TECHNICAL FIELD

THIS INVENTION relates to an apparatus for measuring a characteristic of a fluid and particularly variable characteristics of the fluid such as flow velocity, temperature, humidity, pressure and the like, which are related in some manner.

This invention has particular, but not exclusive, utility in measuring the velocity of air or liquid at a precise location within a duct or pipeline and so can be used for determining velocity profiles of the fluid flow within the duct or pipeline.

The invention also has utility, in a preferred embodiment of being able to compensate for changes in characteristics of the fluid such as temperature, humidity and pressure which may affect determining the velocity of the fluid flow.

BACKGROUND ART

Fluid velocity measurement has previously been performed by utilising anemometers and flow eters of varying designs. These designs have included hot wire anemometers which operate on the principle that heat transfer to air from an object at an elevated temperature is a function of the air speed. Previous types of instruments of this kind have usually consisted of a thin platinum wire heated to approximately 500°C so that its temperature is relatively independent of ambient temperature and environmental variations. Air speed is determined by measuring either the current required to maintain the hot wire at a constant temperature or the resistance variation of the

hot wire while a constant current is maintained. Such instruments are usually very delicate and expensive, and so are utilised principally for precision measuring and scientific applications.

A general characteristic of hot wire anemometers is the need for the hot wire to operate at high temperatures to achieve a temperature co-efficient which is sufficiently high to enable accurate measurements to be obtained and so requires filaments to be used which are not appropriate for industrial use. Even at such temperatures, the temperature co—efficient is still relatively small and hence there is the need to utilise extremely sensitive equipment to detect changes in the voltage of the wire. Another drawback with hot wire anemometers is their use with vaporised fluids where there is a propensity for such fluids to condensate on the hot wire and so drastically affect measurements.

Mechanical anemometers usually rely upon the principle of transduction of fluid velocity into the rotation or movement of an element, the rotation or movement being proportional to the velocity of the fluid applied thereto. These instruments can measure a wide range of fluid velocities, however, they usually tend to be inaccurate since they integrate the velocity variations and require long periods to stabilise particularly at low velocities. In pipes, they also restrict the flow of the fluid due to the relatively large cross-sectional area that the blades require for sensitivity to low fluid flow rates.

In ducts or pipelines, it is common to use orifice plates which rely upon creating a venturi effect to measure pressures and so calculate the velocity of the fluid flow through the plate. Problems with this type of device involve the inevitable build-up of scale, particularly on

the edges of the orifice plate which can have drastic effects on the accuracy of measurements obtained. Additionally, the provision of an orifice plate in a pipeline necessarily causes an interruption in fluid flow to achieve pressure differentials which naturally creates a loss in fluid velocity.

DISCLOSURE OF INVENTION

It is an object of the present invention to provide an improved apparatus for measuring fluid flow to obtain relatively accurate measurements of fluid flow utilising the advantages of solid state technology and at a reduced expense than prior art devices.

In accordance with one aspect of the present invention, there is provided an apparatus for measuring a characteristic of a fluid comprising:-

sensor semiconductor means to be disposed in said fluid; heating means to selectively apply a current to said semiconductor means so as to selectively elevate the junction temperature of said sensor semiconductor means; sampling means to periodically sample the dynamic voltage of said sensor semiconductor means; control means to control the application of said current to said sensor semiconductor means so as to maintain the sampled dynamic voltage at a prescribed level; and processing means to equate the electrical energy expended by said means in maintaining said sampled dynamic voltage at said prescribed level to a characteristic of said fluid.

While said characteristic is preferably related to fluid velocity, the apparatus may provide information relating to the temperature, humidity and/or specific heat of the fluid. Additionally in the case of a gaseous fluid, the apparatus may provide information relating to moisture content.

Preferably, said apparatus includes temperature compensation means to compensate for changes in the ambient temperature of said fluid which may affect the equating of said processing means.

Preferably, said temperature compensation means comprises: an apparatus as claimed at claim 11, wherein said temperature compensation means comprises: a reference semiconductor means substantially identical to said sensor semiconductor means, said reference semiconductor means being disposed proximate to said sensor semiconductor means; said sampling means to supply a substantially low value of forward biased constant current to said reference semiconductor means; and sensing means to sense the dynamic voltage of said reference semiconductor means and apply a temperature compensation signal representative of the same to said control means, whereby changes in the sensed dynamic voltage can be substantially attributable to changes in the ambient temperature; and said control means being adapted to compensate for changes in the dynamic voltage of said sensor semiconductor means attributable to changes in ambient temperature by controlling the application of said heating current to said sensor semiconductor means in accordance with changes in said temperature compensation signal.

Preferably, said control means disables said heating means when said sampled dynamic voltage attains a threshold of said prescribed level representative of an upper

temperature of the junction of said sensor semiconductor means and enables said heating means when said sampled dynamic voltage increases above said threshold beyond a prescribed amount representative of a lower temperature of the junction of said sensor semiconductor means.

Preferably, said control means controls the operation of said sampling means and heating means so that the application of said sampling current and heating current to said sensor semiconductor means is synchronised.

Preferably, said sensor semiconductor means is a junction diode such as zener diode.

Preferably, said sampling means periodically applies a sampling current to said sensor semiconductor means to forward bias the same for sampling purposes and said heating means selectively applies said heating current to said sensor semiconductor means to reverse bias the same, the magnitude of said heating current being greater than said sampling current.

Preferably, said sampling means applies said sampling current to said sensor semiconductor means at a period less than or equal to the period at which said heating means applies heating current to said sensor semiconductor means.

Preferably, said control means comprises reference signal means for generating an upper threshold signal representative of said threshold of said prescribed level, comparator means for comparing said sample signal with said upper threshold signal, and switching means to be selectively switched in response to the result of the comparison by said comparator means to generate an output signal to control the operation of said heating means.

Preferably, said output signal is a digital signal indicating that said heating means is to be enabled at one level of the output signal and disabled at the other level of the output signal.

Preferably, said processing means averages the magnitude of said output signal and correlates said average against a set of predetermined fluid velocity data.

BRIEF DESCRIPTION OF THE DRAWINGS

The invention will be better understood in the light of the following description of several specific embodiments thereof. The description is made with reference to the accompanying drawings wherein:-

Fig. 1 is a perspective view of a zener diode which comprises the sensor semiconductor means;

Fig. 2 is a graph showing the plot of temperature distribution against lead length of a zener diode for a plurality of different air flow velocities;

Fig. 3 is a graph showing the plot of dynamic forward biased resistance of a zener diode against time where a constant current is applied to the diode whilst the junction is being cooled;

Fig. 4 is a graph showing the plot of diode forward voltage against air temperature for a particular current applied to the diode;

Fig. 5 is a graph showing plots of diode current and diode dynamic resistance against time with uncontrolled diode heating;

Fig. 6 is a graph showing the plots of diode dynamic resistance and diode current against time where self regulation diode heating is employed;

Fig. 7 is a graph showing the plot of diode forward voltage against temperature showing the upper and

1ower thresho1ds;

Fig. 8 is a schematic circuit diagram of the fluid flow velocity measuring apparatus which does not show the temperature compensation means nor the processing means;

Fig. 9 is a graph showing the plot of sample signal voltage at the output of the operational amplifier against time;

Fig. 10 are truth tables for the R-S flip flop and

NAND gate of the control means;

Fig. 11 is a schematic circuit diagram showing the temperature compensation means;

Fig. 12 is a truth table of the sensing means of the temperature compensation circuit;

Fig. 13 is a block diagram of an overview of the fluid velocity measuring circuit and the temperature compensation circuit showing the processing means;

Fig. 14 is a graph showing the plot correlating fluid flow velocity with output signal magnitude;

Fig. 14A is a graph showing plots of the HEAT/COOL signal at Q output of the flip flop against time for increasing fluid velocities;

Fig. 15 is a graph showing the plot of the upper and lower temperature compensation signals against temperature;

Fig. 16 is a graph showing the plot of the dynamic voltage of the sensor diode against temperature showing the operating range of the apparatus;

Fig. 17 is a graph of the plots of the dynamic forward voltage for both the sensor and reference diodes against temperature showing how the diodes are calibrated;

Fig. 18 is a schematic circuit diagram showing the velocity and temperature sensing device of the second embodiment;

Fig. 19 is a block diagram demonstrating the calibration process of the diodes to determine their stored characteristic data;

Fig. 20 is a graph of diode forward voltage against temperature, representing the parameters determined in the calibration process of the diodes;

Fig. 21 is a graph of reference sensor and EPROM data representative of voltage against temperature for a sensor in a high ambient temperature fluid;

Fig. 22 is a graph similar to Fig. 21, but for a sensor in a low ambient temperature fluid;

Fig. 23 is a graph of sensor voltage and comparator values against temperature for showing how new threshold values are calculated by the microprocessor system;

Fig. 24 is a graph showing heat to cool ratios versus fluid velocity;

Fig. 25 is a graph of the heat/cool signal at zero fluid velocity conditions;

Fig. 26 is a graph of the heat/cool signal at about maximum velocity condition;

Fig. 27 is a graph of the heating current applied to the sensor diode at zero fluid velocity;

