JPH01318440A - Demodulation circuit for digital transmission - Google Patents

Abstract

PURPOSE:To ensure the output of a data signal containing the decreased number of noises and at the same time to realize the reproduction of a clock signal having a small number of jitters even with a faint input signal by inserting the base band pre-low pass filters of different characteristics into a data decod ing circuit and a clock reproducing circuit respectively. CONSTITUTION:A Gaussian low pass filter 2 serves as abase band pre-low pass filter of a data decoding circuit 3. The circuit 3 reproduces and outputs the data signal. While a Butterworth low pass filter 5 serves as a base band pre-low pass filter of a clock reproducing circuit 4. The circuit 4 reproduces and outputs the clock signal. Both filters 2 and 5 use the operational amplifiers functioning as the active filters. The eye opening of an output signal (a) of a GMSK wave detecting circuit 1 is reduced by the filter 2 and not by the filter 5. Therefore the waveforms suited to the reproduction and the elimination of noises are inputted to both circuits 3 and 4, respectively.

Description

[Detailed description of the invention] [Industrial application field]

The present invention relates to a demodulation circuit for the GMSK system, which is one of the digital data transmission systems in mobile communications, for example.

[Conventional technology]

The demodulation circuit of the GMSK method in conventional digital data transmission is as shown in FIG.
The detection circuit includes a square-shaped low-pass filter 2, a decoding circuit 3, and a clock recovery circuit 4. The GMSK detection circuit 1 includes, for example, a frequency detection circuit, a synchronous detection circuit, and a delay detection circuit. The Gaussian low-pass filter 2 plays the role of a baseband pre-low-pass filter. In the decoding circuit 3, the data signal is reproduced and output. Further, the clock regeneration circuit 4 regenerates and outputs the clock signal. FIG. 7 shows output waveforms at each stage of the demodulation circuit shown in FIG. 6 above. In the same figure, (al) is the waveform of a in the GMSK detection circuit.This wave includes thermal noise (white noise) in addition to the carrier wave and harmonics. This is the waveform of the output A.This output S is input to the decoding circuit 3 and the clock recovery circuit 4.Comparing the waveforms of the output a and the output S, it is found that the output a, that is, the GMSK detection circuit The eye opening of the output b, that is, the output wave of the Gaussian low-pass filter 2, is smaller than the eye opening of the wave. In this way, the eye opening of the detection output a and the eye opening of the baseband pre-low-pass filter output are different. This is due to the following reason: In the GMSK detection circuit l, the cutoff frequency fc of the baseband pre-low-pass filter, that is, the frequency attenuated by 3 dB by the Gaussian low-pass filter, is approximately f
c・0.4X f b (fb is the transmission speed) is optimal. Pit error rate (bit error rate) and cutoff frequency fc
Regardless of the detection method, the relationship between the two is approximately as shown in FIG. 8. Therefore, it is advantageous for the decoding circuit 3 to slightly reduce the eye opening of the baseband pre-low-pass filter output in order to remove white noise and improve transmission efficiency. On the other hand, the clock regeneration circuit 4 regenerates and outputs a clock waveform from the eye pattern of the baseband pre-low-pass filter output. The waveform shown in FIG. 7 fcl is the clock waveform that should be obtained at the output C of the clock recovery circuit 4.

[Problem to be solved by the invention]

However, since the eye opening of the baseband pre-low-pass filter is small as shown in tb+ in FIG. 7, the noise margin for the clock recovery circuit 4 is small. Therefore, when the receiving input level is low, the waveform of output C will have a large amount of jitter in the reproduced clock, as shown in Figure 7 (c-0), which will eventually reduce the pit error rate. . The present invention utilizes GMSK in conventional digital data transmission.
This method was developed to improve the above-mentioned drawbacks of the conventional demodulation circuit, and provides a demodulation circuit that can reproduce a clock signal with extremely low jitter even when the receiving power level is small.

[Means to solve the problem]

A demodulation circuit for digital data transmission to which the present invention is applied to solve the above problems will be explained with reference to FIG. 1 corresponding to an embodiment. As shown in the figure, the demodulation circuit for digital data transmission of the present invention separately includes a data decoding circuit 3 for reproducing a data signal from the output signal of the GMSK detection circuit 1, and a clock regeneration circuit 4 for reproducing a clock signal. , a Gaussian low-pass filter 2 is provided as a pre-baseband low-pass filter for the data decoding circuit 3, and a Butterworth-type low-pass filter 5 is provided as a pre-baseband low-pass filter for the clock recovery circuit 4, respectively.

[Effect]

The eye opening of the output signal a of the GMS K detection circuit l (see Fig. 4 (al)) is reduced by the Gaussian low-pass filter 2 (see fbl in the same figure), but not by the Butterworth-type low-pass filter 5 (see Fig. 4 (al)). Same figure fd
(see l). Therefore, waveforms suitable for reproduction and noise removal are input to the data decoding circuit 3 and the clock reproduction circuit 4, depending on the respective circuits. As a result, data decoding circuit 3
outputs a data signal with less noise. At the same time, even if the input signal is weak, the clock recovery circuit 4 outputs a jitter-free clock signal.

【Example】

Examples of the present invention will be described in detail below. FIG. 1 is a block diagram showing an embodiment of a demodulation circuit for digital data transmission to which the present invention is applied. l in the same figure is GMS
K detection circuit, 2 a Gaussian low-pass filter, 3 a data decoding circuit, 4 a clock recovery circuit, and 5 a Butterworth low-pass filter. The GMSK detection circuit 1 may include, for example, a frequency detection circuit, a synchronous detection circuit, or a delay detection circuit. Gaussian low-pass filter 2
The so-called phase linear low-pass filter plays the role of a baseband pre-low-pass filter of the data decoding circuit 3, and also uses an operational amplifier which is an active filter. The data decoding circuit 3 reproduces and outputs the data signal. Butterworth type low pass filter 5
plays the role of a baseband pre-low-pass filter of the clock recovery circuit 4, and uses an operational amplifier which is an active filter. The clock regeneration circuit 4 regenerates and outputs the clock signal. That is, baseband pre-low-pass filters with different characteristics are inserted on the data decoding circuit 3 side and the clock regeneration circuit 4 side, respectively, to form a two-series GMSK demodulation circuit configuration. FIG. 2 shows the characteristics of the Gaussian low-pass filter 2. A filter having the same characteristics as the Gaussian low-pass filter 2 used in the conventional circuit (see FIG. 4) can be used. The cutoff frequency fC1, that is, the frequency attenuated by 3 dB by the low-pass filter, is approximately fc·0.4×fb (fb is the transmission speed). FIG. 3 shows the characteristics of the Butterworth type low-pass filter 5. The amplitude within the band is flat and attenuates steeply outside the band. Within the band, the phase is relatively linear. The suitable cutoff frequency fc, that is, the frequency attenuated by 3 dB by the low-pass filter, is approximately f c =0.6x f b (f b is the transmission speed), and the optimum value was determined through experiments. FIG. 4 shows output waveforms at each stage of the demodulation circuit shown in FIG. 1. In the figure, fal is the waveform of the GMSK detection circuit power a. This wave includes a carrier wave, harmonics, and thermal noise. FIG. 4 fb) is the waveform of the output of the Gaussian low-pass filter 2. This output is input to the data decoding circuit 3. The waveforms of output a and output are shown in Figure 7 (a),
(Same as bl<The eye opening of the output of the Gaussian low-pass filter 2 is smaller than the eye opening of the output a of the GMSK detection circuit 1. In other words, the waveform with white noise removed and transmission efficiency improved is the data decoding Similarly, in Figure 4, (dl is the waveform of the output d of the Butterworth type low-pass filter 6. The eye opening of the output d is the same as the eye opening of the output a (see figure (a)). Since it is also large, there is a large noise margin.Therefore, even when the reception input level is low, the regenerated clock has extremely small jitter and no deterioration of the pit error rate, as shown in Figure 4 (e-0). Fig. 5 shows the pit error rate characteristics with respect to the received input level.As shown in the figure, the example of the present invention is close to ideal characteristics and has been greatly improved compared to the conventional one. The waveform shown in tel is the clock waveform that should be obtained at the output e of the clock recovery circuit 4, and is the clock waveform shown in the seventh
The waveform is the same as that in Figure fcl. In the above embodiment, the Gauss-type low-pass filter 2 and the Butterworth-type low-pass filter 5 use active filters using operational amplifiers, and can be formed into a hybrid integrated circuit, so that miniaturization is easy. Although the number of elements increases, it is also possible to use a Pudge-Eve filter such as an LC filter.

【Effect of the invention】

As described above, in the demodulation circuit for digital data transmission to which the present invention is applied, optimal waveforms are input to the decoding circuit and the clock recovery circuit from the respective baseband pre-low-pass filters. Therefore, the demodulation circuit outputs a data signal with less noise and also outputs a clock signal with less jitter even when the input signal is weak.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an embodiment of a demodulation circuit for digital data transmission to which the present invention is applied, FIG. 2 is a characteristic diagram of a Gaussian low-pass filter, and FIG. 3 is a characteristic diagram of a Butterworth-type low-pass filter. Figure 4 is an output waveform diagram at each stage of the demodulation circuit, Figure 5 is a pit error rate characteristic diagram with respect to received input level, Figure 6 is a block diagram showing a conventional demodulation circuit for digital data transmission, and Figure 7 is its diagram. FIG. 8, which is an output waveform diagram at each stage of the demodulation circuit, is a diagram showing the relationship between pit error rate and cutoff frequency fc. l...GMSK detection circuit 2...Gaussian low-pass filter 3...data decoding circuit 4...clock regeneration circuit 5...Butterworth-type low-pass filter Komasa-1-
Middle toe Δ-work'j toe) 111'-Δ-work-〇 ^ ^
Ll ○ Anonymous Society

Claims (1)

[Claims] 1. A data decoding circuit that reproduces a data signal from the output signal of the detection circuit and a clock regeneration circuit that reproduces a clock signal are separately provided, and a Gaussian low-pass filter and a clock are used as a baseband pre-low-pass filter of the data decoding circuit. A demodulation circuit for digital data transmission, characterized in that a Butterworth-type low-pass filter is provided as a baseband pre-low-pass filter of a regeneration circuit.