

1 Brief Description of the Drawings 2 The invention will now be described in greater detail, by way of 3 example only, with reference to the accompanying drawings in which:
4 FIG 1 is a diagrammatic axial cross-section of a gravity gradiometer 5 assembly supported on a gimballed mounting within a vacuum can for 6 rotationally stabilised cryogenic operation;

_8_ 1 FIG 2 is an enlargement of part of Figure 1, showing the gradiometer 2 assembly at actual size;
3 FIG 3 is a cross-section on the line 3-3 in Figure 2;
4 FIG 4 is an enlargement (5x magnification) of the flexural pivot bearing by which each of the mutually orthogonal mass quadrupole bars is supported 6 in the gradiometer assembly;
7 FIG 5A is a still greater enlargement (50x magnification) of the bearing 8 in the region of the web;
9 FIG 5B is a view similar to Figure 5A of an alternative pivot bearing;
FIG 6 is a more detailed axial cross-section of one of--------------------------1 the coil/coil holder assemblies;
2 FIG 7 and 8 are respective end elevations of the 3 assembly shown in Figure 6; and 4 FIG 9 is a schematic of the superconducting circuit for the gradiometer.

7 Best Modes of Carrying out the Invention 9 The illustrated apparatus 10 includes a gradiometer assembly 12 supported by a biaxial or triaxial gimballed 11 suspension 14 within a vacuum can 16. Apparatus 10 forms a 12 dewar probe which may be suspended inside a dewar (not 13 shown) and immersed therein in liquid helium. The can 16 14 provides an evacuable enclosure which can thereby be maintained at or near liquid helium temperature for 16 cryogenic operation of gradiometer assembly 12. A thermal 17 shield 17 may be fitted about the gradiometer assembly to 18 reduce radiative and gas conductive heat transfer between 19 the gradiometer assembly and the vacuum can. The entire equipment including the dewar is readily capable of being 21 mounted in an aircraft or other moving vehicle.
22 Gradiometer assembly 12 in fact includes two 23 substantially identical gradiometers, 20, 20' oriented to 24 measure gxy and gyx components of the gravitational gradient tensor. The gradiometers 20, 20' are bolted above and below 26 a central box structure 40 and each includes a pair of 27 rectangular box enclosures 22, 23 e.g, of niobium, arranged 28 one inside the other and outer niobium side plates 60 29 forming a surrounding shield from electromagnetic radiation.
Enclosures 22, 23 are typically niobium and provide two 31 further levels of all-round electromagnetic shielding.
32 A solid bar 25 of superconducting material such as 33 niobium is mounted on a bearing 21 within the inner 34 enclosure 23 for fine pivotal flexure as a mass quadrupole about an axis 8 passing substantially through the centre of 36 mass of the bar. The axes of flexure of the two bars 25, 37 25' are coincident and the bars extend in horizontal planes, 38 mutually orthogonally in the x and y directions. The 1 provision of a pair of orthogonal quadrupole bars permits 2 net elimination of common mode rotational accelerations 3 i.e., rotational noise at each bar. The gradiometers can, 4 of~course, be oriented differently depending on the gradient components of interest.
6 Each gradiometer 20, 20' further includes an array of 7 superconducting coils 30 which are mounted on holders 70 in 8 turn supported by the outer enclosure 22. Coils 30 are 9 positioned in close proximity to quadrupole bar 25.
The dewar (not shown) would typically consist of an 11 outer vacuum container, about 450 mm in diameter and 1.3 m 12 high, and a 300 mm diameter inner well suspended from the 13 mouth in the top of the outer shell by a fibreglass neck 14 tube. The space between the inner well and the outer shell is permanently evacuated and typically fitted with thermal 16 radiation shields surrounded by numerous layers of 17 aluminised Mylar'~ , superinsulation. Vacuum can 16 is 18 supported within the dewar from an aluminium top plate which 19 is attached to the mouth of the dewar. The top plate and' vacuum can are joined by a neck tube 13 through which the 21 vacuum can is evacuated, for example, down to the range of 22 10'8 to 10-10 Torr. Gimballed suspension l4 is attached to 23 a rigid 25 mm thick aluminium plate 15 which is bolted to 24 the bottom flange 15a of the neck tube and also forms a lid for can 16.
26 Gimballed suspension 14 consists of three gimbal rings 27 43, 44, 45 mounted on flexural pivots (not detailed) such as 28 Bendix'~ crossed-web pivots. Suspension 14 provides a 29 ~triaxial~rotational isolation for gradiometer assembly 12 and further incorporates respective fibre optic rotation 31 sensors (not shown) for the x and y axes and associated 32 superconducting electro-mechanical diamagnetic actuators for 33 active stabilisation in a servo circuit controlled by the 34 rotation sensors.
Instead of fibre optic rotation sensors, an optical 36 remote sensing arrangement may be employed, permitting the 37 stabilisation to be physically separated and enable the 38 utilisation of a room temperature gyroscope. In this 1 arrangement (not shown), a collimated beam of light from a 2 laser or luminescent diode attached rigidly to a room 3 temperature gyroscopic inertial reference system is 4 reflected by a plane mirror attached rigidly to the gradiometer assembly. Rotation of the gradiometer assembly 6 about any axis orthogonal to the light beam can then be 7 sensed by measurement of the angle between the incident and 8 reflected beams. This is accomplished by means of a 9 position sensitive photo-detector mounted rigidly to the light source with its planar sensing surface normal to the 11 beam. The detector actually measures the x and y co-12 ordinates of the position of the spot of light from the 13 reflected beam and this is used to monitor the relative 14 orientation of the gradiometer assembly. Isolation against mechanical vibration, is not illustrated but may be provided 16 in established ways.
17 Vibrations travelling along the external 18 instrumentation leads to the dewar may be intercepted by the 19 attachment of all cables near their mid-point to a massive lead block, itself suspended on a soft spring.
21 Each gradiometer 20, 20' is substantially identical and 22 it is therefore now proposed to detail only the construction 23 of gradiometer 20, with particular reference to Figures 2 24 and 3. As already mentioned, enclosures 22, 23 are of rectangular box-like configuration each made up of an 26 assembly of top, bottom, side and end plates. Inner 27 enclosure 23 is a close fit within outer enclosure 22 but 28 arranged to be slid in and out on removal of the bottom 29 plate of~outer enclosure 22. The inner enclosure is provided with multiple circular openings 24 which 31 respectively receive coil holders 70, and on its bottom 32 plate 23a, with a bush 26 for the flexure bearing 21 that 33 supports bar 25.
34 Flexure bearing 21 is detailed in enlarged Figures 4 and 5A, 5H and is formed by electric discharge machining 36 (EDM) an almost continuous cut 27 through bar 25 parallel to 37 axis 8, save for a microscopically thin web 29 extending the 38 width of the bar along axis 8 at the centre of mass of the 1 bar. In the example of Figure 5A cut 27 defines a 270°
2 part-cylindrical core 28 provided with three tapped holes 3 28a at one end for attachment of the core to bush 26. The 4 core may of course be supported at both ends, if desired or necessary.
6 Further tapped holes (not shown) are provided in the 7 bar to contain small screws whose position can be moved to 8 partly achieve mass balance of the bar about axis 8. The 9 radial portions 27a of cut 27 are deviated at their inner ends into right angle segments 27b which are aligned and 11 separated by Web 29. To either side of the web, the cut is 12 bulged slightly at either side at 27c to lengthen the web 13 and reduce its stiffness when acting as a pivot. Web 29 14 defines a micro-pivot some 0.030 mm thick, 0.200 mm long and 30 mm "wide", the width of bar 25. Figure 5B shows an 16 alternative cut to Figure 5A.
17 It will be appreciated in particular from Figures 4 and 18 5A, 5H that core 28 and the adjacent inwardly projecting 19 land 31 define a pair of members with opposed close-spaced regions 28b, 31b or 28b, 31b joined by web 29 in a plane 21 extending the width of the bar. These members are adapted 22 for pivotal mutual flexure about a pivot axis aligned along 23 web 29. It will also be noted that members 28, 31 and the 24 web are comprised of an integral body of substantially uniform material, in this case niobium. More particularly, 26 the quadrupole bar 25 is capable of fine pivotal flexure on 27 micro-pivot web 29 between angular limits determined by 28 contact between the opposed faces of the radial portions 27a 29 of cut 27. This angular limit is about 3 degrees and in any event is about the amount which would give rise to inelastic 31 deformation of the web.
32 The dimensions of bar 25 are selected as 30.00 mm 33 square by 90.0 mm long, thereby producing a gradient sensor 34 with a natural frequency of about 1 Hz in which the sensitivity to accelerations via elastic deformations of the 36 bar and pivot web 29 are made relatively small.
37 The mounting of each superconducting coil 30 is best 38 seen in Figures 6 to 8. Each holder 70, a machined piece of 1 niobium, is of circular cross-section and has an outer 2 peripheral retaining flange 72. The holder further has a 3 co-axial inner recess 71 for a fibreglass coil former 74.
4 The coil 30 is a pancake coil, i.e., a flat spiral wound on the exposed surface of former 74 and held in place by epoxy.
6 The wire 80, necessarily superconducting and conveniently 7 niobium with formvar insulation, enters the centre of the 8 spiral via a diagonal entrance hole 76 in former 74, 9 circulates the former several times and exits through a channel in the former. Holes 75 in former 74 are for 11 temporarily clamping the assembly during winding. Both wire 12 ends pass through a hole 78 in holder 70 and then along 13 various channels (not shown) machined in the outer faces of 14 enclosure 22 and through holes into enclosure 40.
Holders 70 are held in place in registered apertures 24 16 in the enclosures and are covered by one of the shield 17 plates 60, secured in place on the outer enclosure 22 by 18 screws 73 or the like. Plates 60, of which there are four 19 on the sides of each enclosure 22, shield the wires 80 which run from coils 30 to enclosure 40. The inner end of each 21 coil is substantially co-planar with the inner face of the 22 inner enclosure 23, in close proximity to a face of the bar 23 25.
24 The coils 30 are disposed with their axes in a common horizontal plane, three along each side and one at each end 26 of the quadrupole bar. The side coils are arranged in 27 opposed coaxial pairs, one pair with its axis co-planar with 28 axis 8 and the others towards each end of quadrupole bar 25.
29 The end coils 30a, 30b on one side are utilised as push coils for diamagnetically applying a rotational force to, 31 and augmenting the torsional stiffness of, the 32 superconducting bar in the respective rotational directions 33 about axis 8. The two opposite coils 30c, 30d on the other 34 side are utilised for responding by modulation of their inductance to pivotal flexure of bar 25 arising from a 36 gravitational gradient across the bar, the respective coils 37 responding to flexure in the respective rotational 38 directions about axis 8. The remaining four coils are also 1 employed as sense coils, but for detecting translational 2 movement of the bar in the x and y directions. The coils 3 are substantially identical and may therefore be 4 interchangeably employed as either push coils or sense coils, or both.
6 The push coils are required to provide feedback damping 7 and to fine-tune the torsional resonant frequencies of the 8 quadrupole bars to precisely match their response to common 9 mode angular accelerations about the axis 8.
It will be appreciated that quadrupole bars 25 strictly 11 need not be formed in solid superconducting material such as 12 niobium, so long as they include superconducting material 13 for interaction with coils 30. For example, each bar may be 14 an aluminium mass lined with or treated to contain niobium at those parts of its surface which face the operational 16 coils.
17 The eight coils of each set are wired in 18 superconducting circuits as schematically depicted in Figure 19 9 and detailed remarks concerning these circuits are set out hereinafter.
21 The superconducting wires 80 from the coils are fed 22 through machined channels in enclosure 22 to a 23 superconducting joint interface 41 within enclosure 40. The 24 various required transformers are also housed within enclosure 40.
26 Further leads from this interface traverse feedthroughs 27 46 to the exterior of the assembly. The push coils are 28 operated by employing heat switches to enable the insertion 29 of controlled persistent currents while the means to detect inductance changes in the sense coils comprises one or more 31 cryogenic SQUIDs (Superconducting Quantum Interference 32 Devices) to sense differential motion. The heat switches 33 and SQUIDs are housed within vacuum can 16. The switches 34 and current source are typically under computer control.
As the SQUID sensing system is very sensitive to 36 extremely small changes in magnetic flux, all leads and 37 components are shielded by closed superconducting shields, 38 e.g., of fine niobium tubing. External fields are 1 exponentially attenuated as they enter the enclosure 2 provided by the shields: the geometry of the tubing is 3 designed so that the earth's ambient magnetic field produces 4 less than one flux quantum inside the shield.
The illustrated apparatus, operated cryogenically, is 6 capable of measuring angular displacements of the order of 7 10-12 radians. It will be understood that materials other 8 than niobium may be employed in the construction of the 9 illustrated assembly. It is preferred however that the materials chosen have similar coefficients of thermal 11 expansion, and that at least wires, wire shields and bar 12 surfaces are formed in superconducting material. The 13 enclosures, for good temperature control are desirably made 14 in a material which is a good conductor of heat to minimise temperature gradients across the gradiometer. The preferred 16 material for the gradiometer body (bars, enclosures, 17 shields) is niobium.

19 Description of Superconducting Circuits (Figure 9) 21 The preferred circuitry for the gradiometer consists of 22 five circuits of three different types. These are the MAIN
23 READOUT (Figure 9A), the ACCELERATION MONITOR CIRCUITS
24 (Figure 9B) and the PUSH CIRCUITS (Figure 9C). There are two acceleration monitor circuits, for measuring 26 accelerations in the x and y directions, and two push 27 circuits, for the respective bars 25, 25'. Before 28 describing the three circuit types some general notes are 29 appropriate:
31 1. The apparatus can, in principle, be oriented to measure 32 any of the off-diagonal components of the gravity gradient 33 tensor. Throughout the drawings, the figures all show a 34 gradiometer with the z axis parallel to the vertical.
Figure 3, which shows the x axis parallel to the long axis 36 of the bar, is the cross section of the lower coil enclosure 37 as shown in Figure 2. That is, the x axis is parallel with 38 the long axis of the bottom quadrupole bar 25 and the y axis 1 is parallel with that of the top bar 25'.

3 2. In the circuits, the pancake coils used for sensing a 4 superconducting surface of a bar are labelled according to their usage. Thus, PUSH 1 and PUSH 2 are push coils, X and 6 Y are acceleration sense coils and 6~ and 6- are rotation 7 sense coils.

9 3. The circuits consist of several elements. The output of each circuit is from a SQUID whose input is coupled to 11 the rest of its circuit by means of a shielded toroidal 12 air-cored transformer. Hence there are five SQUIDS, one for 13 each circuit.

4. The inductors are of two types: toroidal or flat 16 spiral (pancake). All the coils which face a quadrupole bar 17 surface are pancake coils. The remainder of the inductors 18 are toroidal.

5. In the illustrated instrument, a "heat switch" consists 21 of a heater in close thermal contact with a thin 22 superconducting tube which contains a loop of 23 superconducting wire in good thermal contact with the tube 24 but electrically insulated from it. The tube provides electromagnetic shielding for the loop which is a part of 26 the superconducting circuit. Hy activating the heater, a 27 part of the loop may be heated to a temperature above its 28 superconducting transition temperature. This non-29 superconducting part then becomes an electrical resistor which will dissipate any current passing through the loop 31 and will allow the injection of a new current via the pump 32 leads.

34 In general, the design of heat switch may be refined or replaced by some other method which allows the dissipation 36 and injection of currents in the superconducting circuits.

38 Although in principle, a gravity gradiometer is 1 intrinsically insensitive to linear accelerations, in 2 practice these accelerations may have an effect because of 3 limitations in the achievable common mode acceleration 4 rejection ratio and because of second order effects induced by elastic deformations of the micropivot web 29 and 6 quadrupole bar 25, 25'. Consequently, accelerometers are 7 required for the measurement of accelerations so that the 8 acceleration effects may be appropriately subtracted from 9 the gradient signal and so that the accelerations may be recorded for any subsequent analysis of the data.
11 The motion of the quadrupole bar 25 or 25' as a result 12 of the aforementioned elastic deformations may be used as an 13 accelerometer, or separate accelerometers may be mounted on 14 board the gradiometer package to perform this function. In any case, two accelerometers are used, each measuring the 16 linear accelerations parallel to the long axis of a 17 quadrupole bar. These are labelled R and Y according to the 18 directions of these axes. The two acceleration monitor 19 circuits (a representative one of which is shown in Figure 9B), also labelled R and Y, simply perform the function of 21 providing acceleration data for recording.
22 The two push circuits (one for each bar) are identical 23 and only one is therefore shown in Figure 9C. The following 24 description for one applies equally to the other.
The push circuit loop carries a persistent current 26 which can be adjusted and stored. The resulting magnetic 27 flux in the loop means that the push coils act as magnetic 28 springs thereby increasing the mechanical torsional resonant 29 frequency of the quadrupole bar. This technique is used to match the torsional resonant frequencies of the two bars.
31 The rejection of angular accelerations about the z axis 32 depends on how well these frequencies are matched.
33 Modulati«~s of the current will result due to angular motion 34 of the bar and these are sensed by coupling the push loop to a SQUID. This output can be used in feedback to servo 36 control angular accelerations about the z axis.
37 The main readout circuit depicted in Figure 9A performs 38 the function of combining the angular information from each 1 of the responders together with the x and y acceleration 2 information to provide a temperature compensated output 3 signal proportional to the gravity gradient. There are five 4 loops, in each of which the magnetic flux can be independently set and then locked. These are: the 8 loop 6 for the top bar; the 8 loop for the bottom bar; the X
7 acceleration loop (bottom bar); the Y acceleration loop (top 8 bar); and finally a temperature sensing loop, into which the 9 SQUID input transformer is coupled. Flux in the X,Y loops is trimmed so that the SQUID output is independent of these 11 two accelerations. Similarly the flux in each of the two 8 12 loops is set to cancel the effects of rotational 13 acceleration about the z-axis. The temperature loop flux is 14 adjusted to make a first order cancellation of small temperature inhomogeneities in the gradiometer.

Claims (3)

What is claimed is:

1. A flexural pivot bearing comprising a pair of members with opposed close-spaced faces, which faces are joined by a web, of microscopic thickness, in a plane intersecting the faces, wherein said members and said web are comprised of an integral body of substantially uniform material, and said members are adapted for pivoted mutual flexure about a pivot axis aligned along said web.

2. A flexural pivot bearing according to claim 1, wherein said bearing is cut from a single mass of a superconducting material.

3. A flexural pivot bearing according to claim 1 or 2, wherein said two members of the bearing are a generally annular body and a second body within the annular body.