WO1999010714A1

WO1999010714A1 - A compensated capacitive liquid level sensor

Info

Publication number: WO1999010714A1
Application number: PCT/US1998/017477
Authority: WO
Inventors: Yishay Netzer
Original assignee: Millennium Sensors Ltd.; Netzer, Yohay
Priority date: 1997-08-25
Filing date: 1998-08-24
Publication date: 1999-03-04

Abstract

A fringing-field capacitive level sensor comprising an electrically insulating substrate, having a first and a second pattern of conductive capacitance plates plated on the substrate. Each pattern constitutes a fringing-field capacitor having a capacitance the value of which is both dependent on the liquid level and dependent on the dielectric constant of the liquid. The ratio of the two capacitances, however, is proportional to the liquid level and independent of the dielectric constant of the liquid. The fringing-field capacitive liquid level sensor may be used for measurement of liquids of substantially any dielectric constant without requiring full-scale calibration or knowledge of, and calibration for, dielectric constant of the liquid to be measured. A fringing-field inclinometer sensor application is also provided. The inclinometer uses the liquid level sensor in a circular configuration with four quadrant plates providing the angle sensing of liquid in the sensor liquid container. A two-axis fringing-field inclinometer is provided, enabling tilt angle measurements when the inclinometer is initially tilted with respect to the vertical gravity vector. Improved, side-motion-rejection, liquid-filled inclinometers include a second liquid filling the remaining space in the inclinometer; the filling of the remaining space with the second liquid providing a method of increasing the effective length of the equivalent pendulum. The second liquid is less dense, less viscous, than the first liquid, thereby not decreasing the inclinometer response time.

Description

Title: A COMPENSATED CAPACITIVE LIQUID LEVEL

SENSOR

FIELD AND BACKGROUND OF THE INVENTION The present invention relates to a capacitive liquid level sensor, and more particularly, to a compensated capacitive liquid level sensor, which is insensitive to the dielectric constant of the liquid, the level of which is to be sensed. The resulting dielectric-constant-insensitive capacitive liquid level sensor has the advantage that prior knowledge of the liquid, the level of which is to be sensed, is not required. I.e., a particular vessel and sensor and sensing electronics may be provided, and used for virtually any desired liquid.

Further, the liquid-level sensor of the present invention may also be employed as an inclinometer, measuring the tilt angle relative to gravity.

General Description

It was found by the inventor that a compensated capacitive liquid level sensor can be made, of the fringing-field type, i.e., where the capacitive plates are co-planar, and the effective "area" of the plates is essentially zero and the liquid interacts with the fringing field between the electrodes, as shown in Figure 2. This approach can be implemented on a flat substrate, such as a rigid or flexible printed circuit board, with the resulting two advantages:

1. Simple, batch, manufacturing.

2. The processing electronics can be mounted on the same board, thus saving interconnection wiring and reducing sensitivity to external interferences. 3. The sensor can be made as a rigid printed circuit board and can be snap mounted, e.g., on the fiiel pump assembly in an automotive gas tank. Another aspect of the present invention is a unique geometry for the capacitive plates which enables a measurement range extending the whole sensor length. This geometry provides two capacitances each extending the whole sensor length, and whose ratio is directly proportional to the liquid level, independent of the dielectric constant of the liquid.

Still another aspect of the invention is a signal conditioning electronics that provides a voltage linearly proportional to the liquid level employing a simple analog divider without requiring an analog-to-digital conversion or a microprocessor.

Another object of the invention is a capacitive level sensor for automotive application that is cost competitive with float-type level sensors.

Another object of the invention is a fringing-field capacitive level sensor that is relatively insensitive to contamination deposited between the electrodes of the fringing-field capacitive sensor, and relatively insensitive to liquid splashed over the non-immersed portion of the sensing plates. Such a sensor is typically useful for automotive applications.

In an inclinometer application, the tilt angle relative to gravity may be measured, providing a linear output. Further, a signal proportional to the sine of the tilt angle may be obtained.

Prior Art There are various methods for measuring the liquid level in tanks.

A survey of these methods can be found in: A Look at Level Sensing Sensors, pp. 29-34, August 1990, Vol. 7, No. 9. One well known method is based on the effect of the liquid level on capacitance plates, wherein capacitive plates interacting with the liquid are excited with an alternating voltage to generate a signal current depending on the liquid level.

The basic capacitive level sensor gives an erroneous result if the dielectric constant of the liquid is other than the design value, or is unknown. Compensated capacitive level sensors include an additional, known, reference capacitor that is totally immersed in the liquid. This reference capacitor is intended for measuring the dielectric constant of the liquid. By dividing the measured value of the measurement capacitance by that of the reference capacitor, a normalized output is obtained that is independent of the dielectric constant of the liquid. Level sensors of this type are described, for example, in US Patent No. 4,590,575, US Patent No. 4,667,646, US Patent No. 4,373,389, US Patent No. 4,296,630, and in US Patent No. 4,021,707.

Prior art capacitive liquid level sensors, whether compensated or not, can be divided into uniform-field type and fringing-field type. The first type employs either parallel plates as in Figure 1A, or concentric cylinders as in Figure IB. In these capacitors, it is the parallel (or radial) field between the plates that interacts with the liquid, wherein the fringing field at the edges is insignificant in comparison to the parallel or radial field. The advantage of these two configurations is that by controlling the separation between the plates, the capacitance per unit height can be made relatively large, providing an accordingly large signal current. In the fringing-field type of capacitive sensor the electric field that interacts with the liquid is non-uniform. In the following, the term level sensor will refer to compensated capacitive level sensors. The disadvantage of prior art level sensors, which are invariably of the uniform-field type, are the following:

1. Each of the two capacitors comprises spaced-apart plates, so the construction is relatively complicated since the plates must be isolated electrically, secured mechanically, and wired to the processing electronics.

2. Since the reference capacitor must be totally immersed in the liquid, an erroneous output is obtained for liquid levels lower than the reference capacitor height.

3. The practical requirement that the reference capacitor be located at the bottom of the container makes the measurement sensitive to contaminant liquids at the bottom of the container, since the measured dielectric constant is no longer representative of the intended liquid, for example water in a gasoline tank. Various attempts have been made to provide a capacitive liquid level sensor without the problem of the reference capacitor at the bottom of the measurement vessel. US Patent No. 4,373,389 is intended to overcome the deficiencies of a reference capacitor at the bottom of the container. In this patent there are two capacitive plates, one of them divided into two complementary triangles, resulting in two capacitors, each responding to the liquid level starting from bottom of the container to the maximum level to be measured. By manipulating the measured values of the measured two capacitances, the liquid level can be determined regardless of its dielectric constant. However, since the calculation is not simple, an analog-to- digital conversion, a microprocessor, and, possibly, a digital-to-analog conversion are necessary. German Patent Application DE 42 10737 Al publication describes an automotive fuel level sensor that employs the fringing field lines of a co-planar pair of electrodes that are applied on the outside of the tank ~ which must be non-metallic in order for the lines to penetrate into the tank. However, such sensor has a limited performance because of the following reasons: 1. Since the fuel is separated from the electrodes by the thickness of the plastic wall (e approximately 4), most of the field lines of force are shunted by the wall, and the sensor may respond unpredictably to the wall thickness, rather than to the presence of the fuel (e approximately 1.8).

2. The DE invention attempts to correct this deficiency by independently sensing the dielectric constant of the wall, however, no compensation is made for the thickness variations of the wall. 3. No compensation is made for dielectric constant variations of the fuel such as that result from additions such as of alcoholic additives.

4. The electrodes being on the outside of the tank may respond to soil deposits that adhere to the electrodes, and to fuel splashed on the non-immersed portions of the electrodes.

There is thus a widely recognized need for, and it would be highly advantageous to have, a capacitive liquid level sensor which is independent of the dielectric constant of the liquid, with a measurement range that extends the full length of the sensor plates, with a simple construction, and providing a linear output with no recourse to complex electronic manipulations, and which does not depend on the use of a fully-submerged reference sensor to achieve a reading which is independent of the dielectric constant of the liquid.

