JP2002368700A - Noise eliminator - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a noise eliminator, having for enhanced reliability by improving measurement accuracy of the physical characteristics value. SOLUTION: The noise eliminator, including a measurement means that detects a signal from a prescribed signal line where a noise signal may infiltrate into an object signal to be detected to measure the physical characteristics of the signal, is provided with a period designation means that designates a plurality of sampling periods with a lapse of time, and with a sampling operation execution means that executes the sampling repeated with the lapse of time according to the sampling period designated by the period designation means, so as to sample the prescribed physical measurement values.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a noise eliminator, and more particularly, to a noise eliminator which is preferably disposed around a subscriber circuit of a digital exchange and used by connecting to a subscriber line.

[0002]

2. Description of the Related Art Conventionally, a subscriber circuit of a digital exchange is connected to a telephone circuit mounted on a telephone located at each user's home via a subscriber line.

[0003] The subscriber circuit is a so-called BORSCH.
It has a T function to exchange subscriber line signals with the telephone. The subscriber line signal includes a calling signal, a response signal (sending a dial tone and a telephone bell), a calling signal, a selection signal for sending a dialed number, and the like. In order to ring the telephone at the time of ringing, it is necessary to transmit a ringing signal of considerably large power from the subscriber's circuit, and to activate the handset of the telephone at the time of a telephone call, it is necessary to transmit DC of several tens mA. is there.

The T (Tes) of the BORSCHT function
In the t) function, the capacitance value and the resistance value are inspected between the two-line subscriber lines and between the lines.

For example, when a telephone is not connected to a subscriber line, the subscriber line is insulated both line and ground, and the line-to-ground resistance is substantially infinite (eg, several mega-ohms). ), The leakage current (DC) and voltage are measured, and based on these measurements, the line-to-line,
The resistance value between the ground and the ground is checked to see if the insulation is normal.

That is, in order to confirm the normality of the subscriber line or the switch itself in the digital exchange, it is necessary to provide a function for measuring the DC current or DC voltage in the subscriber line.

[0007] In the digital exchange, such a DC current / voltage measurement function samples an analog value of a current signal (or a voltage signal) detected from a subscriber line, discretizes the analog value in a time axis direction, and performs quantization. The digitalization is performed by discretization in the amplitude direction.

Here, the sampling of the analog value is performed at a constant period using a counter of a constant frequency.

In addition, a low-pass filter for removing high-frequency noise is arranged at a stage preceding the sampling circuit for performing the sampling.

From the subscriber line requirements,

[0011]

However, since the sampling circuit described above performs sampling at a fixed period, a noise signal superimposed on a voltage signal or a current signal to be detected on the subscriber line exists, and the noise signal is generated. The frequency (eg, noise) of a signal (eg, a noise having a pulse-like waveform (ie, a waveform having a rising edge and a falling edge, and having an amplitude value almost zero at times before and after the falling edge)) If the (pulse period) corresponds to the sampling period, the measurement accuracy of the voltage signal or the current signal to be detected decreases.

That is, for the sampling period,
In the case where the frequency of the noise signal is the same or an integer multiple, and the generation phase of the noise signal and the sampling timing overlap, each time adjacent sampling is performed on the time axis (or Since the noise signal is detected by being superimposed on a voltage signal or the like to be detected (at each sampling at a predetermined interval), the sampling value obtained by the sampling has low accuracy.

If the noise signal has a certain width in the time axis direction, or if the signal is not a pulse but a signal that acts continuously on all time positions such as a sine wave, the frequency is sampled. Even if the period does not completely match or is not an integral multiple of the period, the same problem occurs because the same decrease in the accuracy of the sampled value can occur over a long period of time.

If such a decrease in precision occurs at the sampling value stage, the precision of the digital value obtained by the quantization that can be performed following the sampling is necessarily low, so that the final It is inevitable that the measurement accuracy of the obtained DC current and voltage is also reduced.

[0015]

According to the present invention, a signal is detected from a predetermined signal line into which a noise signal may be mixed with a target signal to be detected, and the signal is detected. In a noise elimination apparatus including a measuring unit for measuring a physical characteristic value, a period specifying unit that specifies a plurality of sampling periods over time, and a sampling operation that is repeated over time, Is executed according to the sampling period specified by
Sampling means for sampling a predetermined physical measurement value.

