JP3144649B2 - Distortion compensated quadrature modulator - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a distortion-compensating quadrature modulator, and more particularly to a distortion-compensating quadrature modulator for outputting a quadrature modulated wave with high accuracy.

[0002]

2. Description of the Related Art Multi-level PSK (Phase Shift Keying), QAM
(Quadrature Amplitude Modulation) or the like is used.

[0003] In digital modulation / demodulation, since waveform transmission is performed, a modulator with higher precision than analog modulation / demodulation is required. For digital modulation / demodulation, a quadrature modulator using the in-phase signal amplitude and the quadrature signal amplitude as baseband signal inputs is used, but has the following problems.

(I) Even if the positive / negative of the baseband input signal, that is, the offset level is balanced, it is difficult to balance as an effective input to the quadrature modulator, so that a carrier leak component is superimposed on the modulated wave.

(Ii) Even if the amplitudes of the two baseband input signals are balanced, it is difficult to balance them as an effective input to the quadrature modulator, so that an image component is superimposed on the modulated wave.

(Iii) Two carrier signals having a phase shift of 0 and π / 2 are input to the quadrature modulator, but it is difficult to accurately manufacture a phase shifter for generating a π / 2 carrier signal on an IC. . If the orthogonality is not accurate, an image component is superimposed on the modulated wave.

(Iv) When an IC is used, the above characteristics change due to fluctuations in power supply and temperature.

Conventionally, it has been necessary to adjust a skilled technician in order to solve the above-mentioned problems. In addition, aging was corrected by regular inspection.

Regarding the automation of distortion compensation, two simple construction methods have been conventionally proposed.

The first proposal is disclosed in Japanese Patent Application No. 62-13143, and its schematic configuration is shown in FIG. This is a configuration for compensating for distortion when the 90 degrees are incomplete.

To obtain a modulated wave in a narrow band, a quadrature modulator as shown in FIG. 1 is generally used. This quadrature modulator has input terminals 40 and 41, an output terminal 42, a mixer 4
3, 44, 90 degree phase shifter 45, oscillator 46, synthesizer 4
7. It is composed of a signal point position discriminator 48 and an amplitude calculator 49. In this configuration, the signal point position identifier 48 identifies a signal point composed of the in-phase amplitude input I (t) and the quadrature amplitude input Q (t), and the amplitude calculator 49 obtains the amplitude at the time of identification. Originally, the amplitudes of the in-phase amplitude input I (t) and the quadrature amplitude input Q (t) are equal, but if the phase shift is shifted by 90 degrees, the amplitudes at the two time points do not match. Therefore, the difference or ratio of the amplitude is obtained by calculation, and the phase shifter 45 is adjusted. In this method, (i) the phase shifter 45 is adjusted.
It is difficult to realize a phase shifter that can be fine-tuned with an actual circuit. (Ii) The feasibility is unknown because a specific control method is not described. (Iii) An error occurs due to roll-off shaping of the baseband signal. There are drawbacks such as insufficient accuracy and (iv) other distortions such as DC offset and amplitude imbalance resulting in insufficient accuracy.

The second proposal is disclosed in Japanese Patent Application No. 63-62439, and its schematic configuration is shown in FIG. This is to compensate for DC offset. In this configuration, for each of the in-phase and quadrature components of the modulated wave,
DC according to the sign of the modulation symbol detected by the comparator 51
The modulated wave is rectified while synchronizing the SW of the offset calculation unit 50, and the integrated value is compared. Since the integral value is proportional to the DC offset amount, compensation can be made by forming the feedback loop 52. In this conventional example, compensation is performed for each of the in-phase and quadrature component circuits. However, in an actual high-frequency circuit, local carriers leak and are also superimposed on the in-phase and quadrature component combiner 47 to generate distortion. There is a disadvantage that it is not enough to remove the distortion in front of the vessel 47. Further, this circuit has a disadvantage that the circuit configuration is considerably complicated.

In general, in these conventional modulators, a modulated wave may be deteriorated due to linear distortion and nonlinear distortion. Non-linear distortion is caused by the non-linear response of the diodes used in the mixers 43 and 44, and can be easily reduced by lowering the input level to some extent. Required. Hereinafter, the linear distortion will be specifically described.

In FIG. 12, input terminals 40 and 41 have an in-phase amplitude input I (t) and a quadrature amplitude input Q, respectively.
(T) is input. On the other hand, the carrier r _c of each frequency omega _c output from the oscillator 46 (t) = cos (ω c t)
And r obtained by shifting the phase of the carrier by π / 2 with a 90-degree phase shifter.
_{s (t) = - sin (} ω c t) , respectively mixer 43,4
4, the in-phase amplitude signal a (t) and the quadrature amplitude signal b
(T). This multiplied signal is combined with the combiner 47
And output from the output terminal 42. However, in an actual quadrature modulator, carrier leakage due to stray capacitance, stray inductance, or the like, image generation due to a phase shift error of the 90-degree phase shifter 45, and the like occur. Therefore, the modulator output y (t) is as follows.

Y (t) = y ₁ (t) + y ₂ (t) + y ₃ (t) + y ₄ (t) (1) Next, y ₁ (t) is a DC offset δ _c1 to the baseband signal in the mixer 43. Is added, and is shown as follows.

[0016] _{y 1 (t) = [I} (t) + δ c1] cos (ω c t) (2) y 2 (t) becomes the imbalance of α times the gain of the mixer 44 is phase, DC offset This is a response when the baseband signal when δ _s1 is added is multiplied by a quadrature carrier having a phase shift error θ, and is represented as follows.

[0017] _{y 2 (t) = [-} αQ (t) + δ s1] sin (ω c t + θ) (3) y 3 (t) is the carrier leak of the in-phase component, the size of the carrier leak [delta] _c2, Carrier leak phase θ ₁
And the response is as follows:

[0018] _{y 3 (t) = δ c2} cos (ω c t + θ 1) (4) y 4 (t) is the carrier leak of the quadrature components, the size of the carrier leak [delta] _s2, the phase of the carrier leak theta _Two
And the response is as follows:

[0019] _{y 4 (t) = δ s2} sin (ω c t + θ 2) (5) When thus calculating the modulator output y (t), y (t ) = c (t) cos (ω c t) - _{d (t) sin (ω c} t) (6) , and the phase modulated signal c (t) is, c (t) = I ( t) + δ c sin θ [-αQ (t) + δ s1] and (7) And the quadrature modulated signal d (t) is: d (t) = − αQ (t) cos θ + cos θ · δ _s1 + δ _s (8) The offsets δ _c and δ _s are δ _c = δ _c1 + δ _c2 cos θ ₂ + δ _s2 sin θ ₂ (9) δ _s = −δ _c2 sin θ ₁ + δ _s2 cos θ ₂ (10)

[0020]

However, the output value of the quadrature modulator shown above is an ideal output c (t) = in which the in-phase modulation signal and the in-phase amplitude input and the quadrature modulation signal and the quadrature amplitude input are equal. It is significantly different from a (t) and d (t) = b (t).

FIG. 14 shows a signal space diagram of the modulated wave. In the ideal response when x = 0 to 2π with I (t) = cos (x) and Q (t) = sin (x), modulated wave c
The trajectory on the phase diagram of (t) and d (t) is shown in FIG.
Is drawn in a circle as shown by the solid line, but if not ideal, the center is shifted from the origin and becomes an oblique ellipse as shown by the broken line in FIG. This figure 1
The oblique ellipse of FIG. 4 is actually a trajectory obtained by subjecting the normal trajectory shown in FIG. 15 (a) to a linear distortion due to the DC offset shown in FIG. 15 (b), and an amplitude imbalance shown in FIG. 15 (c). The trajectory subjected to linear distortion and the trajectory subjected to linear distortion due to imperfect 90 degrees shown in FIG. 15D are superimposed. In order to suppress such linear distortion, DC
Adjustment for reducing offsets δ _c1 and δ _s1 , quadrature carrier phase shift error θ, carrier leaks δ _c2 and δ _s2 ,
Alternatively, it is necessary to select balanced parts. Further, there is a problem that a skilled operation, an adjustment time, and an expensive measuring instrument are required to realize such a thing.

As a proposal for solving these problems of the prior art, there is an invention in which a linear conversion means is introduced for compensating for a distortion due to a DC offset. However, in order to derive a linear parameter, an output of a quadrature modulator is converted to an in-phase component. And an IQ detector that extracts the orthogonal components. In this conventional configuration, an IQ detector that is more accurate than the quadrature modulator is required, and its offset adjustment, amplitude adjustment, and orthogonality adjustment are required.

In each of these conventional examples, various linear distortions are individually dealt with, and in practice, 90 degrees cannot be sufficiently compensated unless the DC offset and amplitude balance are sufficiently compensated. Is to leave.

The present invention has been made in view of the above points, and solves the above-described deterioration of a modulated wave due to linear distortion.
An object of the present invention is to provide a highly accurate distortion-compensated quadrature modulator with a circuit having a simple configuration.

[0025]

FIG. 1 is a diagram for explaining the principle of the present invention. A linear compensation parameter for compensating for linear distortion is set, and the in-phase amplitude signal and the quadrature amplitude signal are linearly converted by the linear conversion means (1) to be a baseband signal and linearly converted by the linear conversion means (1). In a distortion compensation quadrature modulator to which a baseband signal is input, a level signal generating means (3) for removing a carrier component from a modulated wave output of the quadrature modulator (2) to generate a level signal; Parameter generation means (4) for deriving a linear conversion parameter using the value of the level signal input from (3) and setting it in the linear conversion means (1)
And

[0026]

According to the present invention, the in-phase amplitude signal and the quadrature amplitude signal are linearly converted by the linear conversion means (1), and the modulated wave output of the quadrature modulator (2) which receives the linearly converted signal as an input is a level signal generation means ( 3), a level signal is extracted, and a parameter of the linear conversion is derived from the level signal by the parameter generation means (4). By using the linear conversion parameter for the linear conversion, a value having a sufficiently small linear distortion is obtained. Can be suppressed.

In the first embodiment described below, the linear parameter is determined based on the level signal using a known test signal.

In the second embodiment described below, the linear parameter is determined based on the level signal according to the state of the in-phase and quadrature amplitude inputs.

Further, since various types of linear distortions are actually mixed, in the present invention, first, a parameter relating to linear distortion due to DC offset is obtained, and then a parameter relating to linear distortion due to amplitude imbalance is obtained while performing DC offset compensation. Finally, all linear parameters are determined by a procedure of obtaining a parameter relating to linear distortion due to imperfect quadrature phase while compensating for DC offset and amplitude imbalance.

[0030]

FIG. 2 shows the configuration of a distortion-compensating quadrature modulator according to a first embodiment of the present invention. The same components as those in FIG. 12 are denoted by the same reference numerals, and description thereof will be omitted. In FIG. 2, the distortion-compensating quadrature modulator according to the present embodiment includes switches 20, 21, 22, 23 having L and C terminals, a linear conversion unit 1 including a linear converter 24, a quadrature modulator 2, and a control circuit. 25, detector 2
6. A low-pass filter 29. The quadrature modulator 2 includes mixers 43 and 44, a 90-degree phase shifter 45, an oscillator 46, and a combiner 47.

In the configuration shown in FIG. 2, the detector 26 is required to have a square detection characteristic. The square detection characteristic can be realized by lowering the input level in a normal detector. In order to surely obtain the square-law detection characteristic, a modulated wave may be input to the RF terminal and the LO terminal of the double balance mixer. Also,
Digital signal processing is advantageous for the control circuit 25. At this time, the output of the low-pass filter 27 is A / D converted and used.