SUMMARY OF THE INVENTION According to the present invention there is provided a capacitive liquid level sensor. According to further features in preferred embodiments of the invention described below, the liquid level sensor does not require the use of a fully-immersed full-scale reference capacitor.

According to yet further features in preferred embodiments of the invention described below, the fringing-field capacitive liquid-level sensor with a relatively high sensitivity to the fuel presence that is relatively insensitive to contamination deposited between the co-planar electrodes of the fringing-field capacitive sensor, and relatively insensitive to liquid splashed on the non-immersed portions of the sensing plates. According to still further features in the described preferred embodiments, the capacitive liquid level sensor is independent of the dielectric constant of the liquid, the level of which is being measured.

The present invention successfully addresses the shortcomings of the presently known configurations by providing a capacitive liquid level sensor which is independent of dielectric constant, and uses either a uniform-field or a fringing-field. The fringing-field version of the sensor may be applied to any non-conducting vessel surface. The preferred embodiment of the sensor is an integrated device that can be applied on a single printed circuit board that includes all the processing electronics — which does not need expensive components for mathematical manipulations, is sensitive to even low dielectric constant fuel like gasoline (e approximately 1.8), and is relatively insensitive to contamination deposited on the electrodes. Further, to reduce the labor cost of the sensor, the sensor may be snap-mounted in place. The present invention discloses a novel capacitive liquid level sensor.

More specifically, the capacitive liquid level sensor of the present invention, measures the level of any liquid independent of the dielectric constant of the liquid, and its fringing-field version is insensitive to contamination deposits between the capacitive sensor measuring electrodes , and insensitive to liquid splashed on the non-immersed portions of the sensing plates.

It was found by d e inventor that a compensated capacitive liquid level sensor can be made where the capacitive plates are co-planar, i.e., the effective "area" of the plates is essentially zero and the liquid interacts with the fringing field as shown in Figure 2. This approach can be implemented on a flat substrate, such as a printed circuit board, with the resulting three advantages:

1. Simple, batch, manufacturing. 2. The processing electronics can be mounted on the same board, thus saving interconnection wiring and reducing sensitivity to external interferences. 3. If the sensor is assembled on a printed circuit board (pcb), then easy assembly by snapping the pcb onto the desired mounting location, without requiring the use of screws, etc.

Another aspect of the present invention is a unique geometry for the capacitive plates, which extends the whole measurement range. This geometry provides two capacitances whose ratio is directly proportional to the liquid level, independent of the dielectric constant of the liquid, and can also be applied to uniform-field type level sensors.

Still another aspect of the invention is a signal conditioning electronics that provides a voltage proportional to the liquid level without requiring an analog-to-digital conversion or a microprocessor.

Another object of the invention is a capacitive level sensor for automotive application that is cost competitive with float type level sensors.

Thus, there is provided, a compensated-type capacitive sensor for measuring the level of a liquid in a container comprising (a) a first sensing capacitor having a first dependence of incremental capacitance on the depth of coverage of said first sensing capacitor by the liquid;

(b) a second sensing capacitor having a second dependence of incremental capacitance on the depth of coverage of said second sensing capacitor by the liquid;

(c) a signal processor for converting said capacitances of said first and second capacitors into respective proportional first and second voltages; (d) a divider for dividing said first voltage by said second voltage, thereby providing an output signal substantially proportional to the depth of coverage of said first and second capacitors by the liquid.

Further the sensor preferably includes sensor capacitors wherein said second dependence of incremental capacitance on the depth of coverage is linear and said first dependence of incremental capacitance on the depth of coverage is constant. This is in contrast with prior art compensated liquid level sensors wherein a voltage proportional to the capacitance of a capacitor with a constant incremental capacitance is divided by a voltage proportional to the capacitance of a fully-immersed capacitor, i.e., with a zero incremental capacitance.

The sensor capacitors may be either substantially uniform-field type capacitors or substantially fringing-field type capacitors.

Further, the sensor electronics preferably includes an implicit divider.

The preferred embodiment of a sensor especially for automotive fuel tank applications further includes a grounded shield plate between said capacitor plates. Another sensor configuration includes a dielectric tube, said electrodes contained inside said tube, thereby providing a probe-type immersible liquid-level measurement probe.

Another application includes a liquid container having a dielectric wall, said dielectric wall having inner and outer surfaces, said electrodes on said outer surface of said liquid container dielectric wall.

The use of the liquid-level sensor of the present invention as an inclinometer involves providing a container partially filled with a dielectric liquid, the container having at least two identical liquid-level sensors fixed relative to the container, and situated at some distance from each other.

When the container is tilted relative to the gravity vector, the readings of the two sensors will differ by an amount that depends on their separation, and on the tilt angle of the container relative to the gravity vector. By taking the difference of the two outputs a signal proportional to the sine of the tilt angle may be obtained. To obtain an inclinometer sensor with an output signal that is linearly dependent on the inclination angle of the container relative to the gravity vector, two liquid level sensors can be wrapped around two semicircles, as will be described with respect to the preferred embodiments. The inclinometer is basically two curved liquid level sensors. Since the dielectric constant of the contained liquid is known a priori, the use of compensation is not needed. Therefore each component liquid level sensor used in the inclinometer is only required to be a simple two-electrode sensor. A capacitive inclination sensor includes

(a) a container partially filled with a liquid, said container having a non-conductive wall, said wall having inner and outer surfaces, said wall having a perpendicular axis, said perpendicular axis defining an inclination axis, said container rotatable about said perpendicular axis, said container having an angle of rotation about said perpendicular axis;

(b) at least one pair of excitation capacitive plates and at least one corresponding signal capacitive plate, forming at least one capacitor pair, one said excitation plate and one said corresponding signal plate forming a capacitor, said excitation and signal plates on said outer surface of said non- conductive wall, said excitation and signal plates symmetrically centered around said perpendicular axis, said capacitors having capacitances depending on said angle of rotation ;

(c) a capacitance measuring device for measuring capacitances of said capacitors, for providing an output signal depending on said angle of rotation. An improved inclinometer sensor further includes grounded electrode plates between each said excitation plate and corresponding said signal plate, thereby diminishing the liquid-independent portion of the mutual capacitance between said excitation and said corresponding signal plate pair. The inclinometers described above, as well as other liquid-based inclinometers, are sensitive to side motions and vibrations that shake the contained liquids (heretofore "the first liquid") and lead to spurious signals. An obvious way to diminish these effects is to use a more viscous liquid or, equivalently, to low-pass filter the output signal. However, these two methods also, undesirably, slow the response of the sensor to changes in the measured inclinations.

The above conflict can be resolved if a second liquid is used to fill the space of the container above the first liquid, resulting in a new liquid- filled inclinometer which is substantially insensitive to side motions and vibrations, while substantially retaining the initial fast inclinometer response time. The properties of the second liquid should be as follows:

1. Density slightly smaller than density of first liquid;

2. Immiscible with the first liquid (stays separate); 3. Different electrical properties.

The first and second properties are needed in order that the second liquid will float above the first liquid. The third property is needed in order that the orientation of the interface between the two liquids relative to the capacitive plates may be determined.