[0016]

DESCRIPTION OF THE PREFERRED EMBODIMENTS (A) Embodiment Hereinafter, an embodiment will be described with an example in which a noise elimination device according to the present invention is connected to a subscriber line of a digital exchange.

(A-1) Configuration of First Embodiment FIG. 1 is a schematic diagram showing a configuration example of a main part of a measurement system 10 of the present embodiment. The electronic circuit board 100 on which the measuring system 10 is arranged is arranged around a subscriber circuit of a digital exchange, and is used by connecting to a subscriber line.

In FIG. 1, the measuring system 10 comprises:
It includes a filter circuit 300, a voltage / current measurement unit 200, a sampling circuit 400, a logarithmic sampling cycle generation circuit 410, and a measurement reading unit 500.

The filter circuit 300 is, for example, a portion having an analog low-pass filter. The filter circuit 300 receives the signal S10 from the above-described subscriber line test terminal and receives the signal S1.
This is the part that outputs 1.

The measurement system 10 originally intends to measure the DC electric signals (DC voltage VL1, DC current) necessary to check the normality of the insulation between the subscriber lines and the ground. (The value of the leakage current IL1), but since a synchronous circuit for generating the above-described subscriber line signal exists around the subscriber line, a certain period (as an example) related to the subscriber line signal , 400 Hz) may be mixed and superimposed on the DC electric signal.

When such a noise signal is superimposed,
The signal S10 is affected by the noise signal. If the noise signal cannot be completely removed by the filter circuit 300, the signal S11 still has the effect.

The voltage / current measuring section 200 receiving the signal S11 is a section for measuring either the voltage VL1 or the current IL1.
1 shall be measured.

In this case, the voltage VL1 may be measured by another measuring device similar to the measuring system 10 of the present embodiment, or may be measured by using a measuring system different from the present embodiment.

Finally, the current value (the value of the leakage current, which is the measured value obtained by the present embodiment) output as the signal S14 from the measuring system 10 of the present embodiment.
The normality of the insulation state is confirmed by checking whether the value of VL1 / IL1 calculated using IL1 and the voltage value VL1 measured by another measurement system is larger than the above-mentioned several megaΩ. be able to.

FIG. 8 shows an example of the configuration of the main part of the voltage / current measuring section 200 (hereinafter simply referred to as "measuring section 200").

(A-1-1) Configuration Example of Voltage / Current Measuring Unit In FIG. 8, the measuring unit 200 includes a first sampling unit 11
, A quantization unit 12 and a period control unit 13.

The first sampling unit 11 outputs the signal S11
(Current value) is discretized in the time axis direction according to a sampling cycle S21 specified by the cycle control unit 13, thereby performing sampling.

The sampled value obtained by the sampling is
The signal S20 is supplied to the quantization unit 12.

The quantization section 12 is a section for digitizing the signal S20 in the amplitude direction and outputting a digital signal S12.

In the first sampling section 11, the sampling period S21
Is a part that changes the sampling period S21 according to the signal S15 supplied from the sampling circuit 400 at the subsequent stage.

As described above, the measuring unit 20
Constitutes a kind of A / D converter.

Next, the sampling section 400 which receives the digital signal S12 from the measuring section 200 performs further sampling on the digital signal S12 obtained by sampling and quantization. A configuration example is as shown in FIG.

(A-1-2) Configuration Example of Sampling Circuit 400 In FIG. 9, the sampling circuit 400 includes a second sampling section 20 and a smoothing section 21.

The second sampling section 400 executes sampling at a sampling cycle that changes every time a sampling operation is performed in accordance with the sampling cycle S17 supplied from the logarithmic sampling cycle generation circuit 410. This is the part that outputs the coded value S22.

Here, assuming that the sampling cycle executed by the second sampling section 20 is T2 and the sampling cycle executed by the first sampling section 11 is T1, the sampling circuit 4
In order to make the multiple sampling performed within 00 effective, it is desirable that the relationship of T1 <T2 always be established.

The smoothing unit 21, which receives the sampled values S22 output in time series from the second sampling unit 20, executes a multiple sampling, and calculates an average value S13 of the plurality of sampled values S22.

The number of sampled values on which the average value is based may be set in any manner.
It is conceivable to set the number to a fixed value. In this case, the smoothing unit 21 calculates the sampling value S2 supplied in time series.
2 is averaged for every 1000 pieces, and the average value S13 is calculated.

