JPH0586082B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a nonlinear distortion removal circuit, and particularly to a nonlinear distortion removal circuit using a traveling wave tube amplifier (hereinafter abbreviated as TWT).

Microwave band digital communications, whether satellite or terrestrial, are important from the perspective of effective use of the frequency band.
It will be mandatory to operate with a higher density transmission method.

That is, it was published on pages 48.4.1 to 48.4.6 of the 1979 [International Conference on Communications] (ICC'79) conference record. Characteristics of (a) High Capacity (Capacity) 16
QAM Digital
Radio System On A Multipath
"Fading Channel" and "Day 1" on pages 35.4.1 to 35.4.3 of the 1979 National Telecommunications Conference (NTC'79) conference record. Distortion Analysis Off 64 QAM As you can see, multilevel quadrature amplitude modulation (QAM) will be used.At this time, the problem arises. is the nonlinear distortion of the transmission amplifier TWT, and the QAM signal is distorted by this distortion.The nonlinear distortion of the TWT is
They differ slightly depending on the situation, but they form a single category. In other words, amplitude saturation characteristics (AM/AM conversion) and output phase rotation θ(x) corresponding to input level x
It is characterized by its characteristics (AM/PM conversion). Therefore, it is generally possible to compensate for this type of distortion to a considerable extent with relatively simple circuitry.

Now, if we consider the case where band limiting of the transmitted signal is not performed in front of the TWT, we can separate the nonlinear effects from the band limiting effects, so we can accurately observe this distortion on both the transmitting and receiving sides. can do.

Based on this idea, various preset type nonlinear distortion compensation circuits have been proposed.
The automatic tracking circuit that automatically guides this circuit to the most desirable operating state is described in "Study of TWT nonlinear compensation using automatic tracking type complex synthesis hybridization" in material CS78-201 of the Communication Systems Study Group of the Institute of Electronics and Communication Engineers. Just look at the precedent. This example was developed for microwave band SSB communication, so it is not very suitable for digital transmission.

An object of the present invention is to provide a nonlinear distortion removal circuit suitable for digital transmission.

According to this invention, a nonlinear distortion correction circuit that passes an input signal through an odd function input/output characteristic circuit and a variable complex coefficient circuit and adds it to the input signal to generate a complex composite distortion characteristic is placed before or after the nonlinear element. , in a method for canceling nonlinear distortion, a change extraction circuit extracts a change from an original value of the input signal after passing through the nonlinear distortion correction circuit and the nonlinear element; An oscillator that provides perturbation, and a correlator that correlates the output of the change extraction circuit with the output of the oscillator, and by changing the variable complex coefficient in a direction opposite to the output polarity of the correlator, A nonlinear distortion removal path characterized in that distortion is canceled by the nonlinear distortion correction circuit is obtained.

Next, the present invention will be explained in detail with reference to the drawings.

Figure 1 shows a block diagram of a conventionally commonly used nonlinear distortion correction circuit, and Figure 2 shows a block diagram of a conventional nonlinear distortion correction circuit.
FIG. 3 is an explanatory diagram of the operation of the circuit shown in the figure.

FIG. 1 shows an odd function input/output characteristic circuit 10 (for example, 3
(multiplicative nonlinear element), a variable phase shifter 11, a variable attenuator 12, and an adder 13 forming a variable complex coefficient circuit.

Let the input x to the input terminal 100 be the vector 200 in FIG. The output of the variable attenuator 12 is changed into vectors 201 and 20, for example, by the phase shift amount φ [rad] of the phase shifter 11, as shown in the vector 208 in FIG.
2 and 203. Since the output γ of the adder 13 is the vector sum 207 of the vector 200 and the vector 208, each vector 20
4,205 and 206. The relative length of the vector 208 with respect to the vector 200 depends on the characteristic f(x) of the odd function input/output characteristic circuit 10, but generally f(x)={ax+bx ³ +cx ⁵ ...}α
If it has the form (b, c≠0), the vector 208 will relatively extend as the input x becomes larger. Therefore, the input/output phase difference θa increases, and the relative output amplitude increases in the range of 0<φ<π/2. FIG. 3 shows the input/output characteristics of the circuit shown in FIG. 1, where a curve 301 shows the amplitude characteristics and a curve 302 shows the phase characteristics. This characteristic is the inverse of the input/output characteristic of TWT.

The problem is how to choose parameters α and φ so that arbitrary
The question is whether it is possible to approximate the inverse characteristic of the TWT characteristic.

Now, suppose that for an input x, the nonlinear distortion of the TWT outputs f(x) as follows: f(x)=x+η·|x| ² ·x. This is the actual
It is a good approximation of TWT nonlinear distortion. Here, η is a complex number η=α′・e ^j 〓′. Also, the second term of f(x) is the first term
Compared to the term, normally |x|》|η・|x| ²・x|

When ^a digital signal X is input to the TWT, its output is as described above ^. α・e ^j 〓) When passed through a nonlinear distortion correction circuit with the following characteristics, its output C ₀ is Co=X+η|X| ²・X+ζ{(X+η|X| ²・
X) ²・(X+η｜X｜ ²・X)} 〓X+η｜X ^{｜ 2 ・}X+ζ｜X｜ ²・X ₌ error E for
becomes E=C ₀ −X=(η+ζ) |X| ²・X. From this formula, by setting η=-ζ, E
=0.

Therefore, let's rewrite ζ as follows.

ζ≡−η±ζd ζd is a parameter that indicates how much ζ currently deviates from the optimal value. Since this error E is usually a complex number, its absolute value |E| is minimized when optimization is performed. As mentioned above, when ζd=0, |E|=0, and otherwise |E|≠0, so in the vicinity of ζd=0, the characteristics are as shown in FIG. 6a. If ζd can be directly observed, η=−ζ±ζd
By setting , E can be minimized. However, we can only observe |E|. At this time, if we can somehow find out whether ζd currently exists on the right or left side of aζd=0 in Figure 6, we can control η by eliminating it. The principle is simple. That is, when it is on the right side, δ|E|/δζd>0, and conversely, when it is on the left side, δ|E|/δζd<0. In order to find the polarity of this differential coefficient, we can actually add a perturbation △ζd to ζ and directly check whether the resulting change in |E| increases or decreases.

Normally, Δζd cannot be made large enough to cause disturbance to the communication signal itself, so it becomes a very small signal.

Therefore, △|E|/△ζd is also a very small signal and is buried in noise. For this purpose, it is necessary to add Δζd over and over again and cumulatively detect the resulting changes in |E|. For that, △ζd and △
An effective method is to observe the correlation with |E|. That is, if the correlation value is R, we will find R such that R = △||
It becomes the polarity of Because sign(△|E|／△ζd)=sign(△|E|・△ζd)
This is because. As perturbation △ζd, ζ ₀・sinω ₀ t
Using the sine wave (shown in Figure 6b), we get |E|
Due to this perturbation, approximately |E|〓|E ₀ |+
It changes in the form sign(ζd)・β・sinω ₀ t (Figure 6b). Here, β is a constant introduced to satisfy the approximate expression.

To find R in the above equation, R=1/2To∫ ^TO _-TO |E|・ζ ₀・sinω ₀ tdt =ζ ₀ /2To∫ ^TO _-TO (|E ₀ |+sign(ζd)・βsin
ω ₀ t) sinω ₀ tdt = ζ ₀ /2To2Tosign(ζd)β・π (However, To》 2π／ω ₀ ) = sign(ζd)ζ ₀ π・β Therefore, from the above equation, the polarity of R is the polarity of ζd sign (ζd)
It turns out that it is the same as This situation is shown in FIG. If ζd reaches the optimal value ζopt as shown in FIG. 6b, R=0. Therefore, to control ζ, it is sufficient to increase or decrease it in the opposite direction to the polarity of the correlation value R determined by the above equation. That is, dζ/dt=α・R: α=minimal coefficient Integrating both sides of the above equation by t gives ζ=−α∫Rdt =−α∫1/2To∫｜E｜・ζ ₀・sinω ₀ tdt・dt Here ζ is a double integral, in which R
Since the integral related to is simply a process of extracting an average value for obtaining a correlation value, the effect does not change even if it is substituted with an integral attached to the outside. Therefore, ζ may be controlled in the following manner.

ζ=−α∫|E|・ζ ₀・sinω ₀ t・dt The above is the principle of the present invention.

FIG. 4 shows an equivalent baseband block diagram of one embodiment of the present invention, and is a concrete example of the above principle.
Block 1 in the figure is the same as the nonlinear distortion correction circuit shown in FIG. In this embodiment, for simplicity, it is assumed that the variable phase shifter 11 has already been optimally set, and the attenuation amount ζ of the remaining variable attenuator 12 is controlled. Block 4
is a nonlinear element such as TWT.

Block 2 is a change extraction circuit that detects the degree E of nonlinear distortion of the input signal. Specifically, the signal discriminator 20 detects four signal points 1+j, -1+j, -1-j, 1 for a four-phase phase modulation signal. −j, a device that outputs the four points closest to the input of the signal discriminator as discrimination values. Specifically, when the input signal is in the first, second, third, or fourth quadrant,
Correspondingly, each of the previous four square values is output. The subtracter 21 detects the difference (X-X^) between the identification value X^ of the output of the signal discriminator 20, which is normally equal to the original signal, and the input signal
and its absolute value |E| is obtained. |E| can be reduced by appropriately selecting the value of the variable attenuator 12 of the nonlinear distortion correction circuit 1. 31 is an oscillator that generates perturbation ζ ₀ sinω ₀ t, 33 is an integrator for control, 34 is a polarity inverting circuit, and each minute adder 32
is for adding a perturbation signal. 30 is a correlator for determining the correlation value R between the perturbation signal and the output of the change extraction circuit, and is composed of a multiplier 300 and an integrator 301. However, this integrator 301 may not be provided.

In this embodiment, the switch 38 has two operating modes. First, when the switch 38 is turned to the a side, the operation according to the principle explained earlier is performed. If a large perturbation signal continues to be applied after the control is stabilized and E is minimized, the jitter due to the perturbation will not be removed from the output of the nonlinear distortion removal circuit of the present invention. Therefore, in order to avoid this, the switch 38 is switched to (b). When tilted to the (b) side, the amount of perturbation ζ ₀
will change in proportion to the average value of |E|. ζ ₀ is given by the attenuation amount δ of the low frequency filter 36 and the attenuator 35, and the multiplier 37 in the form ζ ₀ =δ·||. This allows control to proceed |E
When |→0, the amount of perturbation decreases accordingly, and no disturbance is caused to the signal inadvertently. The relationship between the main elements of FIG. 4 and the principle of the present invention can be summarized as follows: In block 2, the absolute value of E is determined, and the correlation between the change and the perturbation signal (oscillator 31) is multiplied by multiplier 300. The average value is taken by the integrator 31 and added to the control integrator 33 one after another to generate a control signal (ζ=-α∫|E|・ζ ₀・sinω ₀ t・dt)
That is.

FIG. 5 shows a block diagram of another embodiment of the present invention, in which the nonlinear distortion correction circuit includes five
It corrects up to the third-order distortion, and the parameters to be controlled are a variable attenuator 12 for third-order distortion, and a phase shifter 1.
1. Variable attenuator 15 for fifth-order distortion, phase shifter 1
6 of 4. For this purpose, four of the same control circuits 3 as shown in FIG. 4 are used. That is, they are 3, 3', 3'', and 3. However, since the output of the change extraction circuit 2 is used in common, the output signals of the oscillator 31 in the block 3 must be made into four orthogonal functions. In other words, it is necessary to optimize each of the four independent parameters based on changes in one error signal E. To do this, each perturbation signal is For example, {sinω ₀ t, cosω ₀ t, sin2ω ₀ t, ccs2ω ₀ t} or {sinω ₀ t, sin2ω ₀ t, sin3ω ₀ t , sin4ω ₀ t}, etc.

Similarly, by increasing the number of control circuits, higher-order nonlinear compensation can be performed automatically and easily. That is, the control circuits 3, 3', 3'', 3
In order to increase . good.

As explained above, according to the present invention, nonlinear distortion caused by TWT or the like can be automatically and accurately compensated for using a high-order nonlinear compensation circuit.

[Brief explanation of the drawing]

Fig. 1 is a block diagram showing an example of a conventional nonlinear distortion correction circuit, Fig. 2 is an explanatory diagram of the operation of the circuit shown in Fig. 1, Fig. 3 is an input/output characteristic diagram of the circuit shown in Fig.
FIG. 5 is a block diagram of an embodiment of the present invention, and FIGS. 6a, b, and 7 are diagrams for explaining the principle of the present invention. 1... Nonlinear distortion correction circuit, 2... Change extraction circuit, 3... Control circuit, 4... TWT, 30... Correlator, 31... Oscillator.

Claims (1)

[Claims] 1. A nonlinear distortion correction circuit that passes an input signal through an odd function input/output characteristic circuit and a variable complex coefficient circuit and adds it to the input signal to generate a complex composite distortion characteristic is placed before or after the nonlinear element to eliminate nonlinear distortion. In the canceling method, a change extraction circuit extracts a change from the original value of the input signal after passing through the nonlinear distortion correction circuit and the nonlinear element, and an oscillator that perturbs the variable complex coefficient. and,
a correlator that takes a correlation between the change extraction circuit output and the oscillator output, and the nonlinear distortion correction circuit corrects the distortion of the nonlinear element by changing the variable complex coefficient in a direction opposite to the correlator output polarity. A nonlinear distortion removal circuit characterized in that the circuit cancels out the distortion by