BRIEF DESCRIPTION OF THE DRAWINGS

The invention is herein described, by way of example only, with reference to the accompanying drawings, wherein:

FIG. 1A is a prior art cylindrical, compensated, capacitive-type, liquid level sensor; FIG. IB is a prior art parallel plate, compensated, capacitive-type, liquid level sensor;

FIG. 2 is a cross-sectional view of the electric field in a liquid level sensor with co-planar plates;

FIG. 3 A is a front view of a basic capacitive liquid level sensor with co-planar plates, and with a constant capacitance per unit length;

FIG. 3B is a front view of a first co-planar plates capacitor with capacitance per unit length that is increasing with height due to a non- uniform gap;

FIG. 3C is a front view of a second co-planar plates capacitor with capacitance per unit length that is increasing with height due to a non- uniform excitation potential;

FIG. 3D is a side view of a non-parallel planar plates capacitor with capacitance per unit length that is increasing with height, due to a variable separation between the plates; FIG. 4 A is a segmented planar capacitor with capacitance per unit length that is increasing with height, due to an increasing length of the segments;

FIG. 4B is a segmented planar capacitor with a constant capacitance per unit length;

FIG. 4C is a snap mounted single printed board automotive fuel level sensor;

FIG. 5 is a block diagram of the preferred signal-conditioning, including an implicit division scheme; FIG. 5B is a circuit diagram illustrating how a compensation signal is generated;

FIG. 5C illustrates a pair of printed circuit compensation capacitors;

FIG. 6 is a co-planar, non-compensated, capacitive liquid level sensor; FIG. 7 is a co-planar, compensated capacitive liquid level sensor;

FIG. 8 A is a side view illustrating the electric field between two electrodes such as in FIG. 2;

FIG. 8B is a side view of the electric field between two electrodes with a grounded shield between the electrodes for diminishing the effect of contamination between the plates;

FIG. 9A is a planar, multiple-segment, compensated, capacitive liquid level sensor with a grounded shield;

FIG. 9B is a parallel-plates level sensor according to the present invention; FIG. 9C illustrates a simplified version of a primary and secondary capacitance;

FIG. 9D illustrates a fringing field version of FIG. 9C;

FIG. 10A is an inclinometer sensor implementation;

FIG. 10B is an inclinometer sensor implementation with screening; FIG. IOC is an inclinometer sensor implementation with 360 degrees range;

FIG. 10D is an inclinometer sensor as in FIG. IOC. with screening and with increased sensitivity; FIG. 10E is an inclinometer implementation, using a thin tube foππed into the inclinometer shape, here a square;

FIG. 1 OF is an circular formed- ube inclinometer; FIG. 10G is a cross-section of a bent-tube inclinometer; FIG. 10H illustrates a possible electrode placement on the inclinometer of FIG 10E;

FIG. 11 is a plot of output voltage versus tilt angle for the sensot of FIG. 10A;

FIG. 12 illustrates an inclinometer container suitable for mounting the inclinometer sensors of the present invention: FIG. 13 is a probe- type liquid-level sensor;

FIG. 14A shows a two-axis inclinometer; FIG. 14B shows a modified two-axis inclinometer top view; FIG. 15 shows an improved inclinometer; FIG. 16 illustrates a tilt compensated integral liquid sensor; FIG. 17 illustrates two sidc-by-side liquid level sensors separated by a distance d;

FIG. 18 illustrates two level sensors; and FIG. 19 shows a tilt compensated liquid level sensor. FIG. 20 is a fuel tank with two fuel level sensors/two fuel lines. DESCRIPTION OF THE PREFERRED EMBODIMENTS

The present invention is of a capacitive liquid level sensor which can be used to measure the level of any liquid, without prior knowledge of the liquid to be measured and its dielectric constant.

Specifically, the present invention can be used to measure the liquid level of gasoline in an automobile, even if there is sludge or water at the bottom of the tank, or contamination between the plates, which would provide errors in the prior art implementations.

I. Theory of Operation

The invention is based on employing a primary capacitor C,, and a secondary capacitor C₂, each extending from bottom to top of the desired measurement range.

The capacitance dC_x per unit height dh, incremental capacitance, of the primary capacitor, excluding its "dry" component is given by dC_{1 =} f(h) dh. When immersed to a depth h in a liquid with a dielectric constant

€ the integrated capacitance over the entire depth h would be:

Cl = k_x j f(h) € dh where k_x is a constant which depends on the geometry, and e is the dielectric constant of the liquid, j is used herein as the integration symbol.

The capacitance per unit height of the secondary capacitor, again excluding its dry component, is given by dC₂ = g(h) dh and the integrated capacitance over the entire depth h would be:

C₂ = k₂ J g(h) e dh where k₂ is another geometry dependent constant. Our aim is to find two functions f(h) and g(h) such that the ratio of the two integrals is: / C₂ = r(h) i.e. , independent of e and dependent on h in accordance with a prescribed function r(h), e.g., such as may be needed to provide an output signal proportional to the volume of a fuel contained in an automotive fuel tank which has a non uniform cross-section. In the special case when the output signal should be linearly proportional to the liquid level, our aim is a pair of functions, f(h) = k_j h, i.e., the capacitance per unit length is linearly increasing with h, and g(h) = k₂, where the capacitance per unit length is uniform. Performing the integrations for this pair of functions, f(h) and g(h), results respectively in:

and C₂ = k₂ e h. Dividing,

C_x / C₂ = ( Vz k_x € h² ) / ( k₂ e h ) = V2 ( k_t / k₂ ) h. Thus the ratio is linearly proportional to the immersion depth and is independent of e. This pair of functions is a member of an infinite set of pairs of functions that maintain the above ratio, such as the pair of functions f(h) = (k₃ h²) and g(h) = (k₄ h), etc. , however, the function pair of the example integrated and divided above is preferable in being easier to realize, as shown below.

The principles and operation of a dielectric-constant-insensitive capacitive liquid level sensor according to the present invention may be better understood with reference to the drawings and the accompanying description.

II. Preferred Embodiments

Referring now to the drawings, a first implementation of the co- planar primary capacitor with the dependence f(h) is shown in Figure 3B, where the plates are separated by a gap that is substantially inversely proportional to the height h, as a result the capacitance per unit height is linearly proportional to the height.

A second implementation of the primary capacitor is shown in Figure 3C. In this implementation the separation between the two plates is constant, however, one plate is printed using a conductive ink with a relatively high resistivity such as carbon ink. The bottom of this plate is grounded and the excitation is applied to its top, the plate thus serves as a distributed voltage divider, and the excitation voltage applied to each elemental capacitance is proportional to its height, the capacitive current thus behaves similarly to the previous case.

A third implementation of the primary capacitor is shown in Figure 3D. This configuration is similar to the one in Figure 3B in that the separation between the two plates is decreasing with height, however, the two electrodes are not co-planar.

The secondary capacitor should have a constant capacitance per unit height, this could be achieved, for a co-planar sensor, for example, by a set of two or more vertical bands as shown in Figure 3A. A preferred implementation of a primary capacitor is shown in

Figure 4A. The capacitor comprises multiple elements based on discrete printed bands. This is in order to obtain higher capacitance per unit area, resulting in a larger signal level. If the small end of the sensor of Figure 4 A is mounted toward the bottom of the measurement container, then the length of each consecutive band is linearly increasing toward the top, thereby approximating a capacitance per unit height that is increasing with height. I.e., the incremental sensitivity of this sensor electrode pair to liquid height increases with the height of the measured liquid. This provides a non-linear function of sensor sensitivity versus liquid height. In order for the sensitivity increase to be a continuous function, rather than stair-case function, the printed bands are slanted, such that each band is gradually immersed as the liquid level rises.

A preferred embodiment of the secondary co-planar capacitor is shown in Figure 4B. This electrode pair geometry employs the same technique as in Figure 4A, except that the bands are of constant length in order to obtain constant capacitance per unit height. I.e. , the incremental sensitivity of this sensor electrode pair to liquid height is constant with the height of the measured liquid. This provides a linearly responding sensor function. The two electrode pairs of figures 4A and 4B are normally mounted in proximity to each other either on a one side of a printed circuit board, or on the opposing sides of a printed circuit board, to provide a standalone sensor. The electrodes may also be deposited on the inside or on the outside of a dielectric wall of a liquid container. The printed circuit version may also contain the measurement instrumentation electronics to be described below with reference to Figure 5.

