JPH04212015A - Average value rectifying type measuring apparatus - Google Patents

Abstract

PURPOSE:To provide an average value rectifying type measuring apparatus which can correctly measure an input analog signal without adjusting a sampling cycle. CONSTITUTION:An average value rectifying type measuring apparatus comprises AD converting means 1, 2 which sample analog signals to be measured with a specific cycle and convert the sample values into digital signals, an absolute value calculator 6 which admits the digital output signals of the AD converting means to output their absolute values and an arithmetic means 7 which admits the absolute values and applies exponential average calculation to them.

Description

[Detailed description of the invention]

0001

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an average value rectification measuring device for measuring the average value of a rectified waveform of a signal to be measured (input voltage or input current).

0002

[Prior Art] In a device that uses an analog/digital conversion circuit to digitally sample a signal under test and calculate its average value, an integer multiple of one cycle of the signal waveform under test matches an integer multiple of the sampling period. Errors occurred when not doing so. The reason why the error occurs will be explained using drawings. The average value BMEAN of the signal under measurement B was obtained by performing an arithmetic mean calculation as shown in equation (1). BMEAN=(1/N) Σ|Bn|

(1) Bn: Sampling value of the signal under test N: Number of samples Σ: Addition sum when n changes from 1 to N Here, if the signal under test B is a voltage signal, Bn is the sampling voltage value, If it is a current signal, it is a sampling current value.

As shown in the waveform shown in FIG. 5, when one period Tin of the signal under test waveform is not an integral multiple of the sampling period TAD, that is, n1・Tin≠n2・TAD n1=1, 2, 3,... n2 =3, 4, 5,... However, there is a relationship of n2≧2·n1+1. This relationship is meant to prevent continuous sampling of the same phase position of the input waveform. In such a case, a fraction TY occurs as shown in FIG. When a fraction TY occurs in this manner, an error corresponding to the fraction TY occurs in the average value BMEAN obtained by performing the aforementioned arithmetic mean (formula (1) above).

[0004] For this reason, the following methods have conventionally been used to reduce the influence of the above-mentioned fractions. ■ Sampling period TAD of input waveform is n1・Ti
A method of adjusting and controlling TAD so that n=n2・TAD. ■ A method of sampling the input waveform over several cycles until n1・Tin=n2・TAD and calculating the average value. (2) A method of reducing the error by shortening the sampling period TAD and making the fraction of the period Tin of the input waveform extremely small. ■ A method of increasing the number of samples N to relatively reduce the influence of the fraction TY.

[0005]

[Problem to be solved by the invention] However, the above ■~
The conventional method (2) has the following problems. Method (2) requires a phase lock circuit with a wide frequency variable range for adjusting and controlling the sampling period TAD. Therefore, the circuit configuration becomes complicated,
There is also the problem that it is expensive. Method (2) has a problem in that the measurement response time becomes long because measurement cannot be performed until a period that is an integral multiple of one cycle of the input waveform matches a period that is an integral multiple of the sampling period TAD. Method (2) requires a high-speed AD converter that can successively process sampling data taken in a short cycle, and is expensive. The method (2) has the problem that when the input waveform changes from a certain average value BMEAN1 to another value BMEAN2, it is not possible to quickly measure the new value BMEAN2 following this change.

[0006]The reason for (2) will be explained. For example, set the sampling to 1000
It is assumed that a certain average value BMEAN1 is obtained by calculating 1,000 pieces of data obtained by performing the calculation twice using equation (1). Then, if the input waveform changes to another average value BMEAN2 from the 1001st sampling number, then this other value B
MEAN2 does not easily reach the true value BMEAN2 due to the influence of the sum of additions from 1 to 1000 times. In other words, according to equation (1), the number of sampling times is 1 to 1.
The value of Σ up to 000 is very large, so add 10
Even if the sampling values from the 01st time are added, the overall addition sum (Σ) does not change much, and the new value BME
A very large number of samples must be taken to obtain AN2.

An object of the present invention is to solve the above-mentioned problems, and to provide an average value rectification type measuring device that does not require adjustment of the sampling period, can be constructed at low cost, and can quickly measure the average value. It is.

[0008]

[Means for Solving the Problems] In order to achieve the above object, a first invention includes an AD conversion means (1, 2) An absolute value calculator which introduces the digital output signal of this AD conversion means and outputs its absolute value, and an exponential average calculator which introduces this absolute value and adds an indexed average calculation to it ( 7).

The second invention includes an AD converter (20) that samples an analog signal under test at a certain period and converts the sampled value into a digital signal, and an absolute value of the digital output signal of the AD converter. an absolute value calculator (6) for taking out the output data; an arithmetic averaging means (13) for arithmetic averaging a predetermined number of values of the output data of the absolute value calculator; A calculation means (71) is provided for introducing the output data {b(n)} after the output, and performing an exponential average calculation using the arithmetic average value obtained by the arithmetic average means as an initial value. It is something.

[0010] The third invention includes an AD converter (20) that samples an analog signal under test at a certain period and converts the sampled value into a digital signal, and an absolute value of the digital output signal of the AD converter. an absolute value calculator (6) for extracting the output data of the absolute value calculator (6), and an averaging means (6) that takes the maximum value and minimum value from a predetermined number of output data of the absolute value calculator and calculates (maximum value + minimum value)/2. 15) and the output data {b(n)} obtained after the absolute value calculator outputs the predetermined number of output data, and the output data {b(n)} obtained by the averaging means (
A calculation means (71) that performs an exponential averaging calculation using a value of (maximum value + minimum value)/2 as an initial value.

[0011]

[Operation] In the first invention, the AD conversion means samples the analog signal under test at a certain period and converts it into a digital signal. Therefore, the AD conversion means takes in the instantaneous value (Bn) of the signal under measurement B and outputs it. The absolute value calculator takes out the absolute value of the digital output signal of this AD conversion means. That is, the absolute value calculator outputs the absolute value |Bn| of the instantaneous value of the signal under measurement. When considered in terms of an analog circuit, the absolute value calculator corresponds to a rectifier that outputs a full-wave rectified waveform of the signal under measurement. The exponential averaging calculator calculates the absolute value data of the instantaneous value of the signal under measurement |B
Introduce n| and add an exponential averaging operation to it. Therefore, if you repeat the exponential averaging operation, (1/N) Σ|Bn|
A value corresponding to , that is, an average value of the rectified waveform of the signal under test is obtained.

In the second invention, the arithmetic averaging means arithmetic averages a predetermined number of values output from the absolute value calculator. For example, if the predetermined number is 10, {b(1)+b(2)
+...+b(10)} /10 calculation is performed. The average value obtained in this way is the average of a small number of samples, and has a large error compared to the ideal average value P(M), but it is similar to the curve calculated by equation (1) (Figure 8 (c)
(see Figure 8(d)).
reference). In the second invention, the value obtained by arithmetic averaging of the values of a predetermined number of small samples obtained immediately after the start of measurement is set as the initial value P(0), and then the output output from the absolute value calculator is Since the data is subjected to the exponential averaging calculation by the calculation means, the desired average value can be obtained in a short time.

In the third invention, the averaging means takes the maximum value and the minimum value from a predetermined number of output data of the absolute value calculator, and performs the calculation of (maximum value+minimum value)/2. The value obtained in this way has a large error compared to the ideal average value P(M), but it is much larger than the curve calculated by formula (1) (see Figure 9 (C)). The rise is fast (see Figure 9(d)). In the third invention, the maximum value and minimum value are extracted from a small number of sample values obtained immediately after the start of measurement, and the value obtained by calculating (maximum value + minimum value) / 2 is used as an initial value, and then Since the arithmetic means performs an exponential averaging operation on the output data output from the absolute value calculator, a desired average value can be obtained in a short time.

[0014]

[Example] Fig. 1 is a diagram showing an example of the configuration of an average value rectification type measuring device according to the present invention, Fig. 2 is a diagram showing analog signals of each part of the device in Fig. 1, and Figs. 3 and 4 are diagrams showing measurement data and 6 and 7 are diagrams showing another configuration example of the average value rectification type measuring device according to the present invention, FIG. 8 is a time chart of signals of each part of the device in FIG. 6, and FIG. 9 is a diagram showing display timing. 7 is a time chart of signals of each part of the device.

In FIG. 1, the signal to be measured B applied to the terminal P1 may be a voltage signal e or a current signal i. The device in FIG. 1 has a rectified waveform (
This is a device that measures the average value of (not actually rectified).

