JPS58182330A - Modem training system - Google Patents

Abstract

PURPOSE:To make the training successful, to decrease the loss in information transmission time and to increase the amount of data transmission, by decreasing the training time when the deterioration is less and the line characteristics are excellent and increasing the training time when the line characteristics are worse. CONSTITUTION:When a continuous wave inverted for the phase by 180 deg. is transmitted to a reception side 30 at a segment 1, a demodulator 31 detects it to recognize the start of the MODEM training, and rough adjustment of timing and carrier reproduction is done at first. When the shift in the timing and the phase of carrier is within a predetermined range, a sequencer 33 drives a signal transmitter 32 to generate a transmission request signal for a segment 2 (hereinafter called Seg 2 request signal) and returns it to a transmission side 10. In the Seg 2 request signal is not received during a prescribed period (almost several tens of ms.), the sequencer 14 discriminates it as disable reproduction of timing and carrier, the data in the transmission speed is thrown away and the speed is transferred to a transmission speed lower than the speed.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a data transmission system using a modem, and more particularly to a modem tray-end system in a data transmission system suitable for transmitting data over multiple addresses.

When transmitting data such as a 7-axis signal through a telephone line, the telephone line is a rounded type that transmits voice signals that are originally analog signals, and does not necessarily transmit data such as a 7-axis signal. The ninth most optimized version is 7.
&l/s is also a contributing factor, causing various deteriorations in line quality such as delay distortion and attenuation distortion.

A modem (modulator/demodulator) modulates input data such as a facsimile signal on the transmitting side and converts it into a nine-spectrum signal that matches the band of the transmission path, and on the receiving side demodulates the received signal and converts it to the original signal. This is a device that regenerates the data of the 7-axis signal, but since there are various quality deterioration factors in the line, the modem uses 1. A modem relay must be performed before transmitting data.

The modem tracer sends a certain test signal to the 4t modem to compensate for the quality deterioration that has occurred in the line, but CCITT transmits a certain test signal at 9600 bpm,
7 N 0 0 1 eps, 4 8 0
For each transmission speed of 0 kpm and 2400 bpm, the training sequence and the time required for training are predetermined to be constant.

Therefore, the training time cannot be shortened in response to the deterioration of the line, so the training time is too long for transmission lines with little deterioration and relatively good line characteristics, and there is a loss in the overall information transmission time. On the other hand, for transmission lines with a lot of deterioration and poor line characteristics, training may be successful if the training time is a little longer, but since the training time is specified as a constant, Since the data transmission cannot be successfully transmitted at that transmission speed, data transmission at that transmission speed must be abandoned and data transmission must be performed at an even lower transmission speed.
This has the disadvantage of increasing losses in terms of information transmission time.

SUMMARY OF THE INVENTION An object of the present invention is to provide a modem training method which can eliminate the drawbacks of conventional training methods, reduce loss of information transmission time, and increase data transmission amount.

For this reason, the present invention makes the training time variable according to the quality and characteristics of the line. The present invention is characterized by making training successful by lengthening the training time, reducing loss of information transmission time, and increasing the amount of data transmission.

Hereinafter, the present invention will be explained with reference to the drawings.

First, before explaining specific embodiments of the present invention, let us first give an overview! ! will be explained with reference to FIGS. 1 to 3.

Modem training is a day! It causes the demodulator to perform the preparation work necessary for reception, and although it differs depending on the transmission speed, the main contents are the following three items.

(1) Establishment of the main construction regeneration operation, that is, the construction maintenance
・Regenerate Luz and establish synchronization.

(2) Establishing a carrier regeneration operation, that is, performing carrier regeneration for synchronous detection.

(3) Establishing a correction operation of the automatic equalizer, that is, shaping the input data to compensate for nine intersymbol interferences caused by line distortion.

Of these, the most important factor that determines whether modem training is successful or not is the establishment of corrective actions by the automation device.

FIG. 1 is an explanatory diagram of the correction operation of the automatic equalizer, and FIG. 2 is a diagram of the principle of the automatic equalizer. As shown in Figure 1 (, ), when data consisting of Jars codes with a fixed interval T is transmitted from the transmitting side to the receiving side, if there is no transmission distortion on the line, the bandwidth of the line is relatively narrow. Therefore, each nominal pulse becomes as shown in FIG. 1(b). (Figure 1(b) shows only the waveforms for the pulse P.) Each waveform has an amplitude proportional to its respective pulse amplitude value at each timing point, but the amplitude of the pulses of the other pulses At this point, the amplitudes are all zero.

Then, the pulse P has a finite value only at t,
Other timing points t! , tl...'-1y1
−． ...The amplitudes are all expressed as zero. Therefore, time t・
, tl, are sampled with t-1, t-=..., the original Norse code p, e pl a
PI...P-11P-*... will be demodulated. However, when transmission distortion is received from the circuit, distortion occurs in the received waveform as shown in Figure 1 (@), and the pulse P component is present even at timings other than the ninth point t. Become. Other/l Lus PleP! ...
Regarding K, similarly, an output is generated at the timing of other Δ pulses. Therefore, when sampling at time t, in addition to the Δrus P component, P1+Ps...PI.

If components of p and ++ are mixed in, code amount interference will occur. The automatic equalizer is a device that automatically corrects the transmission distortion received from the line in this way to eliminate code amount interference.

In FIG. 2, r is a tagged delay circuit consisting of a plurality of delay lines D1 to D having a delay time Tt, M@-wM,
is a multiplier, and Σ is a synthesis circuit. In this configuration, the output of the multiplier M (B in Fig. 1 (-)) and the output of the multiplier M1 delayed by time T (C) in Fig. 1 0)
When synthesizing, by adjusting the values of the multipliers M and Ml, it is possible to make the waveforms have the same amplitude and opposite phases as C1- at one time point tI, as shown in FIG. 1 (4). Then, the composite waveform becomes zero at time t1, so time t1
The code amount interference in is removed.

Next, by adding the outputs delayed by 2T and adjusting the value of the multiplier M.2M1 as well as the value of the multiplier M, the composite waveform can be made zero at time points t1 and t3.

When the output of the multiplier M is added, a minute output is mixed into the new nine at time tl, so in order to cancel this component, it is necessary to adjust KM·, Ml together with %M1.

Similarly, if the outputs of the multipliers M, ~M, are combined by the synthesizer Σ while adjusting the multiplier values, the time jl*jl
*js...j-1s t-=1 The amplitude of the composite waveform at '-' can be made zero. In theory, by increasing the number of signals without limit, it is possible to compensate so that code amount interference is eliminated even if key multi-transmission distortion that satisfies the convergence condition is received. However, since infinite time is required for this purpose, a value of l is set so that the amount of code amount interference can be practically ignored.

This correction operation involves the adjustment of the multipliers M and %Mn, so it would take a lot of time to do it manually, so an automatic equalizer is an automatic equalizer that automates this process. Automatic equalization operations can be sped up and standardized. However, the worse the line characteristics become, the more time it takes to correct the automatic equalizer, and the time required for tie-setting regeneration and carrier regeneration.The relationship between line characteristics and training time is as shown in Figure 3. Become. As in the past, when the training time is fixed to Ti, the possible range of 1 automatic activation is determined by the range 4C@ of C・~C1, as shown in Fig. 3, and the 4m characteristic is determined from cl. If it deteriorates, Treininda will not be successful.

In the present invention, the training time can be changed according to the line characteristics using a button, that is, when the line characteristics are bad, the training time is lengthened, so that training can be performed successfully even when the line characteristics are bad. It is.

Next, a concrete implementation of the present invention will be explained based on FIGS. 4 to 6.

FIG. 4 is a block diagram showing the configuration of the DM Trayung method of the present invention. In the figure, 10 is the transmitting side;
It is composed of a signal generator 111, a modulator 12, a signal detector 13, and a sequencer 14 that controls the entire operation of the transmitting side. 20 indicates a transmission line. 30 is a receiving side, which includes a demodulator (including an automatic equalizer (not shown) inside) 31, a signal transmitter 32, and a sequencer 33 that controls the entire operation of the receiving side.
Consisted of.

The modem training method using the apparatus shown in FIG. 4 will be explained with reference to the time chart of the modem training sequence shown in FIG. 5 and the flowchart on the transmitting side shown in FIG.

In this embodiment, the modem training sequence is composed of the following three segments.

(1) Segment) 1: Transmits a 180° phase-inverted continuous wave.

(2) Segment 2: 00, transmits a 180° phase inversion code.

(3) Segment 3': Transmits a scrambled signal of the continuous l signal.

The contents of this segment) 1.2.3 are based on CCITT Standard V
This corresponds to segment 3 and 4.5 in 27 TER.

Now, when a 180° phase-inverted continuous wave is transmitted to the receiver 1130 by segment 1, the demodulator 31 detects this and knows the start of modem training, and first performs coarse adjustment of Mayong regeneration and coarse adjustment of carrier regeneration. conduct. When the phase shift between the main ion and the carrier falls within a predetermined range, the sequencer 33 drives the signal transmitter 32 to generate a segment 2 transmission request signal (hereinafter referred to as Ji@g21! request signal). , and returns it to the sending side 10. The timing wave and carrier that fall within a certain deviation range are held at their phase values. Incidentally, since the 180° phase inverted continuous wave is not suitable for automatic equalization operation, automatic equalization is not performed at the stage of segment 1.

The signal detector 13 on the transmitting side 10 checks the reception of the 8-g2 request signal while transmitting the segment 1 signal.
g! When the request signal is received, it is notified to the sequencer 14 and the sequence of segment 2 is started.

If S・g2 within a predetermined period (approximately several I Q ms)
When the 9 request signal is not received, the sequencer 14 determines that the timing and carrier cannot be regenerated, discontinues data transmission at that transmission rate, and shifts to a transmission rate one step lower. When the S order g2* request signal is received,
The sequencer 14 drives the data generator 11 to generate θ°, 1
An 80° phase inversion code is generated, modulated by a modulator 12, and transmitted to a receiving side 30 via a line 2. When the demodulator 31 on the receiving side 30 detects the segment 2 signal, it performs fine adjustment for each phase of the coarsely adjusted timing wave and carrier to establish synchronization. Once the synchronization between the timing wave and the carrier is established, the equalization operation of the automatic equalizer, i.e., the second
The values of the magnification of each multiplier shown in the figure.

! Perform constant setting work. Then, when the code amount interference falls below a predetermined level and the meltening adjustment operation is completed, the sequencer 33 drives the signal transmitter 32 and sends a signal to the segment 3.
A signal detector 1 on the transmitting side 10 that generates a transmission request signal (hereinafter referred to as B@g3 request signal) and returns it to the transmitting side 10.
3 is transmitting segment 2 signal 8・g31! Checking for reception of request signal, 8@g 31! When the request signal is received, it is notified to the sequencer 14 and the sequence of segment 3 is entered.

If the 8.g3 request signal is not received within a predetermined period of time, the sequencer 14 determines that automatic equalization adjustment is impossible, abandons data transmission at that transmission rate, and transmits data at a transmission rate one step lower. Shift to data transmission by 8@g
Upon receiving the three request signals, the sequencer 14 drives the data generator 11 and the modulator 12 to sequentially compress the continuous one signal to generate nine signals and transmit them to the receiving side 30.

When the demodulator 31 on the receiving side 30 receives the segment 3g1, it continues the automatic equalization adjustment tube based on this segment 3 signal. While the segmenter 2 signal has two data 4 inds (phase), the segment 3 signal has two or more data 4 inds (phase).
That is, the same number of data I India in the actual raw Te-l (
For example, according to the CCITT standard, 96 OObps is 16
/in), 7200 bpm, 4800 bps and 2400 bpm are 8 points) 6, so equalize *
The principle of operation of equalization adjustment is the same as in segment 2, except that operation a is performed quickly. When the equalization adjustment operation using the segment 3 signal is completed, the sequencer 3
3 drives the signal transmitter 32 to generate all data transmission routine start signals (hereinafter referred to as data request signals) and sends them back to the transmission side 10.

