JPS5842975B2 - Receiving system adjustment control method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to an adjustment control method for a reception system in a modem reception section, and in particular, after performing initial pull-in adjustment of a carrier phase control circuit, an automatic equalizer and a phase control circuit are independently controlled. The present invention relates to an automatic equalizer configured to operate and capable of pulling in an automatic equalizer while maintaining the state of a phase control circuit that has been once pulled in.

In a data transmission system, before data is actually transmitted and received, it is necessary that the receiving section be adjusted to a state in which it can receive data correctly.

For example, when transmitting and receiving data over a telephone line using a modem, it is not actually possible to transmit and receive data immediately after turning on the power of the device.

When transmitting and receiving data using a telephone line, the line used for sending and receiving data changes every time a connection is made to a public telephone line.

Therefore, the transmission circuit characteristics change each time transmission and reception are performed, so in order to receive data correctly, the receiving side must make adjustments each time it receives data. Specifically, the receiving section's automatic equalizer, phase control circuit, etc. is drawing in.

In other words, the error must be reduced by adjustment.

For example, a high-speed modem (9600bps or 4800bps)
etc.), a training signal is used for the above-mentioned entrainment.

This training signal was developed by the International Telegraph and Telephone Advisory Committee (C
CITT), but its usage is left to the discretion of each device.

According to the above-mentioned CCITT regulations, there are four types of training signals for a 9600 bps modem, and among these, the AB signal and CD signal are used for the above-mentioned pull-in.

The AB signal refers to a signal in which signal points A and B shown in FIG. 1 are sent out alternately, and the CD signal refers to a signal in which signal points C and D are generated pseudo-randomly and sent out.

Here, FIG. 1 will be explained.

When transmitting data at 9600 bps using a telephone line, every 4 bits are converted into one signal point and transmitted.

When transmitting 8-bit information called roooolollJ from Rj, divide it into two 4-bit information, 1'-0000j and ``1011'', and set ``0000'' to signal point C and 11011.
J is expressed as a signal point.

The signal point C is transmitted as a signal wave having a distance of 3 from the origin and a phase angle of 0 degrees with the X axis.

Further, the signal point is transmitted as a signal wave having a distance of 3 d from the origin and a phase angle of 135 degrees with the X axis.

When the receiving side receives these signal waves, it determines that signal points C and D are to be displayed, and displays "0OOO
It is interpreted as "00001011" which is a concatenation of J and "1011".

Therefore, 2'=16 such signal points are determined, and information of every 4 bits is transmitted and received as information of one signal point.

However, due to the characteristics of the transmission line, for example, even if signal point A is sent, when it is actually received, the distance from the origin is not 3 and the phase angle with the X axis is 180 degrees, but these distances or the phase angle may shift, causing an error.

Therefore, if data is received with the above-mentioned error corrected using a phase adjustment circuit or an automatic equalization amount, it becomes possible to receive data accurately.

Furthermore, the cause of the error is not only the difference in distance from the origin or the phase shift.

The signal wave related to the above signal point is transmitted by a carrier of 1700 Hz, but when receiving it on the receiving side, the carrier is received with the frequency increased or decreased from 1700 Hz.

Since a so-called carrier offset occurs, correction for it is also required.

Therefore, an object of the present invention is to provide an adjustment control method for a receiving system that can perform each of these corrections independently of each other.
A receiving section that receives signals transmitted from other stations includes a low-pass filter that filters the received signal, an automatic equalizer that operates to equalize the received signal and reduce errors, and the automatic equalizer. a determination circuit for determining which signal point among the plurality of signal points the transmitted signal corresponds to; A phase control circuit for compensating the phase of the signal is provided, and when receiving an AB signal in which A and B signals of the transmitted training signals are alternately output, the phase control circuit is connected to the low-pass filter. Controlled by the output signal, the above A signal and 180
A C signal having a phase difference of 0 and a D signal having a phase difference of 180° from the B signal are output in a pseudo-random manner.

When receiving the training signal of the IcDIc type, the automatic equalizer and the phase control circuit are controlled with a signal corresponding to the received signal and the output signal of the determination circuit to correct errors in the received signal. It is characterized by being configured to do so.

An embodiment of the present invention will be described below with reference to FIG.

In the figure, 1 is a low-pass filter, 2 is an automatic equalizer, 3 to 6 are multiplication units, 7 is an integrator, 8 is a coefficient circuit, 9 is a multiplication unit,
10 is an adder, 11 is a coefficient circuit, 12 is an integrator, 13 is a coefficient circuit, 14 is a determination circuit, 15 is a reference signal generation circuit,
16 is a reciprocal number generation circuit, 17 is an addition section, 18 is a multiplication section, P
is a phase control circuit, SWl is a changeover switch, and SW2 is a switch.

As the low-pass filter 1, for example, a cyclic digital filter as shown in FIG. 3 is used.

The cyclic digital filter has a well-known configuration consisting of a delay circuit 19, a coefficient circuit 20.21, and an addition section 22.23, and has a filter coefficient α.

It is. The automatic equalizer 2 is, for example, of a transversal type, and includes a plurality of delay circuits, a multiplication section provided for each delay circuit, and a circuit to which the outputs of these multiplication sections are transmitted. It also has a well-known configuration.

The phase control circuit P has the configuration described in the specification of Japanese Patent Application No. 52-89830 filed by the same applicant as the present invention.

The multiplier 4 performs the following operations.

That is, the output of the warpass filter 1 constantly fluctuates even between AB signals due to carrier offset, phase jitter, and the like.

This is multiplied by the output of the integrator 7 so that the output of the multiplier 4 is fixed in the positive direction of the X-axis.

Also, the output of the integrator T is adjusted to such a value.

The multiplier 5 performs the following operations.

Considering the CD signal, the output of the multiplier 4 is the DC component C+D.
The output of the seed extractor 7 is set such that the value is in the positive direction of the X axis.

On the other hand, in the output of the multiplier 3, the C and D signals need to be in the original correct direction, and the DC component C+D is directed in the positive direction of the Y axis.

This has a phase difference of exactly 90° from the output of the multiplier 4,
Since the outputs of the low-pass filter 1 and the automatic equalization amount 2 are in phase, the output of the integrator 7 is multiplied by O+jl in the multiplier 5, rotated by 90°, and input to the multiplier 3.

The reciprocal number generation circuit 16 performs the following operations.

As shown in FIG. is output.

The error signal obtained from this adder 17 indicates a signal shift.

First of all, even if the signal is constant, the error vector differs depending on the amplitude and phase of the signal.

The function of always normalizing this on the X axis is the multiplication of the error signal in the multiplier 18 and the reciprocal signal output from the reciprocal number generation circuit 16.

That is, as shown at the beginning of FIG. 5, if A is a reference signal and R is an automatic equalization amount output, e is an error signal.

By multiplying this error signal e by AN, which is the reciprocal of A, the error signal e can always be a signal on the X-axis regardless of the phase of the reference signal A.

Further, even in the case of the same phase error, when the amplitude of the reference signal A is large, the error signal e is also large, and when the reference signal A is small, the error signal e is also small.

Then, by multiplying the error signal e by the above AN, the magnitude of the error signal due to the amplitude of the signal is also normalized.

The Y component of the standardized error signal eN indicates the phase error, and the phase control circuit is controlled by this.

When actually receiving data, a training signal is first sent out, and then a data signal is sent out, as shown in FIG. 4A.

Therefore, first of all, it is necessary to accurately adjust the receiver while receiving the training signal.

The training signals include an NTE signal that transmits no signal energy, and signal points A, B, and C shown in FIG.
.

Of these, the NTE signal is detected as a signal disconnection in the receiving section, and two signals are used to pull in a descrambler (not shown), and these are not directly related to the present invention.

The AB double signal, a signal representing signal point A and a signal representing signal point B, are each sent out alternately 64 times, for a total of 128 symbols.

Following this AB double signal, a signal representing the signal point C and a signal representing the signal point are transmitted 384 times, that is, 384 symbols.

In the CD signal, signal points C and D are not sent out alternately, but are sent out irregularly, although the order is determined in advance.

When receiving the training signal, first operate the changeover switch SW1 so as to make contact with the contact point a, and leave the switch SW2 in an open state.

Now, the AB double signal is received and transmitted to the low-pass filter 1 through the input terminal (2).

As mentioned above, the AB double signal is one in which signal points A and B are sent out alternately, so if it is received at a normal frequency, it will have a constant output as shown by arrow C in the fourth diagram. , that is, a DC output is output from this low-pass filter 1.

However, in reality, because there is a carrier offset, that is, even if a signal is sent at a frequency of 1700H2, it will be received at a slightly higher or lower frequency at the receiving end. For example, a signal rotating at the carrier offset frequency, that is, an output signal rotating in the force direction of arrow C', as shown by the dotted line in FIG. 4, is output from the low-pass filter 1.

The y component of the output of this low-pass filter 1 becomes a phase error, and if there is a carrier offset as described above, it is sent to the first integrator 12 of the phase control circuit P via the contact a of the changeover switch SW1. This is integrated to obtain a value that is large enough to compensate for the carrier offset.

This is then transmitted to the second integrator 7, and the second integrator 7
The output of the low-pass filter 1 is multiplied by the multiplier 4, and the y component of the output of the low-pass filter 1 is controlled to be zero.

Therefore, it is possible to first adjust so that phase synchronization is achieved completely and the influence of carrier offset is canceled.

The time required for this phase synchronization becomes longer as the carrier offset becomes larger, but in practice it can be sufficiently reduced to within 128 symbols for frequencies up to about 7 Hz.

Next, when the training signal changes from the AB signal to the CD signal, the switch SW2 is closed, but the changeover switch SW,
is in conduction with the contact point a side for a while.

By closing the switch SW2, the automatic equalization amount 2 is now drawn in as follows.

That is, the above signal (1) is received and transmitted to the automatic equalizer 2 from the input terminal IN.

At this time, the determination circuit 14 generates C and D in the same order as the transmitted CD signal.

Therefore, when point C is received, the determination circuit 14 determines that
A reference signal corresponding to the (x, y) coordinate value of the point, that is, (3, 0) is generated from the reference signal generating circuit 15.

This reference signal is then applied to the determination circuit 1 in the adder 17.
This will be subtracted by the signal input to 4.

If the signal now input to the judgment circuit 14 accurately indicates the signal point C and the signal point 14, the judgment circuit 1
4 and the signal generated from the reference signal generation circuit 15 are equal, and in this case, the signal generated from the adder 17 is 0, but in reality, there is a considerable deviation in the received signal. is included, the reference signal generation circuit 15
There is a considerable difference between the signal generated from the signal and the signal input to the determination circuit 14.

Then, the difference, that is, the amount corresponding to the error included in the received signal is transmitted from the adder 17 via the multiplier 6 to the automatic equalizer 2 as a control signal.

Then, as a result of the correction in the automatic equalizer 2, the automatic equalizer 2 is pulled in so that the above-mentioned error is eliminated.

Until the automatic equalizer 2 has drawn in to a certain extent, the adder 17
Since the output signal is thick, the phase control circuit P is controlled by the output of the low-pass filter 1 as before.

In the case of the CD signal, unlike the AB signal, it is difficult to completely detect the direct current component because it does not have regularity. Therefore, the nine frequencies of the low-pass filter 1 are made small.

That is, the filter coefficient α. Make the value close to 1.

At the same time, the coefficient α1 of the coefficient circuits 8 and 11 of the phase control circuit P.
α3 is also small, and the coefficient α2 of the coefficient circuit 13 is
Adjust so that the value of ×α3 remains unchanged.

In this way, the automatic equalizer 2 can be adjusted independently with phase locking.

Then, the automatic equalizer 2 is retracted to a certain degree, and the adder 17
When the error signal generated from the changeover switch SW1 becomes small, the changeover switch SW1 opens the contact a and closes the contact A.

That is, this time, the overall adjustment will be made more precisely.

The output signal from the determination circuit 14 is then used by the reciprocal generator 16 to generate a signal corresponding to (X+j,), which is multiplied by the error signal from the adder 17 in the multiplier 18.

The signal thus obtained becomes a correction signal for phase jitter and carrier offset, which is input to the second integrator 7 via the coefficient circuit 8.

The output of the second integrator 7 is multiplied by a j1 component to give a rotation component in a multiplier 5, and the output from the automatic equalizer 2 is further corrected in a multiplier 3.

In this way, since the received signal is corrected while receiving the AB signal and the CD signal, which are the training signals, the data sent out thereafter can be accurately received and output from the output terminal OU.
It can be taken out from T.

Ultimately, the phase control circuit is controlled by the error signal of the automatic equalization amount output, but the error is large until the automatic equalizer 2 pulls in, and it may not always be accurate at the beginning of the pull-in. I cannot say that it is.

Considering only the phase error, the output of the low-pass filter shows a more accurate error than the output of the automatic equalizer in the initial period of the CD signal.

Therefore, phase control is performed using the output of the low-pass filter until the automatic equalizer has pulled in to a certain extent.

However, the CD signal is not transmitted alternately like the AB signal and contains many frequency components.

Therefore, when detecting an average C+D signal, A
It is necessary to make the cutoff frequency of the low-pass filter smaller than during the B signal.

If the cutoff frequency is made small, the time constant becomes large and the follow-up characteristics are poor, so the cutoff frequency is made somewhat large during the AB signal.

As explained above, according to the present invention, after first correcting the carrier offset of the received signal by the phase control circuit P, the automatic equalizer 2 and the phase control circuit P can be operated independently. Therefore, it is possible to pull in the automatic equalizer 2, correct phase jitter, etc. while maintaining the state in which the carrier offset is corrected.

Therefore, it is possible to reliably draw in the automatic equalizer and the phase control circuit in a stable state while receiving the training signal.

[Brief explanation of the drawing]

Fig. 1 is an explanatory diagram of signal points, Fig. 2 is a configuration diagram of an embodiment of the present invention, Fig. 3 is a configuration diagram of a low-pass filter, Fig. 4 is an explanatory diagram of a received signal, and Fig. 5 is a reciprocal number generation diagram. FIG. 3 is an explanatory diagram of the operation of the circuit. In the figure, 1 is a low-pass filter, 2 is an automatic equalization amount, 3 to 6 are multiplication units, 7 is an integrator, 8 is a coefficient circuit, 9 is a multiplication unit,
10 is an adder, 11 is a coefficient circuit, 12 is an integrator, 13 is a coefficient circuit, 14 is a determination circuit, 15 is a reference signal generation circuit,
16 is a reciprocal number generation circuit, 17 is an addition section, 18 is a multiplication section, P
is a phase control circuit, SWo is a changeover switch, SW2 is a switch, 19 is a delay circuit, 20.21 is a coefficient circuit, 22.
23 is an addition section.

Claims (1)

[Claims] 1. A receiving section that receives signals transmitted from other stations includes a low-pass filter that filters the received signal, an automatic equalizer that operates to equalize the received signal and reduce errors, and the automatic equalizer. a determination circuit for determining which signal point among the plurality of signal points the transmitted signal corresponds to; and a determination circuit for reducing the y component of the received signal of the low-pass filter to zero. AB is provided with a phase control circuit that compensates for the phase of the signal, and in which the A signal and B signal of the transmitted training signals are output alternately.
When receiving a doubled signal, the phase control circuit is controlled by the output signal of the low-pass filter, and the A signal and
C signal having a phase difference of 800 and the above B signal and 1800
When receiving a CD signal of the training signals in which a D signal having a phase difference of 1. An adjustment control method for a receiving system, characterized in that the system is configured to correct errors in a received signal by controlling with a signal corresponding to the signal.