ACTUATOR AND GRAVITY GRADIOMETER

Field of the Invention

This invention relates to an actuator and a gravity gradiometer, and in particular, but not exclusively, to a gravity gradiometer for airborne use. The invention has particular application for measuring diagonal and off- diagonal components of the gravitational gradient tensor.

Background of the Invention

Gravimeters are widely used in geological exploration to measure the first derivatives of the earth's gravitational field. Whilst some advances have been made in developing gravimeters which can measure the first derivatives of the earth's gravitational field because of the difficulty in distinguishing spatial variations of the field from temporal fluctuations of accelerations of a moving vehicle, these measurements can usually be made to sufficient precision for useful exploration only with land-based stationary instruments.

Gravity gradiometers (as distinct from gravimeters) are used to measure the second derivative of the gravitational field and use a sensor which is required to measure the differences between gravitational forces down to one part in 10 ¹² of normal gravity.

Typically such devices have been used to attempt to locate deposits such as ore deposits including iron ore and geological structures bearing hydrocarbons .

International publication WO 90/07131 partly owned by the present applicants associated company discloses a gravity gradiometer . The gradiometer includes a gimbal bearing arrangement comprised of three concentric rings in which is mounted the sensing equipment. The sensing equipment generally comprises two spaced apart bars respectively

located in shielded housings and each mounted on a web bearing. The instrument disclosed in that application is relatively complicated in that it includes a large number of parts and is relatively heavy which is a disadvantage particularly in airborne applications.

Gravity gradiometers of the type described above operate in a cryogenic environment and utilise superconducting circuits and materials . To ensure proper operation of the superconducting circuits, it is necessary to ensure that spurious magnetic fluxes do not produce currents in the superconducting circuits which result in erroneous measurements .

Summary of the Invention

The invention provides a gravity gradiometer for measuring components of the gravitational gradient tensor, comprising: a sensor for measuring the components of the gravitational gradient tensor; a mount for supporting the sensor, the mount being for providing movement about three orthogonal axes; a plurality of actuators for moving the mount about the three axes to stabilise orientation of the mount during use of the gradiometer; each actuator comprising a first housing part for coupling to one of the mounting and a fixed portion of the gradiometer; a moveable member for connection to the other of the mounting and the fixed portion; at least one coil; a reaction member carried by one of the housing part and the moveable member so that when the coil receives a signal, the reaction member is moved relative to the coil to in turn move the moveable member relative to the housing part and therefore the mounting relative to the fixed portion; and

wherein the actuator has an elongate passage for minimising leakage of magnetic flux from the actuator.

Preferably linear and angular accelerometers are provided for detecting movement of the sensor due to angular accelerations for producing signals to counteract error induced by those accelerations .

Preferably the actuators counteract error produced by angular disturbance of the sensor.

Preferably the actuators reduce angular motion to further reduce errors produced due to centrifugal forces or angular acceleration.

This aspect of the invention provides an actuator for producing a force to adjust the position of one component relative to another component, comprising: a first part for coupling to one of the said one component and another component; a moveable member for connection to the other of the said one component and another component; at least one coil carried by one of the first part and the moveable member so that when the coil receives a signal, the moveable member is moved relative to the first part to in turn move the first component relative to the second component; and wherein the actuator has an elongate passage for minimising leakage of magnetic flux from the actuator.

Thus, according to this aspect of the invention, the actuator can be used with an instrument in which magnetic flux contamination is of consideration to minimise spurious magnetic fields which may detrimentally effect operation of the instrument.

Preferably the moveable member has the passage.

Preferably the passage is in the form of a tube.

Preferably the tube has a length to maximum width ratio of about 10:1.

Preferably the tube is substantially circular in cross- section .

Preferably the first part has a rod located in the passage for receiving leads for supplying current to the coil to energize the coil.

Preferably the coil is a pancake coil .

Preferably the moveable part is a hollow disc housing which supports the tube.

Preferably the first part is a disc plate located in the hollow disc housing, the disc plate supporting the rod within the tube, the coil being located on the disc plate, the rod being connected to said one component, and the disc housing being connected to the other component.

Brief Description of the Drawings

Preferred embodiments of the invention would be described, by way of example, with reference to the accompanying drawings , in which :

Figure 1 is a schematic view of a gradiometer of one embodiment of the invention.

