GB2298281A - Differential pressure transducer - Google Patents

Abstract

A capacitive differential pressure transducer has two mechanically similar sensor parts A, B, each comprising a first chamber 18 and a second chamber 19 separated by an electrically conductive diaphragm 15. An electrode 12 is secured in each of the first chambers 18, in close proximity to one side of the respective diaphragm 15 and electrically insulated therefrom to form a capacitance with the diaphragm. The two sensor parts are positioned in like orientation and rigidly conjoined to each other. Fluid corrections 16,17 apply a first fluid pressure to the first chamber 18A of sensor part A and the second chamber 19B of the other sensor part B, and a second fluid pressure to the second chamber 19A of the one sensor part A and the first chamber 18B of the other sensor part B. Because of the mechanical similarity and like orientations of the two parts A, B of the sensor, an electrical measurement of the relationship between the two capacitances (e.g. using a diode pump circuit - Fig.5) provides a differential pressure measurement which is substantially immune to errors induced by forces acting on the diaphragms 15A,B owing to their mass. As shown each part A,B has an upper body part 8 in the form of a cranked conductive ring and a lower body part 14 in the form of a short cylinder with a closed bottom. The parts 8,14 are bolted together with the diaphragm 15 held therebetween. A ring 16 insulates the electrode 12 from the upper body part. The electrodes are connected to a circuit board 20; a conductive plate 21 provides screening.

Description

A DIFFERENTIAL PRESSURE TRANSDUCER This invention relates to a differential pressure transducer of the type used to measure pressure differentials by means of the electrical capacitance formed by a fixed electrode and an electrically conductive diaphragm which is subject to the pressure differential.

The use of symmetrical differential capacitance transducers to measure pressure differential is well known and is described, for example in US2999385, US3232114 and GB-A-2127971. Transducers of this type are generally regarded as the most suitable form of device for measuring very low differential pressure and can provide usable output signals for differential pressures of less than 0.1 micr^warS (0.01 N/tn2 However, there are considerable practical difficulties in the implementation of a satisfactory differential capacitance transducer most of which arise from the need to accurately preserve the small values of capacitance which are obtained and the very small range of diaphragm movements which are permissible to maintain linearity of measurement.

Further problems arise from the need for extremely precise engineering to make the two halves of the transducer and from the affects of stray capacitance caused by movement of any conductive body in the vicinity of the transducer. A further difficulty, which is particularly acute in hand-held instruments, is that of sensitivity to changes in orientation and movements which, due to the finite mass of the diaphragm, can cause deflections of the diaphragm which significantly affect the accuracy of measurements carried out.

Other features of the invention will be apparent from the following description and from the subsidiary claims of the specification.

The invention will now be further described, merely by may of example, with reference to the accompanying drawings in which: Figure 1 is a cross-sectional view of a known symmetrical differential capacitance transducer; Figure 2 is a cross-sectional view of one embodiment of a differential capacitance transducer according to the present invention; Figure 3 is a schematic cross-sectional view illustrating possible fluid connections between parts of the transducer shown in Figure 2; Figure 4 is a plan view of the transducer shown in Figure 2 with various parts omitted for clarity; and Figure 5 is a circuit diagram of one form of sensing circuit that may be used with the transducer shown in Figures 2 to 4.

Figure 1 shows a cross-section of a typical known type of symmetrical differential capacitance transducer. The transducer comprises a pair of body rings 1A and 1B formed of electrically conductive material which are held together by nuts and bolts 2A and 2B with an electrically conductive diaphragm 3 clamped therebetween. Electrically conductive electrodes 4A and 4B are positioned on opposite sides of the diaphragm 3 each being securely mounted in position by attachment to a respective disc 5A and 5B, formed of electrically insulating material, the discs SA and SB in turn being securely bonded to a respective body ring lA, 1B. The transducer thus comprises two chambers 6A and 6B separated from each other by the diaphragm 3.The body rings 1A and 1B, the electrodes 4A and 4B, and the diaphragm 3 are typically formed of brass. The discs SA and 5B are typically formed of glass.