The signal detector 13 on the transmitting side 10 checks the reception of the data request signal during segment signal transmission, and upon receiving the data request signal, notifies the sequencer 14 and proceeds to the normal data transmission routine. If a data request signal is not received within a predetermined period, it is determined that equalization adjustment is impossible and data transmission at that transmission speed is abandoned, which is the same as the Geitance of segment 2. It is unlikely that Trayung will fail at this stage.

In this case, the 8@g2 request signal is the frequency of the gap in the spectrum of the segment 1 signal, the S@g3 request signal is the frequency of the gap in the spectrum of the segment 2 signal, and the data request signal is the frequency of the gap in the spectrum of the segment 3 signal. All you have to do is select them and send them back. The spectrum of the training signal (500-2900 Hz)
Frequencies just outside of may also be used. Further, although this embodiment has been described assuming a half-duplex communication system, a full-duplex communication system may of course be used, and in that case, there is no spectrum restriction on the return signal of the 8.g2 request signal. If the modulation method allows a packed word channel even in a half-duplex communication method, it may be used.

Note that the iJ adjustment sequence of the equalizer of segment 3 (the part surrounded by the dotted line in Figure 6) is performed using the sequence of segment 2. , when the deterioration of the line is not so large, automatic equalization adjustment is performed well in the sequence of segment 2, so in such a case, segment 4 can be omitted.

Also, from a practical standpoint, 9600 bpm
, 7200 kpm, 4800 bps, 2
At a transmission speed of 400 bps, it is orthodox to perform modem training in accordance with the CCITY recommendations.
The present invention is suitable for use with high speed modems, such as 12 Kbpa.

It is considered suitable to apply 14.4 Kbpi K.

As explained above, the present gA has the following effects because the time for preparing for data reception during training is not fixed as in the conventional case.

(1) As soon as the training sequence for each segment is completed, the training sequence for the next segment is started, so the training time can be shortened as much as possible.

(2) Since the automatic equalization adjustment operation can be started immediately after the timing and carrier regeneration operations are completed, more time can be spent on the automatic equalization adjustment, which has the greatest effect on whether or not the training is successful. As a result, data can be transmitted at the highest possible transmission speed given the transmission characteristics of the line.

(3) As a result, the data transmission time, that is, the transmission information can be increased.

[Brief explanation of the drawing]

Fig. 1 is an operating waveform diagram of the automatic equalizer, Fig. 2 is a professional diagram showing the basic configuration of the automatic equalizer, Fig. 3 is an explanatory diagram showing the relationship between frequency 1141 and modem training time, Figure 4 is a block diagram of the modem training method of the present invention, and Figure 5 is a time chart of the modem training method of the present invention.
FIG. 6 is a flow chart of the transmitting side of the modem training method of the present invention. 10... Transmission side, 11... Data generator, 12...
・Modulator, 13... Signal detector, 14. 33-... Sequencer, 20... Transmission line, 30... Receiving side, 3
DESCRIPTION OF SYMBOLS 1... Demodulator (including automatic equalizer), 32... Signal transmitter, F... R, delay line with tag, D1 to Dn... Delay line of delay time T, y0 ~Mm... Multiplier, Σ... Synthesizer. Figure 17 Figure 2 Figure 3 Figure 8-Line Special Note-Good

Claims (1)

[Claims] In a modem training method that is performed prior to transmitting data in a data transmission system using a modem, the modem training sequence is divided into a plurality of segments, and the data to be performed by the demodulator for each segment is determined in advance. The receiving side allocates time for preparatory work for reception, and when the preparatory work is completed while receiving the preceding segment signal among the plurality of segments, the receiving side sends a transmission request signal for the subsequent segment to the transmitting side. The modem train/data mode is such that when the transmitting side receives this transmission request signal, it starts transmitting the subsequent segment signal.