Figure 2 is a perspective view of a first mount forming part of a mounting of the gradiometer of the preferred embodiment;

Figure 3 is a view of a second mount of the mounting;

Figure 4 is a view from underneath the mount of Figure 3 ;

— R •—

Figure 5 is a cross-sectional view along the line IV-IV of Figure 3;

Figure 6 is a cross-sectional view along the line V-V of Figure 3; Figure 7 is a view of the assembled structure;

Figure 8 is a view showing the sensor mounted on the gimbal structure;

Figure 9 is a plan view of a bar of the preferred embodiment; Figure 10 is a diagram showing actuator control;

Figure 11 is a block diagram showing operation of the rotatable support system;

Figure 12 is a view of a gradiometer of the preferred embodiment; Figure 13 is a view of a first mount of a second embodiment;

Figure 14 is a view of part of the mounting of Figure 13 to illustrate the location and extent of the flexural web of the first mount; Figure 15 is a view of the mounting of Figure 13 from beneath;

Figure 16 is a view of the mounting of Figure 13 including a second mount of the second embodiment;

Figure 17 is a cross-sectional view through the assembly shown in Figure 16;

Figure 18 is a view from beneath of the section shown in Figure 17 ;

Figure 19 is a view from beneath of the second mount of the second embodiment; Figure 20 is a view of the second mount of Figure

19 from above;

Figure 21 is an exploded view of the second mount of the second embodiment;

Figure 22 is view of the assembled mounting and sensors according to the second embodiment;

Figure 23 is a perspective view of the gradiometer with some of the outer vacuum container

removed;

Figure 24 is a plan view of a housing for supporting a bar according to a further embodiment of the invention; Figure 25 is a more detailed view of part of the housing of Figure 24;

Figure 26 is a view of a transducer used in the preferred embodiment;

Figure 27 is a view similar to Figure 25 but showing the transducer of Figure 26 in place;

Figure 28 is a diagram to assist explanation of the circuits of Figures 29 and 30;

Figure 29 is a circuit diagram relating to the preferred embodiment of the invention, particularly showing use of one of the sensors as an angular accelerometer;

Figure 30 is a frequency tuning circuit;

Figure 31 is a cross-sectional view through an actuator according to one embodiment of the invention; Figure 32 is a view of part of the actuator of

Figure 31 ;

Figure 33 is a diagram illustrating balancing of the sensors of the gradiometer of the preferred embodiment; and Figure 34 is a circuit diagram of a calibration sensor used when balancing the gradiometer.

Detailed Description of the Preferred Embodiments Figure 1 is a schematic view of a gravity gradiometer according to the preferred embodiment of the invention.

The gradiometer shown in Figure 1 comprises a double walled Dewar 1 which is supported in an external platform 2. The external platform 2 enables adjustment of the Dewar and therefore the contents of the Dewar about three orthogonal axes . The external platform 2 is generally known and its adjustment by suitable motors or the like is

also known. Thus, a detailed description will not be provided.

A vacuum canister 3 is provided in the Oewar and the Dewar is supplied with liquid gas such as liquid helium He so that the gradiometer can operate at cryogenic temperature . The Dewar 1 is closed by an end plate 4 which includes connectors 5a for connecting electrical leads (not shown) to external components (not shown) .

The canister 3 is closed by an end plate 9 which includes connectors 5b for connecting electric leads (not shown) to the connectors 5a. The gradiometer has a main casing 61 formed from a twelve-sided ring 62 and hemispherical domes 63 (see Figure 12) . An internal mounting 5 is connected to the ring 62. The ring 62 carries a support 65 to which a feed through flange 9 is coupled. A neck plug 11 formed of baffles 11a which sandwich foam lib is provided above the canister 3. The baffles 11a are supported on a hollow rod 93 which extends to the canister 3 and which is also used to evacuate the canister 3.

With reference to Figure 2 a first mount 10 of a rotatable mounting 5 (Figure 7) of the gradiometer is shown which comprises a base 12 and an upstanding peripheral wall 14. The peripheral wall 14 has a plurality of cut-outs 16. The base 12 supports a hub 18.

Figures 3 and 4 show a second mount 20 which comprises a peripheral wall 22 and a top wall 24. The peripheral wall

22 has four lugs 13 for connecting the mount to the casing 61. The top wall 24 and the peripheral wall 22 define an opening 28. The peripheral wall 22 has a first part 25, a second part 26 and a third part 27. The second mount 20 is a monolithic integral structure and the first part 25 is formed by making a circumferential cut 19 through the peripheral wall except for the formation of flexure webs

— fi —

as will be described hereinafter. The third part 27 is formed by making a second circumferential cut 29 through the peripheral wall 22 except for flexure webs which will also be described hereinafter. The second mount 20 is mounted on the first mount 10 by locating the hub 18 into the opening 28 and the lugs 13 through respective cut-outs 16 as is shown in Figure 7.

