CA1251948A - Improvements relating to solid state anemometers and temperature gauges - Google Patents

Abstract

- 1a -ABSTRACT OF DISCLOSURE

There is disclosed a solid state temperature and/or anemometer measuring instrument for a fluid medium such as atmosphere the instrument including a pair of semi conductor junction means such as a zener diode one of which is exposed to fluid flow the other being isolated therefrom and providing outputs which are a function of the tempera-ture of the fluid and a measuring device for measuring the respective outputs the differential being an indication of temperature or fluid speed. Such a device has specific application in an optical air pollution monitor to give a continuous indication of fluid flow thus ensuring that the monitor is sampling fresh batches of fluid from a monitored space.

Description

~5 1~

IMPROVEMENTS RELATING TO SOLID-STATE
ANEMOMETERS AND TEMPERATURE GAUGES
The present invention relates to a solid state anemometer or a temperature gauge and in particular to anemometers for use in air-pollution monitoring techniques.
The anemometer is an instrument for measuring the speed of air, whether for meteorological data, ventilation testing or other purposes. The majority of anemometers comprise some form of impeller which rotates due to air 10 movement. The speed of rotation is calibrated to provide a measure of air velocity. Such anemometers consist of mechanical parts and consequently are costly tD manufacture, are subject to wear and damaqe which reduces service life, and are difficult to miniaturize. Furthermore, because 15 energy must be imparted to the impeller to overcome .frictional losses, there is a restriction to flow when mounted within pipes.
More recently, solid-state anemometers have been developed for use in ventilation testing. These 20 utilize a pair o~ thermistor bead elements wired as a compensating circuit. It should be noted that the thermistors are passing current and are consequently dissipating heat, placing each thermistor at a modestly elevated temperature relative to ambient. This is essential 25 so that air flow may cause cooling of the sampling thermistor. One thermistor is exposed to the air flow (sampling element), whilst the other thermistor is protected in ~till air at ambient temperature (reference element).
Ideally, under conditions of zero air speed, the thermistors 30 operate at equal temperatures and thus, the resistance value of each thermistor is equal and the compensating circuit is in balance (zero differential output). When air flows, the sampling thermistor is cooled, which alters its resistance value, causing the circuit to become unbalanced.
35 ~ccordingly the differential output voltage from the circuit can be calibrated in terms of air flow. The intention of the

2 ~

reference thermistor and the compensating circuit is to allow the anemometer to compensate for variations in ambient temperature. However the accuracy of this compensa-tion has been poor, requiring manual adjustment of the zero setting before each series of air speed measurements is taken. In a hand-held anemometer, this is an inconvenience but not a major problem.
Further disadvantages of thermistor beads have been experienced. The fine connecting wires are subject to damage, fatigue and corrosion, whilst the bead resistance has exhibited a susceptibility to ageing. In addition, thermistor beads are quite expensive and are very difficult to handle during anemometer assembly.
A new application for anemometers has given rise to a new method for measuring air speed, which provides adequate co~pens~ation for temperature drift and is compact, robust, inexpensive, simple to manufacture, contains no moving parts, is miniature and provides negligible restric-tion to air flow.
An optional air-pollution monitoring technique has been adapted for the purpose of detecting the earliest traces of smoke from incipient Eires, thereby to reduce the loss of life and property caused by established fires. This is especially applicable to situations where modern synthe-tic materials may become involved, because of the-highly to~ic fumes liberated.
It i~ required that representative samples of the air within each part of the fire detection zone, be passed through the smoke detector, such that the air quality can be continuously monitored. Such monitoring may be achieved by a mechanically-aspirated pipe network or more simply, by coupling to the return-air register of an air-condition-ing system.
In éither case, any interruption to the air flow through the smoke detector must be detected, lest a dangerous fire remain undetected.
In the pipe network case, the possible restric-tion to air flow caused by an anemometer is an important ~2S;~g~8 consideratlon, since an aspirator of minimum energy input is necessary, to enable continued operation from modest-capacity standby batteries, in the event of mains failure.
A solid-state anemometer is known utilizing a ther-mistor compensating circuit to detect a loss of air flow, thence to create an appropriate alarm signal. However such an insitu anemometer must be capable of compensation for wide variations in ambient air temperature, in many of the appli-cations of the smoke detector.
The invention will be described in greater detail having reference to the accompanying figures in which:-Figures 1 to 3 illustrate known thermistor circuits and graphs of differential output voltages plotted against temperatures;
Figure 4 is a graph of test results on zener diodes having a bias current set to produce a uniform power dis-sipation of 80 milli watt (mW);
Figure S is a selector guide to assist selection of a zener diode for a required temperature coefficient;
Figure 6 shows a representative sample of the temp-erature characteristic of numerous zener diodes;
Figure 6(a) is a circuit for temperature measure-ment;
Figure 7 is a graph of differential output voltage against temperature of zener diode compensating circuit;
Figure 7 (a)is a schematic diagram of a zener diode compensating circuit;
Figures 8 and 9 show examples of anenometer con-struction;
Figure 10 shows a typical anemometer calibration curve for a zener diode anemometer;
Figure 11 is a block diagram of a simple arrange-ment for the flow meter with analogue output fed to a thres-hold alarm means (which may be remotely located);
Figure 12 is a circuit diagram of the zener diode ,.~