By the multiple sampling, the influence of a sudden noise signal that lasts for a sufficiently short time as compared with the time for 1000 sampling operations can be significantly reduced.

The average value S13 calculated by the smoothing unit 21
Is supplied to the measurement reading unit 50 in response to a request signal S16 from the measurement reading unit 50 at the subsequent stage.

In this embodiment, the sampling period S17 output by the logarithmic sampling period generation circuit 410, that is, the period T2 corresponds to a predetermined logarithmic function in FIG.
Therefore, when the period T1 is set to a constant value, the period T1 needs to be sufficiently shorter than the shortest T2. However, in the present embodiment, T1 is changed so as to always take a value sufficiently shorter than T2 according to T2.

For this purpose, the second sampling section 20 supplies its own sampling cycle T2 at each time point to the cycle control section 13 using a signal S15, and the cycle control section 13 outputs it at that time point. If the sampling period S21 (T1) is shorter than the received sampling period T2, a new sampling period S21 is created and supplied to the first sampling unit 11.

The logarithmic sampling period generation circuit 410
For example, data corresponding to the logarithmic curve CU1 shown in FIG. 5B is stored in a ROM (Read Only Memory) or the like, and the data is read out as appropriate to output the sampling period S17.

That is, M1, M2, M3,... Arranged at regular intervals on the X axis in FIG. 5B are the address numbers of the ROM, and the storage area specified by each address number is on the Y axis. The arranged data S1, S2, S3,... Are stored. These data S1, S2, S3,... Correspond to each sampling interval executed by the second sampling unit 20,
That is, the period T2 is designated.

The Y corresponding to the straight line SL1 shown in FIG.
As is clear from comparison with the data P1, P2, P3,... On the axis, the values of the data P1, P2, P3,... Are all equal, but the same value as one of the data S1, S2, S3,. not exist.

Of course, actually, M1, M2, M3,... Are set at intervals much shorter than those shown in FIG.
Enough data from one log curve (eg, P1)
To get.

The operation of this embodiment having the above configuration will be described below with reference to FIG.

(A-2) Operation of the First Embodiment FIG. 6A shows an example of the signal S10.

In FIG. 6A, the direct current signal which is the original detection target of the measurement system 10 is TS, and the noise signal is a sine wave NS. Here, the current value of the DC current signal TS is not shown in order to emphasize the influence of the noise signal NS.

FIG. 6B shows measured values (corresponding to S14) corresponding to conventional linear sampling. Where the point PU
1 to PUE indicate the timing of each sampling operation. In the conventional linear sampling, P1 to P in FIG.
Corresponding to 4, the intervals between adjacent points in the points PU1 to PUE are equal.

FIG. 6C shows a measured value S14 corresponding to sampling according to the logarithmic curve of the present embodiment. As in the case of PU1 to PUE in FIG.
SUE indicates the timing of each sampling operation. In the sampling corresponding to the logarithmic curve of the present embodiment, FIG.
, The intervals between adjacent points in the points SU1 to SUE are all different.

As shown in FIGS. 6A and 6B, the period of a single sine wave NS and the sampling period (for example, PU1 and P
U2), the time interval TD shown in FIG.
If multiplex sampling is performed for a sufficiently longer time, the noise component NQ1 shown in FIG. Exists, the noise component NQ1 cannot be completely smoothed even if multiple sampling is performed, and the measured value (S1
(Equivalent to 4).

In order to perform multiple sampling over such a long period of time, it is necessary to provide an enormous capacity of memory for storing a large number of sampled values.
Not always easy.

For these reasons, in the conventional method, the accuracy of the measured value decreases over almost the entire time interval.

On the other hand, in the case of the present embodiment, the sine wave NS
And its harmonics, for example,
The effect of the noise component NQ2 can be reduced to the degree shown in FIG. As is clear from comparison with FIG.
In FIG. 6C, the noise component NQ2 only remains sporadically, there is no need to perform multiple sampling over a long time, and the accuracy of the measurement value S14 is high.

In this embodiment, the logarithmic curve CU1 is used to determine the timings SU1 to SUE of the sampling operation.
Was used, but the function corresponding to the curve is not necessarily a logarithmic function, and any non-linear, non-linear curve can be used.

For example, an exponential curve corresponding to an exponential function as an inverse function of a logarithmic function, or a parabola or hyperbola may be used.

