rus
DOI: 10.22184/1993-8578.2022.15.3-4.216.221
Based on the experimental results of the thermal gas mass flow rate sensor it was shown that the use of various analogue signal processing circuits makes it possible to achieve uniquely high sensitivity and combine measurements of the temperature difference at the front, decline and drop of the temperature of the heater in a single signal.
Based on the experimental results of the thermal gas mass flow rate sensor it was shown that the use of various analogue signal processing circuits makes it possible to achieve uniquely high sensitivity and combine measurements of the temperature difference at the front, decline and drop of the temperature of the heater in a single signal.
Теги: gas flow gas mass flow rate signal extraction and processing thermal mems sensor выделение и обработка сигнала массовый расход газа поток газа тепловой мэмс-сенсор
Received: 20.05.2022 | Accepted: 27.05.2022 | DOI: https://doi.org/10.22184/1993-8578.2022.15.3-4.216.221
Original paper
STUDY OF A THERMAL MEMS SENSOR FOR GAS MASS FLOW RATE
V.T.Ryabov, Cand. of Sci. (Tech), Docent, ORCID: 0000-0002-4781-3186
N.A.Djuzhev2, Director, ORCID: 0000-0002-5205-0304 / v_ryabov@mail.ru
Abstract. Based on the experimental results of the thermal gas mass flow rate sensor it was shown that the use of various analogue signal processing circuits makes it possible to achieve uniquely high sensitivity and combine measurements of the temperature difference at the front, decline and drop of the temperature of the heater in a single signal.
Keywords: gas mass flow rate, thermal MEMS sensor, gas flow, signal extraction and processing
For citation: V.T. Ryabov, N.A. Djuzhev. Investigation of thermal MEMS sensor of gas mass flow rate. NANOINDUSTRY. 2022. V. 15, no. 3–4. PP. 216–221. https://doi.org/10.22184/1993-8578.2022.15.3-4.216.221
INTRODUCTION
Gas mass flow sensors are widely used in vacuum process and research equipment. Recently, the use of such sensors in medical equipment, especially in ventilators, has become particularly relevant.
Application of microelectromechanical systems technology makes it possible to create the miniature mass flow sensing elements (sensors) on silicon crystals (chips). Mass production of a large number of sensors with similar parameters on a single wafer will significantly reduce their cost. A well known sensor from Honeywell measures a temperature difference at the front and decline of a "heat cloud" created by the heater. The technology of this sensor was even used as the basis of a textbook for an undergraduate course at the University of Louisville [1]. This sensor was developed in the early 1980s and is the basis of most of the company’s flux sensors.
A disadvantage of thermal MEMS sensors is the limitation of the gas flow velocity measurements range due to the changing nature of the gas flow: laminar or turbulent. To measure the flow rate in a laminar flow, the calorimetric method is used which is based on measuring the temperature difference in the heat cloud created by the heater. To measure the turbulent flow, the thermo-anemometric method based on the removal of heat from the heater is used.
The purpose of this research is to develop a domestic thermal MEMS sensor with an extended range of gas flow rate measurement with improved accuracy and sensitivity. In addition, we tried to create a domestic analogue of the Honeywell AWM720P1 sensor [2] for use in artificial lung ventilation (ALV).
RESEARCH METHODS
Analytical and experimental studies were carried out on a sample MEMS sensor from a pilot series. A 3 × 4 mm silicon chip was prepared with an etched membrane of about 0.5 × 0.75 mm, consisting of alternating layers of Si3N4 and SiO2, is etched on it. The total thickness of the membrane is about 0.6 µm (Fig.1).
Five platinum thermal resistors are formed on the crystal. One resistor is located on the crystal body (medium resistor, Fig.1, left). It is designed to measure the silicon crystal temperature.
The outputs of this resistor are marked with the symbols Sr in Fig.1. Four other thermistors are located on the membrane and are included in the measuring thermal bridge. It is designed to form a ‘heat cloud’ and measure its temperature at the front and decline of the gas flow. The gas flow is directed from left to right so that the ‘heat cloud’ produces the minimal effect on the media resistor.
The thermostat resistors are located in two sections, in the input In (left on the Fig.1) and in the output Out (right). A schematic of the thermal bridge is shown in Fig.2, a. Each section contains two similar resistors Ri and Ro that form the bridge arm. Resistors Ri are located in the input section and Ro in the output section. The output of the input section resistor is connected to the terminal block Rin and the output of the output section resistor is connected to the terminal block Rout.
Resistors Ri and Ro show arrows indicating whether the resistance value of the respective thermistor increases or decreases when gas flow is applied to the membrane. Between the resistors Ri and Ro are formed diagonal thermo-bridge leads Dhi and Dlow, brought to the respective contact pads (Fig.1). On the Dhi pin, the voltage increases with the increasing gas flow and on the Dlow pin, the voltage decreases.
