JPS5836858B2 - Automatic phase equalization method in FM line - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to an automatic phase equalization method for suppressing phase distortion in an FM line.

In FM lines, it is known that quasi-crosstalk noise is generated due to linear distortion, especially phase distortion, in radio wave propagation paths, waveguide systems, FM signal amplification systems, etc., and degrades signal quality.

This linear distortion can be reduced by adjusting the matching of the waveguide system and the phase equalizer of the amplification system so that no echo occurs in the waveguide system or the amplification system.

However, since fading is often accompanied by multipath propagation paths, especially non-line-of-sight scattering propagation paths, marine propagation paths, and mountain diffraction propagation paths, double to quadruple diversity reception is usually performed. This is extremely uneconomical.

The present invention has been made in view of this point, and detects phase distortion from high-order distorted waves of various pilot signals contained in a baseband signal (demodulated signal), and uses the output to detect phase distortion. The present invention provides an automatic phase equalization method in an FM line that can always maintain a flat overall phase characteristic and suppress quasi-crosstalk noise by automatically controlling the equalizer.

Therefore, according to the present invention, diversity reception for phase distortion correction is no longer necessary, and at the same time, the coherent bandwidth band can be expanded, so it is also possible to increase the number of communication paths.

First, the principle of the present invention will be explained.

Various analyzes have been conducted regarding the transmission characteristics and high-order distortion of FM signals in FM multiplex telephone lines, etc., such as the one by Kuwabara.
"Analysis of the amount of crosstalk in noise-loaded FM or PM waves using the Fourier method" (Journal of the Institute of Electrical Communication Engineers, July 1960)
According to
as a power series, and it is said that the combined waves of second and third orders are usually dominant.

The influence of amplitude characteristics can be removed by a limiter, so if this is excluded, the second-order quasi-crosstalk noise amount is an even-order term in a power series,
The amount of third-order quasi-crosstalk noise is determined by the odd-order terms.

In order to understand this relationship, we will take a cosine wave as an example, and if we consider its high-order distortion, it can be expressed as shown in the following equation, and it can be seen that the second-order distortion is determined by the even-numbered terms, and the third-order distortion is determined by the odd-numbered terms.

Therefore, a phase equalizer for each order is provided before the demodulation means,
By extracting the distorted waves of each order in the baseband signal and controlling the phase equalizer corresponding to each order so as to eliminate the distorted waves, it is possible to correct the phase characteristics of the FM line.

However, even if only 2nd and 3rd order phase equalizers are provided, the 2nd order distortion is caused by 3rd order distortion from higher order terms such as 4th order, 6th order etc.
Since second-order distortion also occurs from higher-order terms such as fifth-order and seventh-order terms, equivalent second-order and third-order distortions generated from higher-order terms are corrected.

Embodiments of the present invention will be described in detail below with reference to the drawings.

In FM multiplex telephone lines, etc., second-order and third-order distortions are usually the determining factors for phase distortion, and as the number of communication channels increases, it is necessary to consider even higher-order distortions.

In this embodiment, a case will be described in which automatic phase equalization is performed for up to fifth-order distortion.

FIG. 1 is a system diagram in this case.

The received intermediate frequency signal applied to terminal 1 is of secondary, tertiary, and 4th order.
Phase equalizers 2, 3, and 4 of the second and fifth orders, that is, phase equalizers 2, 3, and 4, whose phase characteristics with the center frequency as the origin are proportional to the square, cube, fourth, and fifth powers of the frequency, respectively; 5, phase equalization is performed for each order.

Automatic control of the phase characteristics of these phase equalizers 2, 3, 4, and 5 will be described later.

12, 13, 14, 15 are phase equalizers 2, 3, 4, respectively.
FIG. 2 shows an example of the configuration of a phase equalizer and a tap weight generator (within the dotted line frame in FIG. 1), which are tap weight generators for giving desired characteristics to the circuit 5.

In FIG. 2, the second-order and fourth-order phase equalizers 2 and 4 of FIG.
The second and fifth-order phase equalizers 3 and 5 are constructed by cascading transversal filters that are similarly decomposed.

That is, transversal filters 241 and 242
,243,... is αI COSωτ, α2co
It has a phase characteristic of s2ωτ, α3 cos 3ωτ, ..., and the transversal filter 35L352,
353,----... is Asinωτ, β'2 sin
It is constructed to have phase characteristics of 2ωτ, β3Sin3ωτ, . . . .

τ is a unit delay time of a delay element constituting the transversal filter.

The phase characteristic θ (ω) of the transmission line and the phase characteristic φ (1) of the phase equalizer are respectively π2 π4n 3 Since the bow is a constant, it can be excluded, and if the other terms are summarized by sine and cosine terms, the above-mentioned quadratic This means that the 4th, 3rd, and 5th orders are combined, and the following equation is obtained.

However, in reality, low-frequency reference signals with very small amplitudes are superimposed on h2, h3, l14, and h, as will be described later, but they will be omitted here to simplify the explanation.

Therefore, as shown in FIG. 2, h2 is applied from the terminal 27.
The weighting circuit 271 performs weighting according to the above formula for the signal proportional to h4 and the signal proportional to h4 applied from the terminal 47.
272, 273, ... and 2812B2
, 283, . . . and these are added to the adder 26.
Create α by adding 1, 262, 263, . . ., Betzel function generator 251, 252,
253, . . . to determine the tap weights of the transversal filters 241, 242, 243, . . . , thereby obtaining a phase characteristic of α cos nωτ.

Similar calculations can be performed for the h3 signal applied from terminal 37 and the bulk signal applied from terminal 57 (5
β by.

Sin one τ? Phase characteristics can be obtained.

In addition, 381, 382, 383, ... and 391
, 392, 393, ... are weight circuits,
371, 372, 373, ... are adders, 3
61, 362, 363, . . . are Betzel function generators.

A specific example when n = 1 is as follows.

In other words, the phase characteristic in the case of n=1 is φ(o)=−(4h2 +8h4(π2−6
) ) cos ωr+( 2 h3( yr' −
6) +2h5{? -2o (d-6 blade) 8nωτ
Become the commander.

-4 for weight circuit 271, -8 for 281 (7I2-6),
3 8 1 Te2 (yr2-6), 39
1 Te2 {π4-20(J-6)) weighting is performed, and the weight of the first term in the above equation is applied to the adder 37 by the adder 261.
1, the weight of the second term in the above equation can be found.

The Betzel function generator and transversal filter for each order have the same basic structure, and the unit delay time of the delay element only becomes nτ depending on the value of n. For example, when n = 1. I will explain about it.

The relationship between the Betzel function generator (25L361 in this example) and the transversal filter (241,351 in this example) is as shown in Figure 3, and in Figure 3(a), α1 c
The cosine characteristic of osωτ is obtained, and in the same figure (b), β! ,s
A sine characteristic of inωτ is obtained.

This relationship is clear from the transformation of the transfer function as shown below.

The exponential part of the above equation represents the amount of phase shift, so when the delay element Q unit delay time is τ, 2m and (2m+
1) is the tap number from the center tap, and the Betzel function value represents the tap weight.

Therefore, a transversal filter can be constructed immediately from these equations, and when obtaining cosine characteristics, the third
As shown in Figure a, delay elements 121, 122,...
・Weighting circuits 131, 132, ・128 tap outputs each have weighting circuits 131, 132,
After weighting the Bessel function values in step 139, the outputs of odd-numbered taps counting from the center tap are added in adder 140, and the phase is shifted by π/2 radians in phase shifter 141. After that, adder 1 leaves the even number as it is.
Just add 42.

In addition, when obtaining a sine characteristic, as shown in FIG. 3b, weighting circuits 161, 162, . . . 16
The Betzel function values may be weighted in step 9 and then added in adder 143.

Depending on how the center frequency and ωτ are selected, the characteristics of cosine and sine and the relationships shown in FIG. Of course, 3τ may be selected for the transversal filters 243 and 353.

When transversal filters are connected in cascade in multiple stages in this way, as is well known, the convolution of the tap weight outputs of each transversal filter is calculated.
It is also possible to integrate a transversal filter.

Furthermore, in the Fourier expansion formula, the Fourier coefficients of the second-order and fourth-order terms decrease by 1/n2 when n is large, so it is not necessary to use a multi-stage transversal filter until n is large. But the third order,
Since the Fourier coefficient of the fifth-order term decreases by 1/n, a fairly multi-stage transversal filter is required, so the following modification can be considered.

This transformation is effective not only for 3rd and 5th orders, but also for further simplification of 2nd and 4th orders, so regarding this method,
A summary will be given below.

As mentioned above, the 2nd, 3rd, 4th, and 5th order Fourier expansion coefficients are as follows, and the coefficients for each large n are also shown, so with these filters, By taking charge of the characteristics where n is large, the transversal filter itself can be reduced to the value of the original coefficient and the difference thereof.

Therefore, from equations (7) and (15), we get the following.

(Excluding the stationary term) Equation (16) can be further combined into sine terms and cosine terms, as was done in equation (8), so the equalizer configuration can be changed to the one shown in Figure 2. It will look like Figure 4.

In Figure 4, 240, 241, 242, 2 4
3,---゜-゜ are γ4cos T and γlc, respectively
OsG)Tt 7'2CO82G)T, γ3 cos
Phase equalizer with characteristics of 3ωτ, 35
0 , 351 , 352, 353, ... - - are respectively δ1Sln-ti2 δIS1nωτ, δ2S1n2ωτ, δ3 sin 3
It is a phase equalizer with a characteristic of ωτ2°, 250,
251,252,253,...,360,3
61, 362, 363, . . . are Betzel function generators which are weights for providing the respective phase characteristics.

3 1 0 and 320 are weight generators that generate the coefficients An and Bn of equation (19), respectively, and 330 and 340 are weight generators that generate the coefficients An and Bn of equation (19), respectively.
9) Coefficients Cn and D of Eq.

This is a weight generator that generates .

The intermediate frequency signal applied to the input terminal 1 in FIG. , F after Limittsuta 6
It is demodulated into a baseband signal by the M demodulator 7 and sent to the output terminal 8.
is output to.

If there is nonlinearity in the phase characteristics of the FM signal transmission system including the phase equalizers 2, 3, 4.5, coupled waves between various signals in the baseband signal will occur.

From these combined waves, select the combined waves of 2nd, 3rd, 4th, and 5th combined waves of pilot signals that occur when there are two or more pilot signals and have different frequencies, and Combined wave extraction band filter 2 for each order
Extract at 0, 30, 40, 50.

The combined waves of the outputs of these bandpass filters 20 , 30 , 40 , 50 are sent to adders 26 , 36 , 46 .
The rectangular wave (reference signal) from the rectangular wave generator 9 superimposed on the tap weight control voltage through 56 slightly changes the phase characteristics of the phase equalizers 2, 3, and 4.5, respectively. The combined wave amplitude determined by the phase nonlinearity of the FM signal transmission system including the phase equalizers 2 and 3.45 is taken as the average amplitude, and is subjected to minute amplitude modulation by this rectangular wave.

These situations will be explained with reference to FIG.

Figure 5 shows the time waveforms of various parts of a second-order distortion control loop that controls, for example, a second-order phase characteristic equalizer (the situation is almost the same in other higher-order distortion control loops); Vessel 2,
If the coefficient of the second-order component of the phase characteristic of the FM signal transmission system including 3.45 is positive, that is, from equations (5) and (6), f
If 2+h2>0, b is negative, i.e. f2+h
If 2〉0, C is zero, i.e. f2+h2-0
The case is shown below.

In other words, this corresponds to the case where C is completely equalized.

In each case of a, b, and c, S1 is the output waveform of the rectangular wave generator 9 that is added as a reference signal with low frequency and minute amplitude, and S2 is the second-order coefficient in the overall characteristic, that is, f2 + h2 + reference signal, which is the actual distortion. The secondary coupled wave level as a precept is proportional to its absolute value.

This is clear from equations (3) and (4).

Therefore, the extraction level of the secondary coupled wave (the output waveform of the envelope detector 21, 31, 41, 5 for the 3rd, 4th, and 5th orders)
1) becomes like S3, and the DC blocking circuit 22 (3rd, 4th order)
For the next and fifth orders, the output of 32, 42.52) is S4
become that way.

As is clear from comparing the waveforms 81 and 84 in FIG. 5, the relationship between the reference number S1 and the detected waveform S4 is f2+h2>0
In the case of a, the phase is in-phase, in the case of b in f2+h2<0, it is in reverse phase, and in the case of C in f2+h2-0, it is zero.

Therefore, the phase detector 23 (for 3rd, 4th, and 5th order
Positive, negative, and zero voltages are obtained at the outputs of 3.43.53), respectively.

These polarity voltages are transferred to the polarity inversion circuit 24 (third, fourth,
The polarity is inverted through the integrator 25 (35, 4 for the 3rd, 4th, and 5th orders).
5.55) and the adder 26 (36, 46.56 for the 3rd, 4th, and 5th orders) and is automatically controlled so that the combined wave level is reduced.

The integrator 25 is for stabilizing the operation of the feedback system, and may be replaced by a low-pass filter.

The operations of the third-order, fourth-order, and fifth-order control systems are basically the same as in the second-order case.

However, as described above, the fifth-order phase distortion produces a fifth-order coupled wave and a third-order coupled wave, and the fourth-order phase distortion produces a fourth-order coupled wave and a second-order coupled wave.

Therefore, at first, the equalizers of each order start operating at the same time,
The operation of the 5th and 4th order distortion control loops takes priority, and the 3rd and 2nd order distortion control loops operate accordingly.

In this embodiment, we have explained the case of automatically equalizing and correcting distortion up to the 5th order, but it is usually sufficient up to the 3rd order, and even if it is necessary to equalize up to a higher order distortion, it is easy to use this embodiment. It can be expanded to.

If it is sufficient to equalize only the second and third order precepts, simplification as shown in FIG. 6 is also possible.

In FIG. 6, the same numbers as in FIG. 1 are the same circuit elements, 19 is a sine wave generator, and 18 is a π/2 radian phase shifter.

As is clear from the waveform diagram of FIG. 4, even if the rectangular wave generator 9 in FIG. 1 is changed to a sine wave generator, the entire control loop will still operate.

Therefore, in Figure 6, if a sine wave generator is used and the second-order distortion control loop and third-order distortion control loop are operated in orthogonal relation with respect to this sine wave, the second-order combined wave and the third-order combined wave will be the same. Even if the frequency bands that occur in the frequency range are selected, the control loops can be operated independently of each other.

For this reason, in FIG. 6, the coupled wave extraction filter 20, envelope detector 21, and DC blocking circuit 22 are shared by the two control loops.

In other words, in this way, even if the frequency band in which the secondary combined wave and the tertiary combined wave occur at the same frequency as the combined wave extraction frequency can be used to detect the combined wave, the combined wave Selection of extraction frequency becomes easy.

Further, a configuration as shown in FIG. 7 is also possible.

In the case of FIG. 7, the difference from FIG. 6 is that separate sine wave generators 29 and 3 are provided for the secondary distortion control loop and the tertiary distortion control loop.
9 and drive them at different frequencies.

Since these frequencies are different, bandpass filters 28 . 38 is inserted before the phase detector, eliminating the need for a DC blocking circuit.

This configuration can also be applied to the case of correcting higher-order distortion.

