Deviation compensation for flow sensor elements

Abstract

A method for deviation compensation for a flow sensor element comprising a first thermal flow sensor (RH1, RU1, RD1) with a first heating element (RH1) and at least one first temperature-sensitive element (RU1, RD1) and a second thermal flow sensor (RH2, RU2, RD2) with a second heating element (RH2) and at least one second temperature-sensitive element (RU2, RD2), wherein the first thermal flow sensor (RH1, RU1, RD1) and the second thermal flow sensor (RH2, RU2, RD2) are separate and independent of each other, measure the same flow and generate flow-dependent signals, the method comprising: alternately applying heating pulses (i) to the first heating element (RH1) and the second heating element (RH2) such that when the first heating element (RH1) is switched on, the second heating element (RH2) is switched off and when the second heating element (RH2) is switched on, the first heating element (RH1) is switched off; Forming an output signal Vout from the flux-dependent signals such that the contributions of the flux-dependent signals have opposite signs of the flux sensitivity; Measuring and storing the output signal Vout of the flow sensor element at each of the heating pulses (i); and Determining a net output signal Vnet by subtracting at least one previous output signal measurement value V out (i-1) from a last output signal measurement value Vout(i).

Description

Technical area

[0001] The invention relates to the field of thermal anemometers and in particular to thermal anemometers used to measure micro air flows. Background of the invention

[0002] A thermal anemometer measures fluid velocity using heat transfer phenomena. Thermal anemometers, as microcurrent sensors, typically contain identical upstream and downstream temperature-sensitive elements and a heating element in between. The arrangement of such a sensor is usually symmetrical, including the spatial positioning of all functional elements and the distribution of their electrical parameters. Theoretically, this symmetry ensures zero sensor bias and a symmetrical response to bidirectional input signals. In practice, however, variations in physical parameters inevitably occur during the manufacturing process and cause non-zero sensor deviations.This deviation, along with temperature drift and long-term drift caused by temperature and temporal instability of sensor materials, can significantly impact the accuracy of measurements made with flow sensors. In most applications, non-zero deviations must be compensated to ensure the required level of sensor accuracy.

[0003] DE 102 02 210 A1 describes a flow sensor with a heating element and two downstream temperature sensors for detecting fluid flows. The sensors enable fault detection and functional monitoring by comparing temperature changes. The sensor can optionally be operated with a third temperature sensor upstream of the heating element and in pulsed mode to increase accuracy and versatility.

[0004] US 2009 / 0164163 A1 presents an integrated MEMS mass flow sensor for measuring flow rates. The sensor combines A- and C-type technologies, consisting of heaters and temperature sensors on a MEMS chip membrane. It can measure low to high flow rates and is used for higher accuracy and extended measuring ranges. The use of a microcontroller enables signal processing for precise flow and concentration measurement.

[0005] DE 42 08 135 A1 discloses a device for measuring gas or liquid flows. It uses anemometers with temperature-dependent resistance devices as heaters and detectors arranged in a Wheatstone bridge. The device determines flow direction and velocity through changes in the resistance due to changes in temperature distribution. It uses four resistance devices for high measurement sensitivity and a simplified evaluation circuit, ideal for use in silicon microstructure technology.

[0006] US 2009 / 0133490 A1 presents a device and method for measuring the flow velocity of fluids in respiratory systems. It uses two thermal sensor elements that generate and transmit heat signals via a controllable heating element. The sensors are connected so that each received heat signal triggers an electrical feedback signal, which uses the frequency of the feedback signals as an indicator of the flow velocity. Additionally, it enables the analysis of fluid composition and temperature for more precise measurements. Summary

[0007] Described herein is a flow sensor element with deviation compensation and a deviation compensation method, wherein the flow sensor element comprises two separate and independent thermal flow sensors, each containing a heating element and at least one temperature-sensitive element. The components of the two thermal flow sensors are interconnected such that flow-independent contributions of each sensor to a common output signal have opposite signs after passing through a subtraction node. Heating pulses are applied out of phase to the heating elements of the two thermal flow sensors, and an output signal is measured for each applied heating pulse. A net output signal is then determined by calculating a difference between a last output signal measurement and at least one previous output signal measurement.

