JPH06164453A - Automatic equalizer - Google Patents

Abstract

PURPOSE:To provide an automatic equalizer which can converge the variable multiplication coefficient of an equalizer even by an equalizer control signal of a short time obtained in a short training state. CONSTITUTION:An input signal is demodulated by a demodulating part 2 after passing through an AGC circuit 1 and discriminated into the I and Q components through a roll-off filter part 3. The signal equalized by an automatic equalizing part 4 is demodulated at a deciding part 5. The variable multiplication coefficient of the part 4 is controlled by the error signal of an error detecting part 6. When an alternate pattern of ABAB of a segment 1 is received in a short training state, a selecting part 8 selects the signals sent from a level detecting part 7 and controls the variable multiplication coefficient of the part 4. If a low level is detected, the variable multiplication coefficient is increased an vice versa against a high detected level. As a result, the equalization control signal of a segment 2 can converge the variable multiplication coefficient even in a short time.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an automatic equalizer, and more particularly to an automatic equalizer such as a transversal filter which is provided with equalization characteristics by a variable multiplication coefficient.

[0002]

2. Description of the Related Art CCITT Recommendation V.6. In the training of 17, two types of training before TCF and image signal are used. Long training is used for TCF and consists of four types of segments. Segment 1 is ABA
256 are transmitted with the alternating pattern of B as the number of symbol times. The ABAB pattern is a signal space diagram in which the real part is (-6), the imaginary part is (-2), the A point and the real part are (+2), and the imaginary part is (-6). To do. Segment 2 is the equalizer adjustment signal and the number of symbol times is 2976. Segment 3
Is a rate information sequence and the number of symbol times is 64
Is. Last segment 4 was scrambled "
1 ″ is a continuous signal, and the number of symbol times is 48.

CCITT Recommendation V.3 In 17, the variable multiplication coefficient of the equalizer is adjusted by long training to determine the coefficient, and the training before the image signal is performed in a short procedure in order to shorten the communication time. Specifically, segment 1 is transmitted 256 with the ABAB alternating pattern as the number of symbol times. Therefore, it is the same as long training. The equalizer adjustment signal for segment 2 has 38 symbol times, which is a very short time compared to long training. There is no segment 3, and in segment 4, a scrambled "1" sequence signal is transmitted as the number of symbol times of 48.

FIG. 2 is a block diagram of a main part of a facsimile receiving section having a conventional equalizer. In the figure, 1 is AGC
The circuit, 2 is a demodulation unit, 3 is a roll-off filter unit, 4 is an automatic equalization unit, 5 is a discrimination unit, and 6 is an error detection unit. The input signal is subjected to gain adjustment in the AGC circuit 1, demodulated in the demodulation unit 2, and is discriminated into an I component and a Q component through the roll-off filter unit 3. The discriminated output signals are subjected to signal distortion generated on the line by the automatic equalization unit 4. The signal equalized by the automatic equalization unit 4 is decoded by the determination unit 5. The error detection unit 6 detects the difference between the output of the automatic equalization unit 4 and the output of the determination unit 5, and obtains the error between the reception point and the determined signal point.

The automatic equalization unit 4 uses a circuit having a delay characteristic such as a transversal filter or an N-bit shift register, and the variable multiplication coefficient is controlled by the error signal from the error detection unit 6, so that the equalization characteristic is improved. It is determined.

Even during training, the received signal is A
AGC is applied by the GC circuit. The AGC is operated so as to have the same reception level during both long training and short training, but in reality, it does not always have the same reception level. Therefore, the equalizer adjustment signal of the segment 2 for short training may not be able to converge the variable multiplication coefficient of the equalizer.

[0007]

SUMMARY OF THE INVENTION The present invention has been made in view of the above-mentioned circumstances, and can converge the variable multiplication coefficient of the equalizer even with a short-time equalizer adjustment signal during short training. It is intended to provide an automatic equalizer.

[0008]

SUMMARY OF THE INVENTION The present invention provides an automatic equalizer for receiving a training signal and adjusting the equalization characteristic, and detecting the signal level of a predetermined pattern in short training. Means is provided, and the variable multiplication coefficient of the automatic equalization unit is adjusted based on the output of the level detection means.

[0009]

According to the present invention, the variable multiplication coefficient of the automatic equalizer is adjusted in advance based on the signal level of the predetermined pattern in the short training. Therefore, even if the equalizer adjustment signal is short, the variable multiplier of the equalizer is adjusted. The coefficients can be converged.

[0010]

DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 is a block diagram of a main part of a facsimile receiving section having an automatic equalizer according to an embodiment of the present invention. In the figure, 1 is an AGC circuit, 2 is a demodulation unit, 3 is a roll-off filter unit, 4 is an automatic equalization unit, 5 is a discrimination unit, 6 is an error detection unit, 7 is a level detection unit, and 8 is a selection unit. . The gain of the input signal is adjusted by the AGC circuit 1, and the demodulation unit 2
The signal is demodulated by and is discriminated into an I component and a Q component through the roll-off filter unit 3. The discriminated output signals are subjected to signal distortion generated on the line by the automatic equalization unit 4.
The signal equalized by the automatic equalization unit 4 is decoded by the determination unit 5. The error detection unit 6 detects the difference between the output of the automatic equalization unit 4 and the output of the determination unit 5, and obtains the error between the reception point and the determined signal point.

As in the conventional example, the automatic equalizer 4 uses a circuit having a delay characteristic such as a transversal filter or an N-bit shift register, and the variable multiplication coefficient is controlled by the error signal from the error detector 6. And the equalization characteristics are determined. When adjusting the variable multiplication coefficient with the error signal from the error detection unit 6, the selection unit 8 selects the signal from the error detection unit 6.

In segment 1 during short training, when receiving the ABAB alternating pattern,
The selection unit 8 selects the signal from the level detection unit 7 and adjusts the variable multiplication coefficient of the automatic equalization unit 4. The level may be detected by detecting one or both of the A and B signals. When the detected level is low, each variable multiplication coefficient (30 taps in this embodiment) is increased, and conversely, when the level is high, the variable multiplication coefficient is decreased.

As a result, the equalizer adjustment signal of segment 2 can converge the variable multiplication coefficient even in a short time.

[0014]

As is apparent from the above description, according to the present invention, when the path metric value is large, the variable multiplication coefficient is changed with a large change width to improve the convergence accuracy, and the path metric is improved. When the value is not large, there is an effect that the variable multiplication coefficient of the automatic equalizer can be optimally controlled by reducing the change width of the variable multiplication coefficient.

[Brief description of drawings]

FIG. 1 is a block diagram of a main part of a facsimile receiving unit having an automatic equalizer according to an embodiment of the present invention.

FIG. 2 is a block diagram of a main part of a facsimile receiving unit having a conventional equalizer.

[Explanation of symbols]

 1 AGC circuit 2 demodulation unit 3 roll-off filter unit 4 automatic equalization unit 5 discrimination unit 6 error detection unit 7 level detection unit 8 selection unit

Claims (1)

[Claims] 1. An automatic equalizer that receives a training signal and adjusts the equalization characteristic is provided with level detection means for detecting a signal level of a predetermined pattern in short training, and the level detection means is provided based on the output of the level detection means. An automatic equalizer characterized by adjusting a variable multiplication coefficient of an automatic equalizer.