As described above in detail, according to the present invention, the influence of selective fading can be avoided in an FM signal transmission system,
It is possible to expand the coherent band, and also equalize the phase distortion caused by reflections and band limitations that occur within the transmission equipment, eliminating the need for diversity reception and making wideband, high-quality lines economical. There are effects that can be achieved.

[Brief explanation of drawings]

FIG. 1 is a system diagram for explaining one embodiment of the present invention, and FIG. 2 is a system diagram for explaining an embodiment of the present invention.
A system diagram of an order phase equalizer and its tap weight control circuit; FIG. 3 is a configuration diagram of a transversal filter with cosine and sine phase characteristics, which is the basis for forming the phase characteristics of each order in the present invention; FIG. is a configuration diagram for reducing the number of stages of the transversal filter in Figure 2, Figure 5 is a waveform diagram to explain the operation of the embodiment in Figure 1, and Figures 6 and 7 are both secondary , is a system diagram showing another embodiment of the present invention in which only third-order distortion is equalized. 1...Input terminal, 2...Second-order phase characteristic equalizer, 3...Third-order phase characteristic equalizer, 4...
...Fourth-order phase characteristic equalizer, 5...Fifth-order phase characteristic equalizer, 6...Limiter, 7...F
M demodulator, 8...output terminal, 9...square wave generator, 10...terminal, 12, 13, 14
, 15...Tap weight generator, 18...
・π/2 phase shifter, 19...Sine wave generator, 20
,30,40,50...2nd order, 3rd order, 4th order,
Fifth-order coupled wave extraction bandpass filter, 21 , 31 , 41
, 51... Envelope detector, 22, 32, 42,
52...DC blocking circuit, 23, 33, 43, 5
3.1... Phase detector, 24, 34, 44, 54.
...Polarity inversion circuit, 25, 35, 45, 5
5...Integrator, 26, 36, 46,
56... Adder, 27, 37, 47, 57.
...terminal, 28,38...bandpass filter, 29,39...sine wave generator, 121,1
22, 123, 124, 125, 126, 127, 12
8... Delay element, 13L132, 133, 13
4, 135, 136, 137, 138, 139...
... Weight circuit, 140 ... Adder, 141.
... Phase shifter, 142, 143 ... Adder, 151, 152, 153, 154, 155, 156,
157, 158...delay element, 16L162,
163, 164, 165, 166, 167, 168, 1
69... Weight circuit, 241, 242, 243,
244...Transversal filter, 251
,252,253,254... Betzel function generator, 26L262,263,264...
・Adder, 271, 272, 273, 274, 28
1, 282, 283, 284... weight circuit,
350...Transversal filter, 35i
, 352 , 353 , 354 ... Transversal filter, 360 ... Tap weight generator, 361 , 362 , 363, 364 ...
Betzel function generator, 371, 372, 373, 3
74...Adder, 38L382,383 3
84, 391, 392, 393, 394... Weight circuit, 330, 331, 332, 333...
- Adder, 340, 341, 342, 343. 3
81 3, 3823. 3833...Weight circuit.

Claims (1)

[Claims] 1. A variable phase equalizer that can control coefficients for each required order among the orders of the power series characteristic is provided in front of the demodulation means for the FM signal transmitted through the transmission line, and the baseband signal obtained by the demodulation means is provided. The high-order distorted wave of the pilot signal contained in the pilot signal is extracted for each desired order, and the coefficients of the variable phase equalizer are slightly changed using a low-frequency reference signal, and the pilot signal generated by the change is Find the amplitude change in the envelope amplitude of the high-order distortion wave of the signal for each order, detect the correction direction and correction amount of the coefficient of the variable-potential equalizer for each order from the amplitude change, and calculate the detection result. By controlling the coefficients of each order of the variable phase equalizer based on the above, the frequency versus phase characteristic of the transmission system including the variable phase equalizer and the transmission line is configured to be linear. An automatic phase equalization method in an FM line characterized by: