JPS6349927B2 - - Google Patents

Description

[Detailed Description of the Invention] The present invention deals with line signal fluctuations, such as amplitude deterioration,
The present invention relates to a training control method for an automatic equalizer that uses a training signal to initialize the equalization amount in an automatic equalizer that equalizes phase fluctuations, etc.

In general, automatic equalizers are known that accurately obtain received data by equalizing amplitude degradation and phase fluctuations during transmission of received signals, and means for initializing the amount of equalization by such automatic equalizers. As a result, data generated in a predetermined data generation order, so-called training data, can be sent for synchronous suction immediately before data transmission to control the suction.
Specified by CCITT Recommendation V.

Figures 1a and 1b are diagrams for explaining the form of this training signal, Figure 1a is a diagram for explaining the data position on the phase plane of the signal, and Figure 1b is a diagram for explaining the generation order of training data. It is.

The binary data to be transmitted on the transmitting side (see Figure 1a) is grouped into 4-bit units, and the phase and amplitude of the transmitted signal are adjusted so that the data corresponds to the 4-bit data on the phase plane. It is determined.

Training data TR is the data to be sent
Placed immediately before DATA, no signal period NTE,
It is composed of an AB data period AB, a CD data period CD, and a scramble data period Z.

The no-signal period NTE notifies the receiving system of the start of data transmission, and the receiving side modem (modem) detects this and resets various circuits, such as correcting the automatic equalizer. Determine the amount. It is used to reset the so-called tap coefficient.

During the AB period AB, the polarity is reversed with the same amplitude on the amplitude axis (0° to 180°) on the phase plane shown in Figure 1a, and the value is "0" on the phase axis (90° to 270°). This is a period during which data is sent out that alternately repeats points A and B indicating a predetermined phase, and the receiving modem uses this signal to extract initial values such as phase offset to equalize the phase on the line. used for.

The CD data period CD is a period in which points C and D on the phase plane whose phase components are inverted in polarity with respect to the AB period are randomly transmitted. It is used to grow the tap coefficient, which is the equalization coefficient of the equalizer, and to initialize (hereinafter referred to as "drawing in") a correction amount sufficient to equalize the deterioration on the line.

The scramble data period Z is a period in which all types of 4-bit data having each eye pattern indicated by a circle in Figure 1a on the phase plane are randomly transmitted. Equalizer and phase control sections are used to fine-tune the quantities. On the other hand, the conventional method for controlling the equalizer is to take advantage of the fact that the polarity is reversed on the phase axis in the receiving modem when transitioning from the AB data period AB to the CD data period CD. Then, the zero-crossing point of the so-called phase component that crosses the amplitude axis is detected, and this timing is set as the generation start timing of the judgment reference data string used when judging the signal with respect to the equalized signal of the equalizer. I was doing it.

In addition, an equalizer generally performs signal equalization by combining the response fluctuation signal components on the data line, and the received signal has enough information to extract almost all of these components. The period is stored in the equalizer.

However, when the generation activation timing is extracted from the zero crossing point of the training signal as described above, the extraction timing is at the center position of the phase change point.

However, the received signal, including the components to be stored, varies depending on the type of deterioration of the line. For this purpose, the number of taps of the equalizer must be considered so that data for the period in which all line fluctuations are predicted can be stored on both sides of the central position of the extracted signal, and the equalizer itself must be It has the disadvantage of being large.

The purpose of the present invention is to eliminate these conventional drawbacks and
It is an object of the present invention to provide a training control method for an automatic equalizer that can perform waveform equalization and can be downsized with a storage capacity commensurate with the amount of predetermined received data to be stored.

In order to achieve the above object, the present invention uses the data generated in the training data line for equalization as the reference data generation start timing for detecting a received signal containing a fluctuation component and starting tap growth. This will be explained in detail below.

FIG. 2 is a block diagram of an embodiment of the present invention, and FIG. 3 is a time chart.

In the figure, the same parts as those used in FIG. 1 are indicated by the same symbols. Also, DM is a demodulation unit that multiplies the received signal by a carrier signal to create its real and imaginary components, a so-called baseband signal, and EQL is an equalization unit that changes the waveform of the baseband signal. , PU is the change point detection unit,
RGC is a generation control section for a reference signal (hereinafter referred to as reference), and TSR is a reference generation section for a reference data string, which is composed of a circulating shift register, etc., and is sent to its final output stage every time a shift pulse is supplied. , which sequentially supplies reference data strings; D is a determination unit which determines and outputs the position on the plane shown in FIG. 1a based on the equalization output from the equalization unit EQL; TR is a tap register;
EQ is an equalization calculation section, S is an integration circuit, T is a delay section,
A is the addition operation section, Z is the operation section, SL is the slicer,
PG is the pulse generator, G is the gate, rv is the received signal, i and r are the baseband signal of the imaginary number (phase) component and the baseband signal of the real number (amplitude) component, tim is the reception timing, and O is the judgment output. .

Explain the operation.

The received signal is separated into a real component r and an imaginary component i and demodulated in the demodulator DM. Each of these output signals is converted into a digital value, and the tap register TR
is stored in

The tap position data in the tap register TR is supplied to the change point detection unit PU. The change point detection unit PU integrates the input signal using an integration circuit S. Outputs si and sr (shown in FIG. 3) of the integrating circuit S are supplied to an addition operation section A and a delay section T. Addition operation unit A multiplies the signals si and sr delayed by a unit data symbol (signals si and sr shown by dotted lines in FIG. 3) by a value "-1" and then adds them to the signals sr and si, respectively. .

The addition operation unit A thereby outputs a signal having the waveform shown in FIG. 3a.

As explained in Figure 1, from AB period AB
At the time when the CD period changes to CD, the polarity of the complex vector is reversed. Therefore, the output of the addition operation section A has a portion (ch) corresponding to the change point and a line fluctuation component ch'.

The arithmetic unit Z performs an operation to obtain "|R| ² +|I| ² ", where R is the real component of the signal a and I is the imaginary component. The output of this calculation becomes a signal containing a line fluctuation component as shown in FIG. 3b.

The slicer SL of the generation control unit RGC slices this signal at a predetermined slice level sl (shown in FIG. 3b), and when the input signal becomes equal to or higher than the slice level sl, the slicer SL changes the signal from level "0" to level "1". Become.
The pulse generator PG outputs a pulse signal C in FIG. 3 at the change point.

A timing tim of a predetermined period corresponding to a unit data symbol is supplied to the gate section G, and from the moment the pulse c is supplied from the pulse generator PG, a gate (not shown) is opened and this timing tim is shifted to the reference generation section TSR. Supplied as a pulse.

Furthermore, the gate section G becomes closed at the time when timing signals tim corresponding to the number of data existing in the CD period are output.

The reference generation unit TSR stores the data string of the CD period in a corresponding order, and every time a shift pulse is input, it generates reference data as shown in FIG. 3d and supplies it to the determination unit D. .

On the other hand, the equalization calculation unit EQL of the equalization unit EQL uses a known method to calculate the difference between the judgment output of the judgment unit D and the value based on the equalized waveform, and calculates the difference between each tap register TR.
The value to be multiplied by the data at each tap position is adjusted.

After the tap coefficient is adjusted during the CD period,
The value of the equalized signal is determined using a predetermined phase and amplitude determination plane, regardless of the reference data.

On the other hand, in general, the values of tap coefficients corresponding to each tap are distributed in a form corresponding to line fluctuations.

Therefore, as described above, if the tap coefficient is grown from the time when the received signal includes even a slight line fluctuation component at the change point from the AB period to the CD period,
The center tap position corresponds to line fluctuations in the tap register TR.

For this purpose, according to the invention, a tap register is provided.
Even if the storage period of the TR is short, more than 90% of the signals having a center tap at a position corresponding to line fluctuations and having a signal component due to line fluctuations can be stored in the tap register.

In other words, the device can be made smaller.

[Brief explanation of the drawing]

Figures 1a and 1b are diagrams showing the format of training data in the field to which the embodiment of the present invention is applied; Figure 2;
FIG. 3 is a block diagram and a time chart of one embodiment of the present invention. In the figure, DM is the demodulation section, EQL is the equalization section, PU is the change point detection section, RGC is the reference generation control section,
TSR is a reference generation section, and D is a determination section.

Claims (1)

[Claims] 1. An equalizer that equalizes and outputs a received signal by an amount of correction corresponding to a variation in the received signal, and reference data that sequentially generates training data whose generation order is determined in advance. a training control method for an automatic equalizer, which includes a generating section, equalizes a received signal including training data, and initializes the correction amount by comparing it with training data generated from the reference data generating section. The received signal is divided into the first data period (AB) and the first data period (AB).
It includes a second data period (CD) in which the polarity of the data and phase component of the data period is reversed, and the change point between the first data period and the second data period of the received signal is defined as the real part and the imaginary part of the received signal. Detected by detecting that the sum of the components of is greater than or equal to a predetermined value,
1. A training control method for an automatic equalizer, characterized in that training data is sequentially generated from the reference data generator from the detection time point to initialize a correction amount.