JPS6142189Y2 - - Google Patents

Description

[Detailed description of the invention] The present invention uses an averaging processing device, in particular, digitizes the sampling values sampled at each corresponding point of a repetitive signal, weights the values, averages them, and improves the S/N. The present invention relates to an averaging processing device.

The repetitive signal that is input to the measuring instrument contains many noise components in addition to the signal component that is the measurement target, and in order to improve the S/N, it is necessary to repeat the signal that contains many of the above-mentioned noise components. Averaging of signals is generally performed.

Conventionally, for repetitive signals containing noise components,
The sampling values of the analog quantities obtained by sampling at the corresponding positions (corresponding points) for each of the repeated signals are averaged by charging a capacitor with a large time constant, and the averaging is applied to the repeated signals. A so-called Box Car method is known in which the signal-to-noise ratio is improved over the entire area. The boxcar method, which averages analog values, uses an integrating circuit with a large time constant, so the processing speed for averaging is slow, and only the processing results are displayed on the display. There is a drawback that it is impossible to judge whether or not the S/N ratio has been improved and signal components sufficient for measurement have been extracted unless the required number of averaging operations has been reached. Furthermore, the boxcar method has the disadvantage that when the processing speed for averaging is increased, the reproducibility of observed waveforms is lost.

The purpose of this invention is to solve the above-mentioned drawbacks, and digitalizes the sampling values obtained by sampling repetitive signals containing noise components at the corresponding positions for each repetitive signal with an extremely small time constant. It is an object of the present invention to provide an averaging processing device which performs digital averaging over the entire repetitive signal to improve the S/N ratio. Therefore, the averaging processing device of the present invention samples each corresponding point of a repetitive signal containing noise components, averages the sampled values obtained by sampling, and performs S/
The averaging processing device for improving N includes a sample hold circuit and an analog-to-digital converter for digitally converting the sampled values of the repeated signal, and each digitally converted sampling point of the repeated signal up to the m-1th round. The weighted average processing value A(i, m-1) of Ti and the digitally converted value Z(i, m) of the sampling value corresponding to each of the m-th repeated signal are weighted, respectively, and A(i, m)=N-1/NA(i, m-1)+Z(i,
m)/N (N is an arbitrary integer), and a weighted averaging processing circuit that performs averaging processing by digitizing each sampling value of the above-mentioned repeated signal and weighting each value. It is characterized by what it does. This will be explained below with reference to the drawings.

Fig. 1 shows the configuration of an embodiment of the averaging processing device according to the present invention, Fig. 2 is an explanatory diagram explaining the sampling method for averaging, and Fig. 3 shows the signal of the averaging processing process. A time chart is shown that explains why the waveforms are viewed directly on the display.

In FIG. 1, reference numeral 1 is an amplifier, 2 is a sample and hold circuit, 3 is an analog-to-digital converter, 4 is a weighted averaging processing circuit, 5 is an arithmetic circuit, 6 is a RAM, 7 is a digital-to-analog converter, 8 each represent a CRT display tube.

The repetitive signal containing noise components is appropriately amplified by an amplifier 1 and inputted to a sample and hold circuit 2. In the sample-and-hold circuit 2, when performing n samplings of the observed waveform of the repetitive signal, sampling is performed once at the same corresponding position for every n-th repetitive signal. After the sampled values are digitally converted by an analog-digital converter 3, a weighted averaging process is executed by a weighted averaging process circuit 4.
The method of sampling described above will be explained with reference to FIG. 2. One point is sampled during the period T of one repetitive signal. That is, for the first repeated signal, the first sampling point _T1 is sampled, for the second repeated signal, the sampling point _T2 shifted by time T _s is sampled, and in the same manner, for the nth repeated signal, is the sampling point T which is the final point of the repeated signal
Sample _o . Then, for the next n+1st repeated signal, the first sampling point
T ₁ is sampled, sampling point T ₂ is sampled for the next (n+2)th repeated signal, and sampling point T _o is similarly sampled for the 2nth repeated signal.
That is, the same sampling point is sampled every n times.

Each sampled value thus sampled is digitized by an analog-to-digital converter 3 and input to a weighted averaging processing circuit 4. The weighted averaging processing circuit 4 performs the following calculation process (weighted averaging processing calculation).

Let Z(i,k) be the digital conversion value obtained by sampling the sampling point T _i for the k-th repeated signal _. In the second calculation (described below for the sampling point T _i ), the digital conversion value Z (i, 1) is input to the calculation circuit 5, and the digital conversion value Z (i, 1) = A ( i,
1) is written.

The second calculation is the second round digital conversion value Z (i, 2) from the analog-digital converter 3.
is input to the arithmetic circuit 5, and the first arithmetic result A stored at address i from the RAM 6
(i, 1) = Z (i, 1) is read and A (i, 2) = N-1/NA (i, 1) + Z (i, 2)
/N calculation is executed. Then, the calculation result (weighted average processing value) A(i, 2) is written to address i of the RAM 6, and the previous contents are rewritten. Here, N represents an arbitrarily selected positive integer.