Fig. 28 is a graph of the heating current applied to the sensor diode under about maximum fluid velocity conditions;

Fig. 29 is a graph of amplifier output voltage against temperature, showing software computations of upper and lower threshold values under extended dynamic range conditions;

Fig. 30 is a graph of the response of the sensor system operating under extended dynamic range conditions showing sensor diode operating temperature increments related to fluid velocity;

Fig. 31 is a graph of the heating current supplied to the sensor diode under programmable conditions;

Fig. 32 is a graph of the resultant incremental sections of the diode operating temperature family of curves shown in Fig. 30, indicating the overall linearity response of the system to fluid velocity;

Fig. 33 is a schematic circuit diagram of the fluid velocity measuring apparatus in accordance with the fourth embodiment;

Fig. 34 is a graph of the sensor diode forward voltage versus temperature to indicate the effect of hysteresis to provide a lower threshold;

Fig. 35 is a graph of the diode junction temperature versus time in accordance with the fourth embodiment;

Fig. 36 is a graph of the amplifier output voltage versus time in accordance with the fourth embodiment;

Fig. 37 is a schematic circuit diagram showing a modified velocity and temperature sensing device;

Fig. 38 shows waveforms representative of the

HEAT/COOL signal for the second and fourth embodiments;

Fig. 39 is a psychrometric chart plotting grams of moisture per kilogram of air against air temperature for different relative humidities;

Fig. 40 is a series of graphs showing the heat/cool ratio for varying relative humidity and temperature at zero velocity and also with a velocity magnitude;

Fig. 41 is a schematic diagram of the apparatus described in the fifth embodiment;

Fig. 42 shows the effects of varying relative humidity upon a heat/cool signal in fluid environment at zero velocity;

Fig. 43 shows the effect upon magnitude of the heat signal in a heat/cool signal at varying relative humidity to maintain the duration of the signal constant;

Fig. 44 is a fragmentary perspective view of a pipeline or duct with a plurality of sensor and reference zener diodes disposed on an airfoil therein;

Fig. 45 is a close-up of the airfoil; and

Fig. 46 is a schematic drawing of the velocity measuring device for the same.

The first embodiment is directed towards an apparatus for measuring velocity of a fluid flow at a given humidity and pressure and which can compensate for changes in the temperature of the fluid.

To generally describe the principle of operation of the apparatus, reference is made to Figs. 1 to 7 of the accompanying drawings.

Where a semiconductor means, in the form of a zener diode 11 is placed in a fluid such as air which is at zero velocity, which is at stable temperature, pressure and which is totally dry, and where the semiconductor means has applied thereto a reverse biased current, the junction of the diode will heat up until an equilibrium condition is reached where the rate of heat dissipation from the junction, generally along the leads 13 of the diode equals the rate of generation of heat at the junction by the magnitude of the current passing through the diόde and thus the temperature at the junction of the diode and at any particular position along the lead length will become constant.

Where the fluid in which the semiconductor is located is flowing, the temperature gradient along the length of the lead increases with flow velocity as a consequence of heat being removed, thereby lowering the relative temperature at a particular position of the lead length. This is shown in the graph at Fig. 2 of the drawings.

The variation in temperature produced along the lead length affects the dynamic resistance of the diode, so as to make the dynamic resistance and consequently the dynamic voltage of the diode inversely proportional to the junction temperature of the diode.

Fig. 3 shows how the dynamic resistance increases as a function against time as the junction of the diode is allowed to cool, when a constant sampling diode current I _D is applied. As shown, the dynamic resistance and hence the dynamic voltage, increases as the junction temperature progressively decreases with cooling. Conversely, with heating of the diode junction, the dynamic resistance, and hence the dynamic voltage, would progressively decrease as the junction temperature increases.

As shown at Fig. 4, it is a characteristic of most diodes that the forward voltage of the diode changes linearly and inversely with respect to fluid temperature. This in fact applies for both forward and reverse voltage characteristics of the diode.

In. situations where current is periodically applied to the 'diode with an equal mark space ratio as shown as Fig. 5 of the drawings, successive heating and cooling periods of fixed duration are applied to the diode. Accordingly, the dynamic resistance and hence dynamic forward and reverse voltages of the diode will respond as the junction temperature rises. As shown, the dynamic resistance decreases during heating periods and increases during cooling periods. In the case represented by the graph, the diode in question progressively becomes heated with an average increase in its temperature due to its capacity to generate heat more rapidly than it can dissipate this heat, i.e. the cooling period is not sufficiently long to achieve a desired equilibrium condition. Therefore by regulation of the heating and cooling periods, in a manner to be described later, a desired equilibrium condition can be reached whereby the junction temperature of the diode is maintained within a prescribed .range as shown at Fig. 6

of the drawings. Thus as the junction of the diode heats up and reaches a predetermined upper threshold, a control circuit switches off the current applied to the diode, allowing the junction to cool. As the junction temperature cools to a predetermined lower threshold, the control circuit again switches on the current to the diode causing heating of the same, whereinafter the cycle repeats itself. Equating these temperature changes to diode forward voltage, reference should be made to Fig. 7 where it can be seen that as the diode junction temperature oscillates between the upper and lower temperature thresholds, corresponding forward voltage variations are produced which are inversely proportional to the heat contained in the diode.

If the fluid velocity past the diode and its leads increases, heat will be drawn away from the junction causing the junction temperature to decrease at a greater rate. This means that it will take longer for the same heating current to replace the lost heat, and it will take less time to cool to the lower threshold.

This change in duty cycle or mark space ratio of the HEAT/COOL signal is indicative of the electrical energy expended to maintain the dynamic resistance or voltage of the diode within a prescribed range and is a non-linear function of fluid velocity. Hence the expended energy can be equated to the fluid velocity by integrating the current applied to the diode over a prescribed period of time. An indication of the characteristic of fluid velocity versus energy is provided at Fig. 14 of the drawings and shall be described in more detail later.

In addition to fluid velocity past the diode causing a change in the rate of heat dissipation from the diode junction, ambient fluid temperature surrounding the diode

also has an effect. Thus, if the ambient fluid temperature reduces and the diode junction temperature remained the same, then the duty cycle increases as in the manner described in relation to fluid flow past the diode and similarly if the ambient fluid temperature increases, then the duty cycle will decrease as the junction temperature is approached. This may occur until the ambient fluid temperature equals the diode junction temperature, whereupon the diode heating ceases and the method can no longer be used to effect.

In view of the fact that both fluid velocity past the diode and ambient fluid temperature have an effect upon the duty cycle of current applied to the diode, appropriate compensation must be undertaken when it is desired to accurately determine the fluid velocity under all temperatures of the fluid flow past the diode. This shall be described in more detail later.

The apparatus utilised in measuring the fluid velocity past the diode will now be described with reference to Fig. 8 of the drawings. With regard to this arrangement, a sensor zener diode 17 is used as the sensor semiconductor means to sense the fluid flow velocity and is operated in two modes. In the first mode, it is pulsed periodically with a low level constant sampling current to sample its dynamic resistance value in the forward biased condition by sensing the dynamic voltage thereof, and in the second mode it is pulsed selectively with a reverse biasing high level heating current to heat the junction.

The fluid velocity measuring apparatus 15 has the sensor zener diode 17 placed in a fluid stream, the velocity of which is to be measured. A twenty kilohertz clock 19 periodically pulses a 20 microsecond monostable multivibrator 21 for producing a SAMPLE/HOLD signal which

in turn activates sampling means for sampling the forward voltage of the zener diode 17. The sampling means includes a switchable constant current source 23 having a low current forward characteristic. Heating means for applying a heating current to the zener diode 17 to elevate the temperature of the diode junction is provided in the form of a switchable constant current source 25 having a high current reverse characteristic. Thus, the low constant current source 23 when activated by the SAMPLE/HOLD signal output of the monostable multivibrator 21 provides a forward current of approximately 0.5 milliamps through the zener diode 17 when it is forward biased. On the other hand, the high constant current source 25, when activated, provides a relatively large reverse current, (typically 350mA but this will vary according to the power rating of the diode) through the zener diode 17 causing the same to be reverse biased to its reverse voltage.

The output of the sensor zener diode 17 is input to a sample and hold circuit 27 which also forms part of the sampling means. The sample and hold circuit is also controlled by the SAMPLE/HOLD signal output of the monostable multivibrator 21 so as to sample the dynamic voltage of the zener diode 17 during the period that the constant current source 23 is activated and hence during the period that the zener diode is forward biased.

Thus, the voltage sampled by the sample and hold circuit 27 is indicative of the junction temperature of the diode during the SAMPLE period and is stored in the circuit for the duration of the HOLD period until the next SAMPLE period occurs.

The output of the sample and hold circuit 27 is connected to a single stage non-inverting operational amplifier 29

for amplification of the output voltage of the sample and hold circuit 27 with respect to a predetermined Q point offset and slope setting, determined by potentiometers 30a and 30b, respectively (to be described in more detail later) .

The output of the amplifier 29 is in turn connected to a control means which comprises a comparator means in the form of an upper threshold comparator 33 and a lower threshold comparator 31. The control means also includes a switching means in the form of an R-S flip flop 35 which derives its reset input from the output of the upper threshold comparator 33 and its set input from the output of the lower threshold comparator 31. The upper threshold comparator 33 compares the sample signal provided at the output of the amplifier 29 against a preset upper threshold signal provided at the potentiometer 39 to assert the reset ^" input of the flip flop 35 when the sample signal equals the upper threshold signal. Conversely, the lower threshold comparator 31 compares the sample signal output by the amplifier 29 with a lower threshold signal provided by the potentiometer 37 to assert the set input of the flip flop 35 when the sample signal equals the lower threshold signal. The potentiometers 37 and 39 form part of reference signal means connected to a variable voltage output 61 from a temperature compensation means to be described in more detail later.

The output of the flip flop 35 produces an output signal which is indicative of the duration of the heating and cooling periods for the purposes of applying a reverse current through the zener diode 17 by the heating means 25 to effectively heat the junction of the diode whilst it is reverse biased, or to allow it to cool whilst the heating means is disabled.

As shown at Fig. 8, the output of the flip flop 35 is fed back via a gating means in the form of a NAND gate 41 to the high current source 25 so as to selectively operate the same. The NAND gate 41 derives its other input from an inverted output Q* of the monostable multivibrator 21 to disable the heating current source 25 at each sample period of the 20 microsecond SAMPLE/HOLD signal output pulse of the monostable, thereby allowing at any one time, at most, only one of the constant current sources 23 or 25 to apply current to the zener diode 17.

The dynamic temperature variations of the sensor diode can be observed at the output of- the operational amplifier 29 as is shown in the graph at Fig. 9 of the drawings. The comparator outputs indicate logically the state of the diode heating which is detailed in a truth table shown at Fig. 10 the drawings.

The Q* output from the flip flop 35 is fed to another operational amplifier circuit 43 which has variable gain and output offset controls as provided by potentiometers 45 and 47 respectively. The resultant amplified signal is fed to a moving coil meter 49 which integrates the signal to provide the average DC component of the diode heating signal and so provides an analogue reading which can be equated to the flow velocity of the fluid in which the zener diode 17 is placed.

The Q output of the flip flop 25 effectively provides the HEAT/COOL signal for the circuit, and can be connected to a microprocessor system for the computing of fluid velocity detected at output 50.

In order to compensate for variations in the ambient temperature surrounding the sensor zener diode 17, temperature compensation means is provided in the form of

a temperature compensation circuit 51 as shown at Fig. 11 of the drawings. This temperature compensation circuit 51 utilises a reference zener diode 53 substantially identical to the sensor zener diode 17 utilised in the fluid velocity measuring apparatus. The reference zener diode 53 is disposed at a location proximate to the sensor zener diode 17. The reference zener diode is connected in a permanently forward biased arrangement to a further low value constant current source 55. The output of the reference zener diode 53 is in turn connected to a further sample and hold circuit 57, substantially similar to that of the fluid velocity measuring apparatus 15 and is controlled by a modified version of the SAMPLE/HOLD signal output from the monostable 21. The output of the further sample' and hold circuit 57 is connected to the input of a further non-inverting amplifier 59 to amplify the output voltage of the sample and hold circuit 57 with respect to a predetermined Q point offset and slope setting, determined by potentiometers 60a and 60b respectively (to be described in more detail later) and so generate a temperature compensation signal indicative thereof at its output 61. The D.C. output voltage of the amplifier 59 represents the ambient fluid temperature in which the sensors are placed and is simultaneously applied to the upper and lower , threshold potentiometers 37, 39 respectively to correspondingly alter the upper and lower threshold signals in accordance with alterations in the ambient temperature of the fluid stream past the diodes.

The modified version _, of the SAMPLE/HOLD signal output from the monostable multivibrator 21 is obtained by connecting the Q output of the same to an AND gate 62 in conjunction with the Q* output of the R-S flip flop 35, whereby the output of the AND gate 62 is in turn connected to the further sample and hold circuit 57 to control operation of the same so that sampling of the reference diode 53 is

performed in synchronicity with the sampling of the sensor diode 17 during the cooling period of the sensor diode. Accordingly, sampling of the reference diode 53 is not performed during the heating period of the sensor diode 17 since interference occurs if the reference diode is sampled by the sampling means during heating periods of the sensor diode, and can affect measurements obtained by the sample and hold circuit 57 from the reference diode 53. The truth table shown at Fig. 12 helps explain the logic associated with the operation of the temperature compensation means.

An overview of the apparatus is shown at Fig. 13 of the drawings wherein the sensor zener diode 17 and the reference zener diode 53 are respectively connected to the velocity sensing circuit 15, as shown at Fig. 8 of the drawings and the temperature compensation circuit 51 as shown at Fig. 11 of the drawings. For precise measurements of fluid flow velocity, the output signal of the velocity sensing circuit 15 provided at the output of the flip flop 35 is adapted to be connected to a processing means 63.

The processing means 63 incorporates a micro-processor system which runs in accordance with a computer program stored in the memory thereof. The processing means 63 operates in accordance with a prescribed algorithm to equate the electrical energy expended by the heating means, in maintaining the sampled dynamic voltage between the prescribed upper and lower thresholds, to a velocity characteristic of the fluid flow.

Now describing the method of operation of the fluid velocity measuring apparatus 15. As shown at Figs. 8 and 9 of the drawings, the monostable 21 provides a series of sample pulses having a pulse width of 20 microseconds and

a frequency of 25 kilohertz as generated by the clock 19. Consequently, a 0.5 milliamp sampling current is applied to the sensor zener diode 17 by the sampling means for forward biasing the same for the duration of the monostable pulse width during which time the sampling means may sample and amplify the dynamic voltage of the diode. The sampling means holds this sampled value for the remainder of the cycle and compares the same against the upper and lower thresholds. The R-S flip flop 35 operates such that whilst the sample signal falls within the upper and lower thresholds, neither the reset nor set inputs thereof are asserted. Consequently, the output signal follows a level of its previous state until such time as the upper or lower threshold is attained.

In the event that the sample signal reaches the lower threshold, signifying that the diode junction has cooled sufficiently, the set input of the flip flop will be asserted by the lower threshold comparator 31 causing the output signal of the flip flop to be asserted, and subsequently activating the constant current source 25 of the heating means to apply a reverse biasing current through the diode, heating the junction and lowering the dynamic resistance thereof. This activating signal applied to the high current source 25, however, is disabled by means of the NAND gate 41 for the duration that the low current source 23 forward biases the diode to allow the sampling circuit to sample the dynamic voltage thereof. Upon the dynamic resistance decreasing, the sample signal will drop below the voltage value of the lower threshold, preventing the set input of the flip flop 35 from being further asserted. Nonetheless, the output of the R-S flip flop 35 will remain asserted since the output will follow its previous state for the duration that neither the reset or set inputs are asserted. This will remain the case until such time as the sample signal

attains the upper threshold whereupon the reset input of the flip flop will be asserted causing the output signal of the R-S flip flop 35 to be returned to its non-asserted state. Consequently, the activating signal will cease being applied to the current source 25 of the heating means, thereby terminating the heating period thereof and allowing the sensor zener diode to be subject to a cooling period. Accordingly, this cycle of operation will be repeated at each successive assertion and non-assertion of the output signal of the flip flop.

The processing means 63 records the mark space ratio of the activating signal and averages the same to obtain an indication of the electrical energy expended by the heating means in maintaining the sampled dynamic voltage of the sensor zener diode at the prescribed level, i.e. between the upper and lower threshold. This average value is obtained for a prescribed number of sample periods, such as 100 or more, whereupon the overall average for the sample period is obtained. The magnitude of the average mark space ratio is then correlated against fluid velocity data contained in a look up table stored in the memory of the processing means to obtain the corresponding fluid velocity value, or alternatively the fluid velocity values are calculated from linear approximations of the fluid velocity characteristic. For example, as shown at Fig. 14 of the drawings, due to the non-linear characteristic of the fluid velocity versus energy magnitude, the curve may be approximated by a series of linear functions over prescribed portions of the curve. Thus, upon obtaining an energy magnitude value, the appropriate linear function to which the value must be correlated may be determined, and the appropriate fluid velocity value calculated. The particular accuracy of this method can be improved by increasing the number of linear functions representing the curve.

An example of how the average d.c. voltage or resultant energy expended changes with an increase in fluid velocity with respect to the HEAT/COOL signal is shown at Fig. 14A.

Having regard to temperature compensation, as shown at Figs. 15 and 16 of the drawings, the temperature compensation signal provides a voltage to the lower and upper threshold potentiometers 37 and 39 respectively which is divided by the respective potentiometer setting to produce the appropriate lower and upper threshold voltages to the comparator means. As is shown by means of the graph in Fig. 15, the potentiometers are trimmed to produce the appropriate upper and lower threshold signals. Accordingly, as temperature fluctuations occur, corresponding compensation is applied to the control means of the fluid flow measuring apparatus. Fig. 16 effectively shows the range of operation of the apparatus within the prescribed threshold limits where it is necessary to consider the ambient fluid temperature and the diode junction temperature.