~2~

anemometer utilising a comparator for switching of the alarm output.
Description of Prior Art -The graphs of Fig. 1 show the differential output voltage from a compensating circuit, consisting of a typical pair of precision thermistors. Under conditions of zero air flow, the differential output voltage is markedly dependent upon temperature. This dependence becomes yet greater, under conditions of maximum air flow (as determined by the given pipe network and aspirator). The vertical voltage excursion between these two curves, at any given ambient temperature, represents the response of the thermistor anemometer to all air speeds between zero and maximum.
The switching threshold, below which the alarm sig-nal is to be generated, has been placed for convenience and is normally adjustable. This threshold switching function is normally achieved by means of a comparator circuit. At the typical ambient temperature of 20C for example, the alarm would be generated should the differential output voltage fall to about one-third of the possible voltage excursion. It should be noted that because of the non-linear response of the thermistor anemometer to air speed, this one-third volt-age could represent an air speed of only 10 to 20% of the maximum.
Such a span of operation would be acceptable, prov-ided the ambient temperature deviation was small, in the region of 5 to 10C (e.g. air-conditioned area).
However, should the ambient temperature rise by 20C or more, it would become impossible to exceed the alarm threshold and an alarm signal must be generated, irrespective of air speed. Furthermore, should -the ambient temperature fall by 20C or more, it would become impossible to reach the alarm threshold and an alarm signal could not be generated, irrespective of air speed. For practical purposes, the full span of operation of such an anemometer would be rather less ~- .
.:'~ "

- z~

than 40C, rendering it unsuitable for areas which are ex-posed to seasonal temperature changes. To factory-set the al-arm threshold, without knowledge of the product destination, would be largely pointless.
More significantly, the use of dust filters in as-sociation with the smoke detectors has led to a requirement for sensitive setting of the alarm threshold, according to the actual flows achieved in the field, such that the partial-blocking of the filter can be detected. It is not sufficient merely to detect a total air flow loss. To detect for exam-ple, a 30% loss in flow, accurately across a range of -20C
; to +50C, requires a substantial jump in the accuracy of temperature compensation.
~ The reason for the poor performance of the thermis-lS l tor design can be readily understood by reference to Figure 2 of the drawings which shows the individual output voltages from each thermistor, versus bead temperature of each therm-istor, assuming a supply voltage of 15V DC and a series res-istor of 2k Ohm. Curves A and B represent the worst-case tolerance extremities for the precision thermistors used.
Fig. 3 illustra-tes the differential output voltage of a compensating circuit comprising thermistors A and B. In the case where A and B are malntained at the same tempera-ture, the differential output voltage (A-B) is relatively flat. This represents the zero air flow situation. In the case where two identical thermistors are chosen (A-A) or (B-B), the curve would be a horizontal straight line at zero volts. However if one thermistor is exposed to air flow, such that its temperature is reduced by 20C below the other for example, a very steep curve (A-B) results, even when perfect-ly matched thermistors are used (A-A) or (B-B).
In summary, under conditions of imperfectly matched thermistors, the temperature drift relative to flow signal (differential voltage), at an ambient temperature of 20C and high air flow is about 1.6%/C. The difficult task of per-. ~ , fectly matching the therMistors would afford no improvement.
The reason for this poor performance is a funda-mental limitation of the thermistor approach. Namely, it is the inherent non-linearity of the resistance vs. temperature curve. This curve is conveniently expressed in the form:
R = A*exp(K/(T+273)) Where R is the resistance and T is the bead tem-perature (Celsius). A is a constant understood to lie within the range 0.01293 to 0.03884 and K is a constant understood to lie within the range 4096 to 3681, as determined by the manufacturing process of the particular brand of thermistors chosen.
It can be seen that the dynamic resistance is a function of temperature. Because the two thermistors would operate at differlng temperatures (depending upon air speed), they operate at different parts of this curve. Their dynamic resistances are different and consequently it becomes impos-sible for one to compensate for the other.
Intuitively therefore, any solution seemed to req-uire temperature-sensing elements which had a linear res-ponse.
Integrated circuit temperature-sensing elements have been developed, which have a quite linear response.
However, these are costly and are comparatively large, which can present difficulties in miniaturization and restriction to air flow. They also have a significant thermal time-constant. Accordingly, an element of low complexity was sought, such as a semi-conductor junction.
Forward-biased silicon diode junctions are known for their essentially linear temperature characteristic, com-monly 2.2 mV/C. Accordingly they have been used in tempera-ture measurement, although their application to air speed measurement is not previously known. Nevertheless, the for-ward voltage-drop of a silicon diode is approximately 0.6V, and to dissipate sufficient power to raise the chip tempera-ture, such that adequate sensitivity to air flow would be achieved, would require substantial current flow, in the .