In the distortion-compensating quadrature modulator having this configuration, the in-phase amplitude signal I (t) and the quadrature amplitude signal Q (t) are input to the linear converter 24, and the in-phase amplitude input a (t) which is a baseband signal. And a quadrature amplitude input b (t). Baseband signals a (t) and b (t) are input to quadrature modulator 50, and modulated wave y represented by equations (6) to (10).
(T) is obtained. The modulated wave y (t) is input to the detector 26, the carrier component is removed by the low-pass filter 27, and the level signal z (t) is obtained. This level signal z
(T) is input to the control circuit 25, a linear conversion parameter is derived, and the value is set in the linear converter 24.

Next, a specific operation of the distortion-compensating quadrature modulator according to the present embodiment will be described. Modulated wave in-phase / quadrature amplitude c (t), d (t) of modulated wave y (t) output from quadrature modulator 50
Should ideally be equal to the in-phase amplitude signal I (t) and the quadrature amplitude signal Q (t), respectively. Therefore, the following equation c (t) = I (t) (11) d (t) = Q (T) (12) is obtained, and when equations (7) and (8) are simultaneously solved, the in-phase amplitude input a (t) of the baseband signal is given by: a (t) = I (t) + tan θ · Q (t) + a (13) Further, the quadrature amplitude input b (t) of the baseband signal is b (t) = (1 / αcos θ) Q (t) + b (14) a = −δ _c tan θ (δ _s + δ _s1 cos θ) −sin θ · δ _s1 (15) b = (cos θ · δ _s1 + δ _s ) / αcos θ (16) Actually, the values of α, θ, a, and b are unknown, and must be obtained by some method. Assuming that the obtained values are linear conversion parameters α ₀ , θ ₀ , a ₀ , and b ₀ , the linear conversion performed by the linear converter 24 is as follows.

The baseband signal a (t) is as follows: a (t) = I (t) + tan θ ₀ · Q (t) + a ₀ (17), and the baseband signal b (t) is b (t) = (1 / α ₀ cos θ ₀ ) Q (t) + b ₀ (18)

Accordingly, as shown in FIG. 3, the linear conversion performed by the linear converter 24 is a signal obtained by multiplying the in-phase amplitude signal I (t) by the quadrature amplitude signal Q (t) by tan θ ₀ obtained from the parameter θ _0. parameters a ₀ and the added baseband signal a (t) is in phase input signals of the quadrature modulator 2, also quadrature amplitude signal Q (t) to the parameter alpha ₀ and θ 1 / α ₀ determined from ₀ cos theta ₀ Are added to the parameter b ₀ and the baseband signal b (t) is used as the quadrature input signal of the quadrature modulator 2.

The level of the modulated wave extracted through the detector 26 and the low-pass filter 27 is represented by z (t) = c ² (t) + d ² (t) (19). Therefore, the switches 22 and 23 of FIG. 2 are connected to the C terminal side, and the common-mode amplitude input a of the baseband signal is
A linear conversion parameter is derived as follows from (t) and the modulated wave level z (t) when the test signal is input from the control circuit 25 as the quadrature amplitude input b (t). Switch 2
The modulated waves c (t) and d (t) when the terminals 2 and 23 are switched to the C terminal side are expressed by the equations (7) and (8). By taking partial derivatives of (t) and b (t) and setting them to 0, the following simultaneous equations are obtained.

[0037]

(Equation 1) The solution of this simultaneous equation regarding a (t) and b (t) coincides with the right side of equations (15) and (16). Accordingly, the test signals T corresponding to the baseband signals a (t) and b (t)
_a, by obtaining the value when the level signal is varied by the control circuit 25 a T _b is minimized, the value is a linear transformation parameters a ₀ and b _0.

Next, the obtained linear conversion parameter a ₀
And b ₀ , a provisional value of the phase shift error θ ₀ = 0, and a provisional value of the amplitude ratio α ₀ = 1 are set in the linear converter 24, and the switches 22,
When the signal 23 is returned to the L terminal side, the linear converter 24 performs a linear conversion operation. When the converted signals a (t) and b (t) are input to the orthogonal converter 50, the output y (t) of the output y (t) is obtained. The modulation signals c (t) and d (t) are as follows: c (t) = I (t) −α sin θQ (t) (22) d (t) = − αcos θQ (t) (23)

Therefore, when the switches 20, 21 are connected to the C terminal side and the test signals T _I , T _Q are inputted from the control circuit 25, T _I (t) = A (predetermined value), T _Q (t ) =
When 0 is input, the level signal z ₁ at that time is z ₁ = A
It becomes ² . When T _I (t) = 0 and T _Q (t) = A are input, the level signal z ₂ at that time is z ₂ = α ² A ²
Becomes The estimated value α ₀ of α from the level signals z ₁ and z ₂ is obtained by α ₀ = (z ₂ / z ₁ ) ^1/2 (24).

However, in this description, it is necessary to provide a condition that the level signal from the detector 26 is accurately proportional to the square of the level of the modulated wave. In practice, a bias component is superimposed on the output, and the bias component is corrected and used. That is, the calibration is performed so that the level signal value when no modulation wave is applied becomes zero.

Further, when the equations (22) and (23) are satisfied, the level signal z (t) of the modulated wave is partially differentiated with the in-phase amplitude signal I (t) by setting the quadrature amplitude signal Q (t) = A. And
If the partial differential value is set to 0,

(Equation 2) Therefore, if the solution is I ₀ , there is a relationship such as θ = arcsin (I ₀ / αA) (26)

Here, when the test signal T _I (t) corresponding to the in-phase amplitude signal _I (t) is changed, the level signal z
The value I ₀ of T _I (t) when (t) is minimized and the linear transformation parameter α ₀ that is the estimated value of α obtained by equation (24) are substituted into equation (26). For example, an estimated value θ ₀ of the phase shift error θ is obtained.

In summary, the determination of the linear conversion parameter in the control circuit 25 is performed as follows according to the flowchart of FIG.

First, in step 101, switch 2
Switch the 2,23 to C side, to test signal T _a, initial setting T _a (t = 0) of _{T b = a 1, T b} (t = 0) = b 1.

Next, in step 102, the values of the test signals T _a and T _b that minimize the level signal z (t) are obtained, and these are calculated as the optimal values a ₀ and b of the linear conversion parameters a and b.
_Set to ₀ .

Next, in step 201, the optimum values a ₀ and b ₀ are set in the linear converter 24, and the provisional values θ ₀ = 0 and α ₀ = 1 of the linear conversion parameters θ and α are set. The switches 22 and 23 are returned to the L side.

Next, at step 202, switch 2
0, 21 are switched to the C side, and the test signal T _I (t) = A,
T _Q (t) = 0 input to the measured level signal z _1, measuring the level signal z ₂ to input test signal _{T I (t) = 0,} T Q (t) = A.

Next, in step 203, the optimum value α ₀ of the linear conversion parameter α is obtained by calculating α ₀ = (z ₂ / z ₁ ) ^1/2 .

Next, in step 301, the optimum values a ₀ , b ₀ , α ₀ are set in the linear converter 24, and the test signals T _I , T _Q are converted to T _I (t) = I ₁ , T _Q (t). = A is set.

Next, in step 302, the value I ₀ of the test signal T _I (t) that minimizes the level signal z (t) is obtained.

Next, at step 303, the optimum value θ ₀ of the linear conversion parameter θ is calculated as follows: θ ₀ = arcsin (I ₀ / α
₀ A) is calculated and obtained.

Finally, in step 400, the optimum values a ₀ , b ₀ , α ₀ , θ ₀ of the determined linear conversion parameters are set.
Is set in the linear converter 24, and the switches 20 and 21 are returned to the L side.

According to the method described above, the control circuit 25
Obtains a linear conversion parameter only by inputting a level value from a level signal generating means including a detector 26 and a low-pass filter 27, and sets the linear conversion parameter in the linear converter 24. Thereby, the linear distortion can be suppressed to a sufficiently small value.

In the method of FIG. 4 described above, the level value z (t)
Has been described, the values of the test signals T _a , T _b , and T _I that minimize the values of the test signals T _a , T _b , and T _I are described. It is easily obtained using a perturbation method using the principle that the sign of the slope changes. A flow chart of procedure for obtaining the values of T _a when fixing the value of T _b as an example of such a perturbation method shown in FIG.

[0055] In the perturbation method of Fig. 5, first, at step 501, the initial value of the T _a to a _1.

Next, in step 502, the following settings are made.

[0057] _{T a (t) = e ·} sin (t) + a n (27) T b (t) = b (28) where 0 <e«1 perturbation amplitude, b is a constant value, a _n is n is the value of T _a in times eyes of perturbation.

Next, at step 503, equation (27)
Updated based the value of a _n in the following equation (29) in the.

[0059]

(Equation 3) Then, in step 504, steps 502 and 5
03 is repeated until n = N. Here, N is the maximum number of perturbations.

In the optimization shown in FIG. 5, a perturbation is given from the outside, but the time required to converge to the optimum value differs depending on the magnitude of the perturbation amplitude. FIG. 6 shows this state. Therefore, it is preferable to select an appropriate perturbation amplitude that shortens the convergence time.

Step 102 in the method of FIG.
In the case of using the perturbation method, without fixing the value of T _b, T _a, may be adjusted alternately both T _b.

The average bit error rate (Av) at the output of the equalizer using the distortion-compensating quadrature modulator according to the first embodiment of the present invention.
erage BER) and the average bit error rate at the output of an equalizer using a conventional quadrature modulator without distortion compensation are plotted for each of DC offset, amplitude balance, and phase offset as shown in FIGS. It is shown in (c).

FIG. 8 shows the configuration of a distortion-compensating quadrature modulator according to a second embodiment of the present invention. The same components as those in FIG. 2 are denoted by the same reference numerals, and description thereof will be omitted.

In the second embodiment, instead of inputting a test signal from the control circuit 25 through the switches 20, 21, 22, 23 as in the first embodiment, an in-phase and quadrature amplitude input I (t ), Q (t) to the input detector 31
And by reporting this to the control circuit 25,
Appropriate in-phase and quadrature amplitude inputs I (t), Q (t)
The linear conversion parameter is determined by using the test signal in the same manner as in the embodiment. For this reason, in the second embodiment, it is possible to determine the linear conversion parameter in real time.

More specifically, the determination of the linear conversion parameter in the second embodiment is performed as follows.

First, a periodic small-perturbation perturbation signal is superimposed on the linear conversion parameter a, and the level signal at that time is subjected to correlation detection using the perturbation signal, and the optimum value a is set so that the correlation detection signal level becomes 0. _The linear conversion parameter a is adjusted to ₀ , and the optimum value b ₀ is similarly adjusted for the linear conversion parameter b to obtain the optimum parameters a ₀ and b ₀ .

Next, the control circuit 25 sets the optimum parameters a ₀ and b ₀ in the linear converter 24, sets the provisional values θ = 0 and α = 1 for the parameters θ and α, and sets the quadrature amplitude signal Q
The value I of the in-phase amplitude signal I (t) at the moment when (t) becomes 0
₁ and the value z ₃ of the level signal z (t) of the level signal generating means
Was measured, also measures the value z ₄ of the I values Q ₂ and z of the moment when the (t) becomes 0 Q (t) (t) , α 0 = (I 1 / Q 2) · (z ₂ / z ₃ ) ^1/2 (30) is calculated to determine the optimum parameter α ₀ .

Next, the control circuit 25 sets the optimal parameters a ₀ , b ₀ and α ₀ in the linear converter 24, and sets the parameters θ
, A provisional value θ = 0 is set, and the value z ₅ of the level signal z (t) at the moment when the in-phase amplitude signal I (t) and the quadrature amplitude signal Q (t) become equal to the value A ₁ is measured. Also, I (t) =-Q
(T) = the value z ₆ of the moment became A ₂ z (t) is measured, to determine the specific ^{_{β = z 1 A 2 2 /}} z 2 A 1 2 from the measured value θ ₀ = arcsin [(1- β) / (1 + β)] (31) to determine the optimal parameter θ ₀ .

Therefore, in the second embodiment, it is sufficient to simply determine the optimum value by adding perturbation to the DC offset, and it is necessary to confirm that the in-phase and quadrature amplitude inputs are in a predetermined state for the amplitude balance and the orthogonality. An optimum value can be calculated from the level signal at the time of detection.

In the above procedure, correlation detection can be performed by the same method as the perturbation method shown in FIG.

Various forms are conceivable for concrete circuit formation of the distortion compensation adjuster described above. The control circuit 25 is advantageously implemented by a digital circuit due to the nature of processing.
There are two types of the linear converter 24, an implementation using an analog circuit and an implementation using a digital circuit.

In order to form the linear converter 24 with an analog circuit, a high-precision multiplier is required. Since the linear parameter is a digital signal, the multiplier has four quadrants D / A
A converter can be used. The four-quadrant D / A converter is formed of a resistor and a switch as shown in FIG. 9, inputs the multiplied signal to the reference input terminal 33, and outputs the multiplied value to the digital terminal 3.
Enter 2 The digital terminal 32 controls a switch, and the output terminals 34 and 35 can obtain a highly accurate analog multiplied output. At the two output terminals, two current signals, for example, a (t) and a @ (bar) (t), which are inverted with respect to a balanced level, for example, analog ground are generated. These are input to the quadrature modulator 2 via the amplifier.
This quadrature modulator 2 is configured as shown in FIG. 10, and its mixer sections 43 and 44 are formed by differential amplifiers as shown in FIG. 11, and therefore require two inverted inputs. of course,
It is also conceivable that the analog baseband signal generated by the linear converter 24 is inverted by an inverting amplifier to generate two signals.

When the digital circuit is realized by a digital circuit, a digital input signal obtained by A / D conversion of an in-phase amplitude signal and a quadrature amplitude signal is input, and a digital baseband output signal output from the linear converter 24 is D / A converted. Input to the quadrature modulator 2.

Since the linear parameters obtained by the above description are reduced when the power is turned off, they are stored in a non-volatile memory, and when the power is turned on, the values are read and used first. Is desirable.

At the time of actual mounting, the linear conversion means and the parameter generation means are integrally formed as ICs by CMOS, and the quadrature modulator and the level signal generation means are bipolar or GaAs.
It is also possible to realize the distortion-compensating quadrature modulator of the present invention by combining these ICs into an integrated IC.

[0076]

As described above, according to the present invention, it is possible to easily derive a linear conversion parameter by providing a circuit for detecting a level, and to obtain a highly accurate distortion by a circuit having a simple configuration. A compensated quadrature modulator can be realized. Further, in a transmission system performing burst transmission, if the linear conversion parameter is corrected during a time when transmission is not being performed by the method of the present invention, compensation can be optimized in real time at the time of transmission, temperature fluctuation of equipment, change over time, and so on. And the like, a highly accurate modulated wave can be obtained.

[Brief description of the drawings]

FIG. 1 is a diagram for explaining the principle of the present invention.

FIG. 2 is a configuration diagram of a distortion compensation quadrature modulator according to a first embodiment of the present invention.

FIG. 3 is a diagram of a linear conversion performed by a linear converter of the distortion-compensating quadrature modulator of FIG. 2;

FIG. 4 is a flowchart of a procedure for determining a linear conversion parameter by the distortion-compensating quadrature modulator of FIG. 2;

FIG. 5 is a flowchart of a perturbation method used in the procedure of FIG.

6 is a graph showing a relationship between a convergence time and a perturbation amplitude in the perturbation method of FIG.

FIG. 7 shows the average bit error rate of the output of the equalizer using the distortion-compensated quadrature modulator of FIG. 2 and the average bit error rate of the output of the equalizer using the conventional quadrature modulator without distortion compensation. It is a graph shown about DC offset, amplitude balance, and phase offset.

FIG. 8 is a configuration diagram of a distortion compensation quadrature modulator according to a second embodiment of the present invention.

FIG. 9 is a configuration diagram of a four-quadrant D / A converter used in a linear converter of a distortion compensation quadrature modulator according to the present invention.

FIG. 10 is a configuration diagram of a quadrature modulation unit of the distortion compensation quadrature modulator according to the present invention.

11 is a configuration diagram of a double balance mixer of the quadrature modulation unit in FIG.

FIG. 12 is a configuration diagram of an example of a conventional quadrature modulator.

FIG. 13 is a configuration diagram of another example of a conventional quadrature modulator.

FIG. 14 is a signal space diagram of an output modulated wave.

FIG. 15 is a signal space diagram of a modulated wave subjected to normal and linear distortion.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Linear conversion means 2 Quadrature modulator 3 Level signal generation means 4 Parameter generation means 20, 21, 22, 23 Switch 24 Linear converter 25 Control circuit 26 Detector 27 Low-pass filter

Continuation of the front page (56) References JP-A-63-121326 (JP, A) JP-A-63-208330 (JP, A) JP-A-2-254839 (JP, A) JP-A-64-36153 (JP , A) Real opening 63-81549 (JP, U) (58) Fields investigated (Int. Cl. ⁷ , DB name) H04L 27/20 H04L 27/36

Claims (2)

(57) [Claims] 1. A linear conversion parameter for compensating for linear distortion is set, a linear conversion unit converts the in-phase amplitude signal and the quadrature amplitude signal into a baseband signal, and performs a linear conversion by the linear conversion unit. In a distortion-compensating quadrature modulator having a quadrature modulator that receives a baseband signal as input, the linear conversion means includes an in-phase amplitude signal I (t) and a quadrature amplitude signal I (t).
Quadrature from signal Q (t) to in-phase baseband signal a (t)
To determine the baseband signal b (t),
Parameters a, b, α, using a theta, a linear transformation defined by a (t) = I (t ) + tanθ · Q (t) + a b (t) = (1 / αcosθ) Q (t) + b performed, wherein the carrier component is removed from the modulated wave output of the orthogonal transformer, the level signal generation means for generating a level signal, the level signal generating means and said linear transform using the values of the inputted level signal from the parameter a , B, α, θ, and parameter generating means for setting the linear conversion means. 2. A linear conversion parameter for compensating for linear distortion is set, and a linear conversion means for linearly converting the in-phase amplitude signal and the quadrature amplitude signal into a baseband signal, and performing linear conversion by the linear conversion means. In a distortion-compensating quadrature modulator having a quadrature modulator that receives a baseband signal as input, the linear conversion means includes an in-phase amplitude signal I (t) and a quadrature amplitude signal I (t).
Quadrature from signal Q (t) to in-phase baseband signal a (t)
To determine the baseband signal b (t),
Parameters a, b, α, using a theta, a linear transformation defined by a (t) = I (t ) + tanθ · Q (t) + a b (t) = (1 / αcosθ) Q (t) + b The linear conversion parameters a and b relating to the DC offset are first determined, the linear conversion parameter α relating to the amplitude balance is determined next while compensating for the distortion due to the DC offset, and finally the distortion due to the DC offset and the amplitude unbalance is determined. Find the linear transformation parameter θ for orthogonality while compensating,
A distortion generating quadrature modulator comprising parameter generating means for setting the linear conversion means.