The first mount 10 is joined to the second mount 20. The first flexure web 31 is formed in the first mount 10 so a primary mount portion of the mount 10 can pivot about a web 31 relative to a secondary mount portion of the mount 10. This will be described in more detail with reference to the second embodiment shown in Figures 13 to 21.

The lugs 13 connect the mounting 5 in the canister 3 which, in turn, locates in the Dewar 1 for cryogenic operation of the gradiometer .

The Dewar is in turn mounted in a first external platform for course rotational control of the gradiometer about three orthogonal x, y, x axes. The mounting 5 mounts the sensor 40 (which will be described in more detail hereinafter and which is preferably in the form of a mass quadrupole) for much finer rotational adjustment about the x, y and z axes for stabilising the gradiometer during the taking of measurements particularly when the gradiometer is airborne.

The first flexure web 31 allows the first mount 10 to move relative to the second mount 20 about a z axis shown in Figure 7.

Figures 5 and 6 are views along the lines IV and V respectively which in turn are along the cuts 19 and 29 shown in Figure 3. The peripheral wall 22 may be cut by any suitable cutting instrument such as a wire cutter or

the like. Figure S shows the bottom surface 19a formed by the cut 27. As is apparent from Figures 3 and 5 the cut 27 has two inverted v-shaped peaks 34. The apex of the peaks 34 is not cut and therefore form a second flexure web 33 which join the first part 25 to the second part 26. Thus, the second part 26 is able to pivotally rotate relative to the first part 25 about the x axis in Figure 7. The second cut 29 is shown in Figure 6 and again the bottom surface 29a formed by the cut 29 is visible. Again the second cut 29 forms two v-shaped peaks 35 and the apexes of the peaks 35 are not cut and therefore form a third flexure web 37 which connect the second part 26 to the third part 27. Thus, the third part 27 is able to pivotal rotate about the y axis shown in Figure 7.

Figure δ shows sensor 40 mounted on the mounting. The sensor 40 is an Orthogonal Quadrupole Responder - OQR sensor formed of a first mass and a second mass in the form of a first bar 41 and a second bar 42 (not shown in Figure δ) orthogonal to the bar 41 and which is of the same shape as the bar 41.

The bar 41 is formed in a first housing 45 and the bar 42 is formed in a second housing 47. The bar 41 and housing 45 is the same as bar 42 and the housing 47 except that one is rotated 90° with respect to the other so that the bars are orthogonal . Hence only the housing 45 will be described.

The housing 45 has an end wall 51 and a peripheral side wall 52a. The end wall 51 is connected to rim 75 (Figures 2 and 7) of the wall 14 of the first mount 10 by screws or the like (not shown) . The bar 41 is formed by a cut 57 in the wall 51 except for a fourth flexure web 59 which joins the bar 41 to the wall 51. The flexure web is shown enlarged in the top view of the bar 41 in Figure 9. Thus, the bar 41 is able to pivot relative to the housing 45 in

response to changes in the gravitational field. The bar 42 is mounted in the same way as mentioned above and also can pivot relative to its housing 47 in response to changes in the gravitational field about a fifth flexure web 59. The housing 47 is connected to base 12 (Figure 2) of the first mount 10.

The bar 41 and the housing 45 together with the flexure web 59 are an integral monolithic structure.

Transducers 71 (not shown in Figures 2 to 6) are provided for measuring the movement of the bars and for producing output signals indicative of the amount of movement and therefore of the measurement of the differences in the gravitational field sensed by the bars.

Figure 10 is a schematic block diagram showing actuator control to stabilise the gradiometer by rotating the mounting 5 about three orthogonal axes (x, y, z) . A controller 50 which may be a computer, microprocessor or the like outputs signals to actuators 52, 53, 54 and 55. The actuator 52 could rotate the mounting 5 about the x axis, the actuator 54 could rotate the mounting 5 about the y axis and the actuator 54 could rotate the mounting 5 about the z axis. However, in the preferred embodiment, two of the four actuators 52, 53, 54 and 55 are used to rotate the mounting about each axis so that rotation about each axis is caused by a combination of two linear movements provided from two actuators . The linear movement provided by each actuator will be described with reference to Figures 31 and 32. The position of the mounting 5 is monitored so that appropriate feedback can be provided to the controller 50 and the appropriate control signals provided to the actuators to rotate the support 10 as is required to stabilise the support during movement through the air either within or towed behind an aircraft.