i . .

~S~9~3 order of lOOmA or more. This was unacceptable because the smoke detector as a whole must be highly energy-efficient, to enable operation from modest-capacity standby batteries, in the event of mains failure. Accordingly this avenue was abandoned.
It is an objective of the present invention to pro-vide a solid state anemometer which is accurate, has low pow-er consumption and is substantially independent of tempera ture variations over its working range.
A further objective is to provide a solid state temperature measuring instrument.
In one aspect of the invention there is provided a solid state anemometer including a pair of zener diodes con-nected in a compensating circuit configuration, both diodes being under the influence of a fluid, means for generating a fluid flow, the other diode being isolated from said fluid flow, said diodes being adapted to exhibit a voltage output - substantially linearly dependent upon temperature of each diode, means for measuring the voltage differential, if any, existing between said diodes caused by the cooling effect of said fluid flow, if any, said measured differential being an indication of fluid flow speed.
In another aspect of the invention there is pro-vided an optical air pollution monitor comprising the use of a solid state anemometer or temperature measuring instrument.
A solid state temperature measuring instrument for use in an air pollution monitor is also provided.
In a still further aspect, the invention provides a differential temperature measurement circuit comprising:
a bridge circuit;
a source of bias supply connected to opposite cor-ners of said bridge circuit, to thereby define two branches in shunt to said bias supply;
each branch essentially having, in series, a cur-rent limiting resistor and a zener diode poled so that its . .
. .
.

~LZ~
- 8a -P-N junction is back-biased by said bias supply;
means for connecting the conductor between said current limiting resistor and the associated zener diode of each branch to respective output circuits whereat a differ-ential output voltage can be sensed;
the values of the said current limiting resistors, and the breakdown voltage-current characteristics of said zener diodes being substantially identical;
the voltage of said bias source being sufficient to cause a small reverse breakdown current of substantially the same amount to flow through each of said zener diodes;
whereby, when the temperature of said two zener diodes is substantially the same, over a considerable range of ambient temperatures, the said differential output voltage is substantially negligible; and whereby, as the temperature difference between said zener diodes increases, over said same considerable range of ambient temperatures, the said differential output voltage varies substantially and correspondingly in linear relation to said temperature difference.
The use of zener diodes in temperature measurement, let alone air speed measurement, is not previously known to the applicant. They are normally operated in reverse-bias and little detail is published about their temperature char-~5 acteristics. Depending upon the individual-------------------"` ' ~zs~