It is a necessary requirement for the correct operation of the circuit for both the sensor zener diode 17 and the reference zener diode 53 to be substantially identical to each other. In practice, it is almost impossible to achieve two zener diodes of identical characteristics and therefore, it is necessary to compensate for the differences between two diodes so as to achieve accurate measurement. For the purposes of the present invention, the most important differences arise with the transfer characteristics of the respective diodes.

As shown at Fig. 17 of the drawings, the actual transfer characteristic of the sensor zener diode may be represented by the line SD and that of the reference zener

diode by the line RD. As can be seen, the slopes and offsets of these lines are different and hence variations in temperature will produce a different forward voltage detected by the sampling means in each case. Accordingly, the reference diode will respond differently to changes in ambient temperature than would the sensor diode and thereby true temperature compensation will not be achieved.

In order to overcome this problem, an ideal point offset and slope setting is determined and appropriate correcting factors are applied to the respective circuits so that all measurements are equated to this ideal Q point setting. For example, in the case of the sensor zener diode line SD, an ideal Q point is predetermined. Next, the appropriate offset value for shifting the line SD to the ideal Q point is calculated and accordingly the potentiometer 30a is set to provide this appropriate offset value OF1 at the operational amplifier 29. Since this is the first diode to be corrected a preset slope value can be used to provide appropriate gain by way of the potentiometer 30b. Subsequently, the reference zener diode is corrected by similarly determining an appropriate offset value OF2 to shift the line RD to coincide with the Q point. Furthermore, a requisite slope of the line RD must be calculated to achieve the same slope as obtained from the line SD. Accordingly, the potentiometer 60a for the further operational amplifier 59 is adjusted to provide the appropriate offset and the potentiometer 60b is adjusted to provide the appropriate gain to achieve the requisite slope. Once these values are determined, the settings of the operational amplifiers 29 and 59 can be permanently fixed by choosing precision resistors of the required values to replace the potentiometers 30 and 60.

The second embodiment is directed towards a similar apparatus to the preceding embodiment, except that the velocity sensing circuit and temperature compensation circuit is not restricted to being used with a particular pair of zener diodes but can be utilised for determining temperature and velocity measurements from a multitude of zener diode pairs.

In the description of this embodiment, the same reference numerals have been used as in the preceding embodiment to identify similar components used in the embodiments.

As shown at Fig. 18 of the drawings, a velocity and temperature sensing device 71 is provided of substantially similar design to the fluid velocity measuring apparatus 15 and temperature compensation circuit 51 of the preceding embodiment except for the differences outlined below.

Three discrete input terminals 73a, 73b and 73c are provided to connect the device 71 to a pair of sensor and reference zener diodes contained within a sensing module 75 located in a fluid flow or the like. Moreover, the terminal 73a is connected to the respective outputs of the low value constant current source 23 and the high value constant current source 25 of the velocity measuring circuit. Accordingly, the terminal 73a is intended to be connected to the anode of a sensor zener diode 77.

The terminal 73c is connected to the input of the low value constant current source 55 of the temperature compensation circuit and consequently is connected to the anode of the reference zener diode 79.

The terminal 73b is connected directly to ground and

consequently is connected to the common cathodes of the sensor and reference zener diodes 77 and 79 to complete the respective circuit connections involving these diodes.

Additionally, a series of eleven data lines 81 are provided on the apparatus 71 which are connected to an eleven bit digital data bus 83 included in the apparatus. These terminals 81 are intended to be connected to corresponding output lines of a four word by eleven bit or similar EPROM 84 which is incorporated into the sensor module 75 which shall be described in more detail later. A major difference between the apparatus 71 and the preceding embodiment is the inclusion of a micro/processor system 85 interposed between the fluid velocity measuring circuit and the temperature compensation circuit and which provides a digital interface to each of these circuits. Moreover, the microprocessor system 85 includes a main input data bus 87 which is connected to the data bus 83 to receive input information from the EPROM 84 of the sensing module 75 via terminals 81 and to the digital output of an A-to-D Converter 89 which has its analogue input connected to the output of the further operational amplifier 59. The microprocessor system 85 also is provided with an output data bus 91 which connects to the digital input of a D-to-A converter 93 which has its analogue output connected to a pair of sample and hold circuits 95a and 95b, respectively. The sample and hold circuits 95 are respectively connected to the threshold inputs 97a and 97b of the comparators 31 and 33 respectively. The microprocessor system also has discrete control outputs connected to the sample and hold circuits 95 via output control lines 99a and 99b to individually control the sample and hold operation of the same. A discrete input data line 101 is received by the microprocessor system 85 from the output of a logic circuit 103 which is connected to the respective outputs of the comparators 31 and 33 to

provide the resultant HEAT/COOL signal to the microprocessor system for processing. As in the previous embodiment, the output of the logic circuitry 103 is fed back to the high value constant current circuit 25 to control operation of the same.

The operational amplifiers 29 and 59 of the fluid velocity measuring circuit and temperature compensation circuit, respectively, have their gain and offset resistors calibrated to achieve an ideal Q point setting for the circuit. The calibration of these resistors is achieved differently than in the preceding embodiment in view of the fact that the present circuit is intended to be used with a variety of sensor and reference diode pairs. Accordingly, an ideal Q point is preselected and the microprocessor system is utilised to compensate for changes in offset and slope determined for a particular sensor and reference diode pair against the ideal Q point by adjusting threshold levels for the comparators 31 and 33 itself. In this manner, fixed gain values for the amplifiers 29 and 59 are determined by selection of precise resistor values for the gain resistors 30b and 60b respectively. Consequently, potentiometers 30a and 60a can be used to provide fine adjustment of respective offset values for the amplifiers so as to calibrate them to the ideal Q point in each case.

This calibration may be achieved by using predetermined precision resistors 105 which may be packaged in a Q point calibration module 107 to be connected to the input terminal 73 of the apparatus 71, whereby the precision resistors 105a and 105b have been previously determined to provide the correct offset for the operational amplifiers 29 and 59 respectively to achieve the ideal Q points thereof. To perform this calibration, the sensor heating means is temporarily disabled so that inadvertent heating

of the precision resistors is avoided. Thus, the precision resistor 105a would firstly be connected across terminal 73a and 73b and the potentiometer 30a adjusted until an output voltage corresponding to the voltage of the ideal Q point is achieved at which time the resistance of the potentiometer 30a would correspond with the required offset to achieve this Q point. Subsequently, the precision resistor 105b would be connected across the terminal 73b and 73c and the potentiometer 60a adjusted in a similar manner to achieve the correct Q point voltage at the output of the operational amplifier 59. In this manner, the correct calibration of the apparatus 71 can be performed from time to time in the manner described above to ensure accuracy of the measurements obtained by the system.

In view of the fact that the Q point settings of the operational amplifiers 29 and 59 are-no longer calibrated to the slope and offset requirements of a specific pair of sensor and reference diodes, but rather to an ideal Q point setting, it is necessary for appropriate slope and offset compensation data to be provided with each pair of sensor and reference diodes to which the apparatus is connected so that the microprocessor system can make appropriate corrections. Accordingly, this data is stored within the EPROM 84 accompanying the sensor module 75 which can be read by the microprocessor system via its input data bus 87. This information is determined in a laboratory or the like and is equated to the ideal Q point setting to which the operational amplifiers 29 and 59 are calibrated. Accordingly, the EPROM 84 stores four words comprising appropriate slope and offset compensation data for the sensor zener diode 77 and slope and offset compensation data for the reference zener diode 79.

As should be appreciated, each sensor module comprising a different pair of sensor and reference zener diodes would have different data stored within the accompanying EPROM therefore providing the appropriate compensation data for the diodes with respect to the ideal Q point setting.

To determine these stored values, a procedure to calibrate the diodes in a common environment is shown schematically in Fig. 19.

Diodes 110 to be calibrated are placed in a multipin socket assembly 111 of which each socket connection is electrically connected to a switching unit 112 via cable means 113. An environmental chamber 114 is provided to receive the socket assembly 111. The temperature of the environment within the chamber can be regulated by a heating element 115 and a cooling element 116. A temperature probe 117 is arranged to provide an indication of the temperature of the environment within the chamber. A sampling current generator 118 corresponds low value constant current source (of a configuration similar to that which is employed as diode forward bias sampling current described previously) .

Microprocessor 119 performs various functions one of which is to switch a forward bias sampling current to the diodes under test in a predetermined sequence so that the respective forward voltages of the diodes at the instantaneous temperature of the environment within the chamber 114 can be recorded. A further function performed by the microprocessor 119 is to read the temperature of the environment within the chamber 114 using the temperature probe 117. A still further function of the microprocessor 119 is to compute the digital values of the respective offset value of each diode with respect to a Q point value, and also to determine the temperature

coefficient (or "slope" value) of the electrical characteristic of each diode. Relevant data from the microprocessor is printed out on a hard copy printer 120 and also written into a non-volatile memory device (EPROM) 121.

At the end of the calibration process, the microprocessor provides the operator with a message that identification means such as a tag 122 is to be applied to the diode in a designated socket position in the assembly 111 and similarly a further identification means such as a tag 123 to the diode in a further designated socket position thereby to identify the two designated diodes respectively as reference and sensor diodes.

When this procedure has been acknowledged to the microprocessor 119, it then writes the diode calibration data into the EPROM memory device 121 after which the EPROM 121 and two designated diodes (together with the printout from the printer) can be placed in a storage packet 124 for use together later in a measuring apparatus.

When a measuring apparatus fitted with a pair of diode and corresponding EPROM is in operation sufficient digital information relating to the electrical characteristics of the diodes will be known by the data stored in EPROM, and appropriate corrections to the operating parameters can be made based on this stored information.

Reference is now made to the graph of Fig. 20 in which the X-axis represents the operating temperature range of the diode expressed in 11 bit digital values of temperature, as determined by the temperature probe 117 and microprocessor 119 and the Y-axis represents the voltage detected across each diode 110 under test as detected by

the switching unit 112 and microprocessor 119.

A predetermined reference Q point will be defined in digital values to become the reference point in terms of environmental temperature and diode forward voltage as shown.

When the system described is activated, the environmental chamber 114 will be at an ambient air temperature (for this example represented along the vertical projection as shown) , and will produce a corresponding voltage for the particular diode at point A.

These values of diode voltage and ambient temperature are recorded in the microprocessors memory.

The ambient air temperature of the environmental chamber 114 is then raised to a value equal to the Q point reference temperature -as determined by the temperature reference probe 117. At this temperature, the voltage of each diode (represented in this case at B) is then subtracted from the Q point reference voltage value to leave a difference value of voltage equal to the offset value at the Q point reference temperature.

At a later time, when each diode is assigned a reference or sensor role, these values will then become either EHTD or ECOD _Q values (to be explained later) .

The ambient air temperature of the environmental chamber is then lowered, and the voltage of each diode is monitored to determine the temperature at which it attains a forward voltage value equal to the Q point reference voltage. This voltage for each diode will occur at slightly different temperatures. When the temperature value is subtracted from the point reference temperature

a "slope" or temperature coefficient value can then be related to the data obtained. This data will be assigned to the respective diode as either EHTD or ECOD-.

The triangle BCD thus formed from this procedure now has digital data to represent the side CD as offset data, and the side BD as sl ^' ope data.

In later processing in the velocity control circuitry, this data (which is stored in the EPROM) will be read into the control unit and the ratio of these two values computed so as to facilitate the calculation of an unknown side of a similar right angle triangle which will provide a correction to the operation parameters of the circuit.

Referring now to Fig 18 and describing the operation of the apparatus 71, calibration of the operational amplifiers 29 and 59 is performed to ensure that the ideal Q point settings are provided in each case, in the manner previously described. Subsequently, the apparatus 71 is connected to a sensor module 75 and the microprocessor system 85 reads the relevant slope and offset data from the EPROM 84 for the sensor and reference zener diodes 77 and 79 respectively. This data is utilized to initialise appropriate calculating algorithms stored within the microprocessor for the purposes of providing appropriate corrections to temperature data received from the -temperature compensation circuit in determining upper and lower threshold levels, and similarly for calculating fluid velocity data, from the fluid velocity measuring circuit.

Temperature data is converted from an analogue level at the output of the operational amplifier 59 to a digital level by the A to D converter 89. This digital level is supplied on the input data bus 87 to be read by the

microprocessor system. Appropriate multiplexing and de¬ multiplexing may be employed to avoid bus contentions with input data received along the data bus 83.

Fluid velocity data is derived from the outputs of the comparators 31 and 33 by the logic circuitry 103 in a similar manner as to that of the preceding embodiment, whereby the microprocessor system can calculate the mark space ratio and process the same to achieve a fluid velocity measurement by reading data on the input line 101.

The lower and upper thresholds of the comparators 31 and 33 respectively are determined by the microprocessor 85 by software calculation and appropriate digital levels are output on the output data bus 91 to be converted to analogue values by the D to A converter 93. These levels are multiplexed by the microprocessor calculating their values alternately and are input to the respective sample and hold circuits 95a and 95b. The microprocessor demultiplexes these levels by alternately switching the sample and hold circuits 95 by control signals provided on the control lines 99a and 99b respectively. Thus, the sample and hold circuit 95a is enabled to sample the output of the D to A converter 93 at the time when the lower threshold level is being output by the same and subsequently outputs this value along the lower threshold line 97a to the comparator 31. Subsequently, the sample and hold circuit 95a is disabled by the control signal at the output line 99a to hold the lower threshold level at the lower threshold line 97a. Simultaneously with the disabling of the sample and hold circuit 95a, the sample and hold circuit 95b is enabled by the control signal at the output line 99b of the microprocessor to enable the same to sample the analogue output at the D to A converter 93. During this time, the microprocessor has alternated

so that the upper threshold data is output by the D-to-A converter at this time. Consequently, the sample and hold circuit 95b operates in a ^" similar manner to the sample and hold circuit 95a to output the upper threshold level on the upper threshold line 97b to the comparator 33. Thus, the respective sample and hold circuits 95a and 95b are out of phase with each other so as to achieve correct de-multiplexing of the output of the D-to-A converter 93.

Now describing the nature of the calculations performed by the algorithms of the microprocessor 85 in further detail to achieve correct temperature compensation and fluid velocity measurement.

To accurately control the operation of the sensor diode, it is essential that precise fluid temperature information sensed by the reference diode be interpreted correctly, as it forms the basis of temperature compensation.

Two graphs showing the determination of the fluid temperature are shown at Figs. 21 and 22; one for a high ambient fluid temperature, the other for a low temperature.

The rotation listed below is used in the description and drawings to follow. The term "cold sensor" refers to the reference diode and the term "hot sensor" refers to the sensor diode.

Notation

ECOD _Q = Eprom COld sensor Data offset

This is the stored diode OFFSET value that will enable the offset of the actual diode "slope" (or COD _A value) to be moved vertically on the graph to allow the

"compensated slope" to pass through the Q point. At any temperature a point on the "compensated slope" is found by the addition of COD _A + EC0D _o .

ECODg = Epro COld sensor Data slope

During calibration, the environmental chamber's temperature is lowered until the voltage of each diode under test equals the Q point reference diode voltage (say 2FF) . This corresponds to a temperature difference (in digital terms) from the Q point reference temperature of ___£TMP. Each diode will have slightly different values of ECOD _anc j EC _O D . With ECOD _Q and ECOD _s data, the microprocessor system can compute the actual temperature of the fluid as shown in Fig. 21. This is explained more fully later.

CODA _Λ COld sensor Data _Actual

This is the actual digital cold diode voltage value that is representative of the fluid temperature data obtained from A-to-D converter 89 for that diode.

C ^O D _c = cold sensor Data compensated

Firstly a "calibration triangle ratio" requires to be computed so that a ratio between ECOD _Q and ECODg can be determined. This ratio is to be used later in the second triangle calculation to determine an unknown side value.

COD„ = (COD, - ECOD _n ) - 2FF

If the value of COD _c ± _s positive, then the computed value of T _QFF is subtracted from the reference temperature (4FF) and vice versa to determine the final value of T,

A*

^T OFF ⁼ Temperature _0FFset

This value is determined via similar triangles formula from EPROM data values and COD„_

^T OFF ⁼ ECODg x COD _c

ECOD _Q

The decision as to whether the resultant value is positive or negative depends on whether the addition of COD + EC0D _n is less than or greater than the sensor reference voltage (2FF) . If it is greater than 2FF, then the value of T _QFF is made positive as later it affects T _A accordingly.

T _A = 4FF + T _QFF

^T EQ = Temperature _EQuivale nt

This is a point that resides on the compensated slope that passes through the Q point. It also resides at the actual temperature of the fluid (T.) .

T _A = Temperature _Actual

This is the fluid temperature that is to be determined.

TC _REF = Cold sensor Temperature Coefficient REFerence

This is the compensated slope point that has been computed by the addition of

ECOD _Q + 2FF

at a point, a distance of ECOD along from the Q point. Although not a vital reference, it names a point at which the reference triangle is determined. A line drawn between this point and the Q point shows the compensated slope values.

2FF = Q point reference voltage (Hex.)

4FF = point reference temperature (Hex.)

The area within the bold lines in the graph of Fig. 21 indicates the limits within which the diode have been tested, and will produce an expected result. As example, an assumption has been made that most diodes will have a temperature coefficient of between 2.5mV/°C and 3.5mV/°C.

The Q Point established during testing becomes a focal point from where all measurements are referenced.

Firstly, the raw ambient temperature data from the diode (11 bit digital word form) is shown as a vertical value designated COD _A ) . To compensate its value to the Q point, the EPROM offset value ECOD _Q is added to it, thereby producing a vertical value at along that line, somewhere over the range of temperatures.

The slope of temperature of the diode can be determined in the laboratory characteristic by finding the ratio of ECODg and ECOD _Q . Triangle T ₂ is derived in the same manner as triangle T ₁ in Fig 20. As the triangle T ₃ at ^T _E Q is a similar triangle to triangle T ₂ , the value COD _c can be computed from known values and T _0FF can be computed using similar triangles formula. The value of T O__r_ can then be added or subtracted from 4FF to provide the actual temperature _& i _n hexadecimal form for hot diode computing data.

Where:

COD _c = (COD _A + ECOD ₀ ) - 2FF

^T OFF ⁼ ECODg X COD _c

ECOD _Q ⁷

T _A = ^• 4FF + _0Fp negative result of COD _c means T _QFF is ADDED to the 4FF temippeerraattuurree :reference. The reverse occurs if COD result is positive.

Fig. 22 shows the changes in geometry as the ambient fluid temperature reduces .

With reference to Fig. 23, this graph represents the geometry involved in determining the voltages that are applied to the upper and lower reference terminals of the comparators. These values equate to the voltage variations of the sensor diode, produced at the operating temperatures of the diode junction.

The notation listed below is used in the following description.

EHTD _Q = Eprom HoT sensor Data offset

This value is determined in the environmental chamber.

EHTD, Eprom HoT sensor Data ι _ope

This value is determined in the environment chamber.

Temperature _Actual

This is the value found by computations associated with the Cold Diode. It forms the basis for calculating the temperature compensation value from which other calculations are referenced to.

Lower Threshold Temperature

This is the lower temperature threshold value.

U„ Upper Threshold _Teιtιpera t _ure

This is the upper temperature threshold value.

L and U are reference points only

T. Temperature of the Junction offset from the ambient temperature

As a "hot diode sensor" has to work at a temperature elevated above the ambient fluid temperature ⁽ T _A ) , the software value assigned

to this variable becomes the fundamental value that determines the comparator reference voltages.

_ T = Limit variations between lower and upper temperature thresholds.

From the diagram, triangle T^ ± _s again derived in the same manner as triangle T ₁ in Fig 20, being produced from the EPROM Hot Diode sensor Offset EHTD _{Q an} d Slope EHTDg data. The ratio between these two values is only required during processing.

The Offset value is also subtracted from the Q point sensor voltage value, to produce a New Q point. This point- now lies on the actual operating characteristic of the sensor diode, from which the following points are determined.

From the New Q point, triangle T _g ( i _g 23) is derived from the formula for L shown below which determines the lower threshold voltage.

Triangle Tg is determined in a similar manner from the formula for U which also determines the upper threshold value.

The points L and U correspond to the actual voltages the sensor diode will produce, at the desired operating temperatures as shown.

If the comparator voltages correspond to the voltages of the sensor diode at these points, then a known HEAT/COOL ratio can be expected at the threshold values at zero velocity.

The ambient temperature of the fluid that was computed by the reference diode circuitry is shown at point . A "constant" termed T _j i _s added to _A to produce the lower threshold temperature value L _τ . to which the diode will cool. A similar procedure applies to the value at U^.

These operating temperatures of the sensor diode are independent of the amount of heating current applied to the sensor diode. The value of heating current will alter the period of time the diode will take to reach upper threshold temperature from the lower threshold temperature value only. In this embodiment, a fixed value of high value constant current will be supplied to the sensor diode for a known rate of heating.

The benefit of this process is that for. any diode with a differing temperature co-efficient and a different dynamic resistance value at a particular temperature, the comparator reference values- will compensate for the magnitude of these voltages proportionally by computing the respective EPROM data and applying corrective threshold values to the comparators as well as thereby providing a consistant heating pulse width and cooling period between different sensor assemblies. This is an important feature of this processing, as it quantifies all variables associated with the sensors, and continually adjusts the circuit configuration should any changes occur such as ambient fluid temperature variations for example.

Another feature of the hardware configuration (as well as the software processing) is that with the comparators having this essentially constant comparator reference voltage applied via software computation, the circuit is able to self-generate, thereby not requiring total processing dedication to enable sensor operation. This allows the processor free time to attend to other vital tasks associated with the function of the unit.

A method to determine the average DC value the HEAT/COOL signal would be to provide a software 'HEAT counter' and a 'COOL counter* which will commence counting at the beginning of a heating pulse and finish according to their respective times.

These software counter values would then be divided mathematically to acquire a ratio that can be related exponentially to fluid velocity values. A graphical representation of Heat to Cool periods versus fluid velocity is shown at Fig. 24, which is substantially similar to the graph shown at Fig. 14.

To acquire these counter values, the processor would need to be dedicated as a software counter during a HEAT/COOL cycle.

It is envisaged in this embodiment that cycles would alternate between velocity processing (processor is dedicated to this task) and housekeeping tasks.

For example the first HEAT/COOL cycle would involve the performance of housekeeping functions such as:-

(1) reading diode EPROM data

(2) reading reference diode temperature

( 3 ) compute temperature compensation factor

( 4 ) compute offsets

(5) reading the specific heat data of the fluid:

(a) For gaseous fluids: Each gaseous medium will produce a slightly different specific heat value which will take heat away from the sensor diode at a slightly different rate under certain conditions. Therefore, specific heat values may need to be "entered" that compensate for these factors. This data will affect the T _j or W _τ and H _τ values accordingly. Other factors also affecting the precision of the measurements are air or gas pressures and moisture content. If these values are likely to change considerably, then a suitable sensor diode to detect these environmental changes would need to become part of the system.

(b) For liquids: The specific heat value of various liquids such as water and heavy oil will be somewhat different. Either the product application will allow a menu of these values to be selected as required by the operator, so that the accuracy of the sensor between these liquids flows can be maintained.

(6) send offset values to upper and lower threshold controls for operation of the sensor diode at the appropriate junction temperatures.

(7) prepare to read the HEAT/COOL ratio to compute fluid velocity and fluid temperature.

The next HEAT/COOL cycle would involve the performance of actually reading the HEAT/COOL cycle and computation of the fluid velocity.

The third embodiment is directed towards a modification of the preceding embodiment, whereby the dynamic operating range of the sensor diode can be extended.

The hardware and software described in the preceding embodiment is expected to be highly flexible due to the nature of the design.

This feature will allow extended dynamic operating range of the sensor diode through software control of the value

Of _Tj .

The current prototypes have their sensor temperatures set fairly high to enable good sensitivity and a wide dynamic operating range. This, however, causes the sensor diode to be sometimes too sensitive at very low velocities, causing an artificial micro-environment of fluid- flow around the sensor.

However, if at zero velocity the temperature of the sensor diode was reduced to the lowest workable temperature, this would:-

- eliminate any artificial "micro-environments" around the velocity sensor diode that heat transfer instruments are known to produce

- reduce the electrical energy to heat the sensor diode at zero velocity.

The overall benefits of controlling the value T _τ proportional to fluid velocity is to linearise the response of the sensor diode by creating variable operating temperature ranges that can be dynamically controlled proportional to the fluid velocity.

This means the HEAT/COOL signal ratio is reduced back to near the zero velocity value even though the fluid velocity has increased, thereby extending the sensitivity and accuracy of the instruments operation.

Referring to Fig. 25, there is shown a typical HEAT/COOL signal A at a zero fluid velocity condition. The heating pulses are designated A and the signal has an average DC level indicated at the line designated DC1. Cooling of the sensor diode junction from its predetermined upper temperature threshold to its predetermined lower temperature threshold is indicated by the dotted line designated B. At this lower temperature threshold, a second heat cycle commences as before.

Where there is fluid flow, the heating duration of the HEAT/COOL signal increases with fluid velocity, as shown in Fig. 26 where heating pulses are designated by reference chamber C and cooling of the junction diode is indicated by dotted line D. The reason for the increased heated duration is that it takes longer for the junction to heat to the upper temperature threshold. Conversely, it takes less time for the junction to cool to the lower temperature threshold. The overall effect is indicative of an increase in the average DC level of the waveform as indicated by the line designated DC2. The increase in average DC level corresponds to the extra diode heating energy that is required to maintain the diode junction temperature constant at the increased velocity.

Fig. 27 is a plot of diode sensor current against time and shows a typical waveform across the sensor diode at a zero fluid velocity condition (corresponding to Fig 25) . The heating pulses are designated A' and the corresponding average DC diode (reverse) , heating current is indicated by the dotted line designated DC1 ' .

Fig. 28 is a similar plot to Fig. 27 except that it relates to a condition where there is fluid flow. As the velocity of the fluid increases, then additional heating pulses C would normally be required as to raise the average DC level to that at dotted line DC2 ' to provide additional heating energy to replace that taken away by the increased fluid velocity.

The dynamic operating range of the measuring system can be extended by the amount of reverse diode heating current (or the rate at which heat is applied to the diode sensor. This can be achieved using a programmable high value constant current source of the type described in a later embodiment.

To recover the operating dynamic range, involves returning the width of the heating pulses and extending the cooling period between them of the HEAT/COOL signal back to the values that were measured at zero fluid velocity, but at the same time increasing the heating current applied to the sensor diode to replace the amount of heat taken away from the sensor diode at that fluid velocity. This is accomplished through a combination of effects caused by altering the values applied to the programmable high value constant current source, and to both the upper and lower comparator reference voltages. These values will be computed through software means to compensate for the recovered dynamic operating range. A method of controlling the comparator reference values is described in a later embodiment.