The preferred embodiment also includes angular accelerometers which are similar in shape to the bars 41 and 42 but the shape is adjusted for zero quadrupole moment. The linear accelerometers are simple pendulous devices with a single micro pivot acting as the flexural hinge .

Figure 11 is a view of a feedback control used in the preferred embodiment.

Figure 12 is a cut away view of the gradiometer ready for mounting in the Dewar 1 for cryogenic operation which in turn is to be mounted in the external platform. Although Figures 2 to 8 show the gradiometer with the bars 41 and 42 top and bottom, the instrument is actually turned on its side (90°) so that the bars 41 and 42 are at the ends as is shown in Figure 12.

Figure 12 shows the mounting 5 arranged within the casing

61 and formed by the ring 62 and the transparent hemispherical ends 63. The ring 22 has connectors 69 for connecting the internal wiring from transducers 71 (see Figure 8) and SQuID (Superconducting Quantum Interference Device) Electronics located in the casing 61 to the connectors 5b (Figure 1) .

The transducers 71 measure the angle of displacement of the bars 41 and 42 and the control circuitry (not shown) is configured to measure the difference between them.

Error correction can be performed numerically based on digitised signals from the accelerometers and a temperature sensor .

The transducers 71 are SQuID based transducers and the error correction is made possibly by the large dynamic

range and linearity of the SQuID based transducers .

Figures 13 to 21 show a second embodiment in which like parts indicate like components to those previously described.

In this embodiment the first mount 10 has cut-outs 80 which effectively form slots for receiving lugs (not shown) which are connected to the mount 10 in the cut-outs 80 and also to the second mount 20 shown in Figures 19 to 21. In this embodiment the lugs are separate components so that they can be made smaller, and more easily, made than being cut with the second mount section 20 which forms the second flexure web 33 and the third flexure web 37.

In Figure 13 a cut 87 is made to define the part 18a of the hub 18. The cut 87 then extends radially inwardly at 88 and then around central section 18c as shown by cut 101. The cut 101 then enters into the central section 18c along cut lines 18d and 18e to define a core 18f . The core 18f is connected to the central section 18c by the flexural web 31 which is an uncut part between the cut lines 18e and 18d. The part 10a therefore forms a primary mount portion of the mount 10 which is separated from a secondary mount portion 10a of the mount 10 except for where the portion 18a joins the portion 10a by the flexural web 31. The part 18a effectively forms an axle to allow for rotation of the part 18a relative to the part 10a in the z direction about the flexure web 31.

As is shown in Figure 14, the cut line 88 tapers outwardly from the upper end shown in Figure 14 to the lower end and the core 18c tapers outwardly in corresponding shape, as best shown in Figure 17.

As is apparent from Figures 13 to 18, the first mount 10

is octagonal in shape rather than round, as in the previous embodiment.

Figures 19 to 21 show the second mount 20. Figure 16 shows the second mount 20 mounted in the first mount 10. As is best shown in Figures 19 and 20, the second mount 20 has cut-outs 120 which register with the cut-outs 80 for receiving lugs (not shown) . The lugs can bolt to the second mount 20 by bolts which pass through the lugs and into bolt holes 121. The lugs (not shown) are mounted to the mount 20 before the mount 20 is secured to the first mount 10.

In the embodiment of Figures 19 and 20, the peaks 34 and inverted peaks 35 are flattened rather than of V-shape as in the previous embodiment.

In this embodiment, top wall 24 is provided with a central hole 137 and two attachment holes 138a. Three smaller holes 139a are provided to facilitate pushing of the housing 45 off the part lδa if disassembly is required. When the second mount 20 is located within the first mount 10, the upper part of central section lδc projects through the hole 137, as best shown in Figure 16. The mount 20 can then be connected to the mount 10 by fasteners which pass through the holes 138 and engage in holes 139b (see Figure 13) in the part 18a.

Thus, when the first housing 45 and its associated bar 41 is connected to the rim 75 of the housing 10 and the second housing 47 is connected to the base 12, the housings 45 and 47 and their associated bars 41 and 42 are therefore able to move about three orthogonal axes defined by the flexure web 31, the flexure web 33 and the flexure web 37.

As is best seen in Figure 21 which is an exploded view of

the three parts 25, 26 and 27 which make up the second mount 20, an opening extends through the mount 20 which is formed by the hole 137, hole 138 and hole 139. It should be understood that the mount 20 shown in Figure 21 is a monolithic structure and is merely shown in exploded view to clearly illustrate the location of the flexural webs 33 and 35. Obviously the flexural web 33 shown in Figure 21 joins with the part 26 and the flexural web 35 shown in Figure 21 joins with the part 27. The holes 137, 138 and 139 define a passage through which the axle or first portion 18a of the first mount 10 can extend when the second mount 20 is located in the first mount 10.

Thus, when the second mount 20 is fixed to the part 18a, the second mount 20 can pivot with the first portion 10a of the first mount 10 about a z axis defined by the flexure web 31 whilst the second portion formed by the part 18a remains stationary. Movement about the x and y axes is achieved by pivotal movement of the second mount 20 about the flexure webs 33 and 35 as previously described.

Figure 22 shows the linear and annular accelerometers 90 fixed to the housings 45 and 47.

The gravity gradient exerts a torque on a rigid body with any mass distribution provided it has a non-zero quadrupole moment. For a planar body, in the x-y plane and pivoted about the z-axis, the quadrupole is the difference between moments of inertia in the x and y directions . Thus a square or circle has zero quadrupole moment, while a rectangle has a non-zero value.

The torque produced is what constitutes the signal measured by the gradiometer.

There are two dynamical disturbances which can also

produce torques and consequently are sources of error.

The first is linear acceleration.

This produces a torque if the centre of mass is not exactly at the centre of rotation - i.e. the bar is "unbalanced" . The bars 41 and 42 are balanced as well as possible (using grub screws to adjust the position of the centre of mass) but this is not quite good enough, so there is a residual error. This error can be corrected by measuring the linear acceleration and using this to numerically subtract away the erroneous part of the signal .

The second is angular motion.

There are two aspects to angular motion, each of which produces a different error.

The first is aspect angular acceleration.

Angular acceleration produces a torque on the mass distribution through its moment of inertia (even if the quadrupole moment is zero) . This is an enormous error and two preferred techniques are used to counteract it.

The first is to use internal rotational stabilization. This is depicted in the block diagram of Figure 10. Here Ho(s) represents the sensor assembly pivoted about the mounting 5 (as per Figure 9) . The block A(s) represents the actuator, which provides the feedback torque to effect the stabilization by canceling the applied disturbances. T(s) represents the sensor (or transducer) which measures the effect of the applied disturbance. This is the angular accelerometer . Using angular accelerometers in rotational control is unusual - usually gyros and/or highly damped tilt meters are used, but for our purpose

the angular accelerometers are better, as the error is proportional to the angular acceleration disturbance.

The second is to use common mode rejection CMKR - that is why 2 orthogonal bars are needed. For the two bars, the error torque produced by the angular acceleration is in the same direction, but the signal torque produced by the gravity gradient is in opposite direction .

Therefore, by measuring the difference in deflection between the two bars, the gradient is sensed but not the angular acceleration .

Therefore, two separate angular accelerometers 90 (labeled 90" in Figure 22 for ease of identification) are provided. Ne have two independent output signals from the pair of OQR bars 41 and 42. The first is proportional to the difference in deflection, which gives the gradient signal and the second is proportional to the sum of their deflections, which is proportional to the angular acceleration and provides the sensor for the z-axis rotational control .

The x and y axes require separate angular accelerometers . Rotational stabilization about these axes is required because the pivot axes of the two bars are not exactly parallel and also to counteract the second form of error produced by angular disturbance, discussed below.

The second aspect is angular velocity.

Angular velocity produces centrifugal forces, which are also a source of error. The internal rotational stabilization provided by the actuators reduces the angular motion so that the error is below 1 Eotvos .

Figure 23 shows main body 61 and connector 69 with the

hemispherical ends removed.

Figure 24 is a plan view of housing 45 according -bo a still further embodiment of the invention. As is apparent from Figure 24, the housing 45 is circular rather than octagonal , as is the case with the embodiment of Figure 8.

The housing 45 supports bar 41 in the same manner as described via flexure web 59 which is located at the centre of mass of the bar 41. The bar 41 is of chevron shape, although the chevron shape is slightly different to that in the earlier embodiments and has a more rounded edge 41e opposite flexure web 59 and a trough-shaped wall section 41f, 41g and 41h adjacent the flexure web 59. The ends of the bar 41 have screw-threaded bores 300 which receive screw-threaded members 301 which may be in the form of plugs such as grub screws or the like . The bores 300 register with holes 302 in the peripheral wall 52a of the housing 45. The holes 302 enable access to the plugs 301 by a screwdriver or other tool so that the plugs 301 can be screwed into and out of the bore 300 to adjust their position in the bore to balance the mass 41 so the centre of gravity is at the flexure web 59.

As drawn in Figure 24, the bores 300 are a 45° angle to the horizontal and vertical in Figure 24. Thus, the two bores 302 shown in Figure 24 are at right angles with respect to one another .

Figure 24 also shows openings 305 for receiving the transducer 71 for monitoring the movement of the bar 41 and producing signals which are conveyed to the SQUID device. Typically, the transducer is in the form of a coil and as the bar 41 moves slightly due to the gravity difference at ends of the bar, a change in capacitance occurs which alters the current in the coil to thereby provide a signal indicative of movement of the bar 41.

Figure 25 is a more detailed view of part of the housing of Figure 24 showing the openings 305. As can be seen from Figure 25, the openings 305 have shoulders 401 which form grooves 402. A spring 403 is arranged adjacent surface 406.

Figure 26 shows the transducer 71. The transducer 71 is formed by a generally square macor plate 410 which has a circular boss 407. A coil 408 is wound about the boss 407 and may be held in place by resin or the like. The coil 408 may be multi-layer or a single layer coil.

Figure 27 shows the location of the plate 410 in the opening 305 in which the plate locates in the grooves 402 and is biased by the spring 403 against the shoulders 401 to hold the plate 410 in place with the coils 408 being adjacent the edge face 41a of the bar 41.

Thus, the coil 408 and the bar 41 form an Ic circuit so that when the bar 41 moves, the current passing through the coil 408 is changed.

As will be apparent from Figure 24, four transducers 71 are arranged adjacent the ends of the bar 41. The other housing 47 also has four transducers arranged adjacent the bar 42. Thus, eight transducers 71 are provided in the gradiometer .

Figure 28 is a diagram of the bars 41 and 42 showing them in their ^λλ in use" configuration. The transducers which are located in the openings 305 are shown by reference numbers 71a to 71e to equate to the circuit diagrams of Figures 29 and 30.

With reference to Figures 29 and 30, transducers 71a and 71b associated with the bar 41, and transducers 71g and

71e associated with the bar 42 are used to provide the gravity gradient measurements .

Input terminals 361 provide input current to the superconducting circuits shown in Figure 29. Heat switches which may be in the form of resistors 362 are provided which are used to initially set the superconducting current within the circuit. The heat switches 362 are initially turned on for a very short period of time to heat those parts of the circuit at which the resistors 362 are located to stop those parts of the circuit from superconducting. Currents can then be imposed on the superconducting circuit and when the heat switches formed by the resistors 362 are switched off, the relevant parts of the circuit again become superconducting so that the current can circulate through the circuits subject to any change caused by movement of the bars 41 and 42 under the influence of the gravity gradient and angular acceleration, as will be described hereinafter.

The transducers 71a, 71b, 7Ig and 71e are connected in parallel to circuit line 365 and to circuit line 366 which connect to a SQUID 367.

Thus, as the bars 41 and 42 rotate about their respective flexure web, the bars 41 and 42, for example, come closer to the transducer 71a and therefore further away from the transducer 71b, and closer to the transducer 71h and further away from the transducer 71g respectively. This therefore changes the current flowing through the transducers and those currents are effectively subtracted to provide signals for providing a measure of the gravity gradient.

As is shown in Figure 31, transducers 71σ and 71d form a separate circuit and are used for frequency tuning of the bar 41 and transducers 71a and 71b. Similarly, the

transducers 71e and 71f are used for frequency tuning of the bar 42 and the transducers 71g and 71h. Frequency tuning of the bars is important because the bars should be identical in order to reject angular accelerations. The frequency tuning circuits therefore enable electronic tuning of the bars to match resonant frequencies and to achieve mode rejection so that each of the bars does function in an identical manner .

The transducers 71a, 71b, 71g and 71h are also used to form angular accelerometers for measuring the angular movement of the mounting 5 so that feedback signals can be provided to compensate for that angular movement.

To do this, the line 366 is connected to a transformer

370. The polarity of the signals from the transducers 71a and 71b and 7Ig and 7Ih are reversed so that the output of the transducer 370 on lines 371 and 372 is an addition of the signals rather than a substraction, as is the case when the gradient is measured so the addition of the signals gives a measure of the angular movement of the bars. The outputs 371 and 372 are connected to SQUID device 375 for providing a measure of the angular acceleration which can be used in the circuit of Figure 10 to provide compensation signals to stabilise the mounting 5.

Thus, according to the preferred embodiment of the invention, the angular accelerometers 90' provide a measurement of angular acceleration, for example, around the x and y axes, and the angular accelerometer formed by the bars 41 and 42 and the transducers 71a, 71b, 71g and 71h provide a measure of the angular accelerometer around the, for example, z axis.

Figures 31 and 32 show an actuator for receiving the control signals to adjust the mounting in response to

angular movement of the mounting 5.

The actuator shown in Figures 31 and 32 are schematically shown in Figure 10 by reference numerals 52, 53, 54 and 55. All of the actuators are the same and Figures 31 and 32 will be described with reference to the actuator 52 which makes adjustment around the x axis shown in Figure 10.

Actuator 52 shown in Figure 31 has a hollow disc housing 310 which has a mounting bracket 311 for connecting the disc housing 310 to mounting 5. The hollow disc housing 310 therefore defines an inner chamber 312 in which is located coil support plate in the form of a disc 313. The disc 313 has a wide hub section 314 and two annular surfaces 315 and 316 onto which windings Wl and W2 of coils are wound about the hub 314.

The disc 313 is also provided with a radial bore 319 and a hole 320 at the periphery of the disc 313 which communicates with the bore 319. A hole 321 is provided at the hub 314 and communicates with the bore 319 and extends to a hollow rod 328 which locates in a tube 330. The rod 330 is fixed to the disc 313 and also to support frame 340 which is fixed to main body 61 (not shown in Figure 31) . The tube 330 is connected to the disc housing 310 for movement with the disc housing 310 relative to disc 313, rod 328 and frame 340.

The winding Wl provided on the face 315 has a lead 331 which passes through the hole 320 and then through the bore 319 to the hole 321 and then through the tube 328 to the right, as shown in Figure 31. A lead 332 from the other end of the winding Wl passes through the hole 321 and through the hollow rod 328 also to the right so that current can be supplied to the winding Wl through the leads 331 and 332.

The second winding W2 provided on the face 316 has a lead 333 which passes through a radial hole 334 and bore 345 in the disc 313 and then through hole 337 to tube 328 and to the left in Figure 31. The other end of the winding W2 has a lead 338 which passes through the hole 337 into the tube 328 and to the left in Figure 31. Thus, current can circulate through the winding W2 via the leads 333 and 338.

When the windings Wl and W2 are energised or the current passing through the windings changes, the disc housing 310 is moved relative to the disc 313 and frame 340 and because the disc housing 310 is connected to the mounting 5 by the bracket 311, the mounting 5, in the case of the actuator 52, is adjusted. The movement of the disc housing 310 is generally a longitudinal movement (i.e. linear movement) in the direction of the axis of the tube 330 and rod 328. To facilitate such movement, clearance is provided between the ends of the rod 330 and the frame

340 and about the disc 313. The bracket 311 is offset relative to the flexure web (such as the flexure web 37) so that movement of the housing 310 applies a torque to the first part 25 of the mounting 5 to cause rotation of the part 25 about the flexure web 37.

In the preferred embodiment of the invention, four actuators are provided for providing actual adjustment about the various axes and flexure webs and the actuators operate in combination in response to signals received from the angular accelerometers to maintain stability of the mounting 5 when the gradiometer is in use .

For cryogenic operation of the gradiometer, the mounting 5, housings 45 and 47, bars 41 and 42, the hollow disc housing 310, coils, and electrical leads referred to previously, are all made from superconducting material

such as niobium.

In embodiments of the invention where the gradiometer is not cryogenically operated, the components can be formed from other materials such as aluminium.

The angular accelerometers 90 ' have zero quadrupole moment which means that the centre of mass coincides with the flexure web and that consequentially they are insensitive to both gravity gradient and centrifugal force. Linear accelerometers 90' ' (Figure 22) could also be provided. The linear accelerometers 90 ' ' do not apply active compensation but may apply corrections to the final measured gradient data . Thus , data relating to linear acceleration can be recorded and possibly used in later processing.