The electrodes 4A and 4B are positioned in close proximity to the diaphragm 3 and are symmetrically arranged either side of the diaphragm so that each electrode forms an electrical capacitance with the diaphragm. Pressure can be applied to the respective chambers 6A and 6B by inlet ports 7A and 7B. A pressure differential applied between the two ports 7A and 7B will thus cause a deflection in the diaphragm with a consequent increase in the capacitance formed with one electrode and a decrease in the capacitance formed with the other electrode. A variety of electrical techniques are used to provide an output signal proportional to the difference between these two capacitances and thus to provide an inferential measurement of the pressure differential applied between the ports 7A and 7B.

It is preferable to employ a physically and electrically symmetrical arrangement for the circuit used to measure the capacitances so any stray capacitance to the electrodes will be similar and the circuit thus balanced. Further, semiconductor junctions are inherently temperature-sensitive, although with very predictable thermal characteristics. By arranging for the circuit to be symmetrical and for critical circuit elements to be in close thermal communication, thermal drifts can be made to oppose one another, leading to an output signal which is substantially unaffected by temperature changes. The temperature of the critical semiconductor elements may be kept uniform by incorporating them into a single integrated circuit.

As mentioned above, there are, however, considerable practical difficulties in the implementation of a satisfactory differential capacitance transducer due to the small values of capacitance which are obtainable, and the very small diaphragm movement which is permissible for linear measurements. A typical transducer will be cylindrical in form, and 50mm is a commonly chosen outer diameter for which an electrode diameter of around 24mm may be employed. The spaces between the diaphragm 3 and each electrode 4A, 4B, generally known as the airgaps, usually have a width of about 100 microns. These typical dimensions result in capacitance values in the region of 40 picofarads.

It can be shown that the transducer output will be acceptably linear over only a small range of diaphragm movement, and for this reason the calibration limit of a transducer will generally be set at a diaphragm deflection of about 30 microns, giving rise to a capacitance change of 12 picofarads. There are commercially available 50mm transducers which are capable of measuring differential pressures up to 100 millibars with a resolution of 1 microbar, giving a dynamic range of 100,000:1. It can be calculated from these figures that the diaphragm movement for increments of 1 microbar will be nominally 30/100,000 microns, or 3 Angstroms. It can also be calculated that the corresponding change in capacitance will amount to approximately 120 attofarads.It will be apparent from the extreme smallness of these physical quantities that the greatest possible dimensional stability is an essential feature of the transducer.

Thermal expansion of the materials from which a transducer is fabricated can cause serious physical distortion of its structure, leading to changes in the airgaps which can be much greater in magnitude than the pressure-induced changes which it is proposed to rely on for the inferential pressure measurements. This matter is made more complicated by the need to employ a variety of different materials in the structure of a transducer, each of which may have a different coefficient of thermal expansion. As an example, it can be calculated that depending upon its exact composition a piece of steel 25mm long will change in length by approximately 2.7 Angstroms per degree Centigrade temperature change. The corresponding change for brass would be 4.7 Angstroms, for aluminium it would be 5.8 Angstroms and for glass it would be 2 Angstroms.All these materials are commonly used in pressure transducers, and it is the combined effect of these differing coefficients together with the changing stress patterns thereby produced in response to changes in temperature, which determine the thermal effects on the airgap. For a given mechanical design of transducer a combination of materials can usually be determined by experimental methods which will result in the airgap remaining acceptably constant.

If the thermal behaviour of the structure on one side of the diaphragm differs from the behaviour of the structure on the other side of the diaphragm it is very likely that changes in temperature will cause the two airgaps to become unequal, thus altering the null point of the system and thereby causing measurement errors. It is an inherent advantage of the symmetrical differential capacitance transducer that by virtue of its symmetry both sides of the transducer will behave similarly. However, this will only be the case to the extent that the two sides are matched dimensionally and with respect to the stress patterns within its structure. To achieve a high standard of performance the airgaps require matching to within a micron or two, and all mating component faces need to be perfectly parallel and optically flat to avoid unforeseen stresses when the transducer is assembled.Exceedingly precise engineering is required to achieve this.

Even so, production tolerances will mean that some examples of a transducer are better matched than others, leading to inconsistencies in performance between one example and another.

A further inherent advantage of the symmetrical capacitance transducer is that of bidirectional measurement capability. When connected to an appropriate electrical circuit, the electrical output will be driven positively or negatively about a null point depending upon which side of the transducer is subjected to the higher pressure. This feature is of great importance for so-called gauge pressure measurements, in which atmospheric pressure is regarded as zero and pressures which are below atmospheric pressure are said to be negative pressures. If one input port of the transducer is assigned as the signal port and the other port is left open to atmosphere, the electrical output will indicate the gauge magnitude of any pressure applied to the signal port and whether it is positive or negative.The bidirectional performance of the transducer will be seriously impaired if engineering tolerances have caused one airgap to be larger than the other, because the sensitivity of the device will be greater when the diaphragm is deflected in one direction than when it is deflected oppositely.

As already said, the values of capacitance which are obtainable are very small, and the changes in capacitance which it is necessary to detect are very small indeed. Stray capacitance is a frequently occurring problem in a variety of electrical circuits, and is a very important consideration in the art of capacitance transducers. In most practical applications the transducer will be incorporated into an instrument casing which also contains electronic components, and may well contain other electrical and mechanical apparatus which is entirely unconnected with pressure measurement. All conductive bodies have a capacitive link with other conductive bodies unless a grounded electrical screen is interposed between them, in which case each body will instead have a capacitive link with the screen.Although the value of stray capacitance is inversely proportional to the spacing between the bodies, the sensitivity of capacitive measurement being undertaken is such that any physical movement of an unscreened conductive body in the vicinity of the transducer is likely to result in a detectable output signal. The conventional symmetrical transducer is particularly susceptible ot this, because the two electrode terminals are located on opposite sides of the transducer. Any change in stray capacitance which occurs at one side of the transducer will be detected in full measure at the nearest electrode, but will not affect the other electrode at all because it is screened by the body of the transducer.The problem may be overcome by providing an electrical screen around the transducer and the parts of the electrical circuit immediately associated with the electrode terminals, but this has to be done with great care since there will be considerable stray capacitance between the electrodes and the screen itself. The screen must be very rigidly constructed, and everything within the screened space including electrical wiring must be rigidly mounted so that no relative movement can occur. In particular, axial movement of the transducer must be avoided, as any such movement with respect to the screen would give rise to differential outputs. Such a solution is practical but bulky and costly to implement.

Stray capacitance within the screened enclosure can be minimised by keeping the physical size of the electrode terminals and associated wiring as small as possible. As has already been mentioned, thermal stability of the electrical output is enhanced by using a single integrated circuit to detect capacitance differentials. However, these two objectives are in conflict because the electrode terminals are at extreme opposite ends of the transducer, requiring physically long conductors to reach the terminals of an integrated circuit.

A further problem which occurs in hand-held instruments and other mobile applications of differential capacitance transducers is that of sensitivity to changes in orientation with respect to the moment of gravity. This arises because the weight of the diaphragm may cause it to sag towards one electrode or the other depending on its attitude in space. Brass is a commonly used diaphragm material and a typical transducer will have a diaphragm perhaps 50 microns in thickness. It can be calculated that such a diaphragm has a weight of 42 milligrams per square centimetre.If the plane of the diaphragm is vertical, its weight will be of no account in determining its position between the electrodes, but if it is rotated to the horizontal position it will be deflected by its own weight to the same extent as would occur if a differential pressure of 42mg/sq.cm had been applied across the transducer. This represents a pressure of 41 microbars, and would give rise to serious errors in many measurement situations. Changes in orientation of only a few degrees are sufficient to cause errors in instruments calibrated to resolve to 1 microbar, and instruments are available which exceed this resolution by a factor of 10.

It will be appreciated from the above that symmetrical differential capacitance transducers have many advantages but that considerable difficulties arise in the practical implementation of this technology.

The differential pressure transducer described below with reference to Figures 2 to 5 aims to overcome or minimise many of these problems.

References to "upper" and "lower" in the following description relate only to the positions of parts shown in the drawings and should not be interpreted as restricting the description to any particular orientation with respect to the force of gravity.

Figure 2 shows a cross-sectional view of a transducer w which comprises two mechanically similar sensor parts labelled A and B respectively, rigidly conjoined to each other side by side and in like orientation with respect to each other. Like parts in the two sensors parts are given the same reference numeral followed by the letter A or B as appropriate.

Each of the two sensor parts comprises an upper body portion 8 made from a rigid electrically conductive material and is in the form of a short cylinder which is inwardly flanged at the uppermost end and is provided with an annular ring of holes accommodating a number of bolts 9 engaging nuts 10. Within the upper body portion 8 and bonded to its flange lies a centrally perforated insulating disc 11 to which is bonded an electrode 12 fabricated from a conductive material and dimensioned so that it is set back a small distance from the lower end face of the upper body portion 8. The electrode 12 carries a spigot 13 which passes through the insulating disc 11 so that the extremity of the spigot 13 provides an externally accessible electrical terminal. The electrically conductive components may typically be formed of brass and the insulating disc formed of glass.

Each of the two sensor parts also comprises a lower body portion 14 in the form of a short cylinder with its lowermost end closed, with annular holes through which pass the said bolts 9 and which is fabricated from the same material as the upper body portion 8.

Interposed between the upper and lower body portions 8 and 14 is an electrically conductive diaphragm 15 which is perforated to allow passage of said bolts 9 but is otherwise impermeable. Dotted ellipses show the points of entry of tubular pressure ports 16 and 17 (see also Figure 4) through the cylindrical walls of the upper and lower body portions 8 and 14.

Each sensor part is placed in a jig which firmly grips the peripheral extremity of the diaphragm 15. With the nuts and bolts 9, 10 loosely engaged, a known axial force is applied to the assembly to put the diaphragm 15 in a state of radial tension, and the nuts and bolts 9, 10 are then tightened to clamp the diaphragm 15 firmly between the upper and lower body portions 8 and 14. A predictable diaphragm tension will thus be retained when the assembly is removed from the jig. With the diaphragm 15 thus tensioned within the device, the surplus peripheral diaphragm material may be cut off.

Each of the sensor parts forms substantially one half of a "symmetrical" differential capacitance transducer. The upper body portion 8 forms with the diaphragm 15 a chamber 18 within which lies the electrode 12 in close proximity to but electrically insulated from the diaphragm 15 thereby forming a small airgap capacitor. The lower body portion 14 forms with the diaphragm 15 a second chamber 19.

Fluid pressures may be introduced to the chambers 18, 19 by means of the pressure ports 16, 17 causing the diaphragm to be deflected towards or away from the electrode 12 depending on the relative magnitudes of the pressures, thereby altering the value of the electrical capacitance.

The two sensor parts are mounted close together, with their axes in parallel alignment and with the electrode 12 of each sensor uppermost.

The planes of the diaphragms 15A and 15B are thus substantially parallel and, in the arrangement shown, are substantially co-planar. In order to make this twin-sensor assembly function as a differential capacitance transducer, fluid connections are made between the upper chamber 18 of each sensor part and the lower chamber 19 of the other sensor part (this being illustrated schematically in Figure 3) so that when the greater of two pressures is applied to the upper chamber 18A of one sensor part, it is applied to the lower chamber 19B of the other sensor part. This will cause one airgap to increase while the other decreases. By comparison with the conventional transducer of unitary construction as illustrated in Figure 1, such a twin-sensor arrangement offers a number of advantages, as will now be described.

The conventional unitary transducer requires great engineering precision in order to achieve satisfactory matching of the two airgaps, and after such a transducer is manufactured any shortcomings in its performance cannot be rectified except by dismantling it and replacing or modifying components. Manufacturing methods at present in use include assembly of a transducer by welding the two halves together, and an unsatisfactory unit of this type would have to be scrapped.

Other manufacturers bond parts of the transducer together with glass frit, and there is no practical means of dismantling such bonds to reclaim the bonded parts. By constructing each part of the symmetrical pair as a separate assembly it is possible to considerably relax manufacturing tolerances and subsequently to select matched pairs of assemblies from a large number of units. By this means, any desired standard of matching may be achieved simply by connecting each assembly in turn to a testing device which will measure the absolute capacitive value of the airgap, and also the capacitance change in response to a chosen pressure difference.

It is preferable also to match the thickness of the two diaphragms in order to enjoy a further advantage of the twin-sensor assembly transducer, which arises because the electrodes are both located in the upper chambers 18 instead of being symmetrically disposed on each side of a single diaphragm as in the case of the unitary transducer.

As described above, the unitary transducer is inherently sensitive to changes in orientation of its axis with respect to the moment of gravity because the weight of the diaphragm will affect its position between the electrodes, thus causing differential capacitance changes as the orientation changes. The magnitude of measurement error is directly proportional to the weight per unit area of the diaphragm material.

The diaphragms of the twin-sensor assembly transducer also move in response to orientation changes but because the two electrodes 12A, 12B have the same orientation with respect to the planes of the diaphragms 15A, 15B both airgaps will change in the same way instead of oppositely. There will accordingly be no electrical output from the differential detector circuit which by design does not respond to common-mode signals, and the twin-sensor assembly transducer is therefore substantially immune to gravity-induced errors to the extent that the diaphragms are matched in thickness. For the most demanding applications it is possible to select matched pairs of assemblies by noting the capacitance changes of each individual assembly when rotated through 90 degrees while attached to the aforementioned testing device.

Further advantages of the twin-sensor assembly transducer arise from the placement of the electrode terminals 13A, 13B in close proximity to one another and on one side of the arrangement. The desirability of keeping the electrical wiring associated with the electrodes as compact as possible, the advantages of using a single integrated circuit for detecting capacitance differentials, the need for adequate screening of the electrode circuits and the requirement for keeping all adjacent conductive elements rigidly located with respect to the electrodes have been discussed above. All these objectives can be more conveniently achieved with the twin-sensor assembly transducer than with the unitary transducer.

As shown in Figure 2, some of the bolts 9 used to secure the upper and lower body portions 8 and 14 together may also be used to support a circuit board 20 and a conductive electrical screen 21. The electrical terminals 13A and 13B of the electrodes 12A and 12B may also conveniently be attached directly to the circuit board 20. Space is also provided between the circuit board 20 and screen 21 for accommodating other electrical components. The circuit board 20 and electrical screen 21 are thus used to rigidly conjoin the two sensor parts.

Figure 4 is a plan view of the arrangement shown in Figure 2 but with electrical screen 21 omitted for clarity of illustration; said electrical screen 21 being dimensionally similar to circuit board 20. An integrated circuit 23 is mounted on the circuit board 20 which carries symmetrical tracks 22A and 22B to provide electrical connection between the integrated circuit 23 and the terminals 13A, 13B of electrodes 12A and 12B. Figure 4 also shows a fluid passageway 24 for connecting a first pressure source to ports 16A and 17B, and a fluid passageway 25 for connecting a second pressure source to ports 17A and 16B.

It will be clear from Figures 2 and 4 that the twin-sensor arrangement enables compact, well screened connections from the electrodes 12A, 12B to the inputs of the integrated circuit 23 to be provided. The mechanical geometry of the structure is extremely rigid thereby preventing variations in stray capacitance, and the arrangement has mechanical and electrical symmetry to encourage thermal stability and capacitive balance. Although Figure 4 shows only a single integrated circuit, the circuit board 20 and the electrical screen 21 would preferably enclose substantially all the electrical circuits associated with the measuring system, thus screening the entire circuit against electromagnetic interference. As indicated above, the integrated circuit 23 and any other circuit components are preferably mounted symmetrically with respect to the two sensor parts.

In another arrangement (not shown), a second electrode may be provided in the lower portion of each sensor part, i.e. in chambers 19A and 19B. A variety of different techniques may be used to extract a signal from such arrangement. This arrangement has the potential advantage of doubling the output signal but at the expense of greatly increasing the complexity of the transducer.

A variety of electrical circuit techniques are in common use for obtaining an electrical output from a differential capacitance transducer. Most of these techniques employ the transducer capacitances in a balanced bridge where two fixed resistors form the upper arms of the bridge and supply an alternating excitation voltage to the electrodes. The diaphragm of the transducer would normally be grounded so that the two capacitors form the lower arms of the bridge.

Variations in capacitance will cause variations in the amplitude, phase and waveform of the signals appearing at the two electrodes, and any of these parameters may be used to derive an output signal. In order to avoid distortion of the electrode signals it is essential that any circuit connected to the electrodes is of extremely high input impedance, because the resistance-capacitance bridge has a high output impedance unless operated at high frequency. It is normal to buffer the bridge outputs by means of two semiconductor operational amplifiers connected as unity-gain voltage followers. Such amplifiers generally have poor high-frequency performance and therefore demand that the bridge is operated at low frequency so that the amplifiers can accurately follow the electrode waveforms.This increases the bridge impedance to a point where circuit-board leakage currents become a significant problem, and also increases the sensitivity of the circuit to externally generated electromagnetic interference. Further, since the operational amplifiers require electrical power to function, they exhibit warm-up drift which typically takes several minutes to stabilise. Other disadvantages of the resistance-capacitance bridge include the necessity to provide two extremely stable, well-matched, low-noise, noninductive, low-capacitance bridge resistors, and the fact that the output is directly affected by both the frequency and waveform of its excitation signal, thereby increasing the complexity of the signal generator in order to stabilise these parameters.The input capacitance of operational amplifiers is relatively high and appears as a fixed capacitor in parallel with each transducer capacitance, thereby reducing the proportional change of capacitance in response to pressure changes and effectively reducing the available signal swing.

Figure 4 shows a simple electrical circuit which may be used to provide an output voltage the magnitude of which is proportional to the relative magnitude of the capacitances of the two sensor parts. The circuit effectively comprises the combination of a positive-going diode pump with a negative-going diode pump with a common input signal and a common output capacitor to form a diode ring balance detector circuit.

In Figure 4, the capacitances of the two sensor ports are represented by C1 and C2. D1, D2, D3 and D4 are silicon junction diodes having matched characteristics. C3 is a capacitor which is much larger in value than C1 or C2. SI is the signal input terminal and SO is the signal output terminal. G1 and G2 are terminals attached to circuit ground. The input signal is an alternating voltage of constant amplitude, applied between SI and G1. The first negative excursion of the input voltage will cause D1 to become forward biased, charging C1 to a voltage almost equal to the peak value of the input signal.The first positive excursion of the input signal will cause D1 to become reverse biased and D2 to become forward biased, delivering the charge contained in C1 to C3 and causing a positive voltage to appear at output terminal SO. Since C3 is much greater in value than C1, the amount of charge delivered will cause only a small voltage increment across C3. C2, D3 and D4 are arranged in the same configuration as C1, D2 and D1 except that the diodes are oppositely connected. This causes the same circuit operation as already described, but with reversed electrical polarity.Because of the opposing polarity of the otherwise symmetrical circuit, if C1 and C2 are equal in value any charge delivered to C3 via D2 on positive excursions of the input signal will be removed via D3 on negative excursions and the output voltage will therefore remain at zero. If pressure is applied to the transducer causing C1 to increase and C2 to decrease in value, C1 will inject a greater incremental charge into C3 than the decrements extracted by C2. This will cause the output voltage to become progressively more positive and in so doing it will progressively increase the reverse bias on D2 while reducing the reverse bias on D3, thus increasingly opposing the increments and augmenting the decrements as the output voltage across C3 increases.At some point these bias changes will cause the voltage increments and decrements to become equal again notwithstanding the unequal values of C1 and C2, and the output voltage will stablise at this level. By this means an output voltage may be obtained the magnitude of which is proportional to the mathematical ratio of the values of C1 and C2, and the electrical polarity of which is dependent upon which of the two capacitors has the greater value. It should be particularly noted that the circuit functions as a ratio detector, the output being independent of the actual capacitive values of C1 and C2.

This ratio-detecting circuit avoids the difficulties mentioned above.

The diaphragms 13A, 13B of the twin-sensor-assembly transducer are connected together and the input signal is applied thereto. The only connections to each electrode are two low-capacitance diodes, thus maximising the signal swing. The input signal requires only to be controlled in amplitude so that the charges incremented or decremented on each cycle by the transducer capacitances C1 and C2 are controlled; variations in frequency of the input signal will have no effect other than determining the speed of response of the output to capacitance changes; variations in waveform will similarly have no effect provided only that the positive and negative peaks are sufficiently long for the small input capacitors C1 and C2 to charge fully, and for this a simple square wave is ideal.The circuit requires no electrical power other than the input signal itself, so it is economical and will exhibit virtually no warm-up drift. Because the active elements are simple diodes the circuit may be operated at a high frequency withoutloss of accuracy, thus reducing the output impedance with attendant advantages.

Furthermore, the temperature dependence of the four diodes, D1, D2, D3, D4 can be very effectively offset against one another by having all four diodes fabricated as a single integrated circuit. All diodes will then be at the same temperature because the physical size of the silicon chip within its package will typically be less than 1 millimetre square, thus ensuring a negligible thermal gradient across the chip.

Also, fabricating the diodes from a single silicon chip will ensure that characteristics such as forward voltage drop and dynamic resistance are also well matched, leading to further performance benefits. The integrated circuit 23 shown in Figure 4 may include or comprise this electrical circuit.

It will be appreciated that the twin-sensor arrangement described above retains the advantages of the symmetrical differential capacitance transducer whilst overcoming many of the problems associated therewith. In particular, by selecting matched pairs of two sensor parts from a large production batch, the precision with which they need be manufactured is not so critical and wastage of unmatched parts is substantially reduced. Both these factors reduce the manufacturing costs. The side-by-side arrangement of the two sensor parts also enables the electrical connections to the two electrodes to be provided on one side of the device and to be arranged in a compact, rigid and symmetrical fashion to minimise the effects of stray capacitance. The need for bulky and expensive electrical screening is thus reduced. A simple planar electrical screen can instead by provided over the circuit board. By positioning the twin sensors in like orientations, the effect of errors induced by forces acting on the diaphragms due to their mass, e.g. forces due to gravity or acceleration, are also substantially eliminated.

Claims (16)

1. A differential pressure transducer comprising two mechanically similar sensor parts, each sensor part comprising a first chamber and a second chamber separated by an electrically conductive diaphragm, the first chamber in each sensor part including an electrode secured in close proximity to one side of the respective diaphragm and electrically insulated therefrom, so as to form with the diaphragm an electrical capacitance, the two sensor parts being positioned in like orientation and rigidly conjoined to each other, means being provided for connecting a first fluid pressure to the first chamber of one sensor part and to the second chamber of the other sensor part and for connecting a second fluid pressure to the second chamber of the said one sensor part and to the first chamber of the said other sensor part, whereby electrical measurements of said capacitances via said electrodes can be used to determine differences between the first and second fluid pressures, said measurements being substantially immune to errors induced by forces acting on the diaphragms due to their mass because of the mechanical similarity and like orientations of the said two parts of the sensor.

2. A differential pressure transducer as claimed in Claim 1 in which the two sensor parts are formed separately and rigidly connected to each other so that the planes of the diaphragms of the respective sensor parts are substantially parallel and preferably substantially co-planar.

3. A differential pressure transducer as claimed in Claim 2 in which the two sensor parts are conjoined by means of an electrical circuit board.

4. A differential pressure transducer as claimed in Claim 3 in which the two sensor parts are further conjoined by means of an electrically conductive screen member positioned so as to screen the circuit board from external electrical interference.

5. A differential pressure transducer as claimed in any preceding claim in which each sensor part comprises first and second body portions formed of electrically conductive material with the diaphragm clamped therebetween and the electrode mounted in one of the body portions via an electrically insulating member.

6. A differential pressure transducer as claimed in Claim 5 when dependent upon Claim 3 or 4 in which the first and second body portions of each sensor part are clamped together by means of bolts passing therethrough, at least some of the said bolts also securing the electrical circuit board and/or the electrically conductive screen member to the respective sensor parts.

7. A differential pressure transducer as claimed in any preceding claim in which electrical connections to the said electrodes are provided on one side of the transducer.

8. A differential pressure transducer as claimed in Claims 3 and 7 in which the electrodes connect directly to the said electrical circuit board.

9. A differential pressure transducer as claimed in any preceding claim comprising an electrical circuit for carrying out said electrical measurement.

10. A differential pressure transducer as claimed in Claim 9 in which the electrical circuit comprises an integrated circuit symmetrically positioned with respect to the two sensor parts.

11. A differential pressure transducer as claimed in Claims 3, 9 and 10 in which the integrated circuit is mounted on the electrical circuit board and electrical connections between the electrodes and the integrated circuit board comprise symmetrical electrical tracks formed on the circuit board.

12. A differential pressure transducer as claimed in any of Claims 9 to 11 in which the electrical circuit is arranged to measure the ratio of the said capacitances.

13. A differential pressure transducer as claimed in Claim 12 in which the electrical circuit comprises a diode ring balance detector circuit.

14. A differential pressure transducer as claimed in Claim 12 in which the electrical circuit is substantially as hereinbefore described with reference to Figure 5.

15. A differential pressure transducer as claimed in any preceding claim forming part of a hand-portable unit.

16. A differential pressure transducer substantially as hereinbefore described with reference to Figures 2 to 5 of the accompanying drawing.