With reference to Fig. 29, it can be seen that where the diode junction is heated to a value T_,_.» at zero fluid velocity, the sensor diode will be caused to heat up to a temperature at U__, and cool to a temperature at L _τ . The corresponding voltage comparator threshold values are shown accordingly.

With reference to Fig. 30, at zero fluid velocity the diode junction temperature is only a few degrees above ambient fluid temperature on an incremental family of curves on which it will operate on the lower (almost linear) portion of each curve as shown in bold.

As the fluid velocity increases above zero, the average DC value of the HEAT/COOL signal increases as does the average DC heating current applied to the sensor. A microprocessor system (of the type described in a later embodiment) will detect a predetermined DC value of the HEAT/COOL signal at which the diode temperature requires to be incremented to the next range as shown in curve 2 of Fig. 30.

At this point, new comparator voltages are computed by the microprocessor and applied accordingly, together with an incremented value to the programmable high value constant current source. This means that the junction now has an incrementally higher operating temperature which places it at the lower end of operating temperature curve 2 at the current fluid velocity value. At the same time the HEAT/COOL signal cycle is returned to its original heating/cooling periods thereby restoring the dynamic operating range.

In this manner, a higher value of heating current combined with a cooler temperature to which the junction is allowed to cool to (affected by the computed voltage to the lower comparator 31) enables all of the desired features to be achieved.

Fig. 31 shows how the increased heating current A" will compensate for the additional heating pulses C of Fig. 28 while still attaining an average DC level of DC2 ' (which

was achieved in Fig. 28) . Thus the required heating energy can be obtained without additional heating pulses simply by increasing the level of the heating current.

Fig. 32 shows how the sum of the linear portions of each incremental diode heating curve of Fig. 30 are linked together (in software) to produce the resultant, nearly linear, response of the sensor to fluid velocity variations over a broad operating range.

Preferably, software means will provide a hysteresia range around the threshold range increments of Fig. 30 which will ensure stable operation of the system during changes between incremental diode operating temperature ranges.

The by-products of this technique will be to:—

- utilise just the linear portion of each characteristic curve in the family of temperature characteristic curves as much as possible.

- minimise the use of software linearisation algorithm processes to compute the actual fluid velocity

- improve the sensor accuracy especially in the medium velocity ranges

- improve the sensitivity to detect fluid turbulence at medium and high velocities, to a value only seen currently at low velocities.

The fourth embodiment is directed towards a modification to the fluid velocity measuring circuit which simplifies design and operation if extended dynamic range of operation is not required.

As shown at Fig. 33 of the drawings, the fluid velocity measuring circuit 15 is essentially the same as that of

previous embodiments and accordingly the same reference numerals have been used to identify similar components. The principal modification to the circuit involves the removal of the lower threshold comparator 31 so that the output of the operational amplifier 29 is only connected to the non inverting input of the upper threshold comparator 33.

Additionally, the RS flip flop 35 is removed, and the output of the comparator 33 is connected to the input of the NAND gate 41. The output of the NAND gate 41 is connected to the high constant current source 25 and the meter driving circuit 43 in a similar manner as was previously the case with the flip flop 35.

In the case of applying the modification to the first embodiment, the output of the temperature compensation circuit 51 is connected .directly to the upper threshold potentiometer 39 for varying the upper threshold level of the comparator 33 following correct setting of the potentiometer.

In the case of applying the modification to the second embodiment the upper threshold level would be derived directly from the microprocessor system 85 following software calculation of the same in a similar manner to that described in the second embodiment.

Now describing the conceptual operation of the circuit incorporating the modification, previously the sensor zener diode 17 was heated by the high current heating source until the upper threshold value was reached at which time heating power to the sensor diode was removed allowing it to cool until the lower threshold value was reached. In the modification, however, the sensor zener diode is heated until the upper threshold level is reached

at which time the heating power to the sensor diode is removed allowing it to cool until a threshold value slightly below the upper threshold level is detected at which time heating current is again applied to the sensor zener diode.

Although only one comparator is now utilised, since sampling is not performed continuously but periodically in accordance with the SAMPLE/HOLD signal and also due to the intrinsic hysteresis effect displayed by the comparator 33 upon successive switching, a HEAT/COOL signal having discrete heating and cooling cycles will still be generated, albeit having a random grouping of heating current pulses. Thus, when sensor diode heating occurs during the Hold period of the SAMPLE/HOLD signal the temperature of the sensor diode junction will be elevated until it is allowed to exceed the upper threshold level of the comparator 33. After the Hold period has completed, the HEAT/COOL signal is switched to its cooling cycle by the comparator 33 changing state, causing the high current source 25 to cease the application of heating current through the sensor diode. Consequently, the sensor diode will be allowed to cool for the remainder of the hold cycle wherein sampling of the diode forward voltage can be performed during the sample cycle of the SAMPLE/HOLD signal. During the sample period the sensor diode will always be allowed to cool since the application of the HEAT/COOL signal to the high value constant current source 25 is disabled at this time to enable sampling to be performed. As shown at Fig. 35 of the drawings, if during the sample period the output of the operational amplifier 29 is still within the hysteresis region A of the comparator 33, then during the next hold period of the SAMPLE/HOLD signal the high current source 25 will remain disabled allowing further cooling of the sensor diode.

However, if during a particular sample period of the clock, the output of the operational amplifier 29 falls below the hysteresis region threshold, sensor heating shall be initiated again during the next hold period by the comparator 33 changing state again to create a heating period and applying the same to the high current source 25.

It should be appreciated that the hysteresis region represents variations in output voltage of the amplifier in the order of micro volts or fractions of a degree in temperature. This manner of operation is more clearly shown graphically at Fig. 36 of the drawings.

The conditions which affect how many successive heating periods are grouped together and how many SAMPLE/HOLD cycles set these periods apart is determined by several variables of the fluid in which the sensor zener diode is located. If the specific heat value is very high (e.g. very dry air at zero velocity creating a poor heat transfer medium) , the sensor diode will heat up very quickly generating very few successive heating periods , and subsequently have a prolonged cooling period. Conversely, if the specific heat value is very low (e.g. very humid air at high velocity creating a good heat conducting medium) , many heating periods will be required with shorter duration cooling periods separating them.

With regard to temperature compensation, this will be applied in the same manner as was previously the case. Thus, as shown at Fig. 36 of the drawings, as the ambient fluid temperature increases, the average D.C. heating/cooling operating point of the amplifier will increase accordingly towards the negative saturation level of the amplifier output.

Now having particular regard to the application of the modification to the second embodiment and the implications this has insofar as software calculations of changes in threshold level and computation of fluid velocity are concerned reference should be made to Fig. 37. This drawing is substantially the same as Fig. 18 where the only differences involve the removal of the lower threshold comparator 31 and the adoption of a different interface between the microprocessor system 85 and the output of the comparator 33. The latter change is brought about by the fact that the HEAT/COOL signal is no longer of a uniform nature where velocity calculations can be made by merely comparing the ratio of the heat period with the cool period of the signal, but rather these heat and cool periods appear in a non uniform randonly occuring manner requiring an integration to be performed to obtain ^' a value representative of the fluid velocity. This is best ^' shown at Fig. 38 of the drawings which compares typical HEAT/COOL signals generated in the case of the second and present embodiments respectively. Accordingly, a binary up counter 127 is provided to count the number of random heating periods which occur over a given period of time. The binary up counter 127 consequently has its counting input directly connected to the output of the NAND gate 111 and has its various control lines 115 connected to the microprocessor system 85 to control the count process. Additionally, the output data bus 91 of the microprocessor system 85 is connected to the pre-load input lines of the up counter 127 as well as the input to the D-to-A converter 93. Consequently, the output of the data bus can be demultiplexed so as to pre-load a digital value into the up counter 127 which is representative of a predetermined count value corresponding to sensing zero velocity. This predetermined value is determined separately of the operation of the system and is

programmed into the microprocessor during initialisation and/or system calibration. This pre-loaded value in one embodiment is entered as a negative digital value, whereby the up counter 127 increments this preloaded count value with the occurrence of successive heating periods upon commencing a count period. Thus, when the counter value exceeds zero during the count period, this will indicate a fluid velocity greater than zero the magnitude of which is directly proportional to the magnitude of the count in excess of zero. Subsequently, the algorithms stored within the microprocessor system 85 calculates the actual fluid velocity from this data.

The output data bus as shown is still connected to the D-to-A converter to load upper threshold values to the same for the purpose of setting and varying the upper threshold level for the comparator 33 in accordance with variations in the ambient temperature as determined by the temperature compensation circuit. Since only a single comparator 33 is used in this embodiment, it is no longer necessary to de-multiplex the input to the D-to-A converter to obtain upper and lower threshold values, thereby simplifying the design of the circuit for this purpose. However this embodiment should not be restricted to the incorporation of one comparator only. If extended dynamic range of operation was desired, then two comparators would be required.

It should be noted that determination of this threshold level of the comparator will be performed in precisely the same manner as was the case in the second embodiment.

In the description of the preceding embodiments, there is an underpinning assumption that pressure and relative humidity of the air remains constant. In practice, although it is usual to maintain a constant pressure in

performing fluid velocity measurements, however it is not usual for the relative humidity to remain unchanged with changes in the ambient temperature of the fluid flow. As can be seen by the psychrometric chart at Fig. 39 of the drawings, relative humidity is a function of the grams of moisture per unit mass of dry air, and air temperature. Accordingly, if the amount of moisture in the air remains constant, which is generally the case, changes in temperature result in varying levels of relative humidity. However, in _order-..t& achieve a constant relative humidity with changes in temperature, it would be necessary to vary the amount of moisture in the air as shown in the graph for the chosen level of relative humidity, but this is almost an impossible condition to achieve. Thus, in most applications of the invention where fluid velocity is desired to be measured in an environment where the ambient air temperature can change, variations in the ambient temperature cause corresponding changes in the relative humidity or density or specific heat value of the air thus, in the case of the sensor and reference zener diodes, as the relative humidity or density of the air increases or decreases, the ability of the leads of the diodes to distribute heat from the junction thereof correspondingly increases and decreases, thereby affecting the duration of the heating and cooling periods of the heat/cool signal as shown graphically in the drawings at Fig. 40. Accordingly, it is necessary to compensate for changes in the relative humidity or specific heat of the air in addition to compensation of changes in the temperature in which the sensor and reference zener diodes are placed. Accordingly, the fifth embodiment is directed towards a modification of the apparatus as described in the preceding embodiment which is in turn applied to the second embodiment.

As shown at Fig. 41 of the drawings, the modification involves the utilisation of a programmable high value constant current source 131 in place of the high value constant current source 25 to vary the magnitude of the current applied to the sensor zener diode for heating purposes in accordance with a programmed value calculated and applied to the current source 131 by the microprocessor system 85. Thus, the programmable high value constant current source 131 is also connected to the output data bus 91 of the microprocessor system 85 to obtain the programmable value of constant current to be applied to the sensor zener diode 17 from data multiplexed with the pre-loaded count value for the up counter 127 and the upper threshold level for the comparator 33 on the output data bus.

If it is assumed that the grams of moisture in the air emains constant and there is a decrease in the ambient temperature, there will be a corresponding increase in relative humidity, this will result in an increase in the duration of the heating period and a decrease in the duration of the cooling period of the heat/cool signal which causes a net increase in the DC level of the same. Accordingly, if the microprocessor system can calculate the proportional increase in the DC level which is attributed to an increase in the relative humidity alone then the remaining proportion of the DC level can be attributed to fluid velocity. In order to determine this reference relative humidity component, the present modification correlates a measured DC level at zero velocity to that which is expected at zero relative humidity and at zero velocity to arrive at a determination as to the grams of moisture in the air at the given temperature. Knowing this value and by maintaining the duration of the heating period at a constant, a reference level can be obtained and adjusted with ongoing changes in

ambient temperature and relative humidity to enable fluid velocity calculations to be determined.

To assist in understanding the concept, reference should be made to Figs. 42 and 43 of the accompanying drawings. At Fig. 42 the resultant heat/cool signal 1 is shown in an environment of zero relative humidity, and the heating period is therefore at a minimum and the cooling period at a maximum as the dry air has thermal insulation properties. This provides an effective minimum DC level at the line indicated at DC 1. At near to 100% relative humidity, the heating period 2 is at a maximum and the cooling period at a minimum resulting in a net DC level indicated at line DC2; by increasing the diode heating current 3 the duration of the heating period at near 100% relative humidity can be reduced to that at zero relative humidity. This causes a reduction in the heating period which reduces the resultant DC level below DC2 but above DC1. This new level DC 3 can be computed to determine the specific heat value of the air relative to the dry air diode heating value at zero air velocity. Consequently, if the microprocessor system is programming the programmable constant current source with the required value to achieve a heating period commensurate with that for a zero relative humidity heating duration and the resultant DC level of the heat/cool signal is greater than the level at DC 3, then this difference can be attributed to a fluid velocity component which can consequently be calculated by the microprocessor system in the usual manner. If however the microprocessor system has' programmed the programmable constant current source with another value which is now different from the predetermined value it was previously calibrated for to achieve a heating period commensurate with that at zero relative humidity, then this indicates that there has been a change in relative humidity which can be precisely

calculated by the microprocessor given that the grams of moisture in the air has not changed and that the reference zener diode is measuring the ambient temperature by reference to the psychrometric chart which can be stored in the microprocessor system in look-up table form or other conventional means. Thus, calculating the new relative humidity of the environment enables a new DC reference level to be determined representative of the environment at zero velocity. Consequently, any difference in the actual DC level and the reference DC level can be attributed to a fluid velocity component which can subsequently be calculated.

Now describing the process in which the apparatus would require to be . calibrated to enable relative humidity compensation, it is necessary first to create a zero velocity environment around the sensor diode. Upon achieving this condition, the microprocessor system 85 sends a diode, heating value to the programmable current source which applies a calibration heating current at zero relative humidity and zero velocity, which would be required to heat the sensor zener diode to the upper threshold level as computed at the fluid temperature measured by the reference sensor. The microprocessor then determines the duration of the heating cycle which in any environment having greater than zero relative humidity, would result in a heating period of increased duration. Moreover, the microprocessor system determines whether the actual DC level obtained is the same as that which is expected for zero relative humidity. This difference in DC level can then be equated to a particular relative humidity by the microprocessor system whereby specific relative humidity which has been empirically determined for differing DC levels and are stored in a look-up table or the like in a microprocessor system. The psychrometric look-up table can then be referenced to determine the

grams of moisture per kilogram of dry air (or specific heat of the air) in the environment and which value is subsequently stored for future use. The microprocessor system then calculates the actual diode heating value required to be programmed into the programmable constant current source 131 to reduce the expanded heating period to achieve a heating period commensurate with that at zero relative humidity. This will cause the resultant DC level of the heat/cool signal to reduce towards the DC level at zero relative humidity, although the resultant DC level will in fact be higher than that at zero relative humidity if a greater than zero relative humidity condition exists. This resultant DC level becomes the reference level for the purposes of velocity calculations at the given relative humidity. Subsequently, if the sensor zener diode is exposed to a fluid flow at some velocity, resultant changes in the DC level of the heat/cool signal compared with the reference DC level can be attributed entirely to the fluid velocity. Consequently, the fluid velocity can be calculated from this DC level difference in the manner previously described.

If subsequently the ambient temperature changes and these changes being detected by the reference diode and temperature compensation is taking place, and the grams of moisture in the environment remain constant, a corresponding change in the relative humidity will occur causing a change in the DC level. Thus, the previous reference level can no longer be used to compare the actual DC level against for the purposes of calculating velocity, and a new reference level corresponding to the new relative humidity of the environment must be determined. This is achieved automatically by the microprocessor system altering the diode heating value sent to the programmable constant current source to maintain the duration of the heating period commensurate

with that at zero relative humidity. Given that the grams of moisture in the environment have not changed, the temperature measured by the reference zener diode can be used to determine precisely the new relative humidity of the fluid by way of the psychrometric chart stored in the look-up table of the microprocessor system. This relative humidity can then be equated to a new DC reference level empirically determined (the reverse of the previous determination of relative humidity at the calibration stage) for the case where 'the heating period is commensurate to that at zero relative humidity and the actual DC level compared against this new DC reference level to determine the fluid velocity as was previously the case.

It should be appreciated that the invention described in the preceding embodiments has many advantages over previously known devices for measuring fluid characteristics.. In particular, in the case measuring the velocity characteristic of a fluid flow, the velocity of the fluid at a point in the flow can be precisely determined. Accordingly, the apparatuses described in the second and subsequent embodiments lends itself to use with a multiplicity of sensor and reference diodes which can be positioned at precise locations across a fluid flow within a pipe or duct to achieve a velocity profile. In this manner the apparatus would be switched from the output velocity data of one -sensor diode system to another sequentially so that the microprocessor system can determine velocities progressively from one diode system to another. The resultant information can then be stored for later reference or displayed in graphical form on a visual display unit to enable real time or recorded analysis of the cross-sectional fluid flow in the pipeline or duct. An example of such an application is shown at Fig. 44 of the drawings whereby a series of sensor zener

diodes 132 can be mounted in a line across an aerofoil 133 diametrally disposed within the pipe 135. A single reference zener diode 137 may also be mounted within the aerofoil to provide ambient temperature measurements for each of the sensor zener diodes accordingly a ribbon cable (not shown) is interconnected between a microcomputer system 139 and each of the sensor control unit outputs, to enable sequential communications therebetween and calculation of fluid flow velocities from each diode sensor circuit.

It should be appreciated that the scope of the present invention is not limited to the particular embodiments herein described. In particular, the apparatus is not restricted to calculating velocity characteristics but can be adapted to calculate any unknown characteristic of the fluid such as temperature, humidity or pressure, given the other characteristics. Additionally, like a processor system, it is. not necessarily restricted to calculating fluid velocities and compensating for temperature and humidity variations in the manner described herein, but may adopt alternative techniques that would be common knowledge to a person skilled in the field of the invention.