One or both of the bars 41 and 42 can also be used as an angular accelerometer to provide a measure of angular movement of the mounting 5 so that appropriate feedback signals can be generated to compensation for that movement by control of the actuators previously described.

In the preferred embodiment, four angular accelerometers are provided with two of the accelerometers being formed by the bars 41 and 42. The use of four accelerometers arranged at 45° angles with respect to one another enables adjustment about the x, y and z axes by torque supplied from two or more of the actuators at any one time.

The disc 310 prevents flux from the windings Wl and W2 from leaving the actuator and because the leads 331 and 332 and 333 and 338 leave the actuator through the elongate tube 330, the ability of flux to pass out of the actuator is substantially prevented.

Thus, spurious magnetic fields which may detrimentally

effect operation of the instrument are not generated by the actuator and therefore do not influence the sensitivity or operation of the instrument.

The tube 330 preferably has a length to diameter ratio of 10:1 at the least.

The disc plate 316 is preferably formed from macor and the hollow disc housing 310 is formed in two parts 310a and 310b. The part 310b forming a closure panel which enables the disc 313 to be located in the chamber 312 and then the disc housing 310 closed by locating the plate 310b in place . With reference to Figures 33 and 34 , the manner in which the balance of the bars 41 and 42 is achieved will be described. A pair of displacement sensors formed by capacitors 400 and 401 are provided for two main purposes :

1. To measure the residual linear acceleration sensitivity of each bar 41 (and 42) to enable the bars to be mechanically balanced using the grub screws 301 described with reference to Figure 24, before operation at low temperatures ; and

2. To measure the induced linear acceleration sensitivity of each bar 41 and 42.

The bars 41 and 42, in their respective housings, are rotated in a jig (not shown) through 360°. This provides an acceleration range of 2 g _E , which is typically 100 times greater than the accelerations which may be conveniently applied at low temperature. A typically requirement is for the capacitors 400 and 401 to be able to detect 0.1 nm over a period of 1 to 20 minutes . A pair of capacitors 400 and 401 is required for each bar to provide some discrimination against sensor drift, since rotation of the bar 41 will cause one capacitor 400 to increase and the

other capacitor 401 to decrease by the same amount, as is shown in Figure 33, whereas thermal expansion will cause both outputs of the capacitors 400 and 401 to increase. The capacitors 400 and 401 remain in place, even though they are unusable at low temperatures , and therefore their components need to be non-magnetic so as to not interfere with the operation of the gradiometer and, in particular, its nearby superconducting circuitry.

Figure 33 shows that as the bar 41 pivots, the gap applicable to the capacitor 400 decreases and the gap of the capacitor 401 increases.

The capacitors 400 and 401 are formed by the face 41a of the bar 41 (and the corresponding face on the other bar 42) and second plates 405 which are spaced from the face 41a. The gap between the plates of the respective capacitors 400 and 401 must typically be resolved to about 1 ppm.

Figure 34 shows the calibration circuit applicable to the capacitor 400. A circuit for the other capacitor 401 is identical .

The capacitor 400 forms a high Q-factor resonant circuit with inductor 410. The inductor 410 and capacitor 400 are provided parallel to capacitors 411 and 412 and connect via capacitor 413 to an amplifier 414. The output of the amplifier 414 is provided to a frequency counter 415 and also fed back between the capacitors 412 and 411 by line

416. The capacitor 400 therefore determines the operating frequency of the amplifier 414 which can be read to a high precision .

If the bar 41 is out of balance, the frequency counter 45 will tend to drift because of the imbalance of the bar. This can be adjusted by moving the grub screws 301 into

and out of the masses as previously described until balance takes place. The amplifier 414 can then be disconnected from the frequency counter 415 so that the gradiometer can be arranged within the Dewar 1 with the other parts of the circuits shown in Figure 34 in place.

Since modifications within the spirit and scope of the invention may readily be effected by persons skilled within the art, it is to be understood that this invention is not limited to the particular embodiment described by way of example hereinabove.

In the claims which follow and in the preceding description of the invention, except where the context requires otherwise due to express language or necessary implication, the word "comprise" or variations such as "comprises" or "comprising" is used in an inclusive sense, i.e. to specify the presence of the stated features but not to preclude the presence or addition of further features in various embodiments of the invention.