[0008] According to a first general aspect, a method for offset compensation is provided for a flow sensor element comprising a first thermal flow sensor having a first heating element and at least one first temperature-sensitive element, and a second thermal flow sensor having a second heating element and at least one second temperature-sensitive element, wherein the first thermal flow sensor and the second thermal flow sensor are separate and independent of each other, measure the same flow, and generate flow-dependent signals. The method includes alternately applying heating pulses (i) to the first heating element and the second heating element such that when the first heating element is turned on, the second heating element is turned off, and when the second heating element is turned on, the first heating element is turned off. An output signal V is derived from the flow-dependent signals.out formed so that contributions of the flow-dependent signals have opposite flow sensitivity signs. The output signal V out of the flow sensor element is measured and stored at each of the heating pulses (i), and a net output signal V net is calculated by subtracting at least one previous output signal measurement value V out (i-1) from a last output signal measured value V out (i) determined.

[0009] According to a second general aspect, a compensated deviation flow sensor element is provided. The element comprises a circuit comprising a first thermal flow sensor having a first heating element and at least one first temperature-sensitive element, and a second thermal flow sensor having a second heating element and at least one second temperature-sensitive element, wherein the first thermal flow sensor and the second thermal flow sensor are separate and independent of each other, measure the same flow, and produce flow-independent signals of opposite signs that cooperate to produce an output signal V outA heating module is connected to an input of the circuit and is configured to apply heating pulses (i) alternately to the first heating element and the second heating element such that when the first heating element is switched on, the second heating element is switched off, and when the second heating element is switched on, the first heating element is switched off. An output module is connected to the output of the circuit to provide the output signal V out to be measured and stored at each heating pulse (i) and by subtracting at least one previous output signal measurement value V out (i-1) from a last output signal measured value V out (i) a net output signal V net to determine. Short description of the drawings

[0010] Further features and advantages of the present invention will become apparent from the following detailed description taken in conjunction with the accompanying drawings, in which: Fig. 1 is a circuit diagram illustrating a typical embodiment of a prior art deviation compensation circuit with two synchronously switching on and off heating elements; Fig. 2 is a circuit diagram illustrating a typical embodiment of a deviation compensation circuit with phase-shifted heating elements; Fig. 3a, Fig. 3b and Fig. 3c Diagrams showing output signals of microflow sensors according to the Fig. 1 and Fig. 2 circuit diagrams shown; Fig. 4a, Fig. 4b and Fig. 4c typical waveforms from a clock generator corresponding to heating pulses applied to the heating elements; Fig. 5a, Fig. 5b, Fig. 5c and Fig. 5d typical measurement errors of the flow sensor element when using different deviation compensation algorithms for sinusoidal 15 Hz signals; Fig. 6a, Fig. 6b, Fig. 6c and Fig. 6d typical measurement errors of the flow sensor element when using different deviation compensation algorithms for sinusoidal 10 Hz signals; Fig. 7a and Fig. 7b shows a typical flow sensor element response and measurement error for a 60 ms pulse signal with a rise and fall time of 10 ms.

[0011] It will be noted that in all the accompanying drawings, like features are identified by like reference numerals. Detailed description

[0012] In Fig. Figure 1 shows a microflow sensor of a microelectromechanical system (MEMS). This sensor has two identical heating elements (heaters) R H1 and R H2and two pairs of identical upstream and downstream temperature-sensitive elements R U1 , R D1 and R U2 , R D2 , which are arranged on opposite sides of each heating element and mounted on a thin dielectric bridge suspended above the channel in a silicon substrate, which provides high thermal insulation from a solid substrate. The thermal sensors R U1 , R D1 and R U2 , R D2 are configured in a Wheatstone bridge, while the heating elements R H1 and R H2 by a heating current from a heating voltage source V heat be fed.

[0013] The thermal half-bridge sensor elements R H1 , R U1 , R D1 are from the half-bridge sensor elements R H2 , R U2 , R D2 thermally insulated. This means that a temperature change of the heating element R H1a signal from the second half-bridge sensor is not affected and a temperature change of the heating element R H2 a signal from the first half-bridge sensor is not affected. In a microflow sensor constructed in this way, the temperature of the upstream and downstream thermal sensors depends on the heating element temperature and the flow velocity of a fluid of interest. For example, an increase in flow velocity leads to a temperature increase of the downstream thermal sensor and a temperature decrease of the upstream thermal sensor, thus changing their resistance and providing a flow velocity-dependent output signal V out .

[0014] A switch SW controlled by a clock generator controls an ON / OFF state of heating elements R H1 , R H2 . When the switch SW is turned on, current is supplied from a voltage source Vheat or from an optional digital-to-analog converter (DAC) to the heating elements R H1 , R H2 and the output signal V out1 = S2 - S1 + O2 - O1 is generated, where S1, S2 are flux-dependent signals from the first and second half-bridge sensors, respectively, and O1, O2 are deviations from the first and second half-bridge sensors, respectively. In the case of ideal symmetry of all four thermal sensors, O1 = O2 = V br / 2, where V br is the excitation voltage of the bridge.

[0015] When the switch SW is off, the output signal V out2 = O2 - O1. A net output signal V net is the difference between V out1 and V out2 defined: Vnet=Vout1−Vout2=(S2−S1)+(O2−O1)−(O2−O1)=S2−S1 where S2 ≈ -S1 and therefore V net ≈ 2S2 ≈ -2S1.

[0016] A restriction of the Fig. The disadvantage of the microflow sensor circuit shown in Figure 1 is that when the two heating elements are turned off, the sensor's sensitivity drops and tracking of the input signal is lost. This time period is entirely dedicated to the deviation measurement, which can either be repeated automatically before each flow measurement or performed by a user at certain predefined times during system operation.

[0017] To overcome this limitation, Fig. 2 shows an alternative embodiment. As shown, a circuit comprises 200 heating elements R H1 and R H2 which are individually accessible and fed in phase shift so that one of the heating elements is always switched on. In addition, the circuit of the temperature-sensitive elements in the first thermal half-bridge sensor is compared to the configuration of Fig. 1 modified so that the resistance RD1 to the bridge excitation voltage V BR is connected and the resistance R U1 is connected to ground. Such a circuit of the four temperature-sensitive elements results in zero flux sensitivity when both heating elements are switched on, since the flux-independent signals from the two thermal half-bridge flux sensors have the same sign and are subtracted by a subtraction node 206, which can be provided inside or outside the circuit 200. In this configuration, the contributions of flux-dependent signals from the two half-bridge sensors to an output signal after passing through the subtraction node 206 are S2 and -S1, thereby having opposite signs of flux sensitivity. A heating module 202 applies phase-shifted heating pulses to the circuit, and an output module 204 measures and stores the output signal V out .

[0018] The output signal Vnet of the microflow sensor with compensated deviation can then be determined by the output module 204 as the difference between a last output signal measurement and previous output signal measurements according to the characteristics of Table 1 below and Equation 2. A heating element switched off R H1 measured sensor output signal is measured by a heating element R H2 measured output signal. Table 1 Heizelemente V1 V2 V out1 , V out2 R H1 = OFF, R H2 = ON O1 S2 + O2 <h2 style=";text-align:left;direction:ltr">V<h2 style=";text-align:left;direction:ltr"> out1 <h2 style=";text-align:left;direction:ltr"> = V2-V1 = S2+O2-O1 R H1 = ON, R H2 = OFF S1+O1 O2 V out2 = V2-V1 = O2-S1-O1 Vnet=Vout1−Vout2=(S2+O2−O1)−(O2-S1-O1)=S2+S1 where: S1 a signal from the first half-bridge microflow sensor R U1 , R D1 is; S2 a signal from the second half-bridge microflow sensor R U2 , R D2 is; O1 a deviation from the first half-bridge microflow sensor R U1 , R D1 is; O2 a deviation from the second half-bridge microflow sensor RU2 , R D2 is; S2 ≈ S1 and therefore V net ≈ 2S2 ≈ 2S1.