Fig.2, b shows the current flow through the bridge arms and the isotherm of the heat cloud that forms when current I flows through the bridge. Each arm carries half the current of the bridge due to symmetry.
Experimental studies of the sensing element were carried out on an automated test bench [3] equipped with gas flow setters and measuring instruments certified in the Russian Federation. To set the gas flow rate, 12-bit digital-to-analog converters were used and to record the results, 16-bit analog-to-digital converters were used.
RESULTS AND DISCUSSIONS
If the bridge is loaded by a controlled current source I and there is no gas flow, the resistance Rt of each of the thermal bridge resistors will be equal to
. (1)
Here Ro – thermal resistor resistance at normal temperature, α – thermal resistance coefficient (TRC), Δt0 means increasing of thermal resistor temperature relatively to the normal one.
The signal in the diagonal ΔD = (Dhi – Dlow) when gas flow distorts the heat cloud will be equal to:
(2)
where Δtoi – temperature difference between output and input sections of the thermal bridge.
As the current I through the bridge increases, the signal ΔD from the diagonal increases, firstly, because the current enters into this signal as a factor and, secondly, because the value of Rt increases due to additional heating of the resistors. The additional heating is proportional to the square of the current. Thirdly, the increase in current is due to power redistribution between the input In and output Out sections due to distortion of the heat cloud isotherm by the gas flow blowing on the membrane. In the section which temperature is higher, the resistance increases and when the currents are equal more power is released. In other words, there is a positive feedback mechanism in the physical signal formation process of the described DC-operated sensor.
All these factors result in a uniquely high sensitivity of the thermal bridge when powered by a current source. Such circuits with increased sensitivity could be promising for use in gas chromatography.
A current powered sensor was used to develop an analogue of Honeywell’s AWM720P1 mass flow sensor [2]. It was built-in the bypass channel and mounted with the signal processing circuitry in the sensor housing. Tests showed that with a thermo-bridge supply current of about 4.5 mA and a gain of about 60, a signal saturation with supply voltage occurred. The gain was too high.
Figure 3 shows a comparison between the AWM720P1 standard sensor signal and a prototype MEMS sensor when a 0.75 mm orifice choke was fitted to the bypass channel output of the housing. This markedly reduced the magnitude of the output signal.
Figure 3 shows that the flow measurement range with the new MEMS sensor is wider and it is possible to duplicate the AWM720P1 sensor characteristic by adjusting the supply current, throttling the bypass channel and selecting the sensor circuit gain.
Another scheme for acquiring a signal from a measuring thermal bridge is shown in Fig.4. Such a circuit has been used for signal processing of a semiconductor pressure sensor [3]. These are two identical current sources controlled by the signal Ud. The output signal Utm of this circuit is equal to (3).
Here Rt, as in the previous case, is the steady-state resistance of the bridge thermistor when there is no gas flow, α means TRC, Δtoi means the temperature difference between the output and input sections of the thermostat, Ri(t), Ro(t) means the resistance of the input and output section resistors when gas flow occurs, ρtm(t) means the thermostat conductance.
Analysis of formula (3) shows that the output signal of the thermal bridge depends not only on the temperature difference between the inlet and outlet sections Δtoi and resistance Rt, but also on conductivity of the thermal bridge. When the flow turns from laminar to turbulent, the temperature difference signal Δtoi ceases to increase with the higher gas flow rate. Only heat removal increases, the mean temperature of the thermal bridge decreases and its conductivity ρtm(t) increases. The output signal continues to increase. This extends the range of gas flow rates measured by the MEMS sensor.
Figure 5 shows experimental results of the second version of the circuit (Fig.4). The sensor was mounted in the near-wall area of the main sensor channel with a diameter of 0.5’.
The diagram shown in Fig.5 splits clearly into three zones. In zone 1 the flow is laminar, the heat cloud is increasingly pressed against the crystal surface as the flow rate increases, its temperature and the signal from the medium resistor increases. A signal from the thermal bridge is almost linear. In zone 2 the laminar flow starts to break down and the crystal temperature drops. This is the transition zone. The flow becomes turbulent in zone 3. Heat removal from the thermal bridge increases as the flow rate increases and the temperature decreases. A signal from the sensor continues to rise as the thermal bridge conductivity ρtm(t) increases and the thermostat temperature decreases.
CONCLUSIONS
The studied MEMS sensor thermal bridge of four thermistors, arranged in two sections – inlet and outlet by gas flow – makes it possible:
PEER REVIEW INFO
Editorial board thanks the anonymous reviewer(s) for their contribution to the peer review of this work. It is also grateful for their consent to publish papers on the journal’s website and SEL eLibrary eLIBRARY.RU.
Declaration of Competing Interest. The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Original paper
STUDY OF A THERMAL MEMS SENSOR FOR GAS MASS FLOW RATE
V.T.Ryabov, Cand. of Sci. (Tech), Docent, ORCID: 0000-0002-4781-3186
N.A.Djuzhev2, Director, ORCID: 0000-0002-5205-0304 / v_ryabov@mail.ru
Abstract. Based on the experimental results of the thermal gas mass flow rate sensor it was shown that the use of various analogue signal processing circuits makes it possible to achieve uniquely high sensitivity and combine measurements of the temperature difference at the front, decline and drop of the temperature of the heater in a single signal.
Keywords: gas mass flow rate, thermal MEMS sensor, gas flow, signal extraction and processing
For citation: V.T. Ryabov, N.A. Djuzhev. Investigation of thermal MEMS sensor of gas mass flow rate. NANOINDUSTRY. 2022. V. 15, no. 3–4. PP. 216–221. https://doi.org/10.22184/1993-8578.2022.15.3-4.216.221
INTRODUCTION
Gas mass flow sensors are widely used in vacuum process and research equipment. Recently, the use of such sensors in medical equipment, especially in ventilators, has become particularly relevant.
Application of microelectromechanical systems technology makes it possible to create the miniature mass flow sensing elements (sensors) on silicon crystals (chips). Mass production of a large number of sensors with similar parameters on a single wafer will significantly reduce their cost. A well known sensor from Honeywell measures a temperature difference at the front and decline of a "heat cloud" created by the heater. The technology of this sensor was even used as the basis of a textbook for an undergraduate course at the University of Louisville [1]. This sensor was developed in the early 1980s and is the basis of most of the company’s flux sensors.
A disadvantage of thermal MEMS sensors is the limitation of the gas flow velocity measurements range due to the changing nature of the gas flow: laminar or turbulent. To measure the flow rate in a laminar flow, the calorimetric method is used which is based on measuring the temperature difference in the heat cloud created by the heater. To measure the turbulent flow, the thermo-anemometric method based on the removal of heat from the heater is used.
The purpose of this research is to develop a domestic thermal MEMS sensor with an extended range of gas flow rate measurement with improved accuracy and sensitivity. In addition, we tried to create a domestic analogue of the Honeywell AWM720P1 sensor [2] for use in artificial lung ventilation (ALV).
RESEARCH METHODS
Analytical and experimental studies were carried out on a sample MEMS sensor from a pilot series. A 3 × 4 mm silicon chip was prepared with an etched membrane of about 0.5 × 0.75 mm, consisting of alternating layers of Si3N4 and SiO2, is etched on it. The total thickness of the membrane is about 0.6 µm (Fig.1).
Five platinum thermal resistors are formed on the crystal. One resistor is located on the crystal body (medium resistor, Fig.1, left). It is designed to measure the silicon crystal temperature.
The outputs of this resistor are marked with the symbols Sr in Fig.1. Four other thermistors are located on the membrane and are included in the measuring thermal bridge. It is designed to form a ‘heat cloud’ and measure its temperature at the front and decline of the gas flow. The gas flow is directed from left to right so that the ‘heat cloud’ produces the minimal effect on the media resistor.
The thermostat resistors are located in two sections, in the input In (left on the Fig.1) and in the output Out (right). A schematic of the thermal bridge is shown in Fig.2, a. Each section contains two similar resistors Ri and Ro that form the bridge arm. Resistors Ri are located in the input section and Ro in the output section. The output of the input section resistor is connected to the terminal block Rin and the output of the output section resistor is connected to the terminal block Rout.
Resistors Ri and Ro show arrows indicating whether the resistance value of the respective thermistor increases or decreases when gas flow is applied to the membrane. Between the resistors Ri and Ro are formed diagonal thermo-bridge leads Dhi and Dlow, brought to the respective contact pads (Fig.1). On the Dhi pin, the voltage increases with the increasing gas flow and on the Dlow pin, the voltage decreases.
Fig.2, b shows the current flow through the bridge arms and the isotherm of the heat cloud that forms when current I flows through the bridge. Each arm carries half the current of the bridge due to symmetry.
Experimental studies of the sensing element were carried out on an automated test bench [3] equipped with gas flow setters and measuring instruments certified in the Russian Federation. To set the gas flow rate, 12-bit digital-to-analog converters were used and to record the results, 16-bit analog-to-digital converters were used.
RESULTS AND DISCUSSIONS
If the bridge is loaded by a controlled current source I and there is no gas flow, the resistance Rt of each of the thermal bridge resistors will be equal to
. (1)
Here Ro – thermal resistor resistance at normal temperature, α – thermal resistance coefficient (TRC), Δt0 means increasing of thermal resistor temperature relatively to the normal one.
The signal in the diagonal ΔD = (Dhi – Dlow) when gas flow distorts the heat cloud will be equal to:
(2)
where Δtoi – temperature difference between output and input sections of the thermal bridge.
As the current I through the bridge increases, the signal ΔD from the diagonal increases, firstly, because the current enters into this signal as a factor and, secondly, because the value of Rt increases due to additional heating of the resistors. The additional heating is proportional to the square of the current. Thirdly, the increase in current is due to power redistribution between the input In and output Out sections due to distortion of the heat cloud isotherm by the gas flow blowing on the membrane. In the section which temperature is higher, the resistance increases and when the currents are equal more power is released. In other words, there is a positive feedback mechanism in the physical signal formation process of the described DC-operated sensor.
All these factors result in a uniquely high sensitivity of the thermal bridge when powered by a current source. Such circuits with increased sensitivity could be promising for use in gas chromatography.
A current powered sensor was used to develop an analogue of Honeywell’s AWM720P1 mass flow sensor [2]. It was built-in the bypass channel and mounted with the signal processing circuitry in the sensor housing. Tests showed that with a thermo-bridge supply current of about 4.5 mA and a gain of about 60, a signal saturation with supply voltage occurred. The gain was too high.
Figure 3 shows a comparison between the AWM720P1 standard sensor signal and a prototype MEMS sensor when a 0.75 mm orifice choke was fitted to the bypass channel output of the housing. This markedly reduced the magnitude of the output signal.
Figure 3 shows that the flow measurement range with the new MEMS sensor is wider and it is possible to duplicate the AWM720P1 sensor characteristic by adjusting the supply current, throttling the bypass channel and selecting the sensor circuit gain.
Another scheme for acquiring a signal from a measuring thermal bridge is shown in Fig.4. Such a circuit has been used for signal processing of a semiconductor pressure sensor [3]. These are two identical current sources controlled by the signal Ud. The output signal Utm of this circuit is equal to (3).
Here Rt, as in the previous case, is the steady-state resistance of the bridge thermistor when there is no gas flow, α means TRC, Δtoi means the temperature difference between the output and input sections of the thermostat, Ri(t), Ro(t) means the resistance of the input and output section resistors when gas flow occurs, ρtm(t) means the thermostat conductance.
Analysis of formula (3) shows that the output signal of the thermal bridge depends not only on the temperature difference between the inlet and outlet sections Δtoi and resistance Rt, but also on conductivity of the thermal bridge. When the flow turns from laminar to turbulent, the temperature difference signal Δtoi ceases to increase with the higher gas flow rate. Only heat removal increases, the mean temperature of the thermal bridge decreases and its conductivity ρtm(t) increases. The output signal continues to increase. This extends the range of gas flow rates measured by the MEMS sensor.
Figure 5 shows experimental results of the second version of the circuit (Fig.4). The sensor was mounted in the near-wall area of the main sensor channel with a diameter of 0.5’.
The diagram shown in Fig.5 splits clearly into three zones. In zone 1 the flow is laminar, the heat cloud is increasingly pressed against the crystal surface as the flow rate increases, its temperature and the signal from the medium resistor increases. A signal from the thermal bridge is almost linear. In zone 2 the laminar flow starts to break down and the crystal temperature drops. This is the transition zone. The flow becomes turbulent in zone 3. Heat removal from the thermal bridge increases as the flow rate increases and the temperature decreases. A signal from the sensor continues to rise as the thermal bridge conductivity ρtm(t) increases and the thermostat temperature decreases.
CONCLUSIONS
The studied MEMS sensor thermal bridge of four thermistors, arranged in two sections – inlet and outlet by gas flow – makes it possible:
- to simultaneously serve as a heater, i.e. to create a "heat cloud" and measure the temperature difference between its front and its decline;
- to significantly increase the sensor sensitivity when the thermostat is powered from the current source and reduce the requirements for the subsequent sensor signal amplification and the sensor signal normalisation;
- to extend aMEMS sensor measurement range and combine in a single signal the temperature difference and the temperature drop as a function of gas mass flow rate;
- allows of making mass flow rate measurements under laminar, transient and turbulent gas flow conditions.
PEER REVIEW INFO
Editorial board thanks the anonymous reviewer(s) for their contribution to the peer review of this work. It is also grateful for their consent to publish papers on the journal’s website and SEL eLibrary eLIBRARY.RU.
Declaration of Competing Interest. The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Readers feedback