The analog signal under test B applied to the terminal P1 (see waveform in FIG. 2(1)) is led to a sample and hold circuit (hereinafter simply referred to as S/H circuit) 1. Every time the clock signal SC is applied from the clock generator 3, the S/H circuit 1 samples the instantaneous value B(n) of the signal under measurement B at that time, and replaces the value captured in the previous sampling with It outputs the newly imported value. The x marks in FIG. 2(1) indicate each sampling point. Note that n means the number of samplings, and n=1, 2, 3, . . . . In other words, B(1)=B
1,B(2)=B2,...,B(n)=Bn.

An analog-to-digital converter (hereinafter simply referred to as ADC) 2 takes in the analog output signal of the S/H circuit 1 and converts it into a digital signal (hereinafter referred to as ADC).
(simply referred to as AD conversion). This ADC 2
When a signal SD, which is a delayed clock signal SC, is applied, S/H starts from each application time of this signal SD.
The signal applied from circuit 1 is converted into a digital signal and output. Therefore, in the n-th sampling, the instantaneous value Bn is output (see FIG. 2). The AD conversion means described in the claims includes this S/H circuit 1 and the ADC
Consists of 2.

Note that when the current i is the signal under measurement B, this current i is normally converted into a voltage signal Vi and applied to the terminal P1. I will add an explanation to this. When a current to be measured i is passed through a shunt resistor RS (not shown) having a known resistance value, a voltage Vi=RS·i is generated. This voltage Vi is
The shunt resistance value RS is known, and by measuring the voltage Vi, the current to be measured i is measured from i=Vi/RS. In this specification, the voltage signal Vi obtained by converting the current i in this manner is referred to as the current to be measured i.

The absolute value calculator 6 introduces the instantaneous signal value Bn which is the output of the ADC 2, and calculates the absolute value b(n) of this signal.
=|Bn| is output. When considered as an analog circuit, the absolute value calculator 6 corresponds to a rectifier that outputs a rectified waveform of the signal under measurement. It is easy to output the absolute value. That is, it is sufficient to extract only the bit data portion indicating the amplitude of the digital signal Bn. Note that the output of the absolute value calculator 6 is a digital value, and the solid line waveform in FIG. 2(2) is an analog representation of this. The solid line waveform in Figure 2 (2) is the sampling value B1, B2, B3,... (Figure 2 (1)
Since it is the absolute value of ), it becomes step-like. Note that the dotted line waveform in FIG. 2(2) is the ideal absolute value waveform of the waveform in FIG. 2(1).

The exponential averaging calculator 7 exponentially averages the data b(n) output from the absolute value calculator 6, and the output P(n) is expressed in analog form as shown in FIG.
(3) It becomes. Add explanation. The exponential averaging calculator 7 includes, for example, a calculation section 71 and a memory 72. The calculation unit 71 repeats the calculation of equation (2) every time data b(n) is added from the absolute value calculation unit 6. In other words, the calculation unit 71 receives the absolute value b(n
), the previous calculation result P(n-1) is introduced from the memory 72, the calculation of equation (2) is performed, the output P(n) is stored in the memory 72, and the result P(n-1) after a certain period of time is )
Add to CPU9. P(n)=P(n-1)+(1/G)・{b(
n)-P(n-1)}
(2) P(n):
Exponentialized average result processed up to sampling number n (n-1): Indexed averaged result processed up to sampling number (n-1) b(n): Instantaneous of the signal under measurement at the nth sampling number Absolute value of value 1/G: Constant (1/G << 1)

0021
]Here, the constant 1/G is set to 1/G << 1, so each time b(n) is added to an arithmetic means (not shown), the calculation of equation (1) is repeated. (
3) The waveform is as follows. In other words, Equation (2) shows that when this calculation is repeated, P(n) becomes the waveform of b(n) (Fig. 2 (2)
(see). That is, the average value of b(n)=|Bn|, BMEAN=(1/N)·Σ|Bn|, is obtained from the exponential averaging calculator 7.

(2) The meaning of the expression will be explained. Instantaneous absolute value b(n
) and the average calculation result P(n-1) up to the previous sampling number (n-1), and this difference, that is, {b(n)-P(n-1)}, is expressed as a constant 1/ G Multiply. Here, since 1/G <<1, (1/G
)・{b(n)-P(n-1)} is the average calculation result P(n-1) up to the previous time and the current sampling value b(n
) is a “diluted version” of the difference. Then, this "diluted difference" is added to the previous average calculation result P(n-1) to obtain the average calculation result data P(n) up to the current sampling number n, so this calculation is This means that when repeated many times, P(n) in equation (2) becomes the average value of the waveform of b(n). Therefore, the final value in FIG. 2(3) matches the average value of the full-wave rectified waveform in FIG. 2(2).

