JP6003836B2 - X-ray analysis signal processor - Google Patents

Description

  The present invention relates to a signal processing apparatus for X-ray analysis generally called a multi-channel analyzer, and relates to a signal processing apparatus for X-ray analysis having a waveform conversion digital filter for converting a waveform of a differential wave.

  An X-ray fluorescence analyzer irradiates a solid sample, a powder sample, or a liquid sample with excitation X-rays (primary X-rays), and detects the fluorescence X-rays excited and emitted by the irradiated primary X-rays with a spectrometer. Thus, qualitative and quantitative analysis of elements contained in the sample is performed. Such fluorescent X-ray analyzers include wavelength dispersive X-ray fluorescent analyzers and energy dispersive X-ray fluorescent analyzers.

  The wavelength dispersive X-ray fluorescence analyzer has a configuration in which fluorescent X-rays having a specific wavelength are selected by an X-ray spectrometer combining a spectroscopic crystal and a slit and then detected by a detector. On the other hand, the energy dispersive X-ray fluorescence analyzer directly detects fluorescent X-rays with a semiconductor detector or the like without performing such wavelength selection, and then separates the output signal for each wavelength λ (that is, X-ray energy E). It has the structure which performs a process (for example, refer patent document 1). Therefore, when creating an X-ray fluorescence spectrum, wavelength-dispersed X-ray fluorescence analyzers need to perform wavelength scanning, whereas energy-dispersive X-ray fluorescence analyzers can obtain information on multiple wavelengths simultaneously. The fluorescent X-ray spectrum can be acquired in a short time.

FIG. 5 is a schematic configuration diagram showing a configuration of a conventional general energy dispersive X-ray fluorescence spectrometer. An energy dispersive X-ray fluorescence analyzer 101 includes an X-ray tube 10 that _emits primary X-rays to a sample S, and an energy dispersive spectrometer that detects fluorescent X-ray intensities I λ1 to λ2 having a wavelength of λ1 to λ2. (Detector) 30, preamplifier (preamplifier) 41, differentiation circuit 42 including capacitor C and resistor R, A / D converter (ADC) 43, waveform conversion digital filter 61, and peak detector 62 An FPGA (X-ray analysis signal processing device) 160 including a histogram memory 63 and a CPU 150 for controlling the X-ray tube 10, the energy dispersive spectrometer 30, the FPGA 160, and the like are provided.

  In the X-ray tube 10, for example, a target that is an anode and a filament that is a cathode are arranged inside a housing. As a result, a high voltage is applied to the target and a low voltage is applied to the filament, so that the thermal electrons emitted from the filament collide with the end surface of the target, thereby emitting primary X-rays generated on the end surface of the target. It is supposed to be.

In the energy dispersive spectrometer 30, for example, a detection element (lithium drift type Si semiconductor detector) for detecting the fluorescent X-ray intensities I _λ1 to _λ2 is disposed inside the housing. The output signal from the energy dispersive spectrometer 30 is amplified by the preamplifier 41. This output signal has a stepped wave shape, indicating that each step of the stepped wave shape detects fluorescent X-rays, and the height of each step represents the wavelength λ, that is, the X-ray energy E.

The output signal amplified by the preamplifier 41 is sent to the differentiating circuit 42. The differentiating circuit 42 converts the step wave into a differentiated wave represented by the following formula (1). In this way, by changing from a step wave to a differential wave, a wide dynamic range can be obtained and high resolution can be achieved.
y = exp (−nT / τ) = a ^ n (1)
Here, τ is the RC time constant, T is the sampling period, n is the number of samples, and a is the time constant (exp (−T / τ)).

The A / D converter 43 digitizes the differential wave that is an analog signal.
As shown in FIG. 6, the waveform conversion digital filter 61 converts the input differential wave digital signal into a trapezoidal wave digital signal represented by the following formula (2). FIG. 7 is a graph showing a differential wave digital signal and a trapezoidal wave digital signal, with the upper row showing the differential wave digital signal and the lower row showing the trapezoidal wave digital signal. Thus, the peak wave value (peak top value) is accurately calculated by converting the differential wave digital signal into a trapezoidal wave digital signal.

In the above formula, M is a trapezoidal wave top time FL (for example, 0.4 us), and N is a trapezoidal wave rise / fall time PK (for example, 1.8 us).
The peak detector 62 detects a peak in the input trapezoidal wave digital signal, acquires a peak value (peak top value) of each peak, and measures the X-ray energy E corresponding to the peak top value for each peak. The numerical value is incremented and stored in the histogram memory 63.

  Then, the data is sent to the CPU 150. As a result, the CPU 150 creates a wave height distribution diagram (energy spectrum histogram) in which the horizontal axis represents the fluorescent X-ray energy E and the vertical axis represents the element content (intensity).

JP 2007-327902 A

  Incidentally, in the energy dispersive X-ray fluorescence spectrometer 101 as described above, the step wave is converted into the differential wave by the differentiating circuit 42, but the time constant a of the differentiating circuit 42 is the capacitor C generating the differential wave. Since the time constant setting value A of the waveform conversion digital filter 61 is not set accurately (same as the time constant a), the differential wave digital signal and the trapezoid as shown in FIG. It does not become a wave digital signal. FIG. 8A is a diagram when the time constant set value A is set too smaller than the time constant a, and FIG. 8B is a diagram where the time constant set value A is set too larger than the time constant a. It is a figure of time. 8A and 8B, the differential wave digital signal is distorted, and the peak value of the peak is also changed. That is, there is a problem that the energy resolution of the fluorescent X-ray spectrum is deteriorated.

Here, an example of a general differentiating circuit 42 will be described in which a resistance R = 330Ω, a capacitor C = 1 uF, and a sampling period T = 50 ns. The variation between the resistor R and the capacitor C is ± 3% in total, and other factors and aging are ± 1%. Therefore, the maximum range of variation is set to ± 4%. The time constant a is represented by 16 bits.
From RC time constant τ = R × C = 330 × 1e ⁻⁶ = 33 us,
Time constant a = 2 ^ 16 * exp (-T / τ) = 2 ^ 16 * exp (-50 ns / 33 us) = 65437.

When the resistance R and the capacitor C vary to the minimum variation side (−4%), a _min = 65432 from τ = 33 × 0.96 = 31.68 us and 5 smaller than 65437. Further, when the resistance R and the capacitor C vary on the maximum variation side (+ 4%), a _max = 65441 from τ = 33 × 1.04 = 34.32 us and 4 larger than 65437. Accordingly, the time constant a of the differentiating circuit 42 changes between a _min (65432) and a _max (65441) depending on the performance of the capacitor C and the resistor R.

In the case of the energy dispersive X-ray fluorescence spectrometer 101 using the A / D converter 43 with a resolution of several eV / bit, the time constant set value A is ± 2 of the optimum value (time constant a). It is known that the energy resolution of the fluorescent X-ray spectrum hardly deteriorates if it is within the range.
Therefore, the manufacturer or the like uses parts such as a capacitor C and a resistor R with high accuracy (no variation) so that the time constant setting value A does not need to be finely adjusted, or energy-dispersion type during production. The time constant set value A is finely adjusted for each fluorescent X-ray analyzer 101.

Here, when finely adjusting the time constant set value A during production or the like, the time constant set value A is obtained. As described above, when the differential wave is set to the transfer function y = a ^ n, a = It is only necessary to calculate x ^ {(log _x y) / n} (where x is an arbitrary number other than 0, for example, 10). At this time, such a calculation is calculated at a plurality of points of a plurality of differential waves and averaged to obtain the time constant a. However, while such an operation is easy to understand, there is a drawback that it takes time to process due to noise and distortion included in the differential wave, and the amount of calculation and circuit increases.

In order to solve the above-described problems, the present inventors have studied FPGA (signal processing apparatus for X-ray analysis). Therefore, it was considered that the configuration can be simplified and the circuit amount can be reduced by obtaining the time constant set value A using the trapezoidal wave h (z) instead of the differential wave y = a ^ n. As shown in FIG. 8A, when the time constant set value A is smaller than the time constant a, the peak in the trapezoidal wave drags the tail to the positive side. On the other hand, when the time constant set value A is larger than the time constant a, the peak in the trapezoidal wave drags the tail to the negative side as shown in FIG. Focusing on these points, scanning (calculating) several to several tens of values (a plurality of time constant setting values A _x ), and the flat front portion of the peak in the trapezoidal wave of each time constant setting value A _x ( Referring to the difference value ΔH _x between the first N portion shown in FIG. 6 and the rear flat portion (the second portion after N shown in FIG. 6), the difference value ΔH _x is It has been found that a time constant set value A _x that is equal to or less than the threshold value ΔH _th is obtained and selected using the time constant set value A _x .

  That is, the signal processing apparatus for X-ray analysis of the present invention uses a differentiation circuit that converts the step wave detected by the X-ray detector into a differential wave, and the differential wave into a trapezoidal wave or a triangular wave using a time constant setting value. A signal processing apparatus for X-ray analysis comprising a digital filter for conversion and a peak detection unit for discriminating and counting peak values of the trapezoidal wave or triangular wave, wherein a plurality of time constant temporary set values are set by the digital filter. A storage unit that stores trapezoidal waves or triangular waves that are used to convert the differential waves, a flat part calculation unit that obtains a front flat part and a rear flat part of the peak, and a front flat part of the peak; A difference value calculation unit that calculates a difference value with the rear flat portion, and at least one time constant temporary setting value is selected from a plurality of time constant temporary setting values based on the difference value, and the selected time constant Provisional So that and a selection unit for selecting the time constant set value with the value.

Here, the “plural set of time constants A _x ” is a numerical value in an arbitrary range that is expected to include the optimum value (time constant a), and is input using an input device or the like, for example. A numerical value (A _{1 to} A ₉₆ ) or the like shifted by 1 of ± 48 (96 points) of the numerical value A ₄₈ .

According to X-ray analysis for the signal processing apparatus of the present invention, for example, first, the digital filter is a first time with a constant provisional setpoint A _1, is stored by converting the differentiated wave to the first trapezoidal wave . Next, the flat portion calculation unit obtains a front flat portion and a rear flat portion at the peak of the first trapezoidal wave, respectively. Next, the difference value calculation unit calculates a difference value ΔH ₁ between the front flat portion and the rear flat portion.

Then Similarly, the digital filter using a constant provisional setting value A ₂ at the second time, is stored by converting the differentiated wave to the second trapezoidal wave, the flat part calculation unit, the second trapezoidal wave The front flat part and the rear flat part at the peak are obtained, respectively, and the difference value calculation part calculates the difference value ΔH ₂ between the front flat part and the rear flat part. In this way, for example, 96 difference values ΔH _{1 to} ΔH ₉₆ are calculated using ₉₆ time constant temporary setting values A _{1 to} A ₉₆ .
Next, for example, the selection unit obtains a time constant temporary setting value A _x when the difference value ΔH _x becomes equal to or less than the threshold value ΔH _th , and selects the time constant setting value A using the time constant temporary setting value A _x . For example, when the difference values ΔH _{50 to} ΔH ₆₀ become equal to or less than the threshold value ΔH _th , the time constant set value A is selected using the time constant temporary set values A _{50 to} A ₆₀ . That is, the time constant setting value A is selected from the time constant temporary setting values A _{50 to} A ₆₀ determined that the peak in the trapezoidal wave does not drag the tail to the positive side or the negative side.

  As described above, according to the signal processing apparatus for X-ray analysis of the present invention, the time constant set value A can be selected easily and accurately with a small circuit amount. Therefore, the cost increase can be minimized, and calibration at the time of production and operation becomes easy. Further, the parts such as the resistor R and the capacitor C to be used can be inexpensive standard parts instead of expensive ones with small variations. In addition, since any device can obtain the same resolution, the reliability of the device is also improved.

(Means and effects for solving other problems)
In the above invention, the flat part calculation unit obtains a front flat part using a plurality of points on the front side of the peak and obtains a rear flat part using a plurality of points on the rear side of the peak. You may do it.

  In the above invention, the difference value calculation unit may calculate a difference value using a front flat portion and a rear flat portion of a plurality of peaks.

Furthermore, in the above invention, the selection unit may select a median value of a plurality of time constant temporary setting values with the difference value equal to or less than a threshold value as the time constant setting value.
Further, the “threshold value ΔH _th ” is an arbitrary numerical value that determines that the tail is not dragged on the positive side or the negative side, and is, for example, 16 or the like.

1 is a schematic configuration diagram showing an example of an energy dispersive X-ray fluorescence analyzer according to the present invention. The flowchart explaining the selection method using the apparatus of FIG. Explanatory drawing about the selection result using the apparatus of FIG. The figure explaining the verification conditions of the energy dispersive X-ray fluorescence spectrometer of the present invention. The schematic block diagram which shows the conventional general energy dispersion type | mold fluorescence X-ray-analysis apparatus. Explanatory drawing about converting a differential wave into a trapezoidal wave. The graph which shows a differential wave digital signal and a trapezoid wave digital signal. The figure which shows the graph display when the time-constant setting value A is set too small or excessively from the time constant a.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. Note that the present invention is not limited to the embodiments described below, and it goes without saying that various aspects are included without departing from the spirit of the present invention.

FIG. 1 is a schematic configuration diagram showing an example of an energy dispersive X-ray fluorescence spectrometer according to an embodiment of the present invention. In addition, the same code | symbol is attached | subjected about the thing similar to the conventional energy dispersive X-ray fluorescence spectrometer 101 mentioned above.
An energy dispersive X-ray fluorescence spectrometer 1 includes an X-ray tube 10 that _emits primary X-rays to a sample S, and an energy dispersive spectrometer that detects fluorescent X-ray intensities I λ1 to λ2 having a wavelength λ1 or more and a wavelength λ2 or less. (Detector) 30, preamplifier (preamplifier) 41, differentiation circuit 42 including capacitor C and resistor R, A / D converter (ADC) 43, waveform conversion digital filter 61, and peak detector 62 An FPGA (X-ray analysis signal processing device) 60 including a histogram memory 63 and an A set value calculation unit 70, and a CPU 50 for controlling the X-ray tube 10, the energy dispersive spectrometer 30, the FPGA 60, and the like are provided.

The A set value calculation unit 70 according to the embodiment of the present invention includes a trapezoidal wave storage unit 71, a flat part calculation unit 72, a difference value calculation unit 73, a selection unit 74, and an initial setting memory 75.
In the initial setting memory 75, 96 programs (A that are ± 48 of the time constant initial setting value A ₄₈ input by the input device or the like as a program used when the “calculation mode” is input by the input device or the like (A A program for the waveform conversion digital filter 61 to convert to a trapezoidal wave using the time constant temporary setting values A _{1 to} A _{96 of the} value setting scan number x), or “32 (flat portions) on the front side of the peak in the trapezoidal wave A program for obtaining a front flat portion using a point of “sampling number l)” and a rear flat portion using “32 points (number of flat portion sampling number l)” on the rear side of the peak, or “16 number (average number m) "for calculating the difference value [Delta] H _x with the peak of the program and the difference value [Delta] H _x 16 (threshold [Delta] H _th) follows became constant temporarily set value a _x time Program for selecting a time constant set value A is stored in advance the median.

It should be noted that “A value setting scan number x = 96”, “flat part sampling number l = 32”, “acquisition time of flat part sampling number l (part before and after N shown in FIG. 6)” and “average number m = 16 ”,“ threshold value ΔH _th = 16 ”, etc. are stored in advance as“ various settings ”, while“ time constant initial setting value A ₄₈ ”is input after“ calculation mode ”is input by an input device or the like. As “A value initial setting”, an arbitrary numerical value (for example, A ₄₈ = 65437) is input by an input device or the like.

When the “calculation mode” is input by the input device or the like, the trapezoidal wave storage unit 71 uses the x-th time constant temporary setting value A _x by the waveform conversion digital filter 61 to convert the x-th waveform. In order to store the trapezoidal wave, the x-th time constant temporary setting value _Ax is sequentially output to the waveform conversion digital filter 61. Accordingly, for example, the first trapezoidal wave converted from the differential wave using the first time constant temporary setting value A ₁ = 65389 by the waveform conversion digital filter 61 is stored, and then the waveform conversion digital filter 61 stores the first trapezoidal wave. 96 time constant temporary setting values A _{1 to} A ₉₆ are used to store the second trapezoidal wave converted from the differential wave using the time constant temporary set value A ₂ = 65390 of 2, respectively. The 96 converted trapezoidal waves are sequentially stored. Note that the time constant temporary setting value _Ax is increased by 1 when x is increased by 1.

The flat part calculation unit 72 uses the “32 points (the flat part sampling number 1)” points before the m-th peak in the x-th trapezoidal wave stored in the trapezoidal wave storage unit 71 to obtain the m-th front flatness. And the m-th rear flat part is obtained using “32 points (the number of flat part samplings 1)” behind the m-th peak. For example, the time constant first trapezoidal wave provisional setpoint A ₁ was used, 32 points of the front of the first peak (before 1 -th N shown in FIG. 6) (specified time) sequentially The first front flat portion is obtained by adding together and 32 points (specified time) on the rear side (after the second N shown in FIG. 6) of the first peak are sequentially added together to obtain the first The rear flat part is obtained, and the 32 points on the front side of the second peak are sequentially added together to obtain the second front flat part, and the 32 points on the rear side of the second peak are sequentially added together. The front flat portion and the rear flat portion of the 16 peaks are obtained so as to obtain the second rear flat portion. Similarly, the front flat portion and the rear flat portion of 16 peaks are also obtained for 95 trapezoidal waves in which the time constant temporary setting values A _{2 to} A ₉₆ are used.

The difference value calculation unit 73 calculates a difference value ΔH _n between the total front baseline value of the first to m-th peaks and the total rear baseline value of the first to m-th peaks in the x-th trapezoidal wave. calculate. For example, when the constant first trapezoidal wave provisional setpoint A ₁ is used, by adding the first front flat portion of the peak to the front flat portion of the 16 peak, the first total front Baseline Find the value. Further, when the first trapezoidal wave constant provisional setpoint A ₁ is used, by adding from the side flat portion after the first peak to the side flat portion after the peak of the 16, the first total after Determine the side baseline value. Then, the first difference value ΔH ₁ is calculated by subtracting the first total front side baseline value and the first total rear side baseline value. Then Similarly, when the second trapezoidal wave constant provisionally set value A ₂ was used, by adding from the front flat portion of the first peak to the front flat portion of the peaks of the first 16, second total Find the front baseline value. Further, when the second trapezoidal wave constant provisionally set value A ₂ was used, by adding from the side flat portion after the first peak to the side flat portion after the peak of the 16, the second total after Determine the side baseline value. Then, the second difference value ΔH ₂ is calculated by subtracting the second total front side baseline value from the second total rear side baseline value. In this manner, 96 difference values ΔH _{1 to} ΔH ₉₆ are calculated for the 96 trapezoidal waves in which the time constant temporary setting values A _{1 to} A ₉₆ are used.

The selection unit 74 obtains a time constant temporary setting value A _x when the difference values ΔH _{1 to} ΔH ₉₆ are equal to or less than the threshold value ΔH _{th and} selects the median value of the time constant temporary setting value A _x as the time constant setting value A. Are displayed on the monitor and output to the waveform conversion digital filter 61. For example, when the difference values ΔH _{1 to} ΔH ₄₀ exceed the threshold value ΔH _th , the difference values ΔH _{41 to} ΔH ₅₀ become the threshold value ΔH _th or less, and the difference values ΔH _{51 to} ΔH ₉₆ exceed the threshold value ΔH _th , the difference value ΔH ₄₁ The time constant temporary setting value A ₄₅ = 65434 corresponding to the difference value ΔH ₄₅ which is the median value of ΔH ₅₀ is selected as the time constant setting value A and displayed on the monitor. Then, the time constant set value A (65434) is output to the waveform conversion digital filter 61. As a result, when the measurement sample S is placed, the waveform conversion digital filter 61 converts the input differential wave digital signal into a trapezoidal wave digital signal using the time constant setting value A (65434). .

Here, an example of a selection method for selecting the time constant set value A using the energy dispersive X-ray fluorescence spectrometer 1 will be described. FIG. 2 is a flowchart for explaining the selection method.
First, in the process of step S101, it is determined whether or not “calculation mode” is input from an input device or the like. That is, the process of step S101 is repeated until the manufacturer or the like inputs a signal. Therefore, a manufacturer or the like places a certain sample S at the time of production or calibration at the time of operation, inputs a signal using an input device or the like, and further inputs “time constant initial setting value A ₄₈ ”. Will do.

Next, in the process of step S102, the first time constant temporary setting value A _x = A ₁ that is 48 before the “time constant initial setting value A ₄₈ ” is set in the waveform conversion digital filter 61, and the number of peaks m = 1.
Next, in the process of step S103, waveform converting the digital filter 61, a differential wave coming inputted, when using a constant provisional setting value A _x will be converted into a trapezoidal wave of the x.

Next, in the process of step S104, the trapezoidal wave storage unit 71 stores the xth trapezoidal wave. At this time, the trapezoidal wave storage unit 71 latches a waveform (m-th peak) having no VALID before and after and no pile-up. Further, when the flat part before and after the m-th peak can be sufficiently secured (there is no peak immediately before and after), the storage is stopped and the flat part flag is output to the flat part calculation unit 72.
Next, in the process of step S105, the flat part calculation unit 72 sequentially obtains the front flat part of the mth peak by sequentially adding the 32 points in front of the mth peak in the xth trapezoidal wave, The 32 points on the rear side of the m-th peak are sequentially added to obtain the rear flat part of the m-th peak.

Next, in the process of step S106, the flat part calculation unit 72 adds the front flat part of the m-th peak to the front flat parts of the first to (m−1) -th peaks and adds the total front base line. In addition to obtaining the value, the rear flat portion of the mth peak is added to the rear flat portion of the first to (m−1) th peaks to obtain the total rear baseline value.
Next, in the process of step S107, it is determined whether m = 16. If m = 16 is not satisfied, m = m + 1 is set in the process of step S108, and the process returns to step S103.

On the other hand, if m = 16, in the process of step S109, the difference value calculation unit 73 calculates the difference value ΔH _x between the total front side baseline value and the total rear side baseline value in the xth trapezoidal wave. To do.
Next, in the process of step S110, when the selection unit 74 determines that the difference value ΔH _n is equal to or less than the threshold value ΔH _th , the selection unit 74 determines “OK” and determines that the difference value ΔH _n exceeds the threshold value ΔH _th . Sometimes, it is determined as “NG”.

Next, in the process of step S111, the selection unit 74 determines whether or not the _96th time constant temporary setting value A _x = A _{96, which} is ₄₈ times after the time constant initial setting value A ₄₈ . If it is determined that A _x = A ₉₆ is not satisfied, in the process of step S112, A _x = A _{x + 1} is set in the waveform conversion digital filter 61, the number of peaks m = 1, and the process returns to step S103.
On the other hand, when it is determined that A _x = A ₉₆ , in the process of step S113, the selection unit 74 selects the median value of the time constant temporary setting value A _x that is “OK” as the time constant setting value A. Display on the monitor. In this case, when arranging the OK / NG determination result of the constant provisional setting value A _x in turn, OK so that to select the median value of the time constant temporarily set value A _x between those consecutive.

  As described above, according to the energy dispersive X-ray fluorescence spectrometer 1 of the present invention, the time constant set value A can be selected easily and accurately with a small circuit amount. Therefore, the cost increase can be minimized, and calibration at the time of production and operation becomes easy. Further, the parts such as the resistor R and the capacitor C to be used can be inexpensive standard parts instead of expensive ones with small variations. Further, since any device can obtain the same resolution, the accuracy of the device is improved and the reliability is improved.

<Other embodiments>
(1) In the energy dispersive X-ray fluorescence spectrometer 1 described above, the waveform conversion digital filter 61 is configured to convert the input differential wave digital signal into a trapezoidal wave digital signal. It is good also as a structure which converts a signal into a triangular wave digital signal.

(2) In the energy dispersive X-ray fluorescence spectrometer 1 described above, the configuration using the selection method shown in FIG. 2 has been shown. However, the time constant setting value A obtained by repeatedly executing the selection method shown in FIG. It is good also as a structure which averages. Further, among the obtained time constant setting values A, the maximum time constant setting value A and the minimum time constant setting value A may be rejected, and the central time constant setting value A may be averaged. Good. This further improves the reliability.

  Examples The present invention will be described in more detail with reference to examples. However, the present invention is not limited to these examples.

“65437” was input as the time constant initial setting value A _{48 to the} energy dispersive X-ray fluorescence spectrometer 1, and the time constant setting value A was selected ten times. Then, among the ten time constant setting values A, the maximum time constant setting value A and the minimum time constant setting value A were rejected, and eight time constant setting values A were obtained. Further, when the eight time constant set values A were obtained, the number of time constant temporary set values A _x determined by the selection unit 74 as “OK” was obtained. The verification conditions are shown in FIG. 4, and the results are shown in FIG.

The number of eight time constant setting values A and the time constant temporary setting value A _x determined to be “OK” in the same manner as in Example 1 except that “65421” is input as the time constant initial setting value A _48. And asked. The result is shown in FIG.

The number of eight time constant setting values A and the time constant temporary setting value A _x determined as “OK” in the same manner as in Example 1 except that “65453” is input as the time constant initial setting value A _48. And asked. The result is shown in FIG.

As described above, even if “65437”, “65421”, or “65453” is input as the time constant initial setting value A ₄₈ , the A setting value calculation unit 70 selects the time constant setting value A as “65437” This “65437” coincides with the input differential wave. Furthermore, good results with a standard deviation of 1.5 or less were obtained for the degree of variation at each time.

  The present invention can be used in a fluorescent X-ray analyzer or the like that acquires information on elements contained in a sample.

42 differentiation circuit 60 FPGA (signal processing apparatus for X-ray analysis)
61 Waveform Conversion Digital Filter 62 Peak Detection Unit 71 Trapezoidal Wave Storage Unit 72 Flat Part Calculation Unit 73 Difference Value Calculation Unit 74 Selection Unit

Claims (4)

A differentiating circuit for converting the step wave detected by the X-ray detector into a differential wave;
A digital filter that converts the differential wave into a trapezoidal wave or a triangular wave using a time constant setting value;
A signal processing device for X-ray analysis comprising a peak detector for discriminating and counting the peak value of a peak in the trapezoidal wave or triangular wave,
A storage unit for storing a trapezoidal wave or a triangular wave obtained by converting the differential wave using a plurality of time constant temporary setting values by the digital filter, and
A flat part calculating part for respectively obtaining a front flat part and a rear flat part of the peak;
A difference value calculation unit for calculating a difference value between a front flat part and a rear flat part of the peak;
A selection unit that selects at least one time constant temporary setting value from a plurality of time constant temporary setting values based on the difference value, and selects the time constant setting value using the selected time constant temporary setting value; A signal processing apparatus for X-ray analysis, comprising:   The flat part calculation unit obtains a front flat part using a plurality of points on the front side of the peak and obtains a rear flat part using a plurality of points on the rear side of the peak. The signal processing apparatus for X-ray analysis according to 1.   The signal processing apparatus for X-ray analysis according to claim 1 or 2, wherein the difference value calculation unit calculates a difference value using a front flat portion and a rear flat portion of a plurality of peaks.   The said selection part selects the median of the some time constant temporary setting value in which the said difference value became below the threshold value as said time constant setting value, The any one of Claims 1-3 characterized by the above-mentioned. Signal processing apparatus for X-ray analysis described in 1.