[0019] The Fig. 3b and Fig. 3c illustrate the result of a simulation of the response of the sensors for a Fig. 3a shows a half-sinusoidal 15 Hz input signal, with measurements of the input signal being performed at intervals of 5 ms. Fig. Figure 3b shows a reaction of the sensor when the heating elements are switched in phase according to the embodiment of Fig. 1. Every second measurement of the sensor is taken with the heaters switched off for the deviation measurement, and the flow-dependent signal is then subtracted. Fig. Figure 3c shows a response of the sensor with phase-shifted switching of the heating elements according to the embodiment of Fig. 2. The sensor with phase-shifted switching of the heating elements demonstrates more accurate tracking of a time-varying signal.

[0020] Back to the circuit of Fig. 2: there is shown a clock generator which is used to control the switch to apply the excitation voltage V heat1 and V heat2 either to the heating element R H1 or to the heating element R H2 on and off. Examples of the clock signal and the two excitation voltages are shown in the Fig. 4a, Fig. 4b and 4c respectively. If a logical output signal of the generator V clk (i) = 1, then the heating element R H2 switched on and the heating element R H1 is switched off. If V clk (i) = 0, then the heating element R H1 switched on and the heating element R H2 is switched off. Output signal values ​​V out (i) of the sensor are measured and stored with a time interval T for each heating pulse i. This deviation compensation method is based on a subtraction of the last two measured values ​​V out (i) and Vout (i-1) according to equation (2), which can be mathematically represented as follows: Vnet(i)=(Vout(i)−Vout(i−1))(2Vclk(i)−1) where the function 2V clk (i)-1 is equal to 1 if V clk (i)=1, or equal to -1 if V clk (i)=0.

[0021] Compare measurements of the flow sensors with different types of heating elements. If both heating elements of the Fig. 1 are continuously switched on, the sensor operates in traditional DC mode without any offset compensation and with a specific nominal sensitivity. When operating in DC mode, the sensor provides ideal signal tracking. This sensor can also operate in a mode with in-phase switching of the heating elements.

[0022] The Fig. The sensor shown in Figure 2, when both heating elements are switched on, has zero sensitivity due to the subtraction of the signals V2 and V1 from two essentially identical half-bridges by the subtraction node. When only one of the two essentially identical heating elements is switched on, the sensor sensitivity reaches approximately 50% of the sensitivity of the sensor of Fig. 1. Therefore, each measured value contains V out (i) a sensor error component and a component dependent on the flow registered with reduced sensitivity. Subtraction of two consecutive measured values ​​V out (i) and V out (i-1) (according to equation (3)) every T milliseconds compensates for the deviation and effectively sets the sensitivity to the level of the sensor of Fig. 1. An update of the sensor output signal due to changes in the input signal is carried out at time intervals T when either the heating element R H1 or the heating element R H2 is switched on.

[0023] The simulation results of the measurement accuracy are shown in the Fig. 5a-5d and 6a-6d. The filled circles in the Fig. 5a and Fig. 6a represent measured values ​​of the output signal V DC of the sensor with heating elements constantly switched on (according to Fig. 1). The measurements are performed with an interval T = 5 ms for sinusoidal signals of 15 Hz and 10 Hz, respectively. Deviations of the output signals from V DC for sensor 1 (in-phase switching of the heating elements; switching of Fig. 1) and sensor 2 (phase-shifted switching of the heating elements; switching of Fig. 2) are in the Fig. 5b, Fig. 5c (15 Hz) and 6b, 6c (10 Hz). The embodiment according to Fig. 2 demonstrates a lower measurement error than the embodiment of Fig. 1.

[0024] The accuracy of measurements using the proposed technique can be further improved if, according to another embodiment, more than the last two stored measurements V out (i) be used in determining the net output of the sensor. Equation (4) is a mathematical representation of the sensor output resulting from the last three measurements V out (i), V out (i-1) and V out (i-2) is determined: Vnet2=((1+k)Vout(i)−Vout(i−1)−kVout(i−2))(2Vclk(i)−1) where an adjustment coefficient k varies from 0 to 1.

[0025] The Fig. 5d and Fig. 6d show deviations of the sensor output signal according to Fig. 2 of V DCValues ​​based on equation (4) with a coefficient k = 0.5. The coefficient k is used to minimize the time delay between an ideal signal response of the sensor operating in DC mode and the response of the sensor implementing the phase-shifted switching of the heating elements. In some embodiments, the values ​​of the coefficient k range from 0 to 1.

[0026] To understand the influence of k, consider a sensor response for a 60 ms pulse with a rise and fall time τ = 10 ms. Fig. Figure 7a shows a simulated reaction V DC the sensor circuit according to Fig. 1, in which both heating elements are continuously switched on (filled circles) compared to the simulated sensor response using signal processing based on equation (4) with different coefficients k. Note that at k=0, equation (4) agrees with equation (3). Fig.Figure 7b shows deviations of the sensor's measured values ​​from the sensor's output signal with both heating elements switched on at k=0, 0.5, and 1 using an algorithm based on equation (4). A minimum error is achieved at k=0.5. The condition k=1 leads to signal overregulation with an associated decrease in sensor accuracy.

[0027] Although the currently described microflow sensor for implementing the deviation compensation method contains two identical thermal half-bridge sensors with separate heating elements, it should be understood that, in general, the deviation compensation method can be applied to any pair of identical thermal anemometer-type sensors of different designs measuring the same flow. Furthermore, instead of two, only one temperature-sensitive element can be used for each sensor in the pair of largely identical sensors.

[0028] The heating elements and temperature-sensitive elements can, for example, be polysilicon resistors manufactured using a standard CMOS process. They can have different doping levels, providing an optimal resistance value and TCR (temperature coefficient of resistance) for these different functional elements.

[0029] The optimal width of heating element excitation pulses T can be determined by one skilled in the art. The time T can be several times longer than a typical thermal response time for the anemometer-type thermal sensor to allow the output signal to reach its stable value during the heating or cooling process of the temperature-sensitive elements. Response times defined by the geometry, thermal mass, and thermal conductivity of the sensor microstructure are typically about 1-2 ms for thermal anemometers manufactured using known MEMS processes. Therefore, a sampling interval of about 4-5 ms or longer can be used for the described deviation compensation method.

[0030] The deviation compensation method can be used to compensate not only for long-term static and temperature drift, but also for short-term deviation fluctuations. With a sampling interval of 5 ms, deviation fluctuations with a frequency below ~50 Hz can be effectively compensated. This reduces the sensor's low-frequency noise and improves its signal-to-noise ratio.

[0031] It should be noted that the present invention can be embodied as a method and implemented as a system. The above-described embodiments of the invention are intended to serve only as examples. The scope of the invention should therefore be limited solely by the scope of the appended claims.

Claims

[1] A method for deviation compensation for a flow sensor element comprising a first thermal flow sensor (RH1, RU1, RD1) with a first heating element (RH1) and at least one first temperature-sensitive element (RU1, RD1) and a second thermal flow sensor (RH2, RU2, RD2) with a second heating element (RH2) and at least one second temperature-sensitive element (RU2, RD2), wherein the first thermal flow sensor (RH1, RU1, RD1) and the second thermal flow sensor (RH2, RU2, RD2) are separate and independent of each other, measure the same flow and generate flow-dependent signals, the method comprising: alternately applying heating pulses (i) to the first heating element (RH1) and the second heating element (RH2) such that when the first heating element (RH1) is switched on, the second heating element (RH2) is switched off and when the second heating element (RH2) is switched on, the first heating element (RH1) is switched off; Forming an output signal Vout from the flux-dependent signals such that the contributions of the flux-dependent signals have opposite signs of the flux sensitivity; Measuring and storing the output signal Vout of the flow sensor element at each of the heating pulses (i); and Determining a net output signal Vnet by subtracting at least one previous output signal measurement value V out (i-1) from a last output signal measurement value Vout(i). [2] The method of claim 1, wherein the determination of the net output signal Vnet includes consideration of a coefficient k that minimizes a time delay between an ideal signal response of the flow sensor element operating in direct current (VDC) mode and a response of the flow sensor element that implements a phase-shifted switching of the first heating element (RH1) and the second heating element (RH2). [3] Method according to claim 2, wherein the coefficient k is defined by 0 ≤ k ≤ 1 and preferably by k = 0.

5. [4] A method according to any one of claims 1 to 3, wherein determining the net output signal Vnet includes determining the net output signal Vnet from three last measurements Vout(i), Vout(i-1) and Vout(i-2). [5] A method according to any one of claims 1 to 4, wherein measuring and storing the output signal Vout includes performing a measurement with a sampling interval of at least 4 ms. [6] The method of any one of claims 1 to 5, wherein forming an output signal Vout includes inverting and adding signals from the first thermal flow sensor (RH1, RU1, RD1) and the second thermal flow sensor (RH2, RU2, RD2) and generating the output signal Vout. [7] A flow sensor element with compensated deviation, the element comprising: a circuit (200) comprising a first thermal flow sensor (RH1, RU1, RD1) with a first heating element (RH1) and at least one first temperature-sensitive element (RU1, RD1) and a second thermal flow sensor (RH2, RU2, RD2) with a second heating element (RH2) and at least one second temperature-sensitive element (RU2, RD2), wherein the first thermal flow sensor (RH1, RU1, RD1) and the second thermal flow sensor (RH2, RU2, RD2) are separate and independent of each other, measure the same flow and generate flow-dependent signals of opposite signs which cooperate to form an output signal Vout; a heating module (202) connected to an input of the circuit (200) and configured to apply heating pulses (i) alternately to the first heating element (RH1) and the second heating element (RH2) such that when the first heating element (RH1) is turned on, the second heating element (RH2) is turned off and when the second heating element (RH2) is turned on, the first heating element (RH1) is turned off; an output module (204) connected to an output of the circuit (200) for measuring and storing the output signal Vout at each heating pulse (i) and for determining a net output signal Vnet by subtracting at least one previous output signal measurement value Vout(i-1) from a last output signal measurement value Vout(i). [8] Flow sensor element according to claim 7, wherein the output module (204) calculates the net output signal Vnet taking into account a coefficient k that minimizes a time delay between an ideal signal response of the flow sensor element operating in direct current (VDC) mode and a response of the flow sensor element that realizes a phase-shifted switching of the first heating element (RH1) and the second heating element (RH2). [9] Flow sensor element according to claim 8, wherein the coefficient k is defined by 0 ≤ k ≤ 1 and preferably by k = 0.

5. [10] Flow sensor element according to one of claims 7 to 9, wherein the output module (204) determines the net output signal Vnet using at least three measurements Vout(i), Vout(i-1) and Vout(i-2). [11] Flow sensor element according to one of claims 7 to 10, wherein the first thermal flow sensor (RH1, RU1, RD1) and the second thermal flow sensor (RH2, RU2, RD2) each have two temperature-sensitive elements (RU1, RD1, RU2, RD2) and the first and second heating elements (RH1, RH2) are arranged between the two temperature-sensitive elements (RU1, RD1, RU2, RD2), and wherein the temperature-sensitive elements (RU1, RD1, RU2, RD2) are thermal resistors. [12] A flow sensor element according to claim 11, wherein two thermal resistors (RU1, RD1) of the first thermal flow sensor (RH1, RU1, RD1) and two thermal resistors (RU2, RD2) of the second thermal flow sensor (RH2, RU2, RD2) form a Wheatstone bridge by comprising: a set of first terminals of downstream thermal resistors (RD1, RD2) of the first thermal flow sensor (RH1, RU1, RD1) and the second thermal flow sensor (RH2, RU2, RD2) connected to an upper end of the Wheatstone bridge; a set of first terminals of upstream thermal resistors (RU1, RU2) of the first thermal flow sensor (RH1, RU1, RD1) and the second thermal flow sensor (RH2, RU2, RD2) connected to a lower end of the Wheatstone bridge; a set of second terminals of upstream and downstream thermal resistors (RU1, RD1, RU2, RD2) belonging to the same thermal flow sensor (RH1, RU1, RD1, RH2, RU2, RD2) and connected to each other, with two connection points forming a diagonal of the Wheatstone bridge. [13] Flow sensor element according to one of claims 7 to 12, wherein the circuit (200) comprises a subtraction node (206) for inverting and adding signals from the first thermal flow sensor (RH1, RU1, RD1) and the second thermal flow sensor (RH2, RU2, RD2) and for generating the output signal Vout.

Images