Pulse Shaping Circuit for Radiation Detector

1. Field of the Invention

The present invention relates to a radiation detector used for x-ray spectroscopy, for example, and, more particularly, to a circuit for modifying the shape of the output pulse from a radiation detector, such as a semiconductor detector.

2. Description of the Prior Art

A conventional pulse shaping circuit is shown in FIG. 1, where the output signal from a radiation detector 1 is fed to a pulse height analyzer 2 via an integrator 3 and a shaping circuit 4. Specifically, the output signal from the detector 1 is integrated by the integrator 3 and converted into a step voltage signal a (See FIG. 2) proportional to the energy of the incident radiation. The step voltage signal a is shaped into a pulse b having an amplitude proportional to the step voltage by the shaping circuit 4. Pulse b is then fed to the pulse height analyzer 2. The output signal from the radiation detector 1 is shown in FIG. 2(a). The step voltage a delivered from the integrator 3 is shown in FIG. 2(b). When a well known pseudo-Gaussian filter is used as the shaping circuit 4, the output signal b from this Gaussian filter takes a Gaussian waveform as shown in FIG. 2(c).

As shown in FIG. 2(c), it is assumed that radiation falls on the detector 1 at an instant of time t.sub.o. The signal b reaches its maximum value after a lapse of a period T.sub.m. The signal returns to its initial level when a period T.sub.f elapses. The pulse height analyzer 2 makes an analysis of the pulse height when the maximum level is reached.

The next pulse P.sub.2 produced by the radiation detector 1 cannot be measured before a period T.sub.d elapses, T.sub.d =T.sub.f +T.sub.m. The period T.sub.d is called the dead time. The period T.sub.f is equal to the time taken for the pulse b to attenuate below 1/1024 of the maximum level if the pulse height is quantized by means of 1024 quantization levels. Generally, it is necessary that the dead time T.sub.d be several times as long as the period T.sub.m.

V. Radeka has proposed a pulse shaping circuit which has a gated integrator (IEEE, Trans. Nucl. Sci., NS-19, No. 1, 412 (1972)). This shaping circuit is shown in FIG. 3(a), where a gated integrator 5 is inserted between a shaping circuit 4 and a pulse height analyzer 2. A timing controller 6 produces gating signals g.sub.1 and g.sub.2 to the integrator 5. One example of the gated integrator 5 is shown in FIG. 3(b).

Referring to FIG. 3(b), the gated integrator 5 comprises an input capacitor C.sub.1, an operational amplifier and a feedback capacitor C.sub.2. The gated integrator 5 also has switches S.sub.1 and S.sub.2 which are opened and closed in response to gating signals g.sub.1 and g.sub.2, as shown in FIG. 2(d), supplied from the timing controller 6. Therefore, the gated integrator 5 starts integrating its input signal at the instant t.sub.o when radiation falls on the detector 1. After a lapse of the period T.sub.m the Gaussian waveform shown in FIG. 2(c) completely returns to the initial level, and then the switch S.sub.1 is closed, whereby the integration operation ends. The amplitude of the output of the integrator 5 at this time is measured by the pulse height analyzer 2, and then the switch S.sub.2 is closed to discharge a capacitor C.sub.2. Immediately thereafter the output from the integrator 5 goes back to the initial level and so the waveform assumes a shape as shown in FIG. 2(e). As soon as the output settles down to the initial level, the next pulse signal can be measured.

With this pulse shaping circuit shown in FIG. 3(a), signal b is integrated during the measurement and, therefore, noise introduced into signal a is greatly reduced by the averaging action of the integration. Hence, the measuring accuracy is improved.

V. Radeka has also proposed the application of a rectangular wave as shown in FIG. 2(f), instead of the Gaussian waveform shown in FIG. 2(c), to such a gated integrator incorporated in a pulse shaping circuit. The integrator produces a trapezoidal waveform as shown in FIG. 2(g). As the hypotenuse of this trapezoid approaches a straight line, the averaging action approaches an ideal, thus improving the measuring accuracy further.

It is known that a rectangular wave can be created by the use of delay lines, but the resistive component of the delay lines makes it impossible to obtain an ideal rectangular wave. Also, it is not easy to change the delay time. For these reasons, this method has not yet been put into practical use.