In the preferred embodiment for automotive fuel tank level sensor, the electrode pairs in Figures 4A and 4B are printed on the opposite sides of a printed circuit, where the components of the processing electronics are mounted on the top end of the printed circuit. To further reduce the assembly cost of the fuel sensor, the fuel sensor assembly of Figure 4C may be snap-mounted in place, using mounting snaps, as shown in Figure 4C. The two electrode pairs can as well be printed side by side on the same side of a printed circuit. If the printed circuit is flexible, the printed circuit may be wrapped around a cylinder which can preferably be put inside a sealed insulating cylinder such as glass — see Figure 13 - to provide a general purpose level measuring probe. The wall thickness of the cylinder would shunt a portion of the sensing field lines and decrease the sensitivity of the probe. Therefore, the probe is more suitable for aqueous liquids, since the dielectric constant of water is relatively high (e approximately 80).

The division operation needed for obtaining the liquid level can usually be implemented digitally or by using an analog divider, however these two methods are relatively costly. The preferred embodiment of the invention employs an implied-division method which is used, for example, in Signal Conditioning Stops Transducer Errors, EDN, February 18, 1981, pp. 147-150, authored by the applicant. The implementation of the implied division method is shown schematically in Figure 5. The implied division method employs an amplitude controlled oscillator 51, of frequency of typically 10kHz, that excites the primary and secondary capacitors. The current in each capacitor comprises a portion dependent on the liquid level and a portion that depends on the "dry" capacitance, i.e., the capacitance value in the absence of liquid. The currents in the primary and secondary capacitors are converted to voltages by means of charge amplifiers 52 and 53, respectively, and rectified by rectifiers 54 and 55, respectively. The rectifiers can be either diode rectifiers or synchronous rectifiers. The "dry" components of the input voltages to the rectifiers, 54 and 55, are cancelled by subtracting a portion of the excitation voltage via Rl and R2 in Figure 5, for example. This operation constitutes the subtraction of an error component which depends on the excitation.

The source of the "dry" capacitance is interaction between the sensor plates through the air and through the printed substrate. The first constituent is temperature independent, the second one varies in temperature depending on the substrate material. As a result, a fixed compensation as described will be effective only over a narrow temperature range. An effective way for wide temperature variations is to generate a compensation signal that behaves exactly as the actual dry capacitance by deriving it from a capacitance that behaves exactly as the dry capacitance as shown in Figure 5-b. Excitation source 1 is applied to the primary and secondary sensing capacitances Cs_x and Cs₂ that feed charge amplifiers Uld and Ulc. Phase inverted versions of the excitation sources are applied to compensation capacitors Cc_x and Cc₂ through voltage inverters A_t and A₂. The gains of the two voltage inverters are accordingly R₂/Rχ and R₄/R₃ and if the "dry" components of the two sensing capacitors are Csά^ and Csd₂, then in order that the two outputs SI and S2 will be temperature compensated, the following relation should hold:

and

Csd₂ = Cc₂ R₄/R₃

In order that the compensation capacitances will track the dry capacitances of the sensing plates,, they are printed on the same board and are therefore equally dependent on its temperature. However, they should be protected from contact with liquid. Figure 5-c illustrates a pair of printed compensation capacitors Ccl and Cc2. In a typical situation, the dimensions of each are about one square centimeter providing a capacitance of about 4 picofarads . In the preferred embodiment, a metallic cover is soldered to the substrate, not shown, to provide this protection. The same cover is preferably used also for protecting other electronic components from contact with the liquid which is critical in automotive fuel level sensing. It should be noted that the above described sensors mainly with reference to printed circuit board technology may be advantageously manufactured by making the board from injection molded plastic material, hot stamping, where a conductive coating is selectively adhered to the substrate, can preferably be used in making the coating. The advantage in this approach is that other functions, such as mechanical anchoring, can economically be included. Alternatively, the sensing electrodes can be applied to plastic parts that serve different functions, such as liquid containers, pumps, etc.

The same concept of an integral level and tilt sensor described in reference top a fringing-field sensing capacitances can be applied to uniform-field sensing capacitors. One of the advantages of the latter is that it is possible to work with higher viscosity liquids, since thick liquids may, by wetting the plates of a fringing-field sensor, impair its accuracy.

The outputs of rectifiers 54 and 55 are then smoothed by low-pass filters 56 and 57, respectively. The implied division method is based on subtracting a fixed reference voltage V_ref from the output of low-pass filter 56 and integrating the difference with integrator 58 the output of which controls the output amplitude of excitation oscillator 51. Due to the loop gain, the DC voltage at the output of low-pass filter 56 is forced to equal V_ref. If the amplitude of oscillator 51 is V_enc and the combined gain of blocks 52, 54, and 56 is A₂, and the capacitor C₂ = e C₂ ^', where e is the dielectric constant of the liquid then:

* en_c ^f ₂ A₂ = V_ref.

Similarly, if the combined gain of blocks 53, 55, and 57 is A_x and C_x = € C then:

Substituting V_enc from the first equation into the second equation, we obtain:

V₀ = ( A,/A₂ ) ( C//C/ ) ( V_ref ). I.e., V is proportional to the capacitance ratio, and independent of €.

A parallel-plates level sensor according to the present invention is shown in Figure 9-b. This sensor comprises three planar substrates, 100, 200, and 300, which are preferably printed circuit boards. In a typical embodiment of the invention, an electrode plate 201 on a first side of the central substrate, 200, serves as the signal plate, interacting capacitively with an excitation plate, 101, deposited on the substrate, 100; while a second electrode plate 202 on the second side of central substrate 200 interacts with electrode 301 on the facing side of substrate 300. It is evident that electrode plates 101 and 301 could be side by side on a common substrate and facing a common electrode plate on a second substrate whereby the sensor would comprise only two substrates which are wider.

The functions of primary and secondary capacitors fulfilled by the separate triangular and rectangular plates in Figure 9-b can be achieved as in Figure 9-c, i.e., more compactly and using plates only on a single side of the substrate. More specifically, the function of the triangular plate is served by plate 400 and the function of the rectangular plate is served by the sum of the two plates 400 and 401. Uniform-field compensated capacitive sensor with only two printed boards facing each other can thereby be obtained, wherein plates 400 and 401 typically serve as the receiving plates deposited on a first printed board and a second opposing board (not shown) includes the rectangular excitation plate. If the measured capacitance of plate 400 is Cl and that of plate 401 is C2, then the liquid level is lineaτly proportional to C1/(C1 + C2). This construction differs from U.S. Patent 4.373.389 described previously in the following major features:

1. The output level extends to the top of the sensor to its bottom, as opposed to just half of that in 4- ,373,389. 2. The output is linearly proportional to the liquid level, as opposed to parobolic dependence in 4,373,389.

3. Simple division, as opposed to complex mathematical relationships in 4,373,389.

Similarly, in a fringing-field level sensor, triangular plate in Figure 4-a and its rectangular complement in Figure 4-b, or the triaugulai plate. in Figure 9-a and its complement can be replaced with a single pattern shown in Figure 9-d, wherein plate 402 is the equivalent of plate 400 in Figure 9-c, plate 403 is the equivalent of plate 401 in Figure 9-c and plate 404 is the common excitation plate. Planar substrate and electrodes can also be used for implementing a regular non-compeαsated liquid level sensor as in Figure 6, where the measurement capacitor comprises a first excitation set of electrodes, 61, and a second signal set of electrodes, 62.

Such a planar compensated liquid level sensor is shown in Figure 7, where in addition to the measurement capacitor, there is also a reference capacitor at the bottom end of the substrate. The reference capacitor includes an excitation electrode, 71, and a signal electrode, 72. The signal electrical conductor, 73, connecting the signal electrode, 72, to the amplifier (not shown) is protected by two grounded stripes, 74, that serve as a screen (electrical shield). The screen, 74, is intended to reduce the parasitic capacitance of signal conductor, 73, to the excitation conductors, 75. The excitation conductors, 75, serve as excitation plates of the desired sensitive measurement capacitor region, in conjunction with stripes 76. As already mentioned, the division of the electrodes into multiple bands serves to increase the capacitance per unit area for obtaining a larger signal level. However, the closer the bands, the closer the lines-of-force to the substrate, and the smaller is the sensitivity volume of the sensor, i.e. , the volume of the liquid that is affecting the capacitance. In the limit the sensitivity volume is limited to the surface of the substrate. In practical situations, such as in fuel tanks, the liquid may contain contaminants that generate solid deposit coating the surface of the substrate. This solid deposit coating on the surface of the substrate would affect the reading, the liquid level height measurement indication, in proportion to the volume of the solid deposit coating relative to the sensitivity volume of the sensor. The same applies to liquid splashed on the plates.

The reason for this sensitivity is that the electric field is strongest in the gap region co-planar with the electrodes where contamination deposits.

Another object of the screen is to reduce errors due to contaminants deposited between the plates as elaborated later.

Figure 8A is a side view illustrating the electric field between two electrodes such as in FIG. 2, which would be sensitive to surface contamination between the electrodes and splashed liquid. In order to reduce the effect of such deposits of contamination between the electrodes and splashed liquid, a shield electrode is added between the electrodes, as in Figure 8B, which shows a side view of the modified sensor, 82, and the electric field, 83, including lines of force, 5 831 and 832, in the region between two electrodes, 821 and 822, with a grounded shield, 823, between the two electrodes. The result of the shield, 823, is that most of the lines of force, 831, co-planar with the electrodes, 821 and 822, that could interact with the deposit and splashed liquid are shunted to ground, and the only lines of force, 832, coupling the 0 electrodes, 821 and 822, are essentially the lines of force, 832, that pass through the liquid volume.

Figure 9 illustrates a modified printed sensor array that incorporates the shield electrode.

It should be mentioned that all of the planar (fringing field) 5 electrodes capacitive liquid level sensors described can be used where the substrate is the container wall itself in situations where the walls are non conductive, such as glass or plastic. Moreover, since the electrodes can also interact with the liquid at a distance, the electrodes can be deposited on the outside of the container, as in German patent DE 42 10737 Al.

0 The methods discussed above may also be applied to the measurement of tilt angle relative to the gravity vector. This may be accomplished by employing a container partially filled with a dielectric liquid, the container having two identical liquid-level sensors fixed relative to the container, and situated at some distance from each other. When the 5 container is tilted relative to the gravity vector, the readings of the two sensors will differ by an amount that depends on their separation and on the tilt angle of the container relative to the gravity vector. By taking the difference of the two outputs a signal proportional to the sine of the tilt angle may be obtained. To obtain an inclinometer with an output signal that is linearly dependent on the inclination angle of the container relative to the gravity vector, two liquid level sensors can be wrapped around two semicircles. In Figure 10A, two semicircular conductive plates, 101 and 102, share a common circular plate, 104. Semicircular plates, 101 and 102, are separated by axis of symmetry 103, which is normally vertical. Preferably the two semicircular plates are excited by two complementary alternating voltages, i.e., have 180-degrees out-of-phase excitation, and the circular plate serves as a common receiving plate. An advantage of this configuration is that the "dry" component of the capacitances between the common and semicircular plates is nominally equal and is cancelled out. This configuration theoretically responds to tilt angles of the axis of symmetry from zero to +/- 90 degrees relative to the vertical axis. The response of the sensor of Figure 10A is shown in Figure 11. The output voltage is ambiguous for tilt angles greater than +90 degrees and less than -90 degrees.

In the preferred embodiment of the inclinometer sensor, 1200, of the present invention, shown in Figure 12, the container, 1201, is a short cylinder of circular cross-section, oriented such that at least one flat surface of the cylinder, 1202, is in a vertical plane, i.e., parallel to the gravitational lines of force, 1204. The at least one vertically-oriented flat surface is made of a non-conducting material, on which the capacitive plates of sensor, 1203, may be coated.

Similarly, if the flat face of the container is relatively thin, the capacitive plates can be coated on the outer surface of the container, and still respond to the dielectric liquid, thus simplifying the construction by eliminating the requirement of bringing the electrical conductors into the sealed sensor container. For example, the housing may be made from a plastic material and sealed by ultrasonic welding. To further simplify the construction of the inclinometer, the plates do not even have to be coated on the surface of the container, but may be produced on the surface of a printed circuit board which also contains the processing electronics. The printed circuit board is then attached to the face of the container. If the capacitive plates are external to the container, then the active capacitance due to the liquid can be much smaller than that due to the container wall dielectric material, which container wall material is in closer proximity to the plates than the liquid.

Although, as mentioned above, the "dry" component is essentially self-cancelling, an error can result. I.e. , because of an uneven attachment of the printed circuit to the container, therefore, it is advantageous to reduce the dry capacitances by the screening method of Figure 8. This is accomplished by the addition in Figure 10B, of the grounded ring, 105, to the sensor of Figure 10A. The ground ring, 105, essentially increases the proportion of the sensor capacitance which is due to the liquid, compared with that portion of the sensor capacitance which is due to the container wall.

It should be noted that a liquid with a high dielectric constant can be selected, such as water (e approximately 80) or alcohol (e approximately 25) in order to increase the signal level. Alternatively, a conductive liquid such as an electrolytic solution of salt in water or salt in alcohol can be used, as explained below.

There are applications in which the effective range of the inclinometer sensor is required to be the full 360 degrees, without ambiguity. To provide an inclinometer sensor implementation with 360 degrees range, the individual semicircular plates of Figure 10A are subdivided as in Figure IOC. If plates 106 and 109 are commonly excited in opposition to plates 107 and 108, then the operation of the sensor of Figure 10C will be identical to the operation of the sensor of Figure 10 A. Similarly, operation of the sensor of Figure 10C with plates 106 and 107 in opposition to plates 108 and 109, is equivalent to rotating the sensor of Figure 10A by 90 degrees. These two modes of operation may be achieved simultaneously, for example, by exciting common plate 110, and measuring the currents in each of the quadrant plates, 106-109. If the measured four capacitances A, B, C, and D, are converted to a first output voltage, V_1? proportional to (A+C)-(B+D), an equivalent to the sensor of Figure 10 A is obtained, while generating a second output voltage, V₂, proportional to (A+B)-(C+D) is equivalent to rotating the sensor by 90 degrees.

Similarly, a single pair of complementary excitation voltages can be applied to joined electrodes 106, 109; and joined electrodes 107, 108; to obtain an output voltage versus tilt angle identical as shown in Figure 11. Alternatively, a single pair of complementary excitation voltages can be applied to joined electrodes pair 106, 107; and joined electrodes 109, 108; to obtain a complementary output voltage versus tilt angle, shifted by 90- degrees relative to the output voltage versus tilt angle of Figure 11.

Alternatingly switching the electrode connections between the two given interconnection cases, provides the two output voltages versus tilt angle that together define an unambiguous reading over the full 360 degrees. It should be noted that the planar construction of the electrodes and the apparently zero width of the air-gap result in a relatively small tilt- dependent capacitance.

In order to increase the sensitivity of the sensor, several concentric capacitive plates may be employed as illustrated in Figure 10D, where all of the quadrant plates in each given quadrant are connected in parallel.

The inclinometers described above can equally well use a container made of a bent long closed tube having small cross-sectional area, preferably circular, containing a small liquid volume. The long thin tube may be formed into a substantially square shaped inclinometer, 1050, as in Figure 10E, or into a ring shaped inclinometer, 1060, as in Figure 10F. The feature, 1052, at the top of inclinometer, 1050, provides both the filling port for inserting liquid into dielectric tube, 1051, which may for example be glass, and also provides an expansion volume for the case of a two-liquid inclinometer, as is described below, with reference to Figure 5 15. Figure 10G shows a cross-section, through A-A of Figure 10E, of the round tube which is formed into a square in Figure 10E. The two regions, 1071 and 1072, represent capacitor plates on the outside of thin cross-section tube, 1070. The capacitors include one common plate and two, or four, excitation plates electrically equivalent to Figures 10 A and

10 10C, with, or without, a shield electrode. The capacitor plates may be coated along the outside of the tube diametrically opposing ~ when looking at the cross-section of the tube, with the liquid between them, as with cross-section A-A, shown in Figure 10G. The plates 1071 and 1072 are either on the front and rear surfaces of the shape formed by the narrow

15 cross-section tube, or are on the central and outer surfaces of the shape formed by the narrow cross-section tube. In Figure 10H, the inclinometer shape of Figure 10E is drawn, including the features of Figure 10E, except the cross-section B-B, when represented in Figure lOg, is rotated ninety degrees from section A-A. In Figure 10H, the common capacitor

20 plate, 1053, is shown on the central surface of the shape formed by tube 1051, and the four quadrant-electrodes, 1054 - 1057, are located on the outer surface of the shape formed by tube 1051. If the tube, 1051, is bent as a rectangle, as in Figures 10E and 10H, then the output voltage will be proportional to the sine of the tilt angle. The advantage of this

25 embodiment, using the formed thin tube, is that the space available inside the circle, or rectangle is available for packaging other components, when needed, without impairing the performance of the inclinometer. For example, if the inclinometer, 1050, is attached to a vertically-mounted printed circuit board, the board area from the center of the Figure 10H

30 extending to near-to common electrode, 1053, is available for construction of the accompanying electronic instrumentation. Further, the use of the small-diameter tube bent into an inclinometer shape results in a thin inclinometer, which occupies a very small volume in an instrument cluster. A two-axis inclinometer is shown in Figure 14A. This inclinometer comprises a semi-spherical container, 140, partially filled with a liquid, on which container, preferably on the outer surface of the container, four capacitors are plated, in the configuration of two perpendicularly-oriented pairs of diametrically-opposite capacitor pairs, the first pair of capacitors comprising plate-pair 141-142 and opposing plate-pair 143-144; and the second pair of diametrically-opposite capacitors comprising plate-pair 145- 146 and opposing plate-pair 147-148. The differential value of capacitance between plates 141 and 142, and the capacitance between plates 143 and 144, is related to the tilt in one axis, while the differential value of the capacitance between plates 145 and 146, and the capacitance between plates 147 and 148, is related to the tilt on the cross axis. If the incremental capacitance of the four capacitors is constant, then the respective differential values are linearly proportional to the two tilt angles in the respective axes, the axes being the respective perpendicular diameters of the semispherical container. In practice , three capacitors , e . g . , spaced apart at 120-degrees could be used, at the expense of more complicated calculation of the tilt angles.

A modified two-axis inclinometer top view is shown in FIG. 14B, wherein the capacitor plates are planar, and do not conform to the container walls. This implementation is preferable, since the plates may now be made on a printed circuit board, which serves as the carrier of the electronic components as well, The printed circuit board is located below the "south pole", 149, of the illustrated "southern-hemisphere", semispherical container, 140. The center of the printed-circuit-board electrode pattern is located at the "south pole", 149, of the hemisphere, 140, resulting in the electrodes being located below perpendicular diameters, Dl and D2, of the equator of the hemisphere, as shown in FIG. 14B. However, there is now a variable air gap between the capacitor plates and the liquid, which affects the dependence of the differential capacitances on the tilt angles. By properly shaping the gaps between the capacitor plates as in Figure 3B, or using resistive plates as in Figure 3C, the incremental capacitances of the four capacitors can be made to provide a linearly proportional, or other desired dependence of the outputs on the tilt angles.

It should be mentioned that the above implementations of liquid level sensors and inclinometers described in connection with dielectric liquids work equally well when the liquid is electrically conductive such as saline water. If the liquid is separated from the sensor plates by an insulating layer such as a "solder mask", used in printed circuit technology, or by a dielectric container wall when the plates are external to the container, then the impedance between the capacitor plates is the series connection of the liquid resistance and the capacitive reactance of the insulating layer between the plates and the liquid. At operating frequencies in the kiloHertz range the measured impedance is essentially that of the insulating layer capacitance, since the resistive impedance would, typically, be much smaller than the capacitive impedance. For conductive liquids, the measured capacitance depends on the dielectric constant of the insulation and on the insulation thickness, both of which are essentially constant, as well as on the immersed area of the plates, which is the variable to be measured, to provide inclination angle data.

Mathematically operating on the resulting differential capacitance measurements enables compensation for the initial non- vertical orientation of the inclination sensor itself.

The inclinometer capacitors may also include the improvements described above for fringing field capacitors construction. The inclinometers described above, as well as other liquid-based inclinometers, are sensitive to side motions and vibrations that shake the contained liquids (heretofore "the first liquid") and lead to spurious signals. An obvious way to diminish these effects is to use a more viscous liquid or, equivalently, to low-pass filter the output signal. However, these two methods also, undesirably, slow the response of the sensor to changes in the measured inclinations.

A typical application where this sensitivity is critical, is in virtual- reality head-mounted displays that include a head-orientation measuring sensors for controlling the displayed image in accordance with the head orientation. Inclinometers, in conjunction with an electronic compass, have been tried for measuring the head orientation, however, errors due to lateral head motion have caused an image jitter which leads to disorientation and vertigo. The above conflict can be resolved if a second liquid is used to fill the space of the container above the first liquid, resulting in a new liquid- filled inclinometer which is substantially insensitive to side motions and vibrations, while substantially retaining the initial fast inclinometer response time. The properties of the second liquid should be as follows: 1. Density slightly smaller than density of first liquid;

2. Immiscible with the first liquid (stays separate);

3. Different electrical properties.

The first and second properties are needed in order that the second liquid will float above the first liquid. The two liquids do not have to be chemically pure; each liquid may be a solution of more than one chemical. The third property is needed in order that the orientation of the interface between the two liquids relative to the capacitive plates may be determined. Without the third property, for example, if the dielectric constants of the two liquids were the same, or if the two liquids were electrically conductive, then from the standpoint of the capacitive electrodes, the container would be filled with an homogeneous liquid. However, one liquid may be a dielectric liquid, and the other liquid may be a conductive liquid, meeting the third requirement, "different electrical properties". In practice, any other division of the volume between the two liquids can be used, at the expense of more complicated calculation of the tilt angles.

To understand the role of the second liquid in reducing the effect of side motions, we shall assume a single axis inclinometer with a cylindrical container half filled with a first liquid. It is obvious that as long as the volume of the liquid maintains its semicircular shape, the liquid behaves similarly to a pendulum with a solid semi-circular rotor that is pivoted on the axis of symmetry of the cylinder. The pendulum has a certain natural frequency, f, that is dependent on the moment of inertia, I, of the rotor relative to the rotation axis, and dependent on the distance, r, between the rotation axis and the center of mass of the rotor. The relationship is the following: f = 1 / [ 2 * pi * sqrt (gr/T) ], where g is the gravity constant. I is known to be dependent on the square of the radius of the container and on the mass, m, of the liquid. It can also be shown that for a circular container of radius, R, the distance r is approximately equal to R/3. It is further known by those versed in the art that the above pendulum, often referred to as a "compound" pendulum, has a "mathematical" equivalent pendulum having the same natural frequency with the same mass, m, concentrated in a single point separated from the pivot axis by a distance, L, referred to as the equivalent length, where:

L = I / mr. The addition of the second liquid has two effects, both effects increasing the length of the equivalent pendulum compared with the original pendulum with a single liquid: Addition of the second liquid, 1. roughly doubles the moment of inertia, I.

2. shortens r, by shifting the center of mass towards the rotation axis.

U.S. Patents Nos. 4,451,991, and 4,517,750, both granted to the

Applicant, describe single-axis and dual-axis passive pendulums with relatively high immunity to side vibrations, the pendulums having effective lengths which are much longer than the physical lengths of the pendulums, however, the pendulums are relatively complicated. These two patents teach that the longer the effective length of the pendulum, the less the pendulum is affected by side motions above the natural frequency of the pendulum. Therefore, the addition of the second liquid to the liquid-filled inclinometer tends to decrease the sensitivity of the liquid-filled inclinometer to side motion. It should be stressed that the response of the inclinometer to changes in inclination of the inclinometer housing relative to gravity should not be affected by the addition of the second liquid. The response of the inclinometer to changes in inclination of the inclinometer housing relative to gravity will be essentially the sensitivity of the inclinometer to changes in inclination of the inclinometer housing relative to gravity with only the first liquid, if both first and second liquids have roughly the same viscosity, and if this viscosity is low, then the response will be essentially instantaneous. A typical example of two liquids that may be used is (a) water, which has a specific gravity of 1 gm/(cm**3), and a dielectric constant of 80; and, (b) Dow Corning DC-200 type liquid with a viscosity of 2 centistoke, which has a density of around 0.873 gm/(cm**3), and a dielectric constant of 2.45. It should be noted that this method of providing a second liquid filling the space of the container above the first liquid in a liquid-filled inclinometer, can be applied to any kind of liquid filled inclinometer, both of the single-axis and two-axis types. FIG. 15 is a representative drawing of an inclinometer for discussion for the case of the inclinometer filled with two liquids. With reference to Figure 15, an end view of a horizontally-oriented cylindrical liquid-filled inclinometer is shown. The sensor electrodes would be mounted on a flat end surface, as in the previous figures, and are not shown here, in order to avoid confusion in the drawing. The cylinder extends back into the paper. The cylinder has a radius, R, and a horizontal center line passing through point, C, and extending into the paper, about which the inclinometer container body, 1520, is free to pivot. The gravity axis is along the vertical center-line, G. The inclinometer is approximately one-half filled with first liquid, 1551, filling the lower half of the inclinometer body cylinder volume, Region A. When only first liquid, 1551, is contained in the inclinometer body, then the center of gravity is located at r, about R/3 below the center, C, of the cylinder. The volume, Region B, of the upper half of the cylinder is empty. This, so far, describes the "status-quo" case of a half-filled liquid inclinometer, which will be sensitive to side-motions and vibrations. With the addition, now, of a second liquid, 1552, meeting the requirements discussed above, filling the volume of Region B, the sensitivity to side-motions and vibrations is substantially eliminated, as taught above, with the re-location of the center of gravity, now, to a point r' , approximately coincident with, or slightly below, the center line of the cylinder which passes through point C.

As the difference between the densities of the two liquids is made smaller, the center of gravity of the liquid mass is closer to the pivot axis, and the restoring force of gravity against disturbance moment is smaller. Such moment can be generated by side accelerations and by rotational movement coupled to the liquid from the container walls through viscosity. To diminish these disturbances, the diameter of the sensor should be the maximum practical, so as to increase the rotational inertia of the liquid mass. Secondly, the viscosity of the liquids should be at minimum, and thirdly, (in the cylindrical single-axis version) since the viscosity coupled rotation is proportional to the radius and is at minimum close to the rotation axis, the sensing plates should extend, typically, only about half the radius. Figure 15 includes also a bellows, 1525, in bellows housing, 1524, located at the top of the inclinometer container, 1520. The purpose of the bellows arrangement, 1525, is to accommodate expansion of the liquids in the inclinometer, for example, due to expansion when the inclinometer is located in a high ambient temperature environment, when the total cylinder volume is filled by the first and second liquids, thus preventing rupturing of the inclinometer body, 1520.

The closed tube configuration discussed above is especially advantageous with respect to the two liquid inclinometer, since the rotational inertia of the liquids is mainly dependent on the volume of the liquid radially away from the rotation axis. The loss in inertia is therefore small compared to the space freed; also, since the volume of the liquid is reduced, the compensation bellows which allows for thermal expansion of the enclosed liquids in a two-liquids filled inclinometer can be smaller.

A typical application of inclinometers could be the compensation of errors such as those induced in automotive fuel level sensors as a result of an inclination of the car and which result in either increasing or decreasing reading depending on the location of the level sensor relative to the center of the fuel tank. For a box shaped tank and a mild inclination angle, the error would be proportional to LSin θ where θ is the angle and L is the separation between the sensor and the center of the tank. In more complex fuel tanks, such as plastic blown tanks where the cross section of the tank is not constant, die error in the measured amount of fuel will also depend not only on the inclination angle, but also on the actual amount of fuel. Therefore, by measuring the tilt angle and knowing the geometry of the tank, the error can be calculated and compensated. Heretofore, this could not be made economically because of the excessive cost of conventional inclinometers, since in automotive applications, the inclination that is the most importance is the pitch (fore/aft) axis, a single-axis inclinometer would, in practice, be sufficient. Furthermore, since both the fuel level and the inclinometer, which are the subjects of the present invention are adaptable to printed circuit board technology, they can share a common boaτd and thus minimize the total cost. The board could be based on printed circuit technology but in some embodiments described hereinafter, it may not be necessarily fiat and be molded from a plastic material, in fact, it does not have to be rigid and may be printed on a flexible sheet, The only requirement is that the board be normal to the pitch axis. Since it is Sin θ that is needed, an inclinometer wiui the sinusoidal output described hereinabove is preferred.

The board mounted inclinometer could be of die fringing field type as described hereinabove where a sealed liquid container is pressed against the printed circuit board. Alternatively, the printed circuit substrate can serve as part of the container, the rest of which is made of deep-drawn metal sheet which is soldered to the circuit, as described hereinbelow.

Figure 16 illustrates a combination of level and inclination sensors on a common substrate 1, Container 2 is typically metal sheet deep-drawn and soldered to the substrate, thus generating an enclosure that is substantially half filled with a dielectric or conductive liquid that affects the capacitive interaction of planar plates printed on substrate 1 to generate a signal proportional to the Sine of the inclination aagle similar to Figure 10-e. Printed plates 3 and 4 are schematic representations of the level sensor.

In an alternative embodiment of the inclinometer, metal container 2 can be electrically connected to the electronic circuitry and serves as a signal plate in which currents are induced by two plates printed on the substrate and separated from the signal plate by the liquid. In both cases, the liquid can be filled into the container through a hole in the printed circuit board, which is later sealed by soldering, r through an opening in the metal cover that is later sealed. However, an inclinometer based on a liquid container is not necessary for tilt compensation since, as mentioned hereinabove, merely taking the difference between the readings of two spaced apart identical level sensors provides a signal proportional to Sin θ regardless of the actual average level of the liquid, the output will, however, be dependent on the dielectric constant of the liquid unless the level sensors are compensated. Therefore, the vehicle inclination can be obtained without any separate inclinometer by employing two level sensors in combination with the fuel itself. Figure 17 illustrates two side-by-side liquid level sensors separated by a distance d, their outputs V_x and V₂ will differ in proportion to dSin θ as a result of inclination angle θ.

Figure 18 illustrates an embodiment of this concept in which two level sensors, which are dielectric constant compensated as in Figure 9-d, are printed on the same substrate, the difference Vj - V₂ is proportional to Sin θ and the sum V_x + V₂ is proportional to the uncompensated liquid level. Knowing L and d enables compensation of the inclination error.

In another embodiment shown in Figure 19, a tilt compensated liquid level sensor is shown in which there are, in addition to the main one, two auxiliary level sensors, for obtaining tilt information, separated by a distance d. Each of the auxiliary level sensors include a signal plate (S3 and S4), and an excitation plate. They are therefore not compensated against variations in the dielectric constant of the liquid and their output signals are:

£ k h_j and e k h₂ where k is a constant hi and I12 are the respective liquid levels and e is the dielectric constant of the liquid. The Sine of the tilt angle is the difference k (h^-l^) between the two outputs divided by d:

Sinθ = € k (h_rh₂)/d This expression includes the unknown e, which can be eliminated as follows: The average liquid level obtained from the auxiliary level sensors is h =€ k (h_j+h2) 2, but since the output h_m of the main level sensor (which is independent of e) is essentially the same as (h_j+li2)/2, therefore dividing the two provides: h_a/h_m =6 k and

Sinθ=h_a/h_m (h_rh₂)/d

Generally, the fuel level sensor is attached to the fuel pump assembly which is situated near the center of the tank. In situations where the fuel level is low and the vehicle inclined, the fuel level sensor may not be immersed in the fuel and will read zero. This situation is practically equivalent to an empty tank since the pump is incapable of pumping the fuel and is more common as fuel tanks, especially in passenger cars, tend to be shallow. The remedy for this is to use two separate fuel level sensors, sensor 201 in the fore end of the tank and a second sensor 202 in the aft end of the fuel tank 200 as shown in Figure 20. The fuel level sensors can be either tilt compensated or not and the average of then- outputs is displayed in the gauge in front of the driver. In this application, the entrance of the fuel pump 203 should be split into two lines 204 and 205 so that the pump will be able to access both ends of the fuel tank. In order that only the line immersed will be active, the entrance to the pump includes a valve activated by the polarity of the inclination angle of the car to select the specific input line immersed in the fuel.

A probe-type liquid level sensor, 1300, is illustrated in Figure 13. This consists of the two Figure 4 probe pairs, for example, mounted inside a nominally glass tube, 1301. If the Figure 4 A and 4B sensor electrode pairs are mounted, for example, at positions 1303 A and 1303B and are assembled on a flexible circuit substrate, 1302, then the flexible substrate may be curved into a cylindrical shape, and inserted in the liquid measurement container, 1301, as a probe-type sensor. Printed circuit board, 1304, such as in Figure 5 may be also included in the probe tube.

While the invention has been described with respect to a limited number of embodiments, it will be appreciated that many variations, modifications and other applications of the invention may be made.

Claims

WHAT IS CLAIMED IS:

1. A compensated-type capacitive sensor for measuring the level of a liquid in a container comprising

(a) a first sensing capacitor having a first dependence of incremental capacitance on the depth of coverage of said first sensing capacitor by the liquid;

(c) a signal processor for converting said capacitances of said first and second capacitors into respective proportional first and second voltages;

(d) a divider for dividing said first voltage by said second voltage, thereby providing an output signal dependent on the depth of coverage of said first and second capacitors by the liquid in accordance with a predetermined function.

2. A sensor as in claim 1, wherein said capacitors are substantially uniform-field type capacitors.

3. A sensor as in claim 1, wherein said capacitors are substantially fringing-field type capacitors, said capacitors including capacitor plates.

4. A sensor as in claim 3, further comprising a grounded shield plate between said capacitor plates.

5. A sensor as in claim 3, further comprising a dielectric tube, said electrodes contained inside said tube, thereby providing a probe-type immersible liquid-level measurement probe.

6. A capacitive inclination sensor comprising

(b) at least one pair of excitation capacitive plates and at least one corresponding signal capacitive plate, forming at least one fringing field capacitor pair, one said excitation plate and one said corresponding signal plate forming a capacitor, said pairs of excitation and signal plates on said outer surface of said non-conductive wall, said pairs of excitation and signal plates symmetrically centered around said perpendicular axis, said capacitors having capacitances depending on said angle of rotation ;

(c) a capacitance measuring device for measuring capacitances of said capacitors, for providing an output signal depending on said angle of rotation.

7. A sensor as in claim 6, further comprising grounded electrode plates between each said excitation plate and corresponding said signal plate, thereby diminishing the liquid-independent portion of the mutual capacitance between said excitation and said corresponding signal plate pair.

8. A sensor as in claim 4, further comprising a dielectric tube, said electrodes contained inside said tube, thereby providing a probe-type immersible liquid-level measurement probe.

9. A sensor as in claim 1, wherein said capacitors include electrodes plated on at least one printed circuit board, and electromc components of said signal processor and said divider are mounted on said at least one printed circuit board.

10. A single printed circuit board fringing field liquid level sensor comprising a measurement capacitance, a fully submerged reference sensing capacitance, and a signal processor and a divider.

11. A sensor as in claim 10, further comprising snap-mounting on the container containing the liquid, the level of which liquid is to be measured.

12. A capacitor as in claim 3, said capacitor plates further comprising skewed interdigitated fingers.

13. A fringing field two-axis inclinometer comprising a container, partially filled with a liquid, having an instantaneous tilt-angle relative to the vertical gravity vector, on a surface of which said container, four capacitors are plated, in the configuration of two perpendicularly-oriented pairs of diametrically-opposite pairs of capacitors, each said capacitor of each said pair of diametrically-opposite capacitors located at an opposite side of said container; each capacitor having a liquid-level coverage corresponding to said tilt angle, each said capacitor having a capacitance depending on said amount of liquid-level coverage; each said diametrically-opposite pair of capacitors having a differential capacitance corresponding to said tilt angle relative to each said diameter between each said capacitor pair.

14. A fringing field two axis inclinometer comprising

(a) a container, partially filled with a liquid; and,

(b) a set of capacitor plates printed on a flat substrate, said set of capacitor plates including four capacitors, said capacitors in the configuration of two perpendicularly-oriented pairs of diametrically- opposite pairs of capacitors, each said capacitor of each said pair of diametrically-opposite capacitors located below an opposite end of a diameter of said container; each capacitor having a liquid-level vicinity corresponding to said tilt angle, each said capacitor having a capacitance depending on said amount of liquid-level vicinity; each said diametrically-opposite pair of capacitors having a differential capacitance corresponding to said tilt angle relative to each said diameter above each said respective capacitor pair.

15. A liquid-filled type inclinometer, comprising first and second liquids, each said liquid having a density and having electrical properties, each said liquid filling substantially half of the volume of the inclinometer, said first and second liquids having different densities, and having different electrical properties.

16. A method of reducing the sensitivity of a liquid-filled inclinometer to side motions and vibrations, while substantially retaining the initial fast inclinometer response time, of a liquid-containing inclinometer partially-filled with a first liquid, the first liquid having a density and having electrical properties, the inclinometer having a space remaining in the inclinometer volume above the first liquid, comprising filling the space above the first liquid with a second liquid, said second liquid having a slightly smaller density from the first liquid; and said second liquid having different electrical properties from the first liquid.

17. A liquid level sensor comprising

(a) a metal enclosure; and

(b) processing electronics sealed within said metal enclosure to prevent its contact with the liquid.

18. A tilt compensated capacitive liquid level sensor comprising an integral liquid filled inclinometer.

19. A tilt compensated liquid level sensor as recited in claim 18, comprising at least one inclinometer plate and at least one liquid level plate printed on a common substrate.

20. A tilt compensated capacitive liquid level sensor comprising an integral inclinometer operating with the liquid it is to measure.

21. A sensor as recited in claim 20, wherein said inclinometer is a fringing-field type.

22. A sensor as recited in claim 20, wherein said inclinometer is a parallel-field type.

23. A sensor as recited in claim 21, wherein the temperature dependent portion of the sensing capacitance "dry" component is compensated by printing plates sealed from contact with the liquid being measured.

24. An automotive fuel system including:

(a) a fuel tank having an inclination angle;

(b) a display gauge;

(c) an inclinometer providing a signal indicative of the polarity of said inclination angle;

(d) first and second fuel level sensors;

(e) a selector valve controllable by said inclination angle;

(f) first and second fuel pumping lines extending to the bottom level of said fuel tank and connected to said fuel pump through said selector valve; said first fuel level sensor and said first fuel line disposed in the fore end of said fuel tank, and said second fuel level sensor and said second fuel line disposed in the aft end of said fuel tank and wherein said selector valve activates the fuel line that is immersed in the fuel, the average output signal of said first and second fuel level sensors displayed in said display gauge.