Although FIG. 2(3) is drawn so that the value of P(n) becomes constant after a small number of calculations, in reality, for example, by repeating the number of calculations n=1000, P
The value of (n) remains constant. The absolute value calculator 6 and the exponential average calculator 7 having such functions are, for example, digital
It can be configured using a signal processor (hereinafter referred to as DSP) 5. The central processing unit (hereinafter simply referred to as CPU) 9 receives the output signal P(n) of the exponential averaging calculator 7 after a predetermined number of calculations n as shown in FIG. 2(3).
P(n) = (1/N) which reached this steady state by introducing
・Display Σ|Bn| on the display 10. Note that the number of calculations matches the number of clock signals SC, so the number of calculations shown in FIG.
Using the frequency divider 11, when a predetermined number of clocks are generated, a signal is applied to the CPU 9 to notify that a predetermined number of calculations have been performed.

In the apparatus of FIG. 1 as described above, the time TS
Assume that a signal to be measured (which may be a voltage signal or a current signal) having a waveform shown in FIG. 2(1), for example, is applied to the terminal P1 starting from . Here, from the clock generator 3, the S/
Since the clock signal SC with the period TAD is applied to the H circuit 1, the instantaneous data (B1,
B2, B3, ...) are imported. Then, this data is successively converted into digital values by the ADC 2. Note that the delay line 4 slightly delays the timing of the clock signal SC applied to the S/H circuit 1.
/H The value at which the hold state of circuit 1 is stable is sent to ADC 2.
I am trying to make it possible to import it.

The absolute value calculator 6 outputs the absolute value b(n)=Σ|Bn| of the instantaneous value Bn of the signal under measurement B. If this b(n) is represented in analog form, it will have a solid line waveform as shown in FIG. 2(2). The data marked with x in Figure 2 (b(1), b(2),
) is added to the above equation (2) in the exponential averaging calculator 7, the trajectory of the value of P(n) becomes a waveform as shown in FIG. 2(3). The finally settled value of P(n) obtained in this way is the average value BMEAN of the waveform obtained by rectifying the signal under measurement B.
It is. When P(n) shown in FIG. 2(3) is repeated a predetermined number of times (for example, n), a signal is applied from the frequency divider 11 to the CPU 9, and the CPU 9
The output P(n) of the indexed average calculator 7 at that time is read, and the average value is displayed on the display 10 based on that value.

Note that in the apparatus shown in FIG. 1, two
There are ways to display the street. The display method in FIG. 3 displays the average value based on the point (△ mark) where the indexed average output P(n) becomes the final value or near the final value (updates the displayed value),
After that, the contents of the memory 72 are cleared and {P(n-1) is set to 0.
}, the calculation is started again from 0. The display method shown in FIG. 4 is a method of displaying an average value based on the indexed average output P(n) at each given number of calculations without clearing the data in the memory 72.

However, in equation (2), if the exponential averaging is started from the initial value P(0)=0, a constant response time tA (see FIG. 8(C)) is required until the average value P(M) is obtained.
(corresponding to Figure 2 (3)). In Figure 8(C), the indexed average output is shown to be constant with a small number of samples (number of calculations), but in actual operation, it is usually 100
It becomes constant after about 0 calculations. If you increase the value of the constant (1/G) to shorten this response time, the response will be faster, but when measuring low frequency signal under test B, the output after exponential averaging will contain ripples, and the measured value will I have a problem with the movement being unsteady.

FIGS. 6 and 7 show an average value rectification type measuring device that improves this point and shortens the output response time without increasing the output ripple. The apparatus shown in FIG. 6 has an arithmetic averaging means 13 newly added to the structure shown in FIG. 1, and this action improves responsiveness. The arithmetic averaging means 13 arithmetic averages a predetermined number of values of the output b(n) of the absolute value calculator 6. For example, when the absolute value calculator 6 outputs k squared values such as b(1), b(2), b(3), ..., b(k), P(0)={ b(1)+b(2)+b(3)+...+
b(k)}/k
(3) The digital calculation is performed and the result is stored in the memory 72 as the initial value P(0) of the exponential average. Moreover, when the value of this initial value P(0) is obtained, along with the output of this value P(0), the arithmetic unit 71
A signal S1 is output to inform that the initial value P(0) has been output to. In addition, the absolute value calculator 6 having the above-mentioned functions,
The arithmetic unit 71 and the arithmetic averaging means 13 can also be constituted by, for example, a DSP 5.

The operation of the apparatus shown in FIG. 6 as described above will be explained with reference to FIG. Assume that, for example, a signal under test B shown in FIG. 8(a) is applied to the terminal P1 starting from t=0. The AD conversion means 20 samples the instantaneous value of the signal under test every time the clock signal SC is applied from the clock generator 3, and outputs a digital value Bn of this instantaneous value. The absolute value calculator 6 calculates the absolute value b of this instantaneous value Bn.
(n)=|Bn| is output (see FIG. 8(b)). That is, the absolute value calculator 6 successively outputs absolute values b(1), b(2), b(3), . . . in synchronization with the clock signal.

Here, the arithmetic mean means 13 is set to, for example, a numerical value "k", and the absolute value calculator 6 outputs a signal b.
Each time (1), b(2), b(3),... is added,
Subtract 1 from k one by one, and when k=0, perform the calculation shown in equation (3). That is, the arithmetic mean value of data 1 to k is output as the initial value P(0), and this value P(0) is stored in the memory 72. Further, the arithmetic mean means 13 is
A signal S1 indicating that the initial value P(0) has been output is applied to the calculation unit 71. Note that the TU shown in Figure 8(d) is (
3) This is the time required to measure and calculate P(0) in the equation. In the example shown in Figure 8(d), the initial value P(0) is
(b) Absolute values b(1), b(2), b(3) shown in
, b(4). This figure 8
As can be seen by comparing (c) and Figure 8(d), the value P(0) obtained by arithmetic averaging is the curve of equation (2) (Figure 8(C): initial value P (0
) = 0), the rise is much faster (Figure 8(d)
) ).

The calculation section 71 receives the signal S1 from the arithmetic averaging means 13, and starts the exponential averaging calculation according to equation (2) using this signal as a starting point (note that before receiving this signal S1, the calculation section 71 has stopped its calculation operation)
. In this case, in the first operation of equation (2), P(n
-1)=P(0) is used. In other words, using the example of FIG. 8, when the number of samplings is n=5th, the first calculation of the calculation unit 71 is performed, and the content thereof is equation (4). P(5)=P(0)+(1/G) ・{b(5)
−P(0)}
(4) Hereafter, each time there is an output from the absolute value calculator 6, P(6), P(7),...
In order to calculate
As shown in FIG. 8(c), the response waveform of the device shown in FIG. 1 converges to the average value P(M) more quickly. That is, the device in FIG. 1 displays the average value when the number of samplings is tA (FIG. 8).
(c)), and FIG. 6 can be displayed when the sampling number is tB (see FIG. 8(d)). In addition, the constant (1/G)
Since the value of is the same as that of the device in FIG. 1, there is no increase in ripple compared to the device in FIG.

The apparatus shown in FIG. 7 has an averaging means 15 newly added to the structure shown in FIG. 1, and this action improves responsiveness. This averaging means 15 calculates the output of the absolute value calculator 6.
A predetermined number of b(n) are introduced, the maximum value and minimum value are selected from these, and the calculation (maximum value + minimum value)/2 is performed. For example, a numerical value "k" is set in the averaging means 15, and signals b(1), b(2), b(
3) Each time ,... is added, subtract 1 from k, and k
When =0, the calculation shown in equation (5) is performed.

That is, from the absolute value calculator 6, b(1), b
(2), b(3), ..., b(k) Introduce k absolute values and find their maximum and minimum values. The maximum value
b(j) and the minimum value is b(i), then P(0)={b(j)+b(i)}/2

(5) The digital calculation is performed and the result is stored in the memory 72 as the initial value P(0) of the exponential average. Also, the initial value of equation (5)
Once P(0) is obtained, similarly to the arithmetic mean means 13,
Along with outputting this value P(0), the initial value P(
A signal S1 is output to notify that 0) has been output. According to the device shown in FIG. 7, if the maximum and minimum values are known within a quarter cycle of the analog measured signal waveform ((a) in FIG. 9), which is a repetitive waveform, the initial value P(0 )
No further sampling is required to obtain .

