GB2117118A - Compensating gyroscopes for temperature and other variations - Google Patents

Abstract

A gyroscope having a rotor with magnetic poles and a stator with two pairs of coils, in which compensation is provided for variations particularly with temperature, the field strength of the magnet poles and the speed of rotation of the rotor. The former compensation is achieved by monitoring variations in the amplitude of a reference signal (Vref) representative of the sum of the signals induced in one or each pair of coils, and applying an appropriate correction to the gyroscope output signals on terminals (55, 53'). The latter is achieved by monitoring variations in the frequency ( DELTA omega ) of the reference signal, and similarly applying a correction to the gyroscope output signals. The two compensations can be provided simultaneously or one independently of the other. <IMAGE>

Description

SPECIFICATION Gyroscopes This invention relates to gyroscopes and particularly, but not exclusively, to flexure, suspended, free-rotor gyroscopes.

A conventional flexure-suspended gyroscope has a sensitive element serving as a rotor, rotatably mounted within a stator having stator coils. The rotor carries or is formed with magnetic poles which, as the rotor rotates, induce voltages in the stator coils. These induced voltages vary if the sensitive element tilts, and the variation in the induced voltages as a result of such tilt is used to provide the useful output of the gyroscope. A particular type of such a conventional gyroscope (for example as disclosed in European Patent Specification No. 0,009,347) has two pairs of stator coils, with the two coils of one pair being aligned about one diametral axis and the two coils of the other pair being aligned along an axis perpendicular to the coils of said one pair.Each pair of coils may act as combined pick-off and torquer coils in that they not only "pick-off" voltages induced as a result of rotation of the magnetic poles carried by the rotor but they also conduct electric current for applying a torque to the rotor.

It has been known for some time that one of the factors affecting the accuracy of the gyroscope output signal is the temperature of the gyroscope. The general ambient temperature of the gyroscope is important, but so also is the heating effect caused by power dissipation in the gyroscope motor and the heating effect (commonly called joule heating) of the torquer current passing through the combined pick-off and torquer coils. Indeed, the torquer current can give rise to a significant temperature increase within the gyroscope as it is difficult to arrange for adequate heat dissipation. This heating effect is, of course, compounded in the case of a gyroscope of small size which gives rise to close proximity of the gyroscope components within the gyroscope casing.Such small gyroscopes are in demand due to the ever increasing stringency of size and weight limitations in almost all applications but especially in airborne applications. It has been proposed to use direct temperature sensors located at particular places in the gyroscope in order to monitor temperature changes, with the object of compensating the output signal for temperature variations.This proposai gives rise to several problems: first, and most important, a given sensor does not sense the root cause of the change in temperature being monitored, merely the effect thereof as regards the ambient temperature at the location of the sensor, which location is therefore critical; secondly, there is a time lag between the sensing of a temperature change and the implementation of the consequential compensation when using this method; and thirdly, it is undesirable to take further electrical leads through the gyroscope casing; and fourthly, there is frequently no room within the small, tightly packed confines of the gyroscope casing to accommodate a temperature sensor.

According to the present invention there is provided a gyroscope having a motor driven sensitive element and output signal deriving means which includes stator coil means positioned near the sensitive element and which is operable for deriving a gyroscope output signal dependent upon tilt of said sensitive element, said output signal deriving means including two stator coils diametrically opposed on mutually opposite sides of the rotational axis of said sensitive element, combining means connected to said two stator coils and operable for forming a reference signal representative of the sum of respective electrical signals induced in said two stator coils, and compensating means connected to said combining means and operable for monitoring variations in said reference signal and, in response to such variations for adjusting the value of the gyroscope output signal.

The present invention stems from the realisation that in some of Applicants' gyroscopes such as disclosed in European Patent Specification No. 0,009,347 already referred to, there is already present a signal which varies in a predictable manner with temperature. This signal is the so-called reference signal which is obtained by summing the voltages induced in the two coils of at least one pair of stator coils. The crux of the realisation referred to is that the reference signal provides a monitoring point which reflects, virtually instantaneously, any change in temperature experienced by the gyroscope.