zener diode's breakdown voltage, the temperature character-istic is known to change in magnitude and sign.
Tests were conducted and the averaged results are graphed in Fig. 4. Each bias current was set to produce a uniform power dissipation (80 mW), to ensure substantially consistent junction temperatures. This demonstrates that the gradient is negative for a 3V device, becoming zero at about 4.7V, and increasingly positive at higher voltages.
Most importantly, it can be seen that within the accuracy of measurement (lOmV), a linear relationship exists.
Fig. 5 becomes a selector guide, whereb~ the appropriate zener diode operating voltage may be chosen according to a required temperature coefficient. A straight-line approximation to the curve yields a simple relation, the accuracy of which is largely consistent with the spread of results obtained thus far:
G = 1.2*(Vz-4.6) or, Vz = 4.6-~(G/1.2) where G is the therma] gradient (mV/C) and Vz is the nominal voltage of the zener diode.
In view of a 15V DC supply constraint, it was decided to ~select nominally 8.2V zener diodes for more rigorous testing. For any given required power dissipation, the necessary bias current would be nearly 14 times less than for a regular (forward-biased) diode.
Fig. 6 shows a representative sample of numerous zener diodes tested at random. All were nominally 8.2V but two different brands are represented. The linearity and correlation was most promising, with gradients close to 4.0 mV/C. In all cases a bias current of lOmA was used.
To take worst-case tolerance extremities, devices A and H were selected for use in the circuit configuration shown in Fig. 7(a). Fig. 7 shows that the differential out-put voltage (A-H) is quite linear, with virtually no change in gradient. For example when a 20C temperature differen-tial is applied between junctions (because the sampling element is cooled by air flow), there is virtually no change in the gradient (0.38 mV/C becomes 0.37 mV/C as shown).

In the case of perfectly matched devices ((A-A) or (H-H)), the differential output becomes completely independent of temperature. Unlike the thermistor situation~ matching of zener diodes is a relatively simple process (because of the hitherto unconfirmed linearity of voltage variation as against temperature; and also the rugged construction).
This linearity is therein confirmed by the above test results.
The differential output voltage would normally be presented to a differential amplifier, with conventional DC offset adjustment, thence to an alarm threshold switching device (comparator).
In_Summary Under conditions of imperfectly matched zener lS diodes, the temperature drift relative to flow signal (differential voltage), at any ambient temper'ature~in the range of at least 0 to 50C, and high air flow, is 0.4%/C.
This worst-case condition represents a fourfold improvement upon the thermistor method, however, it has been surprising-ly discovered the zener diodes may be relatively simplymatched to whatever accuracy may be necessary to achieve any greater improvement sought by the simple process of selecting matched pairs of zener diodes.
With reference to Figure 8, construction of the anemometer circuit shown results in a sampling element such as a zener diode 10 being exposed to the air stream, and with a reference element such as a zener diode ll~substan-tially protected from air currents yet exposed to ambient temperature. In this arrangement, a sealed tube 13 is inserted into the wall of the pipe 12 conducting the air stream into the smoke detector (not shown). The glass hermetically-sealed zener diodes and connecting wires are mechanically and electrica,lly secured to a printed wiring board 14 fixed within the tube 13. Bias resistors and subsequent electronic circuitry are located remotely from the tube 13 in this example. The reference element 11 is protected deep within the tube 13, whilst the sampling

3~Z5~

element 10 projects into the pipe. Alternatively in Fig.
9 a similar result is achieved using thick-film microcircuit construction where semi-conductor elements, zener diode chips are bonded to a ceramic substrate 20, and conformally coated with a suitable protective material.
In a modified arrangement the zener diodes are operated at different bias currents. In normal operation the sampling element would be cooled by air flow. By contrast, the confined space of the reference element would lead to a rise in temperature (oven effect). Thus it is deemed appropriate to operate the reEerence element at a lower bias, such that in normal operation, the elements attain a similar junctlon temperature. Whilst the tempera-ture characteristic is linear within the accuracy of measurement, maintaining a similar junction temperature could only improve their tracking with ambient temperature.
More importantly, because the bias current is primarily determined by the sampling element in terms of overall flow sensitivity, reduction in reference bias aEfords an oppor-tunity to reduce the overall current drain of the anemome-ter. Experimental results suggest an optimum ratio of 2:1.
Fig. 10 shows a typical anemometer calibration curve, for a zener diode design, with bias currents optimised for air speeds in the region of lm/sec. This curve is quite linear for much of its range, simplifying the adjustmellt of the subsequent alarm threshold circuitry.
Fig. 11 is a schematic block diagram of the anemometer comprising compensating circuit 40 including zener diodes ZS and ZR in a resistive circuit as illustrated in ~igure 12. The differential voltage output is applied to a differential amplifier 41. The output of the amplifier 41 i~ applied to an optional comparator 42. An analog indication of air speed is provided at the output of the differential amplifier 41, whereas the optional comparator 42 provides a set point to generate an alarm which could be used to signal excessive, or inadequate flow.
Naturally the application of this anemometer ~s~

method is not limited to smoke detectors. It may be extended to many diverse fields of fluid flow measurement, including chemical processing, aviation, meteorology, air-condition-ing, automotive, temperature measurement of fluids etc.

Claims (11)

THE EMBODIMENTS OF THE INVENTION IN WHICH AN EXCLUSIVE
PROPERTY OR PRIVILEGE IS CLAIMED ARE DEFINED AS FOLLOWS:

1. A solid state anemometer including a pair of zener diodes connected in a compensating circuit configuration, both diodes being under the influence of a fluid, means for generating a fluid flow, one of said diodes adapted to be exposed to said fluid flow, the other diode being isolated from said fluid flow, said diodes being adapted to exhibit a voltage output substantially linearly dependent upon tempera-ture of each diode, means for measuring the voltage differ-ential, if any, existing between said diodes caused by the cooling effect of said fluid flow, if any, said measured differential being an indication of fluid flow speed.

2. An anemometer as claimed in claim 1 wherein the said other diode is exposed to fluid of substantially the same temperature.

3. An optical air pollution monitor for the purpose of detecting the earliest trace of smoke from incipient fire in an area under surveillance, comprising the use of a solid state anernometer as an air flow detector, the anemometer in-cluding a pair of zener diodes connected in a compensating circuit configuration, one of said diodes adapted to be ex-posed to a fluid flow in the monitor, said fluid being sourced from the area under surveillance which may be subject to pollution, the other diode being isolated from said fluid flow but being in fluid having substantially the same temper-ature as the area under surveillance, means for measuring the voltage differential between said diodes, said differential being an indication of fluid flow speed and therefore an in-dication of continuous fluid sampling by said air pollution monitor of the area under surveillance.

4. For use in an air pollution monitor the use of a solid state temperature measuring instrument comprising a pair of zener diodes each exposed to a fluid medium being monitored, one diode being exposed to fluid flow the other not, each diode adapted to exhibit a voltage output indica-tive of operating temperature, which is substantially linear over at least the operating range, difference amplifier means measuring the voltage differential between the two diodes thereby indicating whether there is adequate flow of fluid to said pollution monitor.

5. The invention as claimed in claim 4 wherein said one exposed diode is positioned in a sampling chamber of the pollution monitor and the other diode is insulated from fluid flow but senses the ambient temperature of the fluid.

6. For use in air pollution monitor the use of a solid state temperature measuring instrument comprising a pair of zener diodes each exposed to a fluid medium being monitored, said diodes drawing a comparatively low current thereby enab-ling long term battery operation, one diode being exposed to fluid flow the other not, each diode adapted to exhibit a voltage output indicative of operating temperature which is substantially linear over at least the operating range, dif-ference amplifier means measuring the voltage differential between the two diodes thereby indicating whether there is adequate flow of fluid of said pollution monitor.

7. The air pollution monitor of claim 3 in which each of said zener diodes include, as circuit elements, a single P-N junction and in which each of said junctions is heated by a small reverse bias current such that the heating effect is substantially the same.

8. The air pollution of claim 7 in which the small reverse bias current is of such value as to dissipate heat each of said P-N junctions on the order of 80 milliwatts.

9. A differential temperature measurement circuit com-prising:
a bridge circuit;
a source of bias supply connected to opposite cor-ners of said bridge circuit, to thereby define two branches in shunt to said bias supply;
each branch essentially having, in series, a cur-rent limiting resistor and a zener diode poled so that its P-N junction is back-biased by said bias supply;
means for connecting the conductor between said current limiting resistor and the associated zener diode of each branch to respective output circuits whereat a differ-ential output voltage can be sensed;
the values of the said current limiting resistors, and the breakdown voltage-current characteristics of said zener diodes being substantially identical;
the voltage of said bias source being sufficient to cause a small reverse breakdown current of substantially the same amount to flow through each of said zener diodes;
whereby, when the temperature of said two zener diodes is substantially the same, over a considerable range of ambient temperatures, the said differential output voltage is substantially negligible; and whereby, as the temperature difference between said zener diodes increases, over said same considerable range of ambient temperatures, the said differential output voltage varies substantially and correspondingly in linear relation to said temperature difference.

10. The subject matter of claim 9 in which the amount of current which flows through the reverse biased P-N junct-ion of each zener diode produces heat at each P-N junction of substantially the same amount.

11. The subject matter of claim 10 wherein the heat produced at each P-N junction is on the order of 80 milli-watts