According to the apparatus shown in FIG. 7, as shown in FIG.
In the example shown in , the maximum value = b(3) and the minimum value =
b(1). As a result, the initial value P(0) obtained from equation (5) becomes P(0) shown in Figure 9(d), and as can be seen by comparing Figures 9(d) and 9(c), The exponential averaging calculation using P(0) obtained by equation (5) as the initial value is performed using the curve of equation (2) (Figure 9 (C)...
The start-up is much faster compared to the device in Figure 1 (Figure 9).
d)...Figure 7 device). That is, the device shown in FIG. 1 displays the average value when the number of samplings is tA (FIG. 9(c)).
(see FIG. 9(d)), the device in FIG. 7 can display this at the sampling number tc (see FIG. 9(d)).

[0035]

[Effects of the Invention] As explained above, according to the present invention,
The following effects can be obtained. (1) Since measurement can be performed regardless of whether an integer multiple of one period of the signal waveform under test matches an integer multiple of the sampling period, a circuit for measuring the period of the signal waveform under test is not required. Also, there is no need for a means for controlling the sampling period to be an integer fraction of the period of the signal waveform under test. Therefore, costs can be reduced. (2) Since calculations are not performed until an integral multiple of one period of the waveform to be measured and an integral multiple of the sampling period match, the measurement response time does not become long. (3) Furthermore, since the sampling period does not need to be shortened, there is no need to use an expensive ADC. (4) The average calculation result data P(n) up to the sampling number n is calculated based on the average calculation result data P(n-1) up to the sampling number (n-1). In other words, since the present invention does not use a method of adding sums {see equation (1)}, it is possible to relatively quickly follow changes in the signal under measurement and obtain new measured values. (5) According to the configurations shown in FIGS. 6 and 7, an initial value P(0) that is closer to the average value is obtained from the values of a small number of samples obtained immediately after the start of measurement, and based on this, the calculation section Since the exponential averaging operation is added, the desired average value can be obtained in a shorter time.

[Brief explanation of the drawing]

[Fig. 1] A diagram showing an example of the configuration of an average value rectification type measuring device according to the present invention.

[Figure 2] Diagram showing analog signals of each part of the device in Figure 1

[
Figure 3: Diagram showing measurement data and display timing

[Figure 4]
Diagram showing different timings of measurement data and display

[Figure 5] Diagram explaining the problems of the conventional example

FIG. 6 is a diagram showing another configuration example of the average value rectification type measuring device according to the present invention.

FIG. 7 is a diagram showing another configuration example of the average value rectification type measuring device according to the present invention.

[Figure 8] Time chart of signals of each part of the Figure 6 device

[Figure 9
] Figure 7 Time chart of signals of each part of the device

[Explanation of symbols]

1 S/H circuit 2 ADC 3 Clock generator 6 Absolute value calculator 7 Exponential averaging calculator 71 Arithmetic unit 72 Memory 13 Arithmetic mean means 15 Average means

Claims (3)

[Claims] Claim 1: AD conversion means (1, 2) for sampling an analog signal under test at a certain period and converting the sampled value into a digital signal; and a digital output signal of the AD conversion means; An average value rectification type measuring device comprising: an absolute value calculator that outputs the absolute value of , and an indexed average calculator (7) that introduces this absolute value and adds an indexed average calculation to it. 2. AD conversion means (20) for sampling an analog signal under test at a certain period and converting the sampled value into a digital signal; and an absolute value for extracting the absolute value of the digital output signal of the AD conversion means. an arithmetic unit (6), an arithmetic averaging means (13) for arithmetic averaging of a predetermined number of values of the output data of the absolute value arithmetic unit, and output data{
b(n)}, and an arithmetic mean value obtained by the arithmetic mean means is used as an initial value, and an arithmetic means (71) performs an exponential mean operation. measuring device. 3. AD conversion means (20) for sampling an analog signal under test at a certain period and converting the sampled value into a digital signal; and an absolute value for extracting the absolute value of the digital output signal of the AD conversion means. an arithmetic unit (6); an averaging means (15) for taking the maximum value and minimum value from a predetermined number of output data of the absolute value arithmetic unit and calculating (maximum value + minimum value)/2; Output data after the absolute value calculator outputs the predetermined number of output data {b(n)
}, and calculating means (71) for performing an exponential averaging operation using the value of (maximum value + minimum value)/2 obtained by the averaging means as an initial value. Value rectification type measuring device.