F. S. Goulding et al. have proposed that a triangular waveform as shown in FIG. 2(h) be applied to a gated integrator (F. S. Goulding and D. A. Landis, IEEE Trans. Nucl. Sci., NS-29 No. 3, 1125 (1982)). They also have noticed that shortening the dead time T.sub.d induces the associated amplifier to produce more delta noise, and have proposed an index given by the product of noise and the dead time, i.e., N.sub.d.sup.2.T.sub.d, to evaluate the characteristics of this kind of pulse shaping circuit. This has been considered highly valuable. N.sub.d.sup.2, called a delta noise index, is given by:

where F (t) is the waveform of the output from a radiation detector. It is considered that as the value of the product N.sub.d.sup.-2.T.sub.d decreases, the characteristics of the pulse shaping circuit are improved.

The theoretically optimum values of the product N.sub.d.sup.-2.T.sub.d of various pulse shaping circuits are listed below in Table 1.

In Table 1, pulse shaping circuit (1) is the conventional circuit shown in FIG. 1. Pulse shaping circuit (2) is the already proposed circuit which uses a gated integrator as shown in FIG. 3(a). Circuit (3) utilizes a triangular waveform as proposed by F. S. Goulding et al. Circuit (4) employs a rectangular waveform as proposed by V. Radeka. In the cases where a gated integrator is used, T.sub.d is given by T.sub.d =Tm.sub.1 +T.sub.r, where Tm.sub.1 is the time period when S.sub.1 is open and T.sub.r is the recovery time as shown in FIG. 2(d).

As can be seen from this table, the use of a rectangular waveform theoretically gives rise to the best value. However, this method has not been put into practical use as mentioned previously.

It is an object of the present invention to provide a practical version of the aforementioned method using a rectangular waveform as proposed by V. Radeka.

In accordance with the invention, appropriate Gaussian filters are used instead of delay lines. The filters are connected in cascade. The interstage outputs and the output from the final stage are weighted in a predetermined manner and summed up. The resulting signal is supplied to a gated integrator and integrated.

The above-mentioned and other objects will be achieved by the embodiments described in the following.

FIG. 1 is a block diagram of a conventional pulse shaping circuit;

FIGS. 2(a), 2(b), and 2(c) are waveform diagrams of the output signals from radiation detector 1, integrator 3, and shaping circuit 4, respectively, shown in FIG. 1;

FIGS. 2(d) and 2(e) are waveform diagrams showing the timing at which switches S.sub.1 and S.sub.2 are opened and closed;

FIGS. 2(f) and 2(g) are diagrams of a rectangular waveform applied to a gated integrator as proposed by V. Radeka and a trapezoidal waveform produced by the integrator, respectively;

FIG. 2(h) is a diagram of a triangular waveform applied to a gated integrator as porposed by F. S. Goulding et al.;

FIG. 3(a) is a block diagram of a pulse shaping circuit using a gated integrator as proposed by V. Radeka;

FIG. 3(b) is a circuit diagram of one example of a gated integrator;

FIG. 4 is a diagram showing the arrangement of the poles of a directly sythesized Gaussian filter with n=7;

FIG. 5(a) is a circuit diagram of a directly synthesized Gaussian filter of the A type;

FIG. 5(b) is a circuit diagram of a directly synthesized Gaussian filter of the B type;

FIG. 6(a) is a circuit diagram of a pulse shaping circuit built in accordance with the invention and based on the A-type pseudo-Gaussian filter shown in FIG. 5(a);

FIG. 6(b) is a circuit diagram of another pulse shaping circuit built in accordance with the invention and based on the B-type pseudo-Gaussian filter shown in FIG. 5(b);

FIG. 7(a) is a waveform diagram of the output from each of the integrator circuits 3 shown in FIGS. 6(a) and 6(b);

FIG. 7(b) is a waveform diagram of interstage signals a.sub.Z, b.sub.Z, c.sub.Z, and d.sub.Z from the circuits shown in FIGS. 6(a) and 6(b);

FIG. 8 is a graph showing the output e from the operational amplifier 11 and the output f from the gated integrator 5 included in the circuit shown in FIG. 6(b), and in which M.sub.o =0.14, M.sub.1 =0.24, M.sub.2 =0.0, and M.sub.3 =0.62;

FIG. 9 is a graph showing the output e from the operational amplifier 11 and the output f from the gated integrator 5 included in the circuit shown in FIG. 6(a), and in which M.sub.o =0.15, M.sub.1 =0.42, M.sub.2 =-0.1, and M.sub.3 =0.53;

FIG. 10(a) is a diagram of another embodiment of the pulse shaping circuit showing the arrangement of the poles of a directly synthesized Gaussian filter with n=5; and

FIG. 10(b) is a diagram of a further embodiment of the pulse shaping circuit showing the arrangement of the poles of a directly synthesized Gaussian filter with n=9.

A directly synthesized Gaussian filter is available as a kind of pseudo-Gaussian filter (Nucl. Instr. and Meth., 138, p. 85 (1976)). This filter is characterized by the waveform going back to the initial level rapidly. Filters of this kind come in various types according to various pole numbers n as listed in Table 2.

FIG. 4 shows the arrangement of poles of a filter with n=7. In FIG. 4, A.sub.o shows a pole of a real number. A.sub.1, A.sub.2, A.sub.3 .+-.W.sub.1, .+-.W.sub.2, and .+-.W.sub.3 show the real parts (A.sub.1 -A.sub.3) and the imaginary parts (.+-.W.sub.1 -W.sub.3) of the remaining six poles. The functions of the circuit are given by A.sub.o, (A.sub.1 .+-.W.sub.1), (A.sub.2 .+-.W.sub.2) (A.sub.3 .+-.W.sub.3). A pseudo-Gaussian filter with n=7 is obtained by connecting active filters N.sub.1, N.sub.2, N.sub.3 in cascade after a differentiating circuit N.sub.o which provides real root A.sub.o. The active filters N.sub.1, N.sub.2, N.sub.3 provide complex roots A.sub.1 .+-.W.sub.1, A.sub.2 .+-.W.sub.2, A.sub.3 .+-.W.sub.3, respectively. FIG. 5(a) shows a Gaussian filter of the A type in which the active filters N.sub.1, N.sub.2, N.sub.3 are connected in cascade in order of decreasing quality factors ("Q factors"), or W/A. FIG. 5(b) shows a Gaussian filter of the B type in which the filters N.sub.1, N.sub.2, N.sub.3 are connected in cascade in order of quality factors.

When a waveform as shown in FIG. 2(b) is applied to these psuedo-Gaussian filters, they usually produce Gaussian waveforms as shown in FIG. 2(c). The present invention is based on this type of pseudo-Gaussian filter. The outputs from the cascaded filters N.sub.o -N.sub.3 are weighted and summed up to obtain an output approximating a rectangular wave.

FIG. 6(a) shows a pulse shaping circuit based on the A-type psuedo-Gaussian filter shown in FIG. 5(a). FIG. 6(b) shows a pulse shaping circuit based on the B-type directly synthesized Gaussian filter shown in FIG. 5(b). The rates M.sub.o, M.sub.1, M.sub.2, M.sub.3, at which the outputs from the filters are respectively added up, are the ratios of input resistance R.sub.o, R.sub.1, R.sub.2, R.sub.3 to a feedback resistance R.sub.f connected to the operational amplifier 11, that is, M.sub.o =R.sub.f /R.sub.o, M.sub.1 =R.sub.f /R.sub.1, M.sub.2 =R.sub.f /R.sub.2, M.sub.3 =R.sub.f /R.sub.3. The weights can be varied by changing the values of the input resistances R.sub.o, R.sub.1, R.sub.2, R.sub.3. The output e from the operational amplifier 11 is give by:

where a.sub.Z, b.sub.Z c.sub.Z d.sub.Z are interstage outputs.

In FIGS. 6(a) and 6(b) C.sub.o is an input capacitor and R is a bias resistor.

If the output from the integrator circuit 3 of the configuration shown in FIG. 6(b) is a step function as shown in FIG. 7(a), the interstage outputs a.sub.Z, b.sub.Z, c.sub.Z, d.sub.Z respond as shown in FIG. 7(b). Since the tails of the waveforms of the outputs a.sub.Z, b.sub.Z, c.sub.Z cancel out, the output d.sub.Z rapidly settles down to the base line.

These interstage outputs were added up under the conditions M.sub.o =0.14, M.sub.1 =0.24, M.sub.2 =0.0, M.sub.3 =0.62. The output e from the operational amplifier 11 was integrated by the gated integrator 5. The output e from the amplifier 11 and the output f from the integrator 5 are shown in FIG. 8. This output e is not a perfect rectangular wave, unlike the wave shown in FIG. 2(f). The output f (FIG. 8) obtained by integrating the output e approximates the trapezoidal wave shown in FIG. 2(g). Especially, the rising portion can be regarded as substantially straight.

FIG. 9 shows the output e from the operational amplifier 11 and the output f from the gated integrator 5 included in the A-type circuit shown in FIG. 6(a), obtained under the conditions M.sub.o =0.15, M.sub.1 =0.42, M.sub.2 =-0.1, and M.sub.3 =0.53.

With the circuit shown in FIG. 6(a) and graphed in FIG. 9 the output f from the gated integrator 5 approaches the trapezoidal wave shown in FIG. 2(g). The rising portion can be regarded as substantially straight.

Table 3 shows the values of N.sub.d.sup.2.T.sub.d of the circuits shown in FIGS. 6(a) and 6(b) calculated under the condition T.sub.r =0.

Comparison between Tables 3 and 1 shows that the novel pulse shaping circuits of the A type and the B type are much superior in performance to the known pulse shaping circuits 1 and 2 with N.sub.d.sup.2.T.sub.d =9.4 and 7.35 respectively. Also, the novel circuit of the A type has a value better than the best value theoretically obtained by the method using a triangular wave.

It is necessary to select the optimum combination of the weight coefficients M.sub.o, M.sub.1, M.sub.2, M.sub.3 so that the rising portion of the trapezoidal wave may be approximated by a straight line.

The switch S.sub.1 of the gated integrator shown in FIG. 3(b) is opened when the detector 1 supplies a signal to it as described above. Therefore, the integrator acts as a gate which initiates an integration. The switch cooperates with a capacitor C.sub.1 to remove DC drift produced by the preceding stage of circuit. As a result, the initial level is stabilized. In reality, switches S.sub.1 and S.sub.2 can be fabricated in the form of an electronic circuit.

Directly synthesized Gaussian filters which can be put into practical use and have odd poles can have five (n=5) and nine (n=9) poles, as well as seven poles (n=7) as in the embodiment described above. As the number of the poles increases, the output approaches a Gaussian waveform more closely, but the circuit becomes accordingly more complex.

FIGS. 10(a) and 10(b) show other pulse shaping circuits according to the invention, the circuits being based on Gaussian filters with n=5 and n=9 respectively. Both shaping circuits are of the A type.

Where n=5, the optimum combination of the weight coefficients is given by M.sub.o =0.24, M.sub.1 =0.01, and M.sub.2 =0.75. In this state, the value of N.sub.d.sup.2.T.sub.d, is 6.1660.

Where n=9, the optimum combination of the weight coefficients consists of M.sub.o =0.12, M.sub.1 =0.14, M.sub.2 =0.16, M.sub.3 =0, and M.sub.4 =0.58. In this case, the value of N.sub.d.sup.2.T.sub.d is 5.8623.

The terms "radiation detector", "pulse shaping circuit", "adder", "integrator", "pseudo-Gaussian filter", "directly synthesized Gaussian filter", "pulse height analyzer", "gated integrator", "differentiator", "filter circuit", "timing controller", "operational amplifier", "switch", "output converter", "filter", "active filter" and other terms identifying features of the disclosed invention are each used generically; that is, the functional requirements that must be met by each of the elements identified by these terms will be apparent to one of ordinary skill in the art, and those terms accordingly are used throughout the specification and the claims to designate any element known to those skilled in the art which meets such requirements.

The present invention is characterized by weighting and summing up the outputs from the cascaded filters and is not limited to the details of the foregoing embodiments but includes various modifications within the scope and spirit of the appended claims. More specifically, the order in cascading the filters is not restricted to the order described in A type and B type embodiments. The cascaded filter can be placed in any order so long as the appropriate weight coefficients are selected. For example, in the case of the pulse shaping circuits with n=7, not only can the filters be ordered, either N1, N2, N3 (A type) or N3, N2, N1 (B type) but the filter can be also ordered N1, N3, N2 or N3, N1, N2, provided the appropriate weight coefficients are selected.