CN101014848B - Sensing apparatus - Google Patents

Abstract

An object of the present invention is to provide a sensing device capable of instantaneously detecting, for example, environmental pollutants present in minute amounts with high accuracy. As a specific solution, the frequency signal from the quartz vibrator is sampled according to the frequency signal from the reference clock generator, the sampled value is output as a digital signal, and the frequency signal corresponding to the output signal is corrected according to the digital signal. In quadrature detection, a rotation vector that rotates at the frequency of the difference between the frequency of the frequency signal and the frequency of the sine wave used for quadrature detection is extracted, and the amount of change in frequency is detected by detecting the speed of the rotation vector from each sample value. Furthermore, by multiplying the rotation vector by an inverse rotation vector corresponding to the speed, the measurement range of the frequency change can be expanded.

Description

Sensing device

technical field

The present invention relates to a sensing device for detecting the natural frequency of a sensor vibrator, such as a quartz vibrator, by using an adsorption layer for adsorbing a sensing object formed on the surface thereof, for example, a sensor vibrator whose natural frequency is changed by the adsorption of the sensing object The amount of change and perceive the object of perception.

Background technique

In order to protect the environment, it is urgent to know the concentration of various environmental pollutants in rivers and soils. Even extremely small amounts of pollutants are highly toxic to the human body. Therefore, it is desired to establish a measurement technology for trace amounts of pollutants. Recently, dioxins are attracting attention as a pollutant. As methods for measuring the dioxins, a method using a gas chromatography mass analyzer and an ELISA method (applicable to an enzyme immunoassay method) are well known. According to the gas chromatography mass analyzer, high-precision microanalysis on the order of 10 ^-12 g/ml can be performed, but there is a problem that the price of the device is extremely high, so the analysis cost is also high, and further analysis requires a long period of time. Shortcomings. Furthermore, the ELISA method has lower equipment and analysis costs than gas chromatography mass analyzers, and the time required for analysis is also shorter, but there is a problem that the analysis accuracy is on the order of 10 ^-9 g/ml.

Therefore, the present inventors focused on a quartz oscillator as a measurement device for pollutants such as dioxins, since when a sensing object adheres to the quartz oscillator, its natural frequency changes according to the amount of adhesion. On the other hand, there is a technique described in Patent Document 1 as a chemical sensor device using a quartz oscillator. The device is a sampling circuit that outputs the absolute value of the difference frequency between the oscillation frequency of the sensor vibrator and the reference frequency generated by the reference vibrator, a frequency division circuit that divides the difference frequency with a desired frequency division ratio, and divides the period of the reference frequency A counter counting the period of the frequency-divided output as a clock, and a calculation device for obtaining the oscillation frequency of the sensor vibrator from the counted period are means for identifying adsorbed gas. According to this chemical sensor device, since the difference frequency is used, the absolute value of the frequency to be measured can be reduced, and there is an advantage that high-resolution measurement can be performed without expanding the measurement range.

However, in the technology of Patent Document 1, the clock frequency of the counter is lowered by using a frequency division circuit, so the above-mentioned difference frequency is high and the number of stages of the frequency divider circuit is large. Since the phase noise increases when the number of stages increases, the above-mentioned difference frequency does not actually increase. If it is too high, it becomes difficult to ensure high measurement accuracy. As a result, the scope of application is limited, and it is difficult to apply it to the case where it is required to detect extremely small amounts of substances such as dioxins with high precision. Furthermore, since a counter is used, there are disadvantages that high resolution is required and measurement time is long.

Patent Document 1: Japanese Patent Laid-Open No. 6-241972

Contents of the invention

The present invention was made under such circumstances, and an object of the present invention is to provide a sensing device capable of detecting sensing objects such as environmental pollutants present in minute amounts with high accuracy. Another object of the present invention is to provide a sensing device capable of instantaneously detecting a sensing object with high accuracy. Still another object is to provide a sensing device capable of detecting a sensing object with high accuracy and capable of expanding the measurement range.

The sensing device of the present invention is characterized in that an adsorption layer for absorbing the object to be sensed is formed on its surface, and a vibrator for a sensor whose natural frequency changes due to the adsorption of the object to be sensed is used. The change perception perception object is characterized in that it has:

an oscillating circuit for a sensor, which oscillates the vibrator for the sensor;

a reference clock generating unit that generates a clock signal for sampling the frequency signal from the vibrator for the sensor;

an analog/digital conversion unit that samples the frequency signal from the vibrator for the sensor with the clock signal from the reference clock generation unit, and outputs the sampled value as a digital signal;

The rotation vector extraction part performs quadrature detection based on the digital signal to the frequency signal corresponding to the output signal from the analog/digital conversion part, and extracts the real part and the imaginary part when the rotation vector is represented by a complex number, wherein the rotation vector is represented by rotation at an angular velocity commensurate with the frequency difference between the frequency in the frequency signal and the frequency specified by the digital signal used for quadrature detection; and

The rotation vector velocity calculating means obtains the angular velocity of the rotation vector based on the respective time-series data of the real number part and the imaginary number part obtained by the rotation vector extraction means.

In addition, the present invention may include means for obtaining the difference between the angular velocity of the rotation vector when the sensor vibrator is in the first environment and the angular velocity of the rotation vector when the sensor vibrator is in the second environment. As the first environment, for example, a solvent such as pure water can be mentioned, and as the second environment, a case where the object to be sensed is contained in the solvent can be mentioned.

The rotation vector extracting means includes means for quadrature-detecting a frequency signal corresponding to an output signal from the analog/digital conversion part, and removing high-frequency components included in data obtained by the means. part. Also, when setting the real part and the imaginary part corresponding to the above sampled value at a certain timing as I(n) and Q(n), respectively, setting the When the real number part and the imaginary number part are I(n-1) and Q(n-1) respectively, the above-mentioned rotation vector speed calculation part is based on {Q(n)-Q(n-1)}·I(n)-{I The calculation of (n)-I(n-1)}·Q(n) obtains the angular velocity of the rotation vector.

The rotation vector velocity calculating means may be configured to include means for obtaining an average value within a predetermined time period of the calculation results obtained by the rotation vector velocity calculating means. As a more specific example, it is configured as means for obtaining a moving average. Furthermore, it is preferable that a correction processing unit that divides the real number part and the imaginary number part from the scalar of the vector determined by the real number part and the imaginary number part is provided in the preceding stage of the above-mentioned rotation vector speed calculation means.

More preferably, the present invention is constituted as follows. That is, constituted as further having:

an inverse rotation vector generation unit that generates a real part and an imaginary part when the complex number represents an inverse rotation vector that rotates inversely to the above-mentioned rotation vector with an angular velocity corresponding to a change amount of the angular velocity obtained by the above-mentioned rotation vector velocity calculation means; and

a frequency range correcting section, which is provided in the front stage of the above-mentioned rotation vector speed calculating section, and calculates the real number part and the imaginary number section of the above-mentioned rotation vector obtained by the above-mentioned rotation vector extracting section, and the reverse rotation generated by the above-mentioned inverse rotation vector generation section. The real part and the imaginary part of the vector, and delay the angular velocity of the above-mentioned rotating vector, and correct the range of the frequency variation.

At this time, the inverse rotation vector generation unit has a data table of sets of real part and imaginary part that define the position of the inverse rotation vector on the complex surface arranged sequentially in the direction of rotation, and the Increment number or decrement number generates the address of the above data table to generate the address control section of the inverse rotation vector. As a more specific example, it can be mentioned that the inverse rotation vector generation unit has a pulse train outputting a duty ratio corresponding to the value of the lower bit when the frequency change amount obtained by the above-mentioned rotation vector speed calculation means is represented by a digital signal. The pulse width control part, and the addition part that adds the value of the upper bit when the above-mentioned frequency change amount is represented by a digital signal and the level of the pulse generated by the above-mentioned pulse width control part, and outputs it to the above-mentioned address control part,

The address generating unit integrates the output value from the adding unit, and uses the integrated value as an address of the data table.

In the present invention, a frequency signal from a sensor vibrator such as a quartz vibrator is sampled based on a clock signal from a reference clock generator, the sampled value is output as a digital signal, and the frequency signal corresponding to the output signal is digitally processed. In the quadrature detection of the signal, the signal rotated at a frequency corresponding to the difference between the frequency of the frequency signal (the output signal of the analog/digital converter) specified by the sampling value and the frequency specified by the digital signal used for quadrature detection is extracted. Rotation vector. Since the angular velocity of the rotation vector corresponds to the oscillation frequency (natural frequency) of the sensor vibrator, by detecting the angular velocity of the rotation vector when the sensor vibrator is immersed in a solvent such as pure water and the The angular velocity of the rotation vector when the liquid is supplied to the medium can detect the amount of change in the angular velocity of the rotation vector, and the adsorption amount of the sensing object adsorbed on the vibrator for the sensor can be known. In addition, in the present invention, it is also possible to check the angular velocity of the rotation vector with the above-mentioned detection amount line by grasping, for example, the correspondence between the various concentrations of the object to be sensed in the solvent and the angular velocity of the rotation vector in advance based on the detection amount line, etc. Estimates the density of the object to be sensed.

Therefore, compared with the method of counting pulses to obtain the frequency, the oscillation frequency of the sensor vibrator can be detected with high accuracy and in an extremely short time. Therefore, the amount of change in the oscillation frequency of the sensor vibrator can be detected with high accuracy and in an extremely short time, and it is useful as a device for detecting extremely minute amounts of environmental pollutants.

And by using the angular velocity equivalent to the frequency change amount obtained by the above-mentioned rotation vector speed calculation means, that is, the angular velocity corresponding to the above-mentioned rotation vector, a reverse rotation vector that rotates in the opposite direction to the rotation vector is generated, and the reverse rotation vector and the above-mentioned rotation Vector multiplication can delay the angular velocity of the above-mentioned rotation vector. As a result, even if the oscillation frequency of the sensor vibrator is high, the angular velocity of the rotation vector can be detected by delaying the angular velocity, and consequently, the frequency measurement range can be expanded.

Description of drawings

FIG. 1 is a block diagram showing the overall configuration of an embodiment of the sensing device of the present invention.

FIG. 2 is a diagram showing the configuration of a carrier remover and a low-pass filter used in the above embodiment.

FIG. 3 is an explanatory diagram showing a rotation vector corresponding to a frequency change amount in a frequency signal of a quartz resonator.

FIG. 4 is a configuration diagram showing a correction processing unit used in the above embodiment.

FIG. 5 is an explanatory diagram showing a situation where a detection error occurs when the rotation vector is delayed.

FIG. 6 is an explanatory diagram showing a phase difference between rotation vectors sampled at timings before and after the phase.

FIG. 7 is a diagram showing a configuration of a speed calculation unit used in the above-mentioned embodiment.

FIG. 8 is a block diagram showing the overall configuration of another embodiment of the present invention.

FIG. 9 is an explanatory diagram showing how a rotation vector and an inverse rotation vector are multiplied in another embodiment described above.

FIG. 10 is an explanatory diagram showing a data table for generating an inverse rotation vector in the other embodiment described above.

FIG. 11 is a block diagram showing in detail the configuration of main parts in another embodiment described above.

FIG. 12 is a timing chart showing a part of operations in the other embodiment described above.

FIG. 13 is a timing chart showing a part of operations in the other embodiment described above.

FIG. 14 is a characteristic diagram showing an example of an input value of a pulse width circuit unit used in the other embodiment described above.

FIG. 15 is a characteristic diagram showing an example of an output value of the pulse width circuit unit.

FIG. 16 is a characteristic diagram showing an example of an output value of an adder used in the other embodiment described above.

FIG. 17 is an explanatory diagram showing a state in which a frequency signal from an oscillation circuit of a crystal resonator is sampled and converted into a digital value in a specific embodiment of the present invention.

FIG. 18 is a characteristic diagram showing the result of verifying the frequency spectrum of the output signal obtained in FIG. 8 .

Fig. 19 is a characteristic diagram showing the I value and Q value obtained by the carrier remover in the above embodiment.

Fig. 20 is a characteristic diagram showing the I value and Q value obtained by the low-pass filter in the above embodiment.

Fig. 21 is a characteristic diagram showing the relationship between the offset frequency and the detection output in the above-mentioned embodiment.

FIG. 22 is a characteristic diagram showing how the detection value of the frequency change amount rises in the above-mentioned embodiment.

Detailed ways

FIG. 1 is a schematic diagram showing the overall configuration of an embodiment of the sensing device of the present invention. 1 is a sensor part, the sensor part 1 has a piezoelectric vibrator as a sensor vibrator, such as a quartz vibrator 11, which is sealed with quartz or the like, and one surface is exposed, and the other surface is exposed as a blue vibrator immersed in a solution 12 in which a substance to be sensed is present. The composition of the Langevin type oscillator. An adsorption layer for adsorbing (capturing) a substance to be sensed, for example, an adsorption layer containing an antibody for adsorbing dioxin, is formed on the surface of the quartz vibrator 11 that is in contact with the solution. 13 is an oscillation circuit for oscillating the crystal resonator 11, and outputs, for example, a high-frequency signal of a sine wave as a frequency signal.

Reference numeral 2 designates a reference clock generation unit that samples a high-frequency signal from the oscillation circuit 13 and outputs a clock signal that is a frequency signal with extremely high frequency stability. 21 is an A/D (analog/digital) converter, which samples the high-frequency signal from the oscillation circuit 13 based on the clock signal from the reference clock generator 2, and outputs the sampled value as a digital signal. The high-frequency signal specified by this digital signal includes high-frequency waves in addition to the fundamental wave. That is, when sampling a sine wave with high-frequency wave distortion, assuming that the high-frequency component is affected by folding, the frequency of the fundamental wave and the frequency of the high-frequency wave overlap on the frequency axis in the spectrum depending on the situation. This overlap must therefore be avoided in order to be able to obtain correct perceptual work.

Generally, when a sine wave signal of frequency f1 is sampled by a clock signal of frequency fs, the frequency f2 of the acquired result is expressed by formula (1). Where mod (,) represents the modulo (module) function.

f2＝|mod(f1+fs/2，fs)-fs/2|...(1)

In this fetched result, since the frequency of the nth high-frequency wave is expressed as n×(fundamental frequency) for the fundamental frequency, if it is placed on f2 and substituted into the above formula (1), it can be calculated as what The frequency is taken into the high-frequency wave. The frequency fs of the high-frequency signal from the oscillation circuit 13 and the sampling frequency (frequency of the clock signal) fs can be set so that the frequency of the fundamental wave and the frequency of the high-frequency wave do not overlap by using this calculation. For example, fs is set to 11 MHz. , fs is 12MHz. At this time, the fundamental wave of the frequency signal specified by the output signal as the digital signal from the A/D converter 21 becomes a 1 MHz sine wave. Also, if fc/fs is set to 11/12, the frequency of the fundamental wave and the frequency of the high frequency wave do not overlap, but fc/fs is not limited to this value.

In the subsequent stage of the A/D converter 21, a carrier remover 31 and a low-pass filter 32 are provided in this order. The carrier remover 31 and the low-pass filter 32 are used to take out the difference between the frequency of the sine wave signal of, for example, 1 MHz specified to pass the digital signal from the A/D converter 21 and the frequency of the sine wave signal used for quadrature detection The frequency of rotation of the rotation vector.

In order to explain easily the action of extracting the rotation vector, the sine wave signal specified by the digital signal from the A/D converter 21 is assumed to be Acos(ω0t+θ). On the other hand, the carrier remover 31 has a multiplication unit 31a for multiplying the sine wave signal by cos(ω0t) and a multiplication unit 31b for multiplying the sine wave signal by -sin(ω0t) as shown in FIG. 2 . That is, quadrature detection is performed by such calculation. The output of the multiplication unit 31a and the output of the multiplication unit 31b are represented by formula (2) and formula (3), respectively.

Acos(ω0t+θ)·cos(ω0t)

＝1/2·Acosθ+1/2{cos(2ω0t)·cosθ+sin(2ω0t)·sinθ}...(2)

Acos(ω0t+θ)-sin(ω0t)

＝1/2·Asinθ-1/2{sin(2ω0t)·cosθ+cos(2ω0t)·sinθ}...(3)

Therefore, since the output of the multiplication unit 31a and the output of the multiplication unit 31b are respectively passed through the low-pass filters 32a and 32b, and the frequency signal of 2ω0 is removed, 1/2·Acosθ and 1/2· Asin theta. In addition, it is described that the low-pass filter 32 is composed of low-pass filters 32a and 32b. The actual digital processing in the low-pass filter 32 calculates a moving average of a plurality of consecutive data, for example, six data, with respect to the time-series data output from the carrier remover 31 .

And when the frequency of the sine wave signal represented by Acos(ω0t+θ) changes, Acos(ω0t+θ) becomes Acos(ω0t+θ+ω1t). where ω1 is sufficiently small compared to ω0. Therefore, 1/2·Acosθ becomes 1/2·Acos(θ+ω1t), and 1/2·Asinθ becomes 1/2·Asin(θ+ω1t). That is, the output obtained from the low-pass filter 32 is a signal corresponding to the frequency change amount ω1/2π of the sine wave signal [Acos(ω0t+θ)]. That is, these values are the real number part (I) and the imaginary number part (Q) when rotating the vector at the frequency of the sine wave signal specified by the digital signal from the A/D converter 21 and the sine wave signal used for quadrature detection. The frequency of the difference between the frequency ω0/2π of the wave signal is rotated.

FIG. 3 is a schematic diagram showing the rotation vector, the length of which is A, and the angular velocity is ω1. ω1t is denoted as φ in FIGS. 2 and 3 . Therefore, if, for example, the frequency signal from the A/D converter 21 when no substance to be sensed is adsorbed on the quartz vibrator 11 has a frequency of ω0/2π, since ω1 is zero, the angular velocity of its rotation vector is zero, but when Since the substance to be sensed is adsorbed on the crystal vibrator 11 and the frequency of the crystal vibrator 11 changes, when the frequency of the sine wave signal changes, the rotation vector rotates at an angular velocity corresponding to the amount of change. Also, when the frequency signal from the A/D converter 21 deviates from ω0/2π when the sensing target substance is not adsorbed on the quartz vibrator 11, the rotation vector rotates at an angular velocity corresponding to the deviated frequency. In any case, since the angular velocity of the rotation vector is a value corresponding to the oscillation frequency of the quartz vibrator 11, for example, when the quartz vibrator 11 is immersed in a solvent and the object to be sensed is added to the solvent to obtain By sensing the angular velocity of each rotation vector when the object is adsorbed on the crystal vibrator 11 and obtaining the difference in angular velocity, the amount of change in the oscillation frequency due to the adsorption of the object to be sensed on the crystal vibrator 11 can be known.

In this way, the carrier remover 31 performs quadrature detection on the above-mentioned sine wave signal, and the low-pass filter 32 removes high-frequency components from the detection result, so that they are means for extracting the real part and the imaginary part of the rotation vector, wherein the rotation vector is represented by The complex number represents the frequency signal from the A/D converter 21 for quadrature detection based on the digital signal, and the angular velocity corresponding to the frequency difference between the frequency in the frequency signal and the frequency ω0/2π of the sine wave signal used for the quadrature detection rotate.

Returning to FIG. 1 , a reduction processing unit 4 is provided after the low-pass filter 32 . The decrement processing unit 4 performs decrement processing (thinning processing) of extracting data every 120, for example, with respect to a digital value group obtained by a clock signal of 12 MHz, which is a time-series digital signal obtained from the low-pass filter 32. installation. In this way, the computational burden of the computer is reduced by reducing the amount of processing. In addition, in this embodiment, as will be described later, the angular velocity of the rotation vector is obtained by grasping how many times the rotation vector has rotated in the sampling interval, so no matter how much the digital value group is thinned out, the detection of the above-mentioned angular velocity Accuracy (detection accuracy of the amount of change in frequency) has no influence.

In the subsequent stage of the decrementing processing part 4, a correction processing part 5 is provided, and this correction processing part 5 divides the above-mentioned rotation vector which is passed through the low-pass filter 32 and undergoes decrement processing by the scalar amount of the rotation vector, respectively. The I value and the Q value as the imaginary part of the above-mentioned rotation vector are processed to obtain the I value and Q value for each unit length of the rotation vector. That is, as shown in FIG. 4 , the correction processing unit 5 obtains the scalar value | V| is formed by dividing the I value and the Q value by |V|.

The reason for correcting the I value and Q value in this way is as follows. In this embodiment, whenever calculating how many times the rotation vector V has rotated in the sampling interval, as shown in FIG. - 1) The factor of the vector ΔV of the rotation vector V(n-1) obtained by sampling is evaluated. Therefore, there is a possibility that when ΔV becomes ΔV′ due to waveform fluctuation of the high-frequency signal from the oscillation circuit 13 so to speak, and ΔV becomes ΔV′, the correspondence relationship between ΔV and the rotation amount Δφ of the rotation vector collapses, impairing the detection of the angular velocity of the rotation vector value reliability. Therefore, by performing the correction processing as described above, the I value and the Q value at each timing are made to coincide with the values corresponding to the unit length of the rotation vector, so that the influence of the delay of the rotation vector can be eliminated.

Further, as shown in FIG. 1 , a speed calculation unit 6 for obtaining the angular velocity of the rotation vector is provided at a subsequent stage of the correction processing unit 5 . Referring to Fig. 6 and Fig. 7, illustrate this speed calculation section 6, as shown in Fig. 6, when setting constant as K, if the angular velocity (frequency) of rotation vector is compared with sampling frequency sufficiently small, then can be with formula (4) Approximate the angle Δφ formed by the rotation vector V(n-1) obtained from the (n-1)th sample and the rotation vector V(n)=V(n-1)+ΔV obtained from the n-th sample . where Δφ is the difference between the phase φ(n) of V(n) and the phase φ(n-1) of V(n-1), and imag is the imaginary part, conj{V(n)} is V(n) Conjugate vector.

Δφ=K·imag[ΔV·conj{V(n)}]...(4)

Here, if the values corresponding to the nth sample are set as I(n) and Q(n) for the I value and the Q value, respectively, when the complex numbers represent ΔV and conj{V(n)}, the formula (5) And formula (6) said.

ΔV=ΔI+jΔQ...(5)

Conj{V(n)}=I(n)-jQ(n)...(6)

where ΔI is I(n)-I(n-1), and ΔQ is Q(n)-Q(n-1). After substituting formula (5) and formula (6) into formula (4) and sorting out, Δφ is expressed by formula (7).

Δφ=ΔQ·I(n)-ΔI·Q(n)...(7)

The above-mentioned speed calculation unit 6 is a device for obtaining an approximate value of Δφ by performing the calculation of the formula (7), and its configuration is shown in FIG. 7 . When the I value input to the speed calculation section 6 is I(n) as a value corresponding to the n-th sample, I(n-1) corresponding to the (n-1)-th sample at one previous timing ) are held in the register 61, they are checked in the checking circuit section 62, and the difference ΔI between I(n) and I(n-1) is taken out, and I(n) and ΔI are input to the calculation section 65. In addition, the Q value is also processed by the register 63 and the collation circuit unit 64 in the same manner, and Q(n) and ΔQ are input to the calculation unit 65 . In addition, in the calculation unit 65, the calculation of the formula (7) is performed to obtain Δφ. The calculation result of the calculation unit 65 is evaluated in detail as Δφ.

Here, if the rotation vectors V(n-1) and V(n) are obtained, the method of obtaining the angle Δφ between them or the method of evaluating it can use various mathematical methods, and as an example, only the formula ( 4) Approximate formula. This formula can also be used as {V(n)+V(n-1)}/ 2. Substitute V(n) into the vector V0 in formula (4). The reason why formula (4) can be approximated in this way is because V0 and ΔV can be regarded as orthogonal, so the length of ΔV can be treated as equivalent to the imaginary value of ΔV when V0 is regarded as a real axis. Moreover, as a method of calculating Δφ, an intuitive and easy-to-understand method may be a method of calculating argV(n) and argV(n-1) and subtracting them. It is a data table corresponding to the phase φ of this vector, so it is a good method to calculate according to the above-mentioned formula (4) in terms of reducing the burden on the computer. Also argV(n) is tan ^-1 (imaginary value/real value).

As shown in FIG. 1 , a time-average processing unit 7 is provided at a subsequent stage of the speed calculation unit 6, and the time-average processing unit 7 takes out and outputs time-series data for Δφ as a calculation result obtained by the speed calculation unit 6. Time averaging processing performed, for example, moving average of a predetermined amount of data. The output value obtained here is input to the frequency difference detection unit 71 . This frequency difference detection unit 71 is for detecting the angular velocity of the rotation vector when the quartz vibrator 11 is placed in a medium such as pure water as the first environment, and when the quartz vibrator 11 is placed in a supply medium as the second environment. A device that senses the difference in the angular velocity of the rotation vector when the object is in a solvent. Since the angular velocity of the rotation vector is a value corresponding to the oscillation frequency of the crystal vibrator 11, the angular velocity detection difference detected by the frequency difference detection unit 71 is a value corresponding to the amount of change in the oscillation frequency caused by the sensing substance being adsorbed on the quartz vibrator 11. , which can be called a value for evaluating the adsorption amount of the sensed substance. More specifically, the frequency difference detection unit 71 has a memory for storing the angular velocity of each rotation vector, means for reading the angular velocity of each rotation vector in the memory to calculate the difference in angular velocity, and the like.

As described above, the configuration of the present embodiment has been described in blocks, and actual calculation and data processing are implemented by software.

The operation of the above-mentioned embodiment will be described below. For example, the frequency signal of 11 MHz is converted by the frequency signal of 12 MHz with the A/D conversion part 21, including the sinusoidal wave as the fundamental wave generated by the oscillation of the quartz vibrator (oscillating circuit 13) of the sensor part 1, and the frequency signal of 11 MHz is converted from the A/D conversion part 21. A signal including a sine wave signal of about 1 MHz as a fundamental wave is output. Here, for the convenience of explanation, when the sine wave signal is set to Acos(ω0t+ω1t+θ) (ω1 is sufficiently small compared with ω0), by performing quadrature detection on the sine wave signal, the lower frequency components are removed, and the A rotation vector that rotates at an angular velocity corresponding to the amount of change in frequency of the sine wave signal. That is, the real part and the imaginary part of the rotation vector are taken out as I value and Q value, respectively. These I values and Q values are decremented by the decrement processing unit 4, further divided by the scalar |V| 6. In addition, in order to avoid the complexity of the description, the same symbol "V" will be added to the rotation vector before and after the correction processing for description.

The rotation vector V rotates at an angular velocity of ω1 which is the difference between the frequency of the sine wave signal Acos(ω0t+ω1t+θ) and the frequency ω0/2π of the sine wave signal used for quadrature detection (see FIG. 3 ). Here, for ease of description, for example, the sensor unit 1 is immersed in a solution for detecting the presence of the sensing target substance, and when the sensor unit 1 is set to be the same as when the sensing target substance does not exist (when it is below the detection limit of the quartz oscillator 11), When the angular velocity corresponding to the natural vibration frequency of 11 is ω0, since ω1 is zero, the rotation vector V is static. Therefore, Δφ, which is the output of the time averaging processing unit 7, becomes zero.

On the other hand, when a substance to be sensed such as dioxin exists in the solution, the oscillation frequency (natural frequency) changes according to the amount of dioxin adsorbed on the quartz oscillator 11 . At this time, the above-mentioned rotation vector V starts to rotate at an angular velocity corresponding to the amount of frequency change. To simplify the description, it is assumed that the rotation vector V rotates once in 1 second when the change amount of the set frequency is 1 Hz. Here, in this embodiment, the sampling is to detect the angular velocity of the rotation vector by grasping the difference Δφ between the phase φ(n) of V(n) before and after the phase and the phase φ(n-1) of V(n-1). Sampling, assuming that the sampling interval is 1/100 second, Δφ becomes 3.6 degrees. In other words, the amount of change in the oscillation frequency of the quartz vibrator 11 can be known only by obtaining the angular velocity of the rotation vector V within the time interval between the rotation vectors.

However, since it is quite rare that the angular velocity corresponding to the oscillation frequency of the quartz vibrator 11 when there is no sensing object substance coincides with the angular velocity of the sine wave signal used for quadrature detection, it is actually necessary to obtain the sum when there is no sensing object The angular velocity of the rotation vector corresponding to the oscillation frequency of the quartz vibrator 11 when the substance is present and the angular velocity of the rotation vector corresponding to the oscillation frequency of the quartz vibrator 11 when the sensing target substance exists, and the difference between the two angular velocities is obtained. Since the difference in the angular velocity of this rotation vector is a value corresponding to the amount of change in the frequency of the crystal oscillator 11 caused by the adsorption of the sensing substance on the quartz oscillator 11, by grasping the relationship between the amount of adsorption of the sensing object substance and the amount of change in frequency , the adsorption amount of the sensing object substance at this time can be obtained. And if the relationship between the concentration of the substance to be sensed in the solution and the amount of adsorption of the substance to be sensed is obtained in advance, then according to the amount of adsorption of the substance to be sensed, that is, the amount of change in the frequency of the quartz vibrator 11, the concentration of the substance to be sensed in the solution can be known. concentration. Therefore, the concentration of the sensing substance in the sample solution supplied with the solution (12 in FIG. 1 ) in which the crystal vibrator 11 is immersed can be known.

Thus, according to the above-mentioned embodiment, the oscillation frequency of the crystal resonator 11 (the frequency of the frequency signal from the oscillation circuit 13 ) is digitally processed, and the angular velocity of the rotation vector is detected from the phase difference corresponding to the sampling interval. As a result, compared with the conventional method of counting pulses, as will be demonstrated in Examples described later, the amount of change in the oscillation frequency of the crystal resonator 11 can be detected with high accuracy and in an extremely short time. In the conventional pulse counting method, since sinusoidal waves are counted as pulses, it takes, for example, 1 second to recognize a 10 MHz pulse with a resolution of 1 Hz. As apparent from the above description, the present invention is extremely useful as a device for detecting trace substances including environmental pollutants.

As the method of use of the present invention, not limited to immersing the sensor part 1 in the solution, the solution may also be dripped on the surface of the quartz vibrator 11, and environmental pollutants other than dioxin such as PCB etc. may be sensed, or may be sensed Virus.

Moreover, in the present invention, for example, by making various changes in the concentration of the substance to be sensed in the solution in contact with the quartz vibrator 11, the relationship between each concentration and the angular velocity of the rotation vector is obtained in advance, and then the solution and the quartz vibrator 11 The angular velocity of the rotation vector at the time of contact and this relationship estimate the adsorption amount of the sensing target substance.

Furthermore, the state in which the quartz vibrator 1 is in contact with a solvent such as pure water may be used as a so-called blank value, but the state in which the quartz vibrator 1 is not in contact with a solvent but exposed to the air may be used as a blank value to capture when the crystal vibrator 1 is used. The amount of change in the rotation number of the rotation vector when the solution comes into contact with the quartz vibrator 1 to absorb the sensing target substance.

Furthermore, the present invention is not limited to sensing substances present in liquids, and can be applied to detecting the smell of petroleum, detecting smoke during fires, detecting poisonous gases such as sarin gas, detecting gases in parts, and detecting purification of semiconductor manufacturing plants, etc. Various fields such as indoor gas.

In addition, the present invention can be used to detect the concentration of the substance to be sensed by indicating the amount of change in frequency itself, but it can also be formed as a method that does not detect the concentration, for example, has a threshold value for the amount of change in frequency, and only knows the presence or absence of the substance to be sensed. The composition of the device. Such a case is naturally included in the technical scope of the present invention because the amount of change in frequency can be captured.

Next, other embodiments of the present invention will be described. The purpose of this embodiment is to further expand the measurement range of the frequency difference in the above-mentioned embodiments. In the previous embodiment, since the angular velocity of the rotation vector was evaluated as the length of a straight line connecting the endpoints of V(n) and V(n-1) as shown in FIG. 6, when the phase difference between these rotation vectors is large measurement error increases. Therefore, in this embodiment, the angular velocity of the rotation vector can be reduced by creating a reverse rotation vector that rotates in the opposite direction at an angular velocity corresponding to the angular velocity of the rotation vector, and multiplying the above-mentioned rotation vector by the reverse rotation vector. Or slowly, the phase difference between V(n) and V(n-1), that is, the angular velocity of the rotation vector, can be detected with high precision, which expands the amount of change in frequency (frequency difference) when the sample solution is supplied to the sensor. range of measurement.

FIG. 8 shows the overall configuration of this embodiment, and an integrator 70 is provided between the output terminal of the frequency difference calculation unit 6 and the time averaging processing unit 7 . On the other hand, a frequency range correction unit 8 is provided between the low-pass filter 32 and the decrement processing unit 4, and a device for generating an inverse rotation vector that rotates inversely to the above-mentioned rotation vector V in accordance with the output of the integrator 70 is provided. Here, the inverse rotation vector generation unit 9 multiplies the generated inverse rotation vector by the frequency range correction unit 8 by the rotation vector specified from the time-series data output from the low-pass filter 32 .

Now, assuming that the frequency difference (frequency variation) is, for example, 500 Hz, in the rotation vector corresponding to this 500 Hz, as shown in FIG. n). When returning this vector to the position along the real axis, an inverse rotation vector V' that rotates inversely to the above-mentioned rotation vector V at an angular velocity corresponding to a frequency difference of 500 Hz can be created and multiplied by the inverse rotation vector V'. The vector I+jQ rotated by the rotation vector V by the inverse rotation vector V' is {I(n)+jQ(n)}×{I'(n)+jQ'(n)}. When this formula is sorted out, it becomes Formula (8), and the frequency range correcting unit 8 performs calculation of this formula.

I+jQ(n)={I(n)·I'(n)-Q(n)·Q'(n)}+j{I(n)·Q'(n)+I'(n)· Q(n)}...(8)

When an inverse rotation vector V' occurs, the values of the real and imaginary parts of the vector are set in such a way that the vector is actually inversely rotated on the complex plane, that is, when the phase of the inverse rotation vector V' is φ, cosφ and The value of sinφ. Fig. 10 shows the I/Q table 90 of the group of cos φ and sin φ of the vector arranged in sequence along the direction of rotation of the vector. The corresponding increment or decrement reads the address of the I/Q table 90 and outputs it to the frequency range correction unit 8 . For example, read addresses from "0" to "11" one by one according to the read timing of the clock, and when returning to "0" again, the 12 clock vector rotates clockwise on the complex number plane once, and the number of increments becomes 2 , every other read address, the angular velocity of the vector is doubled. Therefore, the number of increments is determined according to the size of the integrator 70, and the inverse of the angular velocity (the angular velocity corresponding to the angular velocity of the rotation vector V) corresponding to the frequency difference Δφ calculated by the above-mentioned frequency difference calculation unit 6 (see FIG. 8 ) can be generated. The inverse rotation vector of the rotation.

As mentioned above, when the sine wave signal Acos(ω0t+θ) changes to Acos(ω0t+θ+ω1t), the rotation vector V rotates at the angular velocity of ω1 (refer to Figure 3), so it is determined according to the value of ω1 Whether the clock is turning or counterclockwise. Therefore, the direction of the reverse rotation vector V' is also determined corresponding to the direction of the rotation vector V, and the output of the integrator 70 becomes a positive value or a negative value corresponding to the direction. Regarding the address reading of the I/Q table 90, when the integrator When the output of the integrator 70 is a positive value, it is read with an increment corresponding to the value, and when the output of the integrator 70 is negative, it is read with a decrement corresponding to the value. That is, the I/Q table 90 has a relationship of cos and sin, and the correction direction of the rotation vector V is controlled by the address control unit 103 by incrementing or decrementing the address of the I/Q table 90 . In addition, in order to facilitate the understanding of the present invention, the I/Q table of FIG. 10 is created schematically, and a preferred example of actual table creation is not shown.

A preferred example of the inverse rotation vector generator 9 is shown in FIG. 11 . In this example, the inverse rotation vector generation unit 9 has a bit division unit 100 that divides the output value of the integrator 70 into an upper bit value and a lower bit value. For example, when the output value of the integrator is 16 bits, it is divided into an upper 8-bit value and a lower 8-bit value and output. In this example, by multiplying the value of the upper 8-bit BCD code (binary decimal number) by 1/M (M is 10 to the negative 8th power) in the output value of the integrator 70 represented by 16 bits, ), make the upper 8-bit value (decimal conversion value), subtract the value of the original 8-bit BCD code value by multiplying this value by M from the 16-bit BCD code as the above output value, and make the lower 8-bit value Bit value (decimal conversion value). In addition, the method of dividing the output value of the integrator 70 into the value of the upper bit and the value of the lower bit is not limited to this method, and only the value of the upper bit and the value of the lower bit may be taken out.

Moreover, a pulse width control unit 101 is provided on the output side of the lower bits of the bit division unit 100 (in this example, an output terminal side that outputs a value of the lower 8 bits), and an adder 101 is further provided in a subsequent stage of the pulse width control unit 101. The adder 102 adds the pulse train output by controlling the pulse width according to the value of the lower bit and the value of the upper bit. Furthermore, an address control unit 103 is provided on the rear stage side of the adder 102, and the address control unit 103 integrates the value of the adder 102, and reads the address in the I/Q table 90 according to the integrated value. To control, that is, to control the increment or decrement of the address to form.

The operation of this embodiment will be described below. First, since the object to be sensed is attracted to the crystal vibrator 11, when the frequency of the crystal resonator 11 changes from the state where the object to be sensed is not attracted, the rotation vector rotates at an angular velocity corresponding to the frequency difference that is the amount of change in the frequency. Then, a signal at a level of an upper bit, for example, an upper 8-bit value of the output value of the integrator 70 corresponding to the angular velocity of the rotation vector V is input to the adder 102 . On the other hand, the lower bits of the output value of the integrator 70 , for example, the lower 8 bits, are input to the pulse width control unit 101 . In the pulse width control unit 101, the pulse width is calculated based on the sampling signal generated for every predetermined number of pulses with respect to the clock pulse of the computer, and a pulse train having a duty ratio corresponding to the input value is output.

This pulse train is formed by combining +1 level pulses generated in 1 clock and -1 level pulses generated in 1 clock, assuming that the pulse width is calculated every 20 clocks, when combined with the pulse width control section When the duty cycle corresponding to the input value of 101 is 50%, as shown in FIG. 12 , pulses of +1 level and pulses of -1 level are alternately output every 10. Furthermore, for example, when the level corresponding to the value of the upper 8 bits is "5", between the generation of the sampling signal and the generation of the next sampling signal, "6" and "4" are alternately output from the adder 102, and the address The control unit 103 integrates the output value, and the integrated value becomes the address of the I/Q table 90 . That is, at this time, since the increment of the integral value alternately repeats "6" and "4", the increment number of the address alternately repeats "6" and "4", and the result is used as a value equivalent to the value of the upper 8 bits. The address is accessed by an average increment of "5", and the real part and the imaginary part of the inverse rotation vector V' written in the address are read. That is, an inverse rotation vector V' of an angular velocity corresponding to "5" is generated. In addition, the I/Q table 90 in FIG. 10 is easy to understand and does not correspond to the operation at this time. The values of the real part and the imaginary part thus read from the I/Q table 90 are calculated by the above formula (8) in the frequency range correction unit 8, and the reverse rotation vector V' is multiplied by the rotation vector V. In addition, 81 is a bit processing unit, which performs rounding of the lower digits to an integer in order to reduce the number of calculation bits in the frequency difference calculation unit 6 and later.

And, because the inverse rotation vector generator 9 shown in FIG. 11 constitutes a PLL (Phase Lock Loop: Phase Locked Loop), so if the signal value input to the frequency range correction unit 8 from the low-pass filter 32 side is stabilized, the lock is immediately locked. The angular velocity of the rotation vector V and the angular velocity of the reverse rotation vector V' form a steady state, and each signal is shown in FIG. 12 . That is, the rotation vector V is stabilized by reducing the angular velocity due to the reverse rotation vector V', but because this angular velocity corresponds to the angular velocity when it is not multiplied by the reverse rotation vector V', the angular velocity of the rotation vector V locked by the PLL is the same as that of the quartz oscillator Corresponding to the frequency difference, the measurement range is expanded as a result. That is, even if the amount of change in frequency is large and the angular velocity of the rotation vector is large, the angular velocity can be obtained, that is, the amount of change in frequency can be measured.

Moreover, when the duty ratio corresponding to the input value of the pulse width control unit 101 is greater than 50%, as shown in FIG. Flat pulses and -1 level pulses alternately occur. Therefore, since the increment of the integral value of the pulse train is "6" consecutive at first, the number of increments of the address is "6" consecutively, and "6" and "4" are alternately repeated after a while. Therefore, the average value of the angular velocity of the inverse rotation vector V' at the interval from sampling the input value of the pulse width control unit 101 to the next sampling interval is between the angular velocity corresponding to "5" and the angular velocity corresponding to "6". , becomes a size corresponding to the above-mentioned duty ratio. That is, this angular velocity becomes an angular velocity corresponding to a value below a decimal point added to "5", which means that the input value of the pulse width control unit 101 (the lower-order value of the output of the integrator 70) is interpolated in " Between 5" and "6". In addition, in this example, the timing of pulse width calculation is set to 20 clocks, but the number of clocks can also be set, for example, the number of lower digits, which is 8 in this example, and the pulse width calculation is performed every 8 clocks.

In the case of the method of determining the increment number of the address of the I/Q table 90 according to the output value of the output of the integrator 70 without providing the bit division unit 100 and the pulse width control unit 101, the I/Q table 90 needs The number corresponds to the required accuracy (frequency difference detection resolution), and is larger than that of pulse width control. On the other hand, by configuring as in the present embodiment, it is possible to reduce the storage capacity of the I/Q table 90 only by the amount added during the pulse width control.

Here, as the case where the output value of the integrator 70 does not appear in the upper bits, taking the case where the frequency difference of the crystal resonator is 500 Hz as an example, FIG. 14 and FIG. level changes and output level changes result. As shown in FIG. 14 , it can be seen that the input level of the pulse width control unit 101 is stable for approximately 1 msec, and the angular velocity of the rotation vector V is instantaneously locked by the PLL including the reverse rotation vector generation unit 9 . Moreover, the output of the pulse width control unit 101 is actually a level signal of +1 and -1, and its value is input to the address control unit 103, but it is drawn into a straight line at +1 and -1 in relation to the image resolution to be drawn. . 16 shows the output level of the adder 102 in the case where the output value of the integrator 70 appears in the upper bit, taking the case where the frequency difference of the crystal resonator is 10000 Hz as an example. The output of adder 102 is approximately stable at "81" in this example.

According to the other embodiments described above, in addition to the effects of the above-mentioned embodiments, the following effects are further obtained. Since the rotation vector V' is multiplied by the rotation vector V corresponding to the change in the frequency (natural frequency) of the quartz resonator, the angular velocity of the rotation vector V is reduced. Therefore, for example, as shown in FIG. In the case where the angular velocity is evaluated as the length of a straight line connecting a vector at a certain timing and a vector at the next timing, the angular velocity of the rotation vector V can be obtained with high precision, so even if the angular velocity of the rotation vector V can be fast or slow, Since this angular velocity can also be obtained with high precision, it is possible to measure the amount of change in the natural frequency of the quartz vibrator from a large value to a small value, thereby expanding the measurement range. In addition, the angular velocity of the rotation vector V to be fed back is divided into a high-order value and a low-order value, and the data value of the I/Q table 90 is interpolated by pulse width control for the low-order value, so it can be described in detail. Reducing the size of the I/Q table 90 can reduce the circuit scale.

As an experiment to verify the device of the present invention, the frequency fc of the frequency signal from the oscillation circuit 13 was set to 1 MHz, the frequency fs of the reference clock signal was set to 12 MHz, and data processing was actually performed by a computer. A high-frequency signal (sine wave signal) for testing used as a frequency signal from the oscillation circuit 13 was slightly changed in frequency from 1 MHz. FIG. 17 is a schematic diagram showing a state in which a high-frequency signal for a test is sampled based on a reference clock signal and a digital value is obtained. FIG. 18 shows how the frequency spectrum of a signal specified by the digital value obtained in this way is investigated. Therefore, here, a high-frequency signal having a 1 MHz sine wave signal as a fundamental wave is taken out from the A/D converter 21 .

FIG. 19 shows the I value and Q value output from the carrier remover 31 , and FIG. 20 shows the I value and Q value obtained from the low-pass filter 32 . In this example, the output value of the low-pass filter 32 increases because the frequency is changed with respect to the high-frequency signal for the test. 21 is a schematic diagram showing the relationship between the offset frequency (frequency variation) and the detection output in this embodiment. When the offset frequency increases, the error increases, but it can be seen that in the area where the offset frequency is small, both The correspondence relationship of is good, and the amount of frequency change can be detected with high reliability. 22 is a schematic diagram showing the output of the time averaging processing unit 7 after changing the frequency of the high-frequency signal for the test, and it can be seen that the amount of change in frequency can be detected within 0.1 second. Furthermore, the frequency detection accuracy is equivalent to the amplitude of the pulsation in the region where the output stops rising and is parallel to the time axis, but the value is 0.0035 Hz, which proves that the frequency detection accuracy is extremely high.

Claims (9)

1. sensing device, be formed with the adsorbed layer that is used to adsorb the perceptive object thing on its surface, use the sensor oscillator that makes the natural vibration frequency change owing to the absorption of perceptive object thing, according to the variation perception perceptive object thing of this sensor with the natural vibration frequency of oscillator, it is characterized in that having:

The sensor oscillatory circuit makes described sensor vibrate with oscillator;

The reference clock generating unit produces and to be used for the clock signal of taking a sample with the frequency signal of oscillator to from described sensor;

Analog/digital conversion portion, by from the clock signal of described reference clock generating unit to taking a sample with the frequency signal of oscillator from described sensor, this sampling value is exported as digital signal;

Rotating vector takes out parts, to with the orthogonal detection that carries out from the corresponding frequency signal of the output signal of this analog/digital conversion portion according to digital signal, and real part and imaginary part when taking out by the complex representation rotating vector, wherein this rotating vector with this frequency signal in frequency and by the suitable angular velocity rotation of the difference on the frequency of the specific frequency of the digital signal that is used for orthogonal detection; With

Rotating vector speed calculation parts according to the described real part that is obtained by these rotating vector taking-up parts and each time series data of imaginary part, are tried to achieve the angular velocity of rotating vector, wherein,

Described rotating vector take out parts comprise to carry out the parts of orthogonal detection and remove the parts that are included in the radio-frequency component in the data that obtain by these parts from the corresponding frequency signal of the output signal of analog/digital conversion portion.

2. sensing device according to claim 1 is characterized in that having:

The parts of the difference of the angular velocity of the rotating vector when trying to achieve the angular velocity of the rotating vector when sensor uses oscillator in first environment and using oscillator in second environment when sensor.

3. sensing device according to claim 1 is characterized in that:

Be respectively I (n) and Q (n) when setting real part and the imaginary part corresponding with the described sampling value in certain timing, when setting real part corresponding and imaginary part and being respectively I (n-1) and Q (n-1) with the described sampling value in the timing before this timing, described rotating vector speed calculation parts are tried to achieve the angular velocity of rotating vector according to the calculating of { Q (n)-Q (n-1) } I (n)-{ I (n)-I (n-1) } Q (n).

4. sensing device according to claim 1 is characterized in that having:

The result of calculation of being tried to achieve by described rotating vector speed calculation parts is obtained the parts of the mean value in the stipulated time.

5. sensing device according to claim 4 is characterized in that:

The parts of obtaining the mean value in the stipulated time are parts of obtaining moving average.

6. sensing device according to claim 1 is characterized in that:

In the leading portion of described rotating vector speed calculation parts, be provided with the correcting process portion that removes described real part and imaginary part according to the scalar of the vector that determines by real part and imaginary part.

7. according to each described sensing device in the claim 1～6, it is characterized in that having:

The angular velocity of being tried to achieve by described rotating vector speed calculation parts is carried out the integrating block of integration;

Contrary rotating vector generating unit makes when real part and the imaginary part of complex representation during with the contrary rotating vector that taken out the rotating vector retrograde rotation that parts take out by described rotating vector with the angular velocity corresponding with the output valve of this integrating block; With

Be arranged in the leading portion of described rotating vector speed calculation parts, to rotating vector that obtains by described rotating vector taking-up parts and the parts that carry out multiplying by the described contrary rotating vector that generates against the rotating vector generating unit.

8. sensing device according to claim 7 is characterized in that:

Contrary rotating vector generating unit has along the tables of data of the group of the real part of the position that is limited to the lip-deep contrary rotating vector of plural number that sense of rotation disposes in turn and imaginary part and generates the address control part of contrary rotating vector by the address that produces described tables of data by increment number suitable with the variable quantity of described frequency or decrement number.

9. sensing device according to claim 8 is characterized in that:

Contrary rotating vector generating unit has: pulse width control part, the spike train of the dutycycle that output is corresponding with the next value when representing the frequency variation of being tried to achieve by described rotating vector speed calculation parts with digital signal; With addition portion, will work as the value of the upper position when representing described frequency variation and the level addition of the pulse that makes by described pulse width control part with digital signal, output to described address control part,

Described address control part carries out integration to the output valve from this addition portion, with the address of this integrated value as described tables of data.

Images