GB2102576A - Piezo-electric sensing device - Google Patents

Abstract

A piezo electric device for determining acceleration and angular velocity has at least three piezo-electric arms radially disposed on a rotatable shaft. As shown two dipoles 22 24 are mounted on a shaft 26. The arrangement overcomes disadvantages arising in the single dipole arrangement disclosed in U.S. Patent Specification 4197737. A demodulator circuit 28 is provided. <IMAGE>

Description

SPECIFICATION Information processing means This invention reiates to devices for sensing physical characteristics. In particular, it is related to those devices that can sense such parameters as angular velocity, linear acceleration, magnetic field direction, electric field direction, and air flow data. At present, there exists a multiple sensing device, or multisensor, that is capable of measuring the parameters listed above. This particular apparatus is disclosed in U.S.

Patent No. 4,197,737, for a "Multiple Sensing Device and Sensing Devices Therefor", issued to Roland Pittman on April 15, 1980, which is hereby incorporated by reference.

Figure 1 illustrates the apparatus 10 and a suitable demodulator arrangement 12. Basically, it consists of a rotatable shaft 14 having two radially extending arms of piezoelectric material which form a simple dipole 16. When a physical disturbance is applied, the arms flex, generating a piezoelectric voltage proportional to the degree of flexing. The output is taken from the multisensor by a slip ring arrangement (not shown).

The multisensor as described above senses rates and linear accelerations by rotating piezoelectric elements about a motor spin axis. By appropriately interconnecting the piezoelectric devices, their responses are maximized for the desired input and minimized for the undesired inputs. When the shaft is rotating, any applied acceleration, either linear acceleration in the case of the acceleration sensor or coriolis acceleration in the case of the rotational rate sensor, occurring in the plane of the dipole, is multiplied by cos(o)St), where COs is the spin angular rate, generating a signal at the spin frequency proportional in magnitude to the applied acceleration.However, if an input disturbance occurring at twice the spin frequency (20)5) is applied, it too will produce an output signal at the spin frequency which is thus indistinguishable from the desired signal. Irregularities and imperfections in the bearings can and do introduce accelerations and angular rates that are modulated by the rotating mechanism and result in a significant error in the output. Where the multisensor consists of a single dipole, there is no practical way to compensate for these "spurious" signals.

Given the inability of the apparatus in Figure 1 to discriminate between disturbances at twice the spin frequency and a legitimate input signal, the multisensor has not enjoyed widespread application. It is a goal of the invention, therefore, to utilize the basic multisensor principle in a configuration that is insensitive to bearing noise and other inaccuracies that may result in erroneous output signals.

Summary of the invention The present invention overcomes the deficiencies of the multisensor disclosed in U.S. Patent No. 4,197,737 by incorporating an additional or second dipole which is positioned at right angles to the first. For convenience, the first dipole is referred to as the "real" dipole and the second is referred to as the "imaginary" dipole, being indicative of a rotating coordinate system. The added or imaginary dipole provides an additional source of information that, when demodulated, permits an accurate reconstruction of the information applied originally to the multisensor.

For a better understanding of the present invention together with other and further objects, reference is made to the following description taken in conjunction with the accompanying drawings, and its scope will be pointed out in the appended claims.

Brief description of the drawings Figure 1 is an illustration of the multisensor as disclosed in U.S. Patent No. 4,197,737 and a suitable demodulatorcircuit; and Figure 2 illustrates an improved multisensor in accordance with the invention and its corresponding demodulator circuit.

Description of the preferred embodiment The invention and its associated demodulating circuitry is illustrated in Figure 2. The improved multisensor assembly 20 is almost identical to its predecessor 10 shown in Figure 1. It has two dipoles 22 and 24 of piezoelectric material which are mounted on a shaft 26. The mechanical details of the multisensor, for purposes of construction, are similar to those of the single dipole multisensor. Since they are discussed at length in U.S. Patent No. 4,197,737, they will not be repeated here.

Figure 2 also illustrates, in block diagram form, the necessary circuitry required to demodulate the signal.

The process is similar to that used for the single dipole multisensor with some slight variations. It will be discussed in detail further on.

The following analysis is necessary for an understanding of the operation of the two dipole multisensors.

Consider a sinosoidal disturbance of angular velocity COD in the X-Y plane of the multisensor of amplitude A having a phase angle Q attimet =0. The amplitude of the disturbance is given by: D = A cos(oDt + ) (1) Assuming that the disturbance is along a line at an angle (3 to the X axis of the sensor, the disturbance along the X axis becomes: Dx = D cos 0 (2) and, along the Y axis becomes: Dy = D sin 0 (3) Combining Equations 1 and 2 for the X axis components of the disturbance and Equation 1 and 3 for the Y axis component of the disturbance and expanding provides:: Dx = A/2 [cos (#Dt + Q + 0) + cos (#Dt + (p - 0)] (4) Dy = A/2 [sin (CO0t + Q + 0) - sin (#Dt + 4) - 0)] (5) Considering the sensor to be a pair of dipoles (arbitrarily denoted as real and imaginary) mounted orthogonal to the spin axis and orthogonal to each other, and defining the real dipole as being aligned to the X axis of the instrument at times = 0 and the imaginary dipole as being aligned to the Y axis of the instrument at t = #/(2#s), where COs = the angular rate of shaft rotation, provides, for the real signal: SR = Dx cos (w5t) + Dy sin (#st) (6) and, for the imaginary signal: Sl = Dy cos (#5t) - Dx sin (#st) (7) Expanding provides:: SR = A/2 [cos ((wD ((#D-#s)t - wS)t + # + #) + #) + cos ((#s + (#D)t + # - 0)] (8) Sl = A/2 [sin ( (#D - #s)t + Q + 0) - sin ( (#D + w5)t + # - 0)] (9) In order to recover the original input information for the X and Y axes, it is necessary to multiply the real and imaginary information by sin(#st) and cos(w5t) to obtain four products and combine pairs of these products:: X Axis Signal = SR cos (#st) - Sl sin (#st) (10) Y Axis Signal = Sl cos (#st) + SR sin (#st) (11) Expanding and combining provides for the two signal components: X Signal = A/2 [ cos(Dt + Q + 0) + coslwot + (p - 0)l (12) Y Signal = A/2 [ sin(Dt + # # + + #)- sin(#Dt +#) -6)] (13) These expressions may be rewritten as:: X Signal = A cos(#Dt + (p) cos 0 (14) Y Signal = A cos(#Dt + #) sin 0 (15) The term A cos(#Dt + (p) is the applied disturbance to the sensor, while cos 0 and sin 0 represent the orientation of the disturbance with respect to the instrument axes. Equations 14 and 15 show that, providing that real and imaginary dipoles are employed in the instrument and providing that the signal demodulation is accomplished by sine and cosine multiplication, the output will represent the input to the instrument.

Aliasing of bearing noise at twice the spin frequency into a static bias will not occur.

In the existing multisensor, the information is encoded by means of a single dipole detector rather than by the use of real and imaginary sensors. The effects of the present implementation may be determined by setting the imaginary signal in Equation 10 and 11 equal to zero.

Signal = SO COS (#st) (16) Y Signal = So sin (#st) (17) Expanding and combining provides, for the two signal components: Xsig = A/2 cos(orDt+Q)cos(0)+A/4cos ( (o-25)t+4)+0)+AI4cos ( (D+20l5)t+0) (18) Ysig = A12 cos(CO0t+4))sin(0)A/4sin ( (eoD-20)s)t+Q+0)+A/4sin ((CO0+2CO5)t±0) (19) The first term in Equations 18 and 19 is identical to half of the signal defined by Equations 14 and 15 and represents the correct output for the device. This output, however, is corrupted by the second and third terms.The third term is not overly significant since its lowest possible frequency occurs when COD = O, at twice the spin frequency of the device. It represents a high frequency noise which will normally not affect the system. The second term in Equations 18 and 19 presents difficulty. It shows that disturbances at twice the spin frequency will produce a static bias in the output at a sensitivity (as compared to the static sensitivity of the instrument) of 50%. Since the effect of bearing noise is high at twice the spin frequency, construction of the instrument without the use of both real and imaginary dipoles imposes a significant limitation on its bias stability.

The multisensor is essentially an encoder or modulator. To derive useful information from the multisensor, the output from the dipoles is taken by a slip ring arrangement (not shown) and processed in a demodulator circuit. The demodulation process is simply the reverse of modulation, and as such is governed by the relationship set forth in equations 1 through 15, above. The signal from each dipole is fed into separate channels on lines 30 and 32, respectively. Initially, the signals are amplified and filtered in stages 34 and 36. Then, each signal is simultaneously fed to two demodulators, one being referenced at zero degrees and other at ninety degrees.Mathematically, this is equivalent to multiplying each signal by cos(ozst) and sin(5t). To complete the process, the outputs of demodulator 38 and demodulator 42 are added (the output of demodulator 42 is actually inverted) at summing junction 46 and the outputs of demodulator 40 and demodulator 44 are added at summing junction 48. Output amplifiers and filters are provided at 50 and 52.

The outputs of these stages are the x axis andy axis indicators, as shown mathematically in equations 14 and 15. It should be noted that for a steady state input, tOD is zero and that these signals are not time varying functions, but instead constant DC voltages.

The two dipole multisensor is only one possible configuration that will overcome the deficiency inherent in the single dipole multisensor. Theoretically, any system that samples linear acceleration and angular velocity using a two dimensional system will achieve the desired result. For instant, a sensor having three radially extending arms spaced at 120 with respect to each other will discriminate against the 2Ons term.

Similarly, five arms having 72 spacing is also permissible. The corresponding demodulator would require the appropriate reference frequencies for decoding the sensor output.

In the preceding discussion, it was assumed that the signal from the instrument would be multiplied by a sine or a cosine function to perform the demodulation process. In principle, this could be accomplished by sampling the signals at a high rate, after a suitable anti-aliasing filter, performing on A/D conversion, and multiplying the resultant signal by the appropriate sine or cosine function as stored in a ROM and addressed by the counter on a phase lock synchronization loop. Simple digital low pass filtering of the resultant data would provide information ready for use by a computer. In practice, to take such approach effectively would seem to require an A/D converter of 16 to 18 bit accuracy and resolution which is capable of operating at perhaps a 100 KHZ rate to permit multiplexing of the AID device.At present, such device does not seem remotely approachable. (The accuracy of available analog multipliers also is woefully inadequate for the application.) Currently, the only implementation for the demodulation function which seems adequate is the switching demodulator. The switching demodulator operates by multiplying the signal by + 1 rather than by sin(ost) or cos (east). If a switching demodulator is employed, odd harmonics, characteristic of a square wave, will be introduced into the output and will require filtering.

While there has been described what is believed to be the preferred embodiment of the invention, those skilled in the art will recognize that other and further modifications may be made thereto without departing from the spirit of the invention, and it is intended to claim all such embodiments as fall within the true scope of the invention.

Claims (13)

1. A sensing device comprising: a rotatable shaft; a first dipole comprising two piezoelectric crystalline arms affixed at its midpoint to said shaft and having a signal output; a second dipole comprising two piezoelectric crystalline arms affixed at its midpoint to said shaft, and having a signal output, said second dipole being angularly displaced relative to said first dipole; and demodulation means.

2. A sensing device as set forth in claim 1 wherein said first and second dipoles are oriented in planes normal to the axis of said shaft.

3. A sensing device as set forth in claim 2 wherein said second dipole is oriented at right angles with respect to said first dipole.

4. A sensing device as set forth in claim 3 wherein said first and second dipoles are coplanar.

5. A sensing device as set forth in claim 4 further comprising: slip ring means coupled to said first and second dipoles for providing an output.

6. A sensing device as set forth in claim 5 wherein said demodulation means comprises synchronous demodulator means.

7. A sensing device as set forth in claim 5 wherein said demodulation means comprises: first, second, third and fourth synchronous demodulators, each having a signal input, a reference input, and a signal output; a first signal generator phase referenced to 0"; a second signal generator phase referenced to 90"; a first summing junction having a non-inverting input, an inverting input, and an output; a second summing junction having first and second non-inverting inputs and an output;; wherein said output of said first dipole is connected to said signal inputs of said first and second synchronous demodulators, said output of said second dipole is connected to said signal inputs of said third and fourth synchronous demodulators, the output of said first signal generator is connected to said reference inputs of said first and fourth synchronous demodulators, the output of said second signal generator is connected to said reference inputs of said second and third synchronous demodulators, said output of said first synchronous demodulator is connected to said non-inverting input of said first summing junction, said output of said second synchronous demodulator is connected to said first non-inverting input of said second summing junction, said output of said third synchronous demodulator is connected to said inverting input of said first summing junction, and said output of said fourth synchronous demodulator is connected to said second non-inverting input of said second summing junction.

8. A sensing device comprising: a rotatable shaft; first, second, and third arms of piezoelectric material attached to said shaft and extending outward from said shaft in a normal direction thereto, said arms being oriented at 120" with respect to each other; and demodulation means.

9. A method of sensing linear acceleration and angular velocity, comprising the steps of: spinning a first dipole of piezoelectric crystalline material at its midpoint about an axis of rotation normal to said dipole to produce an output signal from said first dipole; spinning a second dipole of piezoelectric crystalline material at its midpoint about said axis of rotation where said second dipole is angularly displaced with respect to said first dipole to produce an output signal from said second dipole; and demodulating said output signals of said first and second dipoles.

10. A method of sensing linear acceleration and angular velocity as set forth in claim 9 wherein said demodulating further comprises the steps of: synchronously demodulating said output of said first dipole at 0" and 90" phase reference to produce first and second demodulated signals, respectively; synchronously demodulating said output of said second dipole at 0" and 90" phase reference to produce third and fourth demodulated signals, respectively; inverting said third demodulated signal; summing said first demodulated signal and said inverted third demodulated signal to produce an X-output; and summing said second and fourth demodulated signals to produce a Y-output.

11. The method of measuring acceleration and angular rate of a body having a spin axis on at least two nonparallel axes rotating about said spin axis, said at least two axes being essentially perpendicular to said spin axis.

12. A sensing device substantially as hereinbefore described with reference to and as shown in the accompanying drawings.

13. A method of sensing linear acceleration and angular velocity substantially as hereinbefore described