JPS6320939A - Quantization feedback type integration detection circuit - Google Patents

Abstract

PURPOSE:To reduce the occurrence in identification error by clamping an input signal to an identification means to a normal voltage at the time of code inversion just after the occurrence of an error to cancel an erroneons voltage by the feedback of a regeneration signal of identification error so far thereby suppressing the identification error from being propagated after the next code inversion. CONSTITUTION:A discharge time constant (product between a resistor R2 and a capacitor C2) of a clamp device 11 is equalized to a time constant (product between a resistor R1 and a capacitor C1) of a primary low-pass filter to a regeneration signal input of an adder means, and an error propagation detector 13 consists of two voltage comparators CP2, CP3 detecting it as a positive or negative error propagation if an output signal of an adder 10 exceeds a positive or a negative constant voltage respectively. When the error propagation is detected, the peak value of the input signal of the identifier 12 is clamped to a normal voltage corresponding to 1 or zero. Since the time constant of the primary low-pass filter of the clamp means and the adding means is identical to each other, the output signal of the adding means is coincident with a decreasing curve of the error voltage due to the propagation of error. Thus, in applying clamping, the error due to the feedback of the preceding identification error is cancelled out.

Description

DETAILED DESCRIPTION OF THE INVENTION Field of the Invention The present invention relates to a quantization feedback integral detection circuit for identifying received signals during transmission, recording and reproduction of digital signals.

BACKGROUND ART Electromagnetic conversion systems in magnetic recording generally exhibit low-frequency cutoff characteristics, and even in transmission lines, it may be necessary to cut off the low-frequency range for purposes such as supplying power to repeaters. It is well known that passing through a system with such low-frequency cutoff characteristics causes code amount interference due to low-frequency cutoff, and one way to improve this is to use quantization feedback as shown in Figure 6. There is a way.

When the characteristics of the electromagnetic conversion system or transmission line and the low-frequency cut-off characteristics including the high-frequency equalizer are represented by the bypass filter 1 in Figure 6, this transfer characteristic is approximated by a first-order bypass filter, and the transfer function is is expressed as follows.

Gh(s)=ST/(1+ST)...
・・・・・・・・・・・・・・・0) On the other hand, feedback filter 2
The transfer function of is as follows: 0 Gl(s)=1/(1+ST)...
(2) Therefore, if there is no error in the reproduced signal identified by the discriminator 4, the output of the adder 3 will be a completely reproduced signal of low frequency components.

Problems to be Solved by the Invention As is well known, the biggest drawback of DC regeneration using quantized feedback is the problem of error propagation.

The problem of error propagation will be described below with reference to the waveform diagram in FIG. A signal as shown by the solid line of waveform A becomes a signal with the low frequency cut off as shown by the solid line of waveform B at the input of the adder 2 due to the low frequency cutoff characteristics of the electromagnetic conversion system and the receiving side. If a signal like the broken line during Tm of waveform B is received due to dropout or noise, the sign reversal at point a will naturally be undetectable, and the output signal of the discriminator 4 will be incorrect.0 Equation (1) The error voltage Ve at the output of the adder 3 has a waveform C by feeding back the error through a low-pass filter such as . In this way, if the normal signal amplitude is 1, then the error voltage Ve during the period Tn is Ve=1-e-t/T...
・・・・・・Consideration) increases. Note that t is the elapsed time from time a, and T is the time constant of the low-pass filter 11.

Therefore, even if the influence of dropout or noise on the input signal disappears at point b, the error voltage Ve at the input of the discriminator 4 is ye=1-e-Tm/T...
...(4), and the output signal of the discriminator 4 remains erroneous.

Furthermore, the error at the input of the discriminator 4 increases due to feedback of the error, and the error is propagated until the next sign inversion time point C.

Next, if there is no sign reversal of the original signal for a while after time C, as shown by the solid line of waveform A, the error voltage Ve at the input of the discriminator 4 is as follows, where t' is the elapsed time from time C, Ve = (1- e-T0/T) ・e-t7”-・
・It can be expressed as -(5), and decreases as shown in waveform C. However, when there is the next sign reversal near time point C as shown by the dashed line in waveform A, the error voltage Ve remains large as is clear from the expression (phrase), and the input waveform of the discriminator 4 is As a result, an identification error occurs.

In the method of DC regeneration using quantized feedback as described above,
If one code reversal cannot be identified, not only will identification errors continue until the next code reversal, but if the period of identification error is relatively long, the feedback of the error will cause errors to occur after the next code reversal. There is a problem in that there is a high probability of causing identification errors.

Therefore, the present invention provides a quantization feedback type integral detection circuit which suppresses the propagation of identification errors after the next sign inversion and reduces the occurrence of identification errors.

Means for Solving the Problems The present invention provides an addition means for adding high frequency components of a received signal and a low frequency component of a reproduced signal after identification, an error propagation detection means for detecting error propagation, and an error propagation detecting means for detecting error propagation. Clamping means for peak value clamping the input signal of the discriminating means to a normal voltage corresponding to 1 or zero when error propagation is detected by the propagation detecting means;
The discharging time constant of the clamping means is made equal to the time constant of the first-order one-pass filter for the reproduction signal input to the adding means.

Effect With the above configuration, in the present invention, the identification means input signal is clamped to the normal voltage at the time of sign reversal immediately after an error occurs, and the error voltage due to the feedback of the reproduced signal of the identification error up to that point is canceled out. This cuts off error propagation. This effect will be described in more detail below◇ It is assumed that an identification error occurs from the time an error occurs until the next sign reversal, and that period is Te, and 1 of the output of the adding means is
Alternatively, if the normal voltage corresponding to zero is ±E, and the time constant of the first-order low-pass filter for the input of the reproduction signal of the adding means is T, then the adding means output signal voltage Va immediately after the next sign inversion where an identification error occurs is It is as follows.

Va = ±(E +2E (1-e-TV/T) )-
(6) Next, when error propagation is detected by the error propagation detecting means, the adding means output signal is clamped by the clamping means to a voltage level ±E corresponding to a normal one or zero. Tsumarikura y7”K!ri formula (6) no ±2E (1
= e-''/T).The time constant of this clamping means is the same time constant T as that of the first-order low-pass filter for the reproduction input signal of the adding means, and the discharge time from the clamping point is t. ', then the discharge becomes a curve of et'/T, which coincides with the reduction curve of the error voltage due to error propagation of the output signal of the adding means as shown in equation (5).Therefore, when clamping is performed by the clamping means, the previous All errors due to feedback of identification errors are canceled out.In the above explanation, noise included in the received signal was ignored, but even if noise is included in the received signal, it has no effect on the basic operation. There is no change.

However, since errors occur in the clamping means due to noise components, it is better to change the sensitivity setting of the error propagation detection means depending on the amount of noise.

EXAMPLE A first example of the present invention will be described below with reference to FIG. 1, but the description will be made while ignoring the high frequency deterioration caused by the transmission path, which does not have much influence on the present invention. This embodiment includes an i-day calculator 1o that adds the high-frequency component of the received signal and the low-frequency component of the reproduced signal after identification, an error propagation detector 13 that detects error propagation, and an error propagation detector a clasp device 1 that peak-clamps the input signal of the discriminator 12 to a normal voltage corresponding to 1 or zero when i-response propagation is detected by the device 13;
1 and the output signal of the clamper 11 is compared with the reference voltage.
・A discriminator 12 for identifying zero is provided, and the discharging time constant (product of resistor R2 and capacitor 02) of the clamper 11 is connected to a Roux-order low-pass filter 〇 time constant i (W anti-R1) for the reproduction signal input of the adding means. and capacitor 01), and when the adder 1 output signal exceeds a certain voltage on the positive side or negative side, two voltage comparators CP2 and CP2 detect error propagation on the positive side or negative side, respectively. This is a quantization feedback type integral detection circuit in which the error propagation detector 13 is configured with CF2. The operation of this embodiment having the configuration shown in FIG. 1, where BFO is a buffer amplifier, and MM1 and MM2 are monostable multivibrators, will be described with reference to the waveforms of each part in FIG. 2.

The adder 1o is configured as shown in the figure, and the transfer function Gh(s) from the received signal input end to the output end is expressed by the above equation (1).
Similarly, the transfer function Gl(S) from the reproduced signal input end to the output end after identification is as shown in equation (2), and T in the same equation is the product of resistor R1 and capacitor C1. It is.

In this way, the adder 1o has a composite configuration of a first-order low-pass filter for the reproduced signal input, which is a quantization feedback filter, and a bypass filter, which is a pre-filter. Therefore, when transmitting an original signal such as waveform E, if the high frequency component of the received signal as shown by the broken line of waveform F is not obtained due to dropout or noise during the period Tn, and the result is as shown in the solid line, as shown in Fig. 7 above. The waveform will look like waveform G. Note that a to d in the figure indicate time points, and Te is a period from time point a when an error occurs to the next sign inversion time point.

Next, the error propagation detector 13 is configured as shown in the figure,
Reference voltage vh, vz of two voltage comparators CP2 and CF2
As shown in waveform G, the voltage is approximately twice the normal voltage amount E, which corresponds to 1 or 0. Therefore, when the output signal of the adder 1o exceeds one constant voltage VhsVe on the positive side or negative side, it is detected as error propagation on the positive side or negative side, respectively, and a constant pulse is generated in the monostable multivibrators MMl and MM2. Outputs width detection signal. The monostable multivibrator MM1 and MM2 are used to limit the period during which the peak value clamp is operated and to reduce the influence of noise.The output voltage level of the detection signal of the multivibrator MM1 and MM2 is When error propagation is detected, the waveform G is set to -E and +E, and when no error propagation is detected, the waveform G is set to +E and -E.

Next, in the clamper 11, the forward voltage of the diodes D1 and D2 is set to twice 1±E1 of the waveform G, and when error propagation is detected by the error propagation detector, the clamper 11
The output signal of 1 is clamped to a normal voltage corresponding to the voltage of ±E of waveform G and a voltage of 1 or 0, resulting in a waveform such as waveform I. At this time, the voltage waveform across capacitor 02 is as shown in waveform H, and as mentioned in the explanation of the operation above, the voltage charged in capacitor C2 cancels out the error due to feedback of the reproduced signal due to identification error. .

Therefore, as shown by the dashed line in each waveform in Figure 2, even in cases where there is another sign reversal immediately after a relatively long misidentification, which cannot be identified using conventional methods, it is possible to The signal is input to the discriminator 12 and is successfully discriminated. Note that the capacitor Co and variable resistor vR1 of the discriminator 12 are used to create the reference voltage (Vr of waveform I) of the discriminator 12.
Further, the output level of the voltage comparator is set to a voltage equivalent to 1 or 0 (±E of waveform G) during normal operation in order to satisfy the conditions for quantization feedback.

As described above, according to this embodiment, it is possible to detect the occurrence of error propagation with an error propagation detector with a simple configuration, and cancel the amount of error due to feedback of the reproduced signal of an identification error, and the quantization feedback type integral detection circuit Error propagation can be reduced.

In this example, a pre-filter for cutting off low frequencies in conventional quantization feedback, a first-order low-pass filter (which is a feedback filter), and an adder are combined, and are simply referred to as an adder, but they can be used independently. However, there is no problem. Furthermore, if there is a low-frequency cut-off in a system such as a VTR that necessarily passes through, there is no problem in applying this embodiment even if it is regarded as a pre-filter. Conversely, it is also possible to remove the buffer amplifier BFO, resistor R2, and capacitor 02 of this embodiment, and to configure the adder and clamp device of this embodiment in a more complex manner.

The second embodiment will be described below with reference to FIG. As shown in FIG. 3, this embodiment is newly equipped with a second adder 14 that adds the high frequency component of the received signal and the low frequency component of the reproduced signal after identification. The crossover frequency is set higher than the crossover frequency of the first adder 10, and the output signal of the second adder 14 is used as the input signal of the error propagation detector 13, and the rest is the same as the first embodiment. It is structured as follows. As shown in the figure, the second adder 14 is composed of a resistor R3 and a capacitor C3.

The operation of this embodiment will be described below. first adder 10
Assuming that the crossover frequency of the second adder 14 is 1/T1 and the crossover frequency of the second adder 14 is 1/T2, each adder 10. The output signal voltage of 11 is expressed by formula (6)
The time constant T of is replaced with T1 and T2, respectively. Also, as mentioned above, <1/T2 is thicker than 1/T1. Therefore, as is clear from equation (6), even if error propagation occurs over a short period of time, the output of the adder 14 contains a large amount of error voltage due to feedback of identification errors. In other words, short-term error propagation can also be detected. Therefore, the error voltage at the input of the discriminator 12 due to the feedback of the reproduced signal of the identification error after the first sign inversion after the occurrence of the identification error becomes extremely small.

As described above, according to this embodiment, it is possible to increase the detection sensitivity of error propagation by adding a simple circuit without changing the crossover frequency of the system that identifies the received signal, and as a result, the following code This has the effect that the probability that an error propagates beyond the inversion becomes very small.

A third embodiment of the present invention will be described below. As shown in FIG. 4, this embodiment includes a voltage divider 15 which is a coefficient multiplier that multiplies at least one of the output signal voltage of the second adder 14 and the output signal voltage of the discriminator 12, and a voltage comparator CP4 that compares the voltages after multiplying at least one of the output signal voltage of the detector 14 and the output signal voltage of the discriminator 12 by a coefficient using the coefficient multiplier; The error propagation detector 13 is constituted by the gates G1 and G2 which are gated by the output signal of the discriminator 12, and the other configuration is the same as that of the second embodiment. Note that the voltage divider 15 simply connects the output signal voltage of the second adder 14 to the resistor R.
4.The partial pressure is halved by R6.BFl.B
F2 is a buffer 1-amplifier, and variable resistor VR2 and capacitor 04 are simply used to create a reference voltage for voltage division.
The operations other than 3 are the same as those in the second embodiment, and the error propagation detector 13 will be described below with reference to the waveform diagram in FIG. Note that FIG. 6a-C shows time points, To-Tn shows periods, and ±E shows voltage levels. When transmitting or recording/reproducing an original signal such as waveform I, the adder 14
If the received signal component in the output signal of is as shown in the waveform (dropout or noise during the period of Tn, it should be as a solid line but should be as a broken line), the sign reversal at time a can be identified as described above. This will result in an error in identifying the period of Te.

Therefore, the output signal waveform of the adder 14 whose voltage is divided in half by the voltage divider 15 is as shown by the solid line of waveform I. Note that the amplitude direction of the waveform (■) is shown to be twice as large as that of the waveforms (M to O) for convenience of drawing.

Next, the output signal of the voltage divider 16 and the output signal of the discriminator 12 as shown in the dashed line of the waveform J are added to the voltage comparator CP4, and the output signal of the voltage divider 16 as shown in the waveform M is applied from the voltage comparator CP4.
A determination signal indicating whether the output signal voltage of is higher or lower than the output voltage of the discriminator 12 is obtained. This is because when there is no identification error, the output signal amplitude of the voltage divider 15 is smaller than the output signal amplitude of the discriminator 12, so the output signal of the voltage comparator CP4 becomes the inverted version of the output signal of the discriminator 12. In other words, when the error voltage included in the output signal of the adder 14 exceeds a certain value (IEI) due to feedback of the reproduced signal of the identification error, the polarity of the output signal of the voltage comparator CP4 changes to the discriminator 1.
It matches the output signal of 2.

Next, the output of the voltage comparator CP4 is gated by the output signal of the discriminator 12 using the NAND gate G1, so that error propagation on the positive side is detected for a period of Tc, as shown by waveform N. Note that waveform 0 is the output of G2, and if there is error propagation on the negative side, the level will be +E.

As shown in waveforms N and O, the outputs of gates G1 and G2 are as follows:
Whiskers occur due to group delays in gates, voltage comparators, etc., but this does not pose a practical problem if the pulse width of the whiskers is small. If the group delay is very large,
It is also effective to insert a series resistor into the clamp diode of the clamp circuit.

According to the embodiments described above, only one voltage comparator is required and the circuit configuration is simplified.

Effects of the Invention As described above, according to the present invention, a quantization feedback type integral detection circuit that suppresses the propagation of identification errors after the next sign inversion and reduces the occurrence of identification errors can be realized with a simple circuit configuration.

[Brief explanation of drawings]

FIG. 1 is a circuit diagram showing a first embodiment of the present invention, FIG. 2 is a waveform diagram for explaining the same embodiment, and FIG. 3 is a circuit diagram showing a second embodiment of the present invention.
4 is a circuit diagram showing a third embodiment, FIG. 6 is a waveform diagram for explaining the third embodiment, FIG. 6 is a block diagram showing a conventional example, FIG. 7 is a waveform diagram for explanation of a conventional example. 1...Bypass filter, 2...Low pass filter, 3...Adder, 4...
...Discriminator, 1o... Adder, 11...
・Clamp device, 12... Discriminator, 13...
...Error propagation detector, 14...Adder, 15.
...Voltage divider. Figure 1 Figure 2 Figure 3 ・1 [71 Figure 5

Claims (4)

[Claims] (1) Adding means for adding the high-frequency components of the received signal and the low-frequency components of the reproduced signal after identification, error propagation detection means for detecting error propagation, and error propagation detected by the error propagation detection means clamping means for peak value clamping the output signal of the adding means to a regular voltage corresponding to 1 or 0 when 2. A quantization feedback type integral detection circuit comprising means, wherein a discharge time constant of said clamp means is made equal to a time constant of a first-order low-pass filter for inputting a reproduced signal of said addition means. (2) The error propagation detection means is composed of two voltage comparators that detect error propagation on the positive side or negative side when the output signal of the adding means exceeds a certain voltage on the positive side or negative side, respectively. A quantization feedback integral detection circuit according to claim 1. (3) A second adding means for adding the high-frequency component of the received signal and the low-frequency component of the re-corrected signal after identification, and the crossover frequency of the second adding means is set to the cross-over frequency of the first adding means. 2. The quantization feedback type integral detection circuit according to claim 1, wherein the quantization feedback integral detection circuit is set higher than the overfrequency, and the output signal of the second addition means is used as an input signal of the error propagation detection means. (4) a coefficient unit that applies a coefficient to at least one of the output signal voltage of the second addition means and the output signal voltage of the identification means;
a voltage comparator that compares the output signal voltage of the second adding means and the output signal voltage of the discriminating means after multiplying at least one of them by a coefficient using the coefficient multiplier; 4. The quantization feedback type integral detection circuit according to claim 3, wherein the error propagation detection means is constituted by a gate gated by an output signal.