(A-3) Effects of the First Embodiment As described above, according to the present embodiment, the effects of all frequencies including the noise signal and its harmonics are reduced, and the effect is finally obtained. It is possible to improve the accuracy of the measurement value (S14) and improve the reliability.

(B) Second Embodiment Hereinafter, only the points of this embodiment different from the first embodiment will be described.

In the first embodiment, the sampling period S17 output from the logarithmic sampling period generation circuit 410 is the signal S1
Although it was determined independently of the state of 0, in the present embodiment,
According to the state of S10 (noise signal included in S10),
It is characterized in that the sampling cycle (S30) is changed.

(B-1) Configuration and Operation of Second Embodiment FIG. 2 shows a configuration example of a main part of the measurement system 30 of the present embodiment.
Shown in

In FIG. 2, the functions of the components and signals assigned the same reference numerals as those of the components and signals shown in FIG. 1 are the same as those in FIG.

Therefore, the present embodiment is different from the first embodiment only in the portions related to the sampling circuit 400A and the variable sampling period generating circuit 420.

The sampling circuit 400A is basically the same as the sampling circuit 400, but includes a dispersion fluctuation detecting section 22, as shown in FIG.

The variance fluctuation detecting unit 22 performs a predetermined statistical process on each continuous average value obtained for every 1000 sampled values by the smoothing unit 21A, and when the variance is equal to or less than a predetermined threshold value TH1, the noise is calculated. Is determined to have no effect, and when the threshold value TH1 is exceeded, it is determined that there is an effect of noise, thereby outputting the signal S31 in a state according to the determination result.

The output destination of the signal S31 is the variable sampling cycle generation circuit 420. When the signal S31 indicates "no influence of noise", the variable sampling cycle generation circuit 420 maintains the sampling cycle S30 (corresponding to S17) which has been supplied to the second sampling unit 20 until then. If "influence of noise" is indicated, it is changed.

As long as the signal S31 indicates "influence of noise", such a change in the sampling period S30 is repeated any number of times. Therefore, an optimal sampling period capable of sufficiently reducing the effect of noise is selected. Will be done.

Various methods can be considered as to how to generate a new sampling period after the change in the case of changing. For example, the logarithmic sampling period generating circuit of the first embodiment is used as an example. RO that 410 was equipped with
A ROM similar to M may be prepared, and new data may be read out by specifying a new address number only when changing.

In this case, for example, if the data S1 has been read by designating the address number M1, the data S2 is read by designating M2 as a new address number.

The data corresponding to the logarithmic curve (S
1), a random number generated in advance may be stored.

In this embodiment, for example, in contrast to FIGS. 7A and 7B similar to FIGS. 6A and 6B, a highly accurate current value IL1 as shown in FIG. , Measured value S14.

(B-2) Effects of the Second Embodiment According to the present embodiment, the same effects as those of the first embodiment can be obtained.

In addition, in the present embodiment, since the optimal sampling period can be set according to the state of noise (period, etc.), the measurement accuracy can be further improved.

(C) Third Embodiment Hereinafter, only the points of the present embodiment that are different from the first embodiment will be described.

This embodiment is different from the first embodiment in that the sampling period S17 specified according to the logarithmic function is specified according to a random number.

(C-1) Configuration and Operation of Third Embodiment FIG. 3 shows an example of the configuration of the main part of the measurement system 40 of this embodiment.
Shown in

In FIG. 3, the functions of the components and signals assigned the same reference numerals as those of the components and signals shown in FIG. 1 are the same as those in FIG.

Therefore, the present embodiment is different from the first embodiment only in the portion related to the random number sampling period generating circuit 430.

The random number sampling cycle generation circuit 430
The sampling period S41 output by the image randomly changes like white noise according to a random number.

In the case of a logarithmic function or the like, the value (the value of the Y component) changes gradually and continuously with the change of the X component, but in the case of a random number, mathematically, there is no discontinuity. Indicates a continuous change.

The random numbers need not necessarily have perfect white noise in a mathematically strict sense. In many cases, it is considered that so-called pseudo-random numbers having almost white noise characteristics are sufficient. Can be

Various methods are conceivable for generating such a random number as the sampling period S41. One of the methods is to write a previously generated pseudo-random number in a ROM or the like, and to generate a new sampling period. Each time S41 becomes necessary, a different address number of the ROM may be designated. In this case, the previously generated pseudo-random value is replaced with the above-described data (for example, S1) of FIG.
The above-mentioned address number (for example, M1) is used as the address number.

As another method, a random number generator is mounted on the random number sampling cycle generation circuit 430, and a pseudo random number generated from the random number generator is used as the sampling cycle S41.

Although this is considered to be a very rare case,
If a phenomenon occurs in which the period of the noise signal is not constant and fluctuates, for example, in a logarithmic function, and a phenomenon occurs in which the change in the sampling period S17 coincides with the change in the noise signal period, the first In the embodiment, it is possible that the high-accuracy measurement value S14 may not be obtained indefinitely. However, even in such a case, in the present embodiment, the high-accuracy measurement value S14
It is possible to obtain

(C-2) Effects of Third Embodiment According to the present embodiment, the same effects as those of the first embodiment can be obtained.

In addition, in the present embodiment, since the sampling period (S41) can be changed discontinuously using random numbers, the period of the noise signal is not constant but varies, for example, in a logarithmic function. Even in a simple case, a highly accurate current value (S1
4) can be obtained, and the reliability is improved.

(D) Fourth Embodiment Hereinafter, only the points of the present embodiment that are different from the second embodiment will be described.

The present embodiment is characterized in that the accuracy of the current value (S14) obtained as a measurement result is improved by switching the filter to be used, instead of changing the sampling period (S30).

(D-1) Configuration and Operation of the Fourth Embodiment FIG. 4 shows an example of the configuration of the main part of the measuring system 50 of the present embodiment.
Shown in

In FIG. 4, the functions of the components and signals assigned the same reference numerals as those of the components and signals shown in FIG. 2 are the same as those in FIG.

Therefore, the present embodiment is different from the second embodiment only in the portions related to the filter circuit (A) 300, the filter circuit (B) 310, and the filter switching circuit 320.

The filter circuits 300 and 310 are the same in that they are analog low-pass filters, but have different filter specifications.

Therefore, even if the same signal S10 is received, signal S1 output from each of filter circuits 300 and 310 is output.
1A and S11B are different.

When the signal S31 output from the sampling circuit 400A indicates "no noise effect", the filter switching circuit 320 maintains the selection of the filter circuit (for example, 300) selected so far, When “influence of noise” is indicated, the selection is switched.

Thus, it is possible to efficiently reduce the influence of noise signals having various frequencies and obtain a highly accurate measurement value S14.

In this example, there are two filter circuits 300 and 310 having different filter characteristics, but the number may be three or more.

Further, even if the circuit is a single circuit, a variable filter circuit capable of switching the filter characteristics can be used. In this case, the switching of the filter characteristics can be replaced by the switching of the filter characteristics.

(D-2) Effects of the Fourth Embodiment According to the present embodiment, it is possible to efficiently reduce the influence of noise signals having various frequencies and obtain a highly accurate current value (S14). It becomes possible.

(E) Other Embodiments In the first to fourth embodiments, the noise signal shown in FIG.
Although the sine wave shown in FIG. 7A and FIG. 7A is assumed, the present invention is not limited to such a sine wave noise signal, and when pulse noise of a very short time is periodically generated, for example, It can also respond to spike noise.

Since the first to fourth embodiments are not necessarily in an exclusive relationship, they can be used in a combined manner.

For example, if the first embodiment and the fourth embodiment are combined and a highly accurate measured value S14 cannot be obtained in the sampling cycle corresponding to the logarithmic curve, the filter circuit (or (Filter characteristic) may be switched.

In the first to third embodiments, the sampling period (S17 and the like) of the second sampling unit 20 is changed. However, the sampling period (S17) of the second sampling unit 20 is changed. S
17) Instead of the sampling period S21 of the first sampling unit 11
Can be expected to obtain substantially the same effect.

In particular, as in the first embodiment and the third embodiment, the sampling period is determined according to the logarithm or random number irrespective of the state of the measured value S14 (whether or not there is an influence of noise). Is changed, it is considered that the effect is substantially the same even if the sampling period S21 of the first sampling unit 11 is changed.

In this case, the second sampling section 20 in the sampling circuit 400 (400A) can be omitted.

In the first to fourth embodiments, the case where the measuring system is connected to the subscriber line of the digital exchange has been described as an example. However, the present invention is applicable to other use environments. Is also possible.

[0105]

As described above, according to the present invention, it is possible to improve the measurement accuracy of the physical characteristic value and increase the reliability of the noise elimination device.

[Brief description of the drawings]

FIG. 1 is a schematic diagram illustrating a configuration example of a main part of a measurement device according to a first embodiment.

FIG. 2 is a schematic diagram illustrating a configuration example of a main part of a measurement device according to a second embodiment.

FIG. 3 is a schematic diagram illustrating a configuration example of a main part of a measuring device according to a third embodiment.

FIG. 4 is a schematic diagram illustrating a configuration example of a main part of a measuring device according to a fourth embodiment.

FIG. 5 is an operation explanatory diagram of the first embodiment.

FIG. 6 is an operation explanatory diagram of the first embodiment.

FIG. 7 is an operation explanatory diagram of the second embodiment.

FIG. 8 is a schematic diagram illustrating a configuration example of a main part of a voltage / current measurement unit used in the first embodiment.

FIG. 9 is a schematic diagram illustrating a configuration example of a main part of a sampling circuit used in the first embodiment.

FIG. 10 is a schematic diagram illustrating a configuration example of a main part of a sampling circuit used in the second embodiment.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 10 ... Measurement apparatus, 11 ... 1st sampling part, 12 ... Quantization part, 13 ... Period control part, 20 ... 2nd sampling part, 21 ... Smoothing part, 22 ... Dispersion fluctuation detection part, 200 ... Voltage / Current measuring unit, 300: filter circuit, 320: filter switching circuit, 400, 400A: sampling circuit, 410: logarithmic sampling cycle generation circuit, 420: variable sampling cycle generation circuit, 430: random number sampling cycle generation circuit, 500: measurement reading Department.

 ──────────────────────────────────────────────────続 き Continued from the front page (72) Inventor Juichi Saito 1-7-12 Toranomon, Minato-ku, Tokyo Oki Electric Industry Co., Ltd. (72) Inventor Satoshi Matsumoto 3-2-2 Shibaura, Minato-ku, Tokyo Shares Company Oki Comtech F term (reference) 5K050 AA02 BB06 BB12 BB14 DD21 DD30 EE23 FF13 FF14 FF17 GG02 GG10 5K052 AA01 AA03 BB11 GG47 GG48

Claims (7)

[Claims] 1. A noise eliminator provided with a measuring means for detecting a signal from a predetermined signal line into which a noise signal can be mixed with a target signal to be detected and measuring a physical characteristic value of the signal. A period specifying means for specifying a plurality of sampling periods as time elapses, and a sampling operation repeated as time elapses in accordance with the sampling period specified by the period specifying means to perform a predetermined physical measurement. A noise eliminator comprising: a sampling operation executing means for sampling a value. 2. The noise elimination apparatus according to claim 1, wherein said measurement means includes a quantization processing unit for outputting said physical characteristic value as a digital signal. 3. The noise elimination device according to claim 2, wherein said sampling operation execution means includes an averaging unit that executes a multiple sampling operation by calculating a time average of a plurality of sampling values. Removal device. 4. The noise elimination apparatus according to claim 1, wherein said period specification means includes a non-linear function corresponding unit that changes a specified sampling period according to a predetermined non-linear function. 5. The noise removing apparatus according to claim 1, wherein the cycle specifying means changes a specified cycle based on a variation in variance of a multiplex sampling value obtained as a result of the time averaging from the averaging unit. A noise elimination device comprising a dispersion correspondence unit. 6. The noise removing apparatus according to claim 1, further comprising random number generating means for generating a random number, wherein said cycle specifying means changes a designated sampling cycle in accordance with the random number generated by said random number generating means. A noise removing device comprising: 7. A noise eliminator provided with measuring means for detecting a signal from a predetermined signal line into which a noise signal can be mixed with a target signal to be detected and measuring a physical characteristic value of the signal. A sampling operation execution unit that executes a sampling operation that is repeated with time according to a supplied sampling cycle to sample a physical measurement value measured by the measurement unit; and that the sampling operation execution unit performs the sampling operation. Averaging means for performing a multiple sampling operation by obtaining a time average of a plurality of sampling values obtained by performing the above-mentioned, a plurality of filter means having different filter characteristics arranged before the measuring means, and Means based on the variation of the variance of the resulting multiple sampling values of said time average. Noise elimination device being characterized in that a filter selection means for selecting a filtering means to enable from among the plurality of filter means.