Thus, the reference signal (which does not vary substantially with tilt of the sensitive element) can be monitored and used as the basis for the correction signal which is applied to the gyroscope output signals.

Two of the most significant parameters of the gyroscope which vary with temperature are the magnetic field strength of the rotor magnet poles (although the extent of this depends on the material chosen for the magnets), and the rotational frequency or speed of the rotor, the latter being temperature sensitive because the frequency of the oscillator used to drive the rotor varies with temperature. For many materials, the field strength of a magnetised site or pole in the material varies with temperature because domains of the magnetic material become more randomly oriented with increase in temperature, causing a consequential reduction in field strength.

Compensation for variations in magnetic field strength can conveniently be effected by making the compensating means operative to monitor variations in the amplitude of the reference signal. This can be achieved because variation in the magnetic field strength of the rotor magnets results in a corresponding variation in the amplitude of the reference signal. Compensation for variation in the rotational frequency of the sensitive element can conveniently be achieved by making the compensating means operative to monitor variations in the frequency of the reference signal. This result can be achieved because the frequency of the oscillator driving the synchronous motor means which in turn drives the rotor, and the rotational speed of the rotor itself, cause a corresponding variation in the frequency of the reference signal.

As will be explained, compensation can be provided for magnetic field strength variation alone, for rotor speed variation alone, or for both of these variations simultaneously.

In a gyroscope sensing angular rate, the torquer current is related to the angular rate (as sensed by the gyroscope output signal) by the torquer scale factor. Commonly, such a gyroscope will be used in applications where the angular rate will be integrated to give angular displacement, the latter being used to control or navigate a vehicle. Hence, errors in the output signal due to temperature variations will cause positional or displacement errors to occur in the navigation. The amount of joule heating of the torquer coils increases with increasing rate demands. Thus, at high rates when minimum torquer scale factor error is required, the maximum changes in magnetic field strength occur. By recourse to the invention, such errors in the output signal due to changes in the torquer scale factor, can be reduced or substantially eliminated.

A flexure-suspended, free-rotor gyroscope in accordance with the present invention will now be described in greater detail, by way of example, with reference to the accompanying drawings, in which: Figure 1 is an elevational view of the gyroscope in cross-section, Figure 2 is a fragmentary cross sectional view taken at ninety degrees to the section of Figure 1, Figure 3 is a block diagram relating to the basic operation of the gyroscope, Figure 4 is a block diagram illustrating how the gyroscope output signal is corrected both for variations in magnetic field strength and rotor speed with temperature, Figure 5 is a block diagram illustrating how the gyroscope output signal is corrected for variations in magnetic field with temperature, and Figure 6 is a block diagram illustrating how the gyroscope output signal is corrected for variations in rotor speed with temperature.

Referring to Figures 1 and 2, the gyroscope is indicated at 1 and is generally similar in mechanical construction to the gyroscope disclosed in the Applicant's European Patent Specification No.

0,009,347. The gyroscope comprises a casing or housing 2 within which are mounted anti-friction bearings 3 and 4 which journal a hollow drive shaft 5 about its geometrical spin axis. The drive shaft 5 is rotated by a generally conventional synchronous hysteresis spin motor 6 having a stator 7 and a rotor 8.

The gyroscope has a sensitive element in the form of a wheel 9 radially suspended by a flexure support spider 11 having four equiangularly disposed resilient, thin, flat arms 12, 12, 13, 13 which flexibly support the wheel 9 for rotation about the normal spin axis. The wheel 9 is supported in the spin axis direction by a further flexure support or strut 14 extending through an aperture 1 5 in the spider 1 The flexure support 14 has a cylindrical extension 1 6 received by the hollow shaft 5, whereby it is affixed within the latter.

The flexure support 14 consists of a machined cylindrical rod having three flat flexure elements 18, 19 and 21 milled in its active flexure region. The cylindrical portion 16 of the rod is affixed within the shaft 5 in the central bore 23, while its opposite cylindrical portion 24 is fixed to a spoke 25 (Figure 1) of the gyroscope wheel 9 through a tubular projection 26. The end flexures 18 and 21 lie in the same diametral plane of the support 14, a plane perpendicular to the diametral plane of the intermediate flexure 1 9. The flexure element 1 9 is preferably twice as long as either of the equal length end flexure elements 18 and 21.

The suspension system for the gyroscope wheel 9 provides translational rigidity along three mutually perpendicular axes and a low torsional restraint in the plane of the wheel in a simple, low cost configuration having an inherently low sensitivity to twice-rotor-speed vibration. The use of the series of three flat flexure elements 1 8, 1 9 and 21 results in a desirable and significant reduction in the flexural rigidity of the support 14. The use of the three flat flexure elements 1 8, 19 and 21 also advantageously keeps the centre of flexing constant, no matter what the direction of deflection of the gyroscope wheel 9. The intermediate flat flexure element 1 9 is centred in the aperture 1 5 in the spider 11.

In operation, all radial and drive motor torque loads on the rotating system are carried by the spider 11 which accommodates tilt of the gyroscope wheel 9 with respect to the drive shaft 5 by twisting deflection of its crossed arms 12, 13. In fact, the central part of the spider 11 may be likened to the intermediate gimbal of a Hooke's universal joint. Because this effective gimbal is formed from extremely thin metal sheet, it is inherently mass-balanced with respect to the two mutually perpendicular pivot axes and the gyroscope thus has a low inherent sensitivity to twice-rotor-speed vibration.

All axial loads on the rotating system are carried by the triple flexure support 14. As previously noted, the support 14 is proportioned with the fiat intermediate flexure element 1 9 twice as long as each of the two flat and flexure elements 18 and 21. Such a configuration has equal flexural stiffness in any deflection direction, as well as equal columnar strength. Although the axial support of the effective central gimbal portion of the radial suspension is soft, the extremely low mass of the effective gimbal prevents an excessive anisoelastic acceleration sensitivity.

It is seen that the axially disposed triple flexure support 14 is affixed at one end through the tube 26 to the spoke 25, and fixed at its opposite end in the bore 23 in the hollow drive shaft 5. The opposite end of the hollow shaft 5 is provided with a screw 28 mating with a thread internally of the hollow shaft 5. The races of the ball bearings 3 and 4 are thus confined between a flanged portion 29, from which a bridge or yoke 31 extends, and the head of the screw 28 when the latter is tightened. There may be used a magnetic suspension spring compensation system (not shown) of the type disclosed in British Patent Specification No. 722,492.

The gyroscope wheel 9 includes a ring-shaped or annular channel 32 at its periphery. The open end of the annular channel 32 faces the hysteresis motor 6 and provides an air gap region generally indicated at 33, the channel 32 being constructed of soft magnetic material and having integrated sides or legs 34, 35 and 36 for providing a magnetic circuit, including the air gap 33. Within the air gap 33 and affixed by a conventional adhesive to the inner surface of the outer leg 34 of the annular channel 32 for rotation therewith is a ring-shaped permanent magnet 37, which magnet may be constructed as a flat cylinder of a conventional magnetic alloy such as a platinum-cobalt, samarium-cobalt, or other permanent magnetic alloy having similar characteristics.The magnetic material of the ring 37 is permanently magnetised in the radial direction, for example, at eight equiangularly spaced sites all of which are polarised in the same radial sense. Between adjacent poles, the magnetisation of the ring 37 falls to a low value or preferably even to zero. Thus a unidirectional magnetic field resides in the air gap 32 between the ring magnet 37 and the second or inner leg 36 of the annular channel 32, the amplitude of the field varying in a generally sinusoidal or undulating manner around the air gap 32.

Whilst eight permanently magnetised sites are provided on the permanent magnet 37, they are arranged to cooperate with four equiangularly spaced air core pick-off coils 38 (only two being seen in Figures 1 and 2) disposed in a cylindrical shell 39 of electrically-insulating material such as a conventional synthetic plastics composition The coils 38 are disposed generally conformally within the cylindrical shell 39, so that they may be supported by the shell partly in the annular air gap 33. In this mariner, the four air core coils 38, serving as stator coils, are mounted in the shell 39 for fixed support with respect to the housing 2, the edge of each coil being inserted into a sector of the air gap 33, between the permanent magnet 37 and the inner leg 36 of the channel 32.In view of the use of four coils 38 and of the eight permanently magnetised sites in the magnet 37, the angular length of each coil 38 along the air gap 33 is approximately equal to two and a half times the angular distance between the centres of the magnetised sites in the permanent magnet 37. It will be understood that the number of magnetised sites in the magnet 37 was chosen merely by way of example, and that this number may be changed as circumstances dictate.

Referring to Figure 3, the four coils 38 are arranged in two orthogonal pairs, the two coils of each pair being mutually opposite and being connected in series opposition. Any difference between the voltages induced in a first pair of the coils 38 as a result of tilt of the wheel 9 is fed by a lead 40 to a first preamplifier 42. A signal representing the sum of the voltages induced in the first pair of coils 38 is fed by a lead 43 to an electronic block 44, and a signal representing the sum of the voltage induced in the second pair of coils 38 is fed by a lead 41 to the same block 44. The block 44 provides a sum of the signals on the leads 43 and 41, this being the previously mentioned reference signal, shown as Vref in Figure 3.The reference signal, after processing in a conventional phase-locked loop 45, is used to synchronously demodulate the output of the preamplifier 42, this demodulation occurring in a demodulator 46. The demodulated signal, after shaping in a shaping circuit, is fed to a torquer (voltageto-current) amplifier 48 which delivers current by a lead 49 to the second pair of coils 38, this current serving to torque the wheel and hence cause precession of the sensing element to counteract any tilt of the latter about a first axis, so that the sensing element is brought back to its null position. A torquer feedback circuit 52 is provided.

The first pair of coils 38, the leads 40, the preamplifier 42, the demodulator 46, the shaping circuit 47, the torquer amplifier 48, the lead 49 and the torquer feedback circuit 52 constitute a first channel.

An identical second channel is formed by the second pair of coils 38, and associated duplicate components 40', 42', 46', 47', 48', 49' and 52'. It will be appreciated that any difference in the signals induced in the second pair of coils 38 is processed in the second channel and fed back as a torquer current into the first pair of coils 38, in order to null the tilt of the sensitive element about a second, orthogonal axis by precessing the gyroscope. The first and second channels provide gyroscope rate output signals at terminals 53, 53', i.e. across precision resistors 51, 51' connected between ground and one end of the respective pairs of coils 38. These output signals are not corrected for any temperature variation experienced by the gyroscope.

Figure 4 shows how the rate output signals on the terminals 53, 53' are corrected for variations with temperature in magnetic field strength of the magnetised sites or poles of the magnet 37, and for variations with temperature of rotational speed of the gyroscope wheel 9. The former correction is derived from the reference signal Vref which is taken from terminal 55 in Figure 3. The latter correction is derived from the phase-locked loop 45, shown both in Figure 3 and 4. The phase-locked loop 45 operates by phase locking a square wave to the gyroscope reference signal, and as a result the control voltage at the output 56 of the phase detector of the phase locked loop 45 gives a measure of the frequency difference 9w by which the gyroscope reference signal differs from its normal value at a reference temperature.

The reference signal Vref which is a sinusoidal voltage, is amplified in an amplifier 57, a.c. coupled to remove any unwanted d.c. bias using 'any conventional a.c. coupler which may be a unity gain buffer amplifier 58, half-wave rectified in a rectifier 59 and averaged in an averaging circuit 60 which may be a low pass filter. The resultant d.c. signal is fed to a summing junction 62 which acts as a differencing means by combining three signals with the polarities indicated. The junction 62 derives the difference between the d.c. signal from the circuit 60 and a d.c. voltage from a temperature stabilised d.c. source 63. To the summing junction 62 is also fed, after amplification in an amplifier 64, the signal Aw representing the variation of the frequency of the wheel 9 from its frequency at a reference temperature.

The output of the summing junction 62 is a d.c. signal having a magnitude and polarity respectively representative of the magnitude and sense of the correction to be applied to the gyroscope output signals on the terminals 53, 53'. The d.c. output of the summing junction is amplified and scaled (as will be explained) in an amplifier 65 to constitute the correction signal. The correction signal varies the gains of two variable gain, voltage controlled amplifiers 66, 66' to which the respective rate output signals from the terminal 53 or 53' are applied. Hence, the outputs of the amplifiers 66, 66' on terminals 67, 67' are the rate output signals corrected for the two variations mentioned resulting from changes in temperature.

As previously explained, the torquer current is related to the angular rate of the gyroscope (as detected at the terminals 53, 53') by the torquer scale factor.

The torquer scale factor (TSF in radians per Ampere) of the gyroscope is given by: 377 Bur 2 CB TSF ' 8 lo 8 l X 3tr r2 C=----n 81 where B is the magnetic flux density (Tesla), n is the number of turns of the coils, r is the radius of the wheel 9 in metres, I is the amount of inertia of the wheel 9 in Kg.m2, and w is the angular frequency of the wheel 9 in radians per second.

The gyro reference voltage is given by Vref = 2rlnBw cos (wot) (volts) where 1 is length in metres of conductor cutting magnetic field and wc is the carrier frequency (radians per second) of the pick-off signal induced in the coil pairs 38.

Thus the maximum value of V,, = 2rlnBw = KBw K=2rin The change in torquer scale factor is given by:gives

The change in reference voltage is given by:

gives Typically, the variation of CA) with temperature is much less than the variation of B. Thus for less accurate applications it may be treated as a secondary effect and ignored. Thus we can say: C ATSF= - AB w and AVref = KwAB If AVret is scaled by a factor proportional to C n)2K a correction for torquer scale factor can be obtained. In addition as the varation of B is linear with temperature the correction signal is a direct measure of the temperature and joule heating effects.

For more accurate applications it is clear from equations (1) and (2) that account must be taken of the wheel speed variation. If the variation of the wheel speed is monitored and scaled by a factor proportional to -2K this can be added to the reference voltage to correct for the wheel speed variation.

Thus when the summed voltage is scaled by a factor proportional to C a)2K correction for torquer scale factor can be obtained.

From the foregoing, it can be seen that the amplifier 64 shouid multiply the As9 signal from the phase-locked loop 45 by a factor proportional to -2K and tnat the summed voltage from the summing junction 62 should be multiplied by a factor of C/o2K in the amplifier 65, to give a correctly scaled correction signal for application to the amplifiers 66, 66'.

Whilst Figure 3 shows the signals on leads 41 and 43 being fed to a common block 44 and phaselocked loop 45, it is possible for the leads 41 and 43 to feed separate blocks 44 and phase-locked loops 45. There are then Vref and Aw signals for each channel, the circuit of Figure 4 being duplicated to provide independent compensation for the rate signals on terminals 53, 53'.

Figure 5 is similar to Figure 4, but shows the circuit if compensation for magnetic field strength only is desired. It will be seen that Figure 5 is identical to Figure 6, but the correction for Aw is omitted.

Should compensation for wheel speed variation (i.e. bo) only be required, the circuit of Figure 6 can be used. Figure 6 is identical to Figure 4 extept that correction for variations in the amplitude of Vref are omitted, enabling the two amplifiers 64 and 65 of Figure 4 to be replaced by a single, equivalent amplifier 70.

It will be appreciated that in gyroscopes where Vref already exists, it is a relatively simple matter to take advantage of the present inventive concept. Even in cases where Vref does not exist, the modification required to provide at least one set of stator pick-off coils to produce Vref is not a major one. It will also be appreciated that because the variation in temperature is monitored not as an absolute value at one or more specific locations (with the attendant problems referred to hereinbefore) but as it affects one or more parameters of the gyroscope, then the compensation for temperature variation is far more accurate and dynamic with the result that the performance of the gyroscope is greatly enhanced. This represents a significant advance in the art.

Claims (11)

1. A gyroscope having a motor driven sensitive element and output signal deriving means which includes stator coil means positioned near the sensitive element and which is operable for deriving a gyroscope output signal dependent upon tilt of said sensitive element, said output signal deriving means including two stator coils diametrically opposed on mutually opposite sides of the rotational axis of said sensitive element, combining means connected to said two stator coils and operable for forming a reference signal representative of the sum of respective electrical signals induced in said two stator coils, and compensating means connected to said combining means and operable for monitoring variations in said reference signal and, in response to such variations for adjusting the value of the gyroscope output signal.

2. A gyroscope according to claim 1 comprising a sensitive element rotatable by synchronous motor means with respect to at least one pair of stator coils, the two coils of the or each pair being diametrically opposed on mutually opposite sides of the rotational axis of the sensitive element, the stator coils of the or each pair being operative to produce an electrical signal from which is derived a gyroscope output signal representative of the tilt of the sensitive element, combining means for obtaining from the, or at least one of the, pair of coils a reference signal representative of the sum of the signals induced in the or each pair of coils, and compensating means for monitoring variations in the reference signal with temperature and deriving therefrom a correction signal which is applied to the or each gyroscope output signal to compensate the latter for temperature variations.

3. A gyroscope according to claim 1, wherein the compensating means are operative to monitor variations in the amplitude of the reference signal so as to compensate the or each gyroscope output signal for variations in the magnetic field strength of magnetic poles carried by or formed on the sensitive element.

4. A gyroscope according to claim 3, wherein the compensating means comprise an amplifier for amplifying the reference signal, a half-wave rectifier for rectifying the amplified reference signal, an averaging circuit for converting the rectified signal to a d.c. signal, an electrical signal source of stabilised magnitude, and difference means for obtaining the difference between the output of the averaging circuit and the signal source of stabilised magnitude, the output of the difference means being applied as the correction signal to the or each gyroscope output signal.

5. A gyroscope according to claim 4, wherein the compensating means additionally comprise an a.c. coupler connected between the first-mentioned amplifier and the half-wave rectifier in order to remove any unwanted d.c. bias.

6. A gyroscope according to claim 4 or 5, wherein the compensating means further comprise variable gain amplifiers to which the gyroscope output signals are respectively applied and the gain of each of which is altered by the correction signal.

7. A gyroscope according to claim 1, wherein the compensating means monitor variations in the frequency of the reference signal so as to compensate the or each gyroscope output signal for variations in the rotational speed of the sensitive element.

8. A gyroscope according to claim 7, wherein the compensating means derive from a phaselocked loop, to which the reference signal is applied, a signal representative of the change in frequency of the reference signal attributable to variation in rotational speed of the sensitive element.

9. A gyroscope according to claim 8 as appendant to claim 3, wherein the compensating means are operative to apply said signal representative of the change of frequency to the difference means.

1 0. A gyroscope according to any of the preceding claims, wherein the gyroscope is a flexuresuspended free-rotor gyroscope.

11. A gyroscope constructed and arranged substantially as herein particularly described with reference to Figures 1 to 4, or as modified by Figure 5 or Figure 6 of the accompanying drawings.