The third calculation is the third round digital conversion value Z(i,
3) is input to the arithmetic circuit 5, and the RAM 6
A (i, 3) = N-1/NA (i, 2) + Z (i, 3) is read from the second operation result A (i, 2) stored at address i from
/N calculation is performed. And the calculation result A(i, 3)
is written to address i in RAM6, rewriting the previous contents.

Similarly, for the m-th calculation, the m-th digital conversion value Z(i, m) from the analog-digital converter 3 is input to the calculation circuit 5, and
The m-1st operation result A(i, m-1) stored at address i from RAM 6 is read out and A(i, m)=N-1/NA(i, m-1)+Z( i,
m)/N. And the calculation result A(i, m)
is written to address i in RAM6, rewriting the previous contents.

Such weighted averaging processing is performed over the entire repetitive signal from the sampling point T ₁ to T _o , and the calculation results are written to the corresponding addresses of the RAM 6, that is, addresses 1 to n. Therefore, the contents of the RAM 6 are completely rewritten every n repeated signals.

By the way, the digital conversion value Z(i,k) obtained by sampling the sampling point T _i for the k-th repetition signal can be expressed by the following equation.

Z (i, k) = X (i, k) + Y (i, k) where X (i, k) and Y (i, k) represent the signal component and noise component at the sampling point T _i , respectively. , for the signal component X(i,k), the same point T _i of the repeated signal is sampled every n times, so X(i,1)=X(i,2)=X(i,3)=...
X(i,m), and the magnitude of the signal component does not change.

However, the distribution of the noise component has a normal distribution,
If the variance is σ ² , the magnitude of the noise component in the first operation Y ² (i, 1) is Y ² (i, 1) = σ ² , and the magnitude of the noise component in the m-th operation is S N ² (i, m) is It is.

Therefore, if the weighted averaging processing circuit 4 performs the averaging process m times, the compression ratio of the noise component, that is, the S/N can be expressed as follows.

As can be seen from the above equation, the S/N decreases exponentially as m increases. In other words, the signal component remains unchanged and only the noise component decreases.

In this way, each time the repeated signal is sampled, the i address of the RAM 6 corresponding to the sampling point T _i is rewritten, and the noise components are converged. The newly rewritten contents of the RAM 6 are sequentially read out, converted into analog data by a digital-to-analog converter 7, amplified as appropriate, and displayed on, for example, a CRT display tube 8.

FIG. 3 is a time chart explaining why the signal waveform in the averaging process is directly visible on the display device. During the time T _w of the period T of the repetitive signal,
The digitally converted value obtained by sampling a certain sampling point of the repetitive signal and the data of the corresponding address stored in the RAM 6 are averaged with the weighting as described above, and the calculation result is written to the corresponding address of the RAM 6. Then, for a time T _R
The rewritten contents of the RAM 6 are successively read out and displayed on the CRT display tube 8 as explained above. That is, during the period T, new writing (time _Tw ) and reading (time _Tr ) are repeated as shown in FIG.

For example, if the repetition period T of the repetition signal is 200μ
s, if the number of sampling points is set to 1024, it takes 200 μs × 1024≒ to collect all the data of the repetitive signal.
This will require 200ms. Therefore, if the time T _w is selected to be several microseconds, it will appear that the averaging of the repeated signals on the screen of the CRT display tube 8 is progressing in about 0.2 seconds. As explained above, according to the present invention, averaging is performed by weighting the averaging process, so the magnitude of the preceding and succeeding signal waveforms changes so as to reduce only the noise component. This means that the signal component has been extracted, and the S/N ratio is improved exponentially. In addition, the signal waveform whose S/N ratio is improved as the repeated signal averaging process progresses can be directly observed on the CRT display tube.
Necessary measurements can be made even during the averaging process, there is no need to average values more than necessary, and the time required for on-site measurements can be shortened. Furthermore, by reducing the number of sampling points, the processing speed required for averaging processing can be easily increased.

[Brief explanation of the drawing]

Fig. 1 shows the configuration of an embodiment of the averaging processing device according to the present invention, Fig. 2 is an explanatory diagram illustrating the sampling method for averaging, and Fig. 3 shows the signal of the averaging processing process. A time chart is shown that explains why the waveforms are viewed directly on the display. In the figure, 1 is an amplifier, 2 is a sample and hold circuit, 3 is an analog-to-digital converter, 4 is a weighted averaging processing circuit, 5 is an arithmetic circuit, and 6 is an arithmetic circuit.
RAM and 7 represent digital-to-analog converters, respectively.

Claims (1)

[Claim for Utility Model Registration] Averaging processing that improves S/N by sampling each corresponding point of a repetitive signal containing noise components and averaging the sampled values obtained by sampling. In the apparatus: a sample hold circuit and an analog-digital converter for digitally converting the sampled value of the repetitive signal;
The weighted average processing value A(i, m-1) of each sampling point Ti obtained by digital conversion of the repeated signal up to the m-1th round and the digital sampling value corresponding to each of the m-th repeated signal. The converted values Z(i, m) are weighted respectively, and A(i, m)=N-1/NA(i, m-1)+Z(i,
m)/N (N is an arbitrary integer), and a weighted averaging processing circuit that performs averaging processing by digitizing each sampling value of the above-mentioned repeated signal and weighting each value. An averaging processing device characterized in that: