JP2001057578A - Linear compensation circuit - Google Patents

Abstract

(57) [Summary] [Object] To improve the degree of distortion improvement in a Cartesian feedback circuit for reducing distortion in a wireless transmitter. A local in-phase demodulation signal I4 and a local quadrature demodulation signal Q4 obtained by branching a part of a transmission output and performing local demodulation are subtracted from the in-phase signal I2 and the quadrature signal Q2, respectively. Is calculated. This signal is converted into a digital signal and input to the correction signal generation circuit 12 to generate a correction signal. The correction signal is converted into an analog signal and added to the in-phase error signal I3 and the quadrature error signal Q3 using the adder 3 and the adder 4, thereby improving the degree of distortion improvement. While maintaining the distortion improvement effect, the bandwidth in which distortion can be improved is widened, and distortion due to transmitter noise, characteristic changes due to temperature and power supply voltage changes, etc. can be improved.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a linear compensation circuit, and more particularly to a linear compensation circuit having improved distortion compensation in a Cartesian feedback circuit for reducing distortion in a radio transmitter.

[0002]

2. Description of the Related Art In a transmitter for transmitting a linearly modulated signal, if the linearity is poor, not only does the modulation accuracy deteriorate, but also excessive out-of-band unnecessary radiation is generated, causing interference to adjacent channels. There is a risk. For this reason, it is necessary to ensure linearity over the entire operating range of the transmitter. Conventionally, one of the methods for improving the linearity of a transmitter is so-called Cartesian feedback. In this technique, a part of the output of the transmitter is locally demodulated into a quadrature baseband signal, and negative feedback is applied to the modulated baseband signal.

FIG. 4 is a schematic block diagram showing an example of a Cartesian feedback circuit which is a conventional linear compensation system. In the conventional example shown in FIG.
Is subtracted from the local in-phase demodulated signal I4 subjected to the local quadrature demodulation to obtain an in-phase error signal I3. The quadrature input Q1 is obtained by subtracting the local quadrature demodulated signal Q4 subjected to local quadrature demodulation to obtain a quadrature error signal Q3. The in-phase error signal I3 and the quadrature error signal Q3 are input to the loop filter 5 and the loop filter 6, respectively, band-limited, and then quadrature demodulated by the quadrature modulator 7. The quadrature-modulated signal is output after being amplified by the high frequency amplifier 8. A part of the output is branched and subjected to local quadrature demodulation by a quadrature demodulator 9 to obtain a local in-phase demodulated signal I4
And the local quadrature demodulated signal Q4. As described above, the Cartesian feedback performs the linear compensation of the high-frequency amplifier by applying the feedback in the baseband.

FIG. 5 shows the principle of the Cartesian feedback circuit. The input signal Vin is input to a subtractor,
The feedback signal Vmon is subtracted and becomes an error signal Verr. The error signal Verr is amplified by an amplifier having a gain G1. The distortion component Vd of the amplifier is added to the amplifier output and becomes Vout. Vout is input to the attenuator to generate a feedback signal Vmon, which is subtracted from the input Vin. The error signal Verr is as follows: Verr = Vin−Vmon = Vin−G2 × Vout. Since the output Vout is Vout = Verr × G1 + Vd, solving the above equation for Vout gives: Vout = (G1 × Vin) / (1 + G1 × G2) −Vd / (1+
G1 × G2). As described above, the distortion component is 1
It is improved by a factor of / (1 + G1 × G2).

FIG. 11 is a diagram showing phase characteristics of a circuit showing the principle of the Cartesian feedback circuit shown in FIG. In order for the Cartesian feedback circuit not to oscillate, at a frequency where the loop phase becomes -180 [deg],
The loop gain must be less than 0 [dB]. Therefore, the band loop gain is linear compensation is performed in the case of G _H becomes to 0 to F _H, the band loop gain is linear compensation is performed in the case of G _L is up to 0 to F _L. As described above, when the Cartesian feedback is used, when the loop gain is increased, the bandwidth in which the distortion is improved is reduced. Since the degree of distortion improvement is proportional to the loop gain, when the degree of distortion improvement is increased, the bandwidth in which distortion is improved is reduced.

Next, a predistortion system, which is one of the linear compensation systems, will be described. First, the distortion of the high-frequency amplifier will be described, and a method of removing the distortion of the high-frequency amplifier by preliminarily correcting the input signal will be described. FIG. 6 shows input / output characteristics of the high-frequency amplifier. P
out represents the output of the high-frequency amplifier when the input power of the high-frequency amplifier is Pin. The output phase θ ₀ indicates the output phase of the high-frequency amplifier when the input power is Pin. The output phase when the input Pin is sufficiently small is defined as θ ₀ . At this time, dθ (Pin) represents the difference between the output phase and θ ₀ .

FIG. 7 is a graph showing the input / output reverse characteristics of the high-frequency amplifier. This graph shows the inverse function of the input / output characteristics of the high-frequency amplifier, and the vertical axis and the horizontal axis are reversed.
Pin (Pout) represents the input power Pin required to obtain the power output of Pout. From the above definition, when Pin (Pout) is input to the high frequency amplifier, the output power becomes Pout
Becomes When this is represented by an equation, Pout (Pin (x)) = x.

FIG. 8 shows a graph of the distortion correction characteristics of the high-frequency amplifier. This graph is obtained by dividing the horizontal axis Pout of the graph of the input / output inverse characteristic of the high-frequency amplifier by the gain Gpa at the time of low power and converting it into the input power. PinPrd (Pin) represents the corrected input power necessary for the high-frequency amplifier output to have the same output as in the linear operation when the input power is Pin.

Since PinPrd (Pin) = Pin (Pin × Gpa), when the corrected input power is applied to the high-frequency amplifier, the output power becomes: Pout (PinPrd (Pin)) = Pout (Pin (Pin × Gpa)) = Pin × Gpa (∵Pout (Pin (x)) = x), and the same output power as when the high-frequency amplifier is operating linearly is obtained.

Based on the above relationship, the distortion correction method will be described separately for the amplitude correction method and the phase correction method. P
inPrd is represented by a gain correction value Gprd, and Gprd (x) = PinPrd (x) / x. At this time, Gprd (Pin) × Pin is applied to the high frequency amplifier.
Is input, the output becomes Pout (Gprd (Pin) × Pin) = Pout (PinPrd (Pin) / Pin × Pin) = Pout (PinPrd (Pin)) = Pin × Gpa, and the input amplitude is multiplied by Gprd (Pin). Then, the output power at the time of linear operation is obtained from the output of the high-frequency amplifier.

At this time, the phase of the output of the high-frequency amplifier is dθ
The phase is rotated by (PinPrd). In order to correct this, the input phase may be rotated by −dθ (PinPrd). When the above operation is represented by an operation on the I and Q signals, the amplitude correction is as follows: (I ′) = (Gprd (Pin) 0) (I) (Q ′) (0Gprd (Pin)) (Q) The phase correction is as follows: (I ′) = (cos (−dθ) −sin (−dθ)) (I) (Q ′) (sin (−dθ) cos (−dθ)) (Q)

The above operations can be summarized as follows: (I ′) = (Gprd × cos (−dθ) −Gprd × sin (−dθ)) (I) (Q ′) (Gprd × sin (−dθ) Gprd × cos (−dθ)) (Q) (1)

The distortion of the high-frequency amplifier can be eliminated by correcting the in-phase input I1 and the quadrature input Q1 according to equation (1).

FIG. 10 shows an example of the correction circuit. The in-phase input I1 and the quadrature input Q1 are input to a multiplier 19 and a multiplier 20, respectively, and are squared. The result of the squaring is input to the adder 21, and I1 × I1 + Q1 × Q1 is output. The correction table outputs Gprd × cos (−dθ) and Gprd × sin (−dθ) according to the square of the input amplitude I1 × I1 + Q1 × Q1.

Gprd × cos (−dθ) which is a correction table output
Is input to a multiplier 15 and a multiplier 18 and is multiplied by the in-phase input I1 and the quadrature input Q1.
cos (−dθ) × I1 is output.
× cos (−dθ) × Q1 is output. Gprd × sin (−dθ), which is the output of the correction table, is input to the multiplier 16 and the multiplier 17 and is multiplied by the in-phase input I1 and the quadrature input Q1, and Gprd × sin (−dθ) × Q1 is output from the multiplier 16. Gprd × sin (−dθ) × I1 is output from the multiplier 17.

The output of the multiplier 15 and the output of the multiplier 16 are added to each other.
23, and from the adder 23, Gprd × cos (−dθ) × I1-Gprd × sin (−dθ)
× Q1 is output. The output of the multiplier 17 and the output of the multiplier 18 are input to the adder 24. From the adder 24, Gprd × sin (−dθ) × I1 + Gprd × cos (−dθ)
× Q1 is output. By the above operation, the operation according to the equation (1) is performed. By using this correction table to correct the input signal, an output equal to that when the high-frequency amplifier output is operating linearly can be obtained.

FIG. 9 is a schematic block diagram showing an example of a predistortion circuit which is a conventional linear compensation system. The in-phase input I1 and the quadrature input Q1 are input to a correction circuit 7, and corrected so as to cancel the distortion of the high-frequency amplifier 8,
The signals are output as the in-phase signal I2 and the quadrature signal Q2. The in-phase signal I2 and the quadrature signal Q2 are input to the quadrature modulator 5, and the output of the quadrature modulator is input to the high-frequency amplifier 8 and output after amplification.

[0018]

However, in the conventional Cartesian feedback system as described above,
When the degree of distortion improvement is increased, there is a problem in that the bandwidth in which distortion can be reduced becomes narrow.

Further, in the predistortion method, there is a problem that distortion due to noise of a transmitter, characteristic change due to temperature / power supply voltage change, and the like cannot be improved.

The present invention solves the above problems,
It is an object of the present invention to provide a linear compensation circuit capable of widening the bandwidth in which distortion can be improved while maintaining the distortion improvement effect, and also capable of improving distortion due to a change in characteristics of the transmitter due to noise, temperature and power supply voltage changes.

[0021]

According to the present invention, there is provided a local in-phase demodulated signal I which is locally demodulated.
4 and a local quadrature demodulation signal Q4 are subtracted from the in-phase signal I2 and the quadrature signal Q2, respectively, to calculate an in-phase error signal I3 and a quadrature error signal Q3. An addition circuit 3 for adding a common-mode error signal I3 to the digital
An adding circuit 4 for adding the quadrature error signal Q3 to the output of the analog converting circuit 14, a loop filter 5 and a loop filter 6 for band limiting the outputs of the adding circuit 3 and the adding circuit 4;
A quadrature modulator 7 for quadrature-modulating the outputs of the loop filter 5 and the loop filter 6 with a carrier, a high-frequency amplifier 8 for amplifying the quadrature-modulated signal, and using the amplified signal as a transmission output and branching a part. And a quadrature demodulator 9 for demodulating the signal into a local in-phase demodulation signal I4 and a local quadrature demodulation signal Q4 using the carrier wave whose phase has been adjusted and outputting the demodulated signal to a subtraction circuit. An analog-to-digital conversion circuit 10 and an analog-to-digital conversion circuit 11 for converting the in-phase error signal I3 and the quadrature error signal Q3 output from the subtraction circuit into digital signals,
A configuration including a correction signal generation circuit 12 which receives the outputs of the analog / digital conversion circuit 10 and the analog / digital conversion circuit 11, generates a correction signal, and inputs the correction signal to the digital / analog conversion circuit 13 and the digital / analog conversion circuit 14 And

With this configuration, the loop gain can be reduced, and the bandwidth over which distortion can be improved can be widened while maintaining the distortion improvement effect. When the distortion improvement bandwidth and the loop gain are constant, the degree of distortion improvement is improved by adding a distortion correction signal in addition to the distortion improvement for the loop gain. Further, it is possible to improve distortion caused by a noise of a transmitter, a characteristic change due to a temperature / power supply voltage change, and the like, which cannot be performed by the conventional pre-distortion method.

[0023]

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

(First Embodiment) In a first embodiment of the present invention, a local in-phase demodulation signal and a local quadrature demodulation signal obtained by branching a part of a transmission output and performing local demodulation are used as input signals. A phase signal and a quadrature signal are subtracted from each other to calculate an in-phase error signal and a quadrature error signal, a digital conversion is performed to generate a correction signal, and an analog conversion is performed. This is a linear compensation circuit of a Cartesian feedback circuit for correcting distortion by coordinates.

FIG. 1 is a block diagram showing the configuration of the linear compensation circuit according to the first embodiment of the present invention. In the present embodiment, one of the signal channels of the quadrature modulation radio having two channels is called an in-phase signal, and the other is called a quadrature signal. In FIG. 1, a subtraction circuit 1 is a circuit that subtracts a local in-phase demodulated signal I4 from an in-phase signal I2 and outputs an in-phase error signal I3. The subtraction circuit 2 is a circuit that subtracts the local quadrature demodulation signal Q4 from the quadrature signal Q2 and outputs a quadrature error signal Q3. The loop filter 5 outputs the common-mode error signal I
3 is a filter for limiting the band. Loop filter 6
Is a filter for limiting the band of the quadrature error signal Q3. The quadrature modulator 7 is a modulator that performs quadrature modulation on the output of the loop filter. The high-frequency amplifier 8 is an amplifier that amplifies an output signal of the quadrature modulator. The quadrature demodulator 9 is a demodulator that demodulates a signal obtained by branching a part of the output of the high-frequency amplifier and outputs a local in-phase demodulated signal I4 and a local quadrature demodulated signal Q4. The analog / digital conversion circuit 10 is a circuit for converting the in-phase error signal I3 into a digital signal. The analog / digital conversion circuit 11 is a circuit that converts the quadrature error signal Q3 into a digital signal. Correction signal generation circuit
A circuit 12 calculates a correction signal according to the distortion of the high-frequency amplifier from the digitally converted in-phase error signal I3 and quadrature error signal Q3. The digital / analog conversion circuits 13 and 14 are circuits for converting the output of the correction signal generation circuit into an analog signal. The addition circuits 3 and 4 are circuits that add the output of the correction signal generation circuit converted into the analog signal to the in-phase error signal I3 and the quadrature error signal Q3.

FIG. 2 is a block diagram showing the configuration of the correction signal generation circuit. In FIG. 2, a multiplying circuit 19 is a circuit for calculating the square of the in-phase input I1. The multiplication circuit 20 is a circuit that calculates the square of the orthogonal input Q1. Adder circuit 21
Is a circuit for calculating the sum of the output of the multiplication circuit 19 and the output of the multiplication circuit 20. The correction signal generation table 22 indicates that I1 × I1 + Q1 which is the square of the amplitude of I1 and Q1 output from the adder circuit 21.
According to × Q1, Gprd × cos (−dθ) −1 and Gprd × s
It is a table for outputting in (−dθ). Multiplication circuit 1
5, 16, 17 and 18 are the correction signal generation table output and the in-phase input I
1. A circuit for calculating the product of the orthogonal input Q1. Adder circuit
23 and 24 are the outputs of the multiplication circuits 15 and 16 and the multiplication circuit 17
And a circuit for calculating the sum of the outputs of the two.

The principle of the present invention will be described with reference to FIG.
The input voltage of the amplifier is Vin2, and the output voltage of the amplifier is Vout
2. The distortion voltage of the amplifier is Vd, the noise voltage of the amplifier is Vn,
The distortion correction voltage is Vprd. Since Vprd is based on the result of measuring the characteristics of the amplifier in advance, it is not possible to correct distortion due to changes in characteristics due to changes in temperature, time, and power supply voltage, noise in the amplifier, and the like. For this reason, distortion due to temperature, time, characteristic fluctuation due to power supply voltage, amplifier noise, etc. is included in Vn.

First, the case where the noise voltage Vn is 0 will be examined.
Assuming that the input / output characteristics of the amplifier are Vout2 = Vout2 (Vin2), Vout2 (Vin2) = G1 × Vin2 + Vd where G1 is the gain of the amplifier. This is expressed as a value Verr before distortion correction, and Vin2 = Verr + Vprd, so that Vout (Verr + Vprd) = G1 × Vin2 + Vd. Vprd is a correction voltage for producing the same output as when the amplifier operates linearly, and the amplifier output has a gain G1 times that of Verr. This can be expressed as follows. Vout (Verr + Vprd) = G1 × Verr From this, the equation of the loop is obtained as follows: Vin−G1 × Verr × G2 = Verr Vout = G1 × Verr Solving this results in Vout = G1 / (1 + G1 × G2) Vin. Thus, when there is no noise voltage Vn, the distortion voltage component Vd can be completely compensated.

Next, a case where the noise voltage Vn is not 0 will be examined. Since the noise voltage Vn is sufficiently smaller than the signal component, a correction signal generation circuit and a small signal equivalent circuit of an amplifier are used. In the following, the signal in the case where the noise Vn = 0 is represented by Ver
r, Vprd, Vin2, Vout2, and Vout, and the signal when the noise Vn ≠ 0 is represented by Verr ', Vprd', Vin2 ', Vout2',
Vout '. The change of each signal is represented by {Verr,}
Vprd, ∂Vin2, ∂Vout2, ∂Vout. The relationship between the variables is as follows: Verr '= Verr + ∂Verr Vprd' = Vprd + ∂Vprd Vin2 '= Vin2 + ∂Vin2 Vout2' = Vout2 + ∂Vout2 Vout '= Vout + ∂Vout

The small signal input / output characteristics of the correction signal generation circuit are as follows: Vprd '= Vprd (Verr') = Vprd (Verr + ∂Verr) = Vprd (Verr) + (∂Vprd) / (∂Verr) ∂∂Verr . Therefore, ∂Vprd = (∂Vprd) / (∂Verr)
・ ∂Verr.

Since Vin2 = Vprd + Verr, ∂Vin2 = ∂Vprd + ∂Verr = (∂Vprd) / (∂Verr) ∂∂Verr + ∂Verr = ((∂Vprd) / (∂Verr) +1) ∂∂Verr. The input / output characteristics of the amplifier are as follows: Vout2 '= Vout2 (Vin2') = Vout2 (Vin2 + ∂Vin2) = Vout2 (Vin2) + (∂Vout2) / (∂Vin2) ∂∂Vin2 ≒ Vout2 (Vin2) + G1∂∂Vin2 ∂Vout2 = G1∂∂Vin2 = G1 ・ ((∂Vprd) / (∂Verr) +1) ∂∂Verr.

Since Vout = Vout2 + Vd + Vn, ∂Vout = ∂Vout2 + Vn = G1 ・ ((∂Vprd) / (∂Verr) +1) ∂∂Verr + Vn. Here, it is assumed that the change in the distortion signal Vd due to the noise voltage Vn is sufficiently small and constant.

Since Verr = Vin−G2 · Vout, ∂Verr = −G2 · ∂Vout. From this, Vout = −G1 · G2 ((∂Vprd) / (∂Verr) +
1) · ∂Verr + Vn Therefore, ∂Vout = 1 / (G1 · G2 ((∂Vprd) / (∂Ver
r) +1)). Vn. Since Vprd is a correction signal, it is small, and (prVprd) / (∂Verr) ≒ 0, so that Vout ≒ 1 / (G1 · G2) · Vn. As described above, the noise component becomes 1 / (G1 · G2) times, which is almost 1 / loop gain.

The distortion caused by the imperfect pre-distortion can be included in Vn, and as described above, in the present method, it can be improved by the amount of the loop gain. For this reason, in the present system, the distortion of the amplifier is the sum of the improvement effect by the predistortion and the improvement effect by the Cartesian feedback.

As described above, the distortion component is the sum of the distortion improvement effect of the predistortion and the distortion improvement effect of the Cartesian feedback, and the distortion due to noise, power supply voltage, and temperature fluctuation is improved by the loop gain of the Cartesian feedback. Is done.

Next, the operation of the linear compensation circuit according to the first embodiment of the present invention will be described with reference to FIG.
The in-phase error signal I3 and the quadrature error signal Q3 are digital
Analog conversion circuit 10, digital / analog conversion circuit
According to 11, it is converted into a digital signal. This signal
The signal is input to the correction signal generation circuit 12. Assuming that the input is I, Q and the output is I ′, Q ′, the correction signal generation circuit has (I ′) = (Gprd × cos (−dθ) −1 −Gprd × sin (−dθ)) (I) ( Q ′) (Gprd × sin (−dθ) Gprd × cos (−dθ) −1) (Q) is calculated. Therefore, the correction signal generation circuit calculates I ′ = (Gprd × cos (−dθ) −1) × I3−Gprd × sin
(−dθ) × Q3 Q ′ = Gprd × sin (−dθ) × I3 + (Gprd × cos (−d
θ) -1) × Q3 is output. This output is a digital-to-analog converter
13. The analog signal is input to the digital / analog conversion circuit 14, and is input to the addition circuits 3 and 4. Since the addition circuit 3 adds I3 and I ', the output of the addition circuit 3 is Gprd × cos (−dθ) × I3−Gprd × si
n (−dθ) × Q3.

The adder circuit 4 adds Q3 and Q '.
The output of the addition circuit 4 is Gprd × sin (−dθ) × I3 + Gp
rd × cos (−dθ) × Q3. Therefore, the outputs of the adder circuit 3 and the adder circuit 4 become equal to the signal subjected to the distortion compensation by the predistortion method.

The outputs of the adder circuits 3 and 4 are input to the loop filter 5 and the loop filter 6, and are band-limited. The outputs of the loop filter 5 and the loop filter 6 are input to the quadrature modulator 7, and are quadrature-modulated and output. The output of the quadrature modulator 7 is input to the high frequency amplifier 8 and output after amplification. A part of the output is branched and input to the quadrature demodulator 9 for quadrature demodulation and output as a local in-phase demodulation signal I4 and a local quadrature demodulation signal Q4. Local in-phase demodulated signal I4,
The local quadrature demodulated signal Q4 is input to the subtraction circuit 1 and the subtraction circuit 2, and is subtracted from the in-phase signal I2 and the quadrature signal Q2 to output an in-phase error signal I3 and a quadrature error signal Q3. As described above, the Cartesian feedback is configured.

As described above, in the first embodiment of the present invention, the linear compensation circuit of the Cartesian feedback circuit is provided with a local in-phase demodulation signal and a local quadrature demodulation signal obtained by branching a part of the transmission output and performing local demodulation. Is subtracted from the in-phase signal and the quadrature signal, which are input signals, respectively, to calculate an in-phase error signal and a quadrature error signal, generate a correction signal by digital conversion, convert the analog signal, and convert the in-phase and quadrature error signals. Since the distortion is corrected by adding the baseband rectangular coordinates, distortions due to changes in characteristics due to fluctuations in temperature, time, power supply voltage, amplifier noise, etc., were impossible with the conventional predistortion method. Can be corrected. Further, since the distortion is corrected by the pre-distortion, the same distortion improving effect can be obtained even if the loop gain is reduced. For this reason, the phase margin of the loop is increased, and the frequency band in which distortion can be improved can be widened.

(Second Embodiment) In a second embodiment of the present invention, an in-phase input and a quadrature input, which are input signals, are converted into analog signals to generate an in-phase signal and a quadrature signal. The local in-phase demodulation signal and the local quadrature demodulation signal, which are partially demodulated, are subtracted from the in-phase signal and the quadrature signal, respectively, to calculate the in-phase error signal and the quadrature error signal. This is a linear compensation circuit of a Cartesian feedback circuit that generates and converts the signal into an analog signal, and corrects the distortion in the baseband orthogonal coordinates that adds the in-phase error signal and the quadrature error signal.

FIG. 12 is a block diagram showing a configuration of a linear compensation circuit according to the second embodiment of the present invention. FIG.
2, the digital / analog conversion circuits 25 and 26
This circuit converts the in-phase input I1 and the quadrature input Q1 into analog signals. The difference between the second embodiment and the first embodiment is that the input of the correction signal generation circuit 12 is the in-phase input I1 and the quadrature input Q1, and the digital-analog conversion circuit 10 and the digital-analog conversion circuit 11 What is omitted is that the inputs of the correction signal generation circuit 12 are the in-phase input I1 and the quadrature input Q2.

The operation of the linear compensation circuit according to the second embodiment of the present invention having the above configuration will be described. The in-phase error signal I3 and the quadrature error signal Q3 have a loop gain of GL.
Then, I3 = − (I1) / (GL + 1) Q3 = − (Q1) / (GL + 1) If the correction signal generation table 22 in the correction signal generation circuit 12 is represented by I1 and Q1 according to the above relationship, the same value is output from the correction signal generation circuit 12 at the same input. Thus, it is possible to operate in the same manner as in the first embodiment.

As described above, in the second embodiment of the present invention, the linear compensation circuit of the Cartesian feedback circuit converts the in-phase and quadrature inputs, which are input signals, into analog signals to convert the in-phase and quadrature signals. A local in-phase demodulated signal and a local quadrature demodulated signal, in which a part of the transmission output is generated and locally demodulated, are subtracted from the in-phase signal and the quadrature signal, respectively. And a correction signal is generated according to the quadrature input and converted to an analog signal, and the distortion is corrected by the baseband quadrature coordinates for adding the in-phase error signal and the quadrature error signal, respectively, so that the phase margin of the loop increases, A frequency band in which distortion can be improved can be widened.

(Third Embodiment) In a third embodiment of the present invention, a signal is corrected according to an in-phase input and a quadrature input, and is converted into an in-phase signal and a quadrature signal by analog conversion. And a local in-phase demodulation signal and a local quadrature demodulation signal are subtracted from the in-phase signal and the quadrature signal, respectively, to calculate an in-phase error signal and a quadrature error signal.
This is a feedback-time path-linear compensation circuit.

FIG. 13 is a block diagram showing a configuration of a linear compensation circuit according to the third embodiment of the present invention. Third
This embodiment is different from the first embodiment in that an analog / digital conversion circuit 10, an analog / digital conversion circuit 11, a correction signal generation circuit 12, a digital / analog conversion circuit 13, a digital / analog conversion circuit 14 , And the adder circuit 3 and the adder circuit 4 are omitted. Instead, the in-phase input I1 and the quadrature input Q1 are input to the correction circuit 7, and the output of the correction circuit 7 is a digital / analog conversion circuit 25 and a digital / analog conversion circuit 26. It is input to.

The operation of the linear compensation circuit according to the third embodiment of the present invention having the above configuration will be described. The in-phase error signal I3 and the quadrature error signal Q3 have a loop gain of GL.
Then, I3 = − (I1) / (GL + 1) Q3 = − (Q1) / (GL + 1) If the correction table 27 in the correction circuit 7 is represented by I1 and Q1 in accordance with the above relationship, it is possible to operate similarly to the first embodiment.

As described above, in the third embodiment of the present invention, the path linear compensation circuit of the Cartesian feedback circuit corrects the signal according to the in-phase input and the quadrature input,
The in-phase signal and the quadrature signal are converted into analog signals, and the local in-phase demodulation signal and the local quadrature demodulation signal obtained by locally demodulating a part of the transmission output are subtracted from the in-phase signal and the quadrature signal, respectively. Since it was calculated and corrected for distortion with the baseband rectangular coordinates,
The phase margin of the loop is increased, and the frequency band in which distortion can be improved can be widened.

(Fourth Embodiment) In a fourth embodiment of the present invention, an amplitude / angle error correction circuit corrects a signal in accordance with an in-phase input and a quadrature input, and converts the analog signal to an in-phase signal and a quadrature signal. And a local in-phase demodulation signal and a local quadrature demodulation signal, in which a part of the transmission output is locally demodulated, are respectively subtracted from the in-phase signal and the quadrature signal to calculate an in-phase error signal and a quadrature error signal. This is a Cartesian feedback path-linear compensation circuit for correcting distortion.

FIG. 14 is a block diagram showing a configuration of a linear compensation circuit according to the fourth embodiment of the present invention. FIG.
In 4, the amplitude angle error correction circuit 28 is a circuit for multiplying an input signal by a coefficient to correct the input signal. The fourth embodiment is different from the third embodiment in that the correction circuit 7 is replaced by an amplitude angle error correction circuit 28.

FIG. 15 is a block diagram of the amplitude angle error correction circuit. The amplitude angle error correction circuit includes a multiplication circuit 15, a multiplication circuit 16, a multiplication circuit 17, and a multiplication circuit in the correction circuit 7.
The input of 18 is not a correction table but a coefficient 1, a coefficient 2, a coefficient 3, and a coefficient 4.

The operation of the linear compensator according to the fourth embodiment of the present invention will be described. If the Cartesian feedback circuit has an amplitude error or an angle error of the quadrature modulator 7 or the quadrature demodulator 8, an error occurs in the amplitude or angle of the Cartesian feedback as a whole.

[0052] Amplitude and angle errors cause image leaks and consequently degrade modulation accuracy. Here, the amplitude error of the entire Cartesian feedback is d
G, the angle error is dθ, the output signal converted to baseband is the in-phase output I5, the quadrature output Q5, and the gain of the entire loop is G. In-phase input I1 and quadrature input Q1
And the in-phase output I5 and the quadrature output Q5 are: (I5) = GＩ (1 + dG0) ・ (cos (dθ) -sin (dθ)) ・ (I1) (Q5) (01) ( sin (dθ) cos (dθ)) (Q1).

When M = (1 + dG 0) · (cos (dθ) −sin (dθ)) (01) (sin (dθ) cos (dθ)), the inverse matrix of M is And Coefficients in the amplitude angle error correction circuit of FIG. 15 are coefficient 1 = IM1, coefficient 2 = IM2, coefficient 3 = IM3, and coefficient 4.
= IM4, the input / output relationship of the circuit shown in FIG. Becomes As described above, the amplitude error and the angle error are corrected by using the amplitude angle correction circuit.

Generally, since the amplitude error and the angle error of the orthogonal modulator 7 and the orthogonal demodulator 8 are constant for each block, if the orthogonal modulator 7 and the orthogonal demodulator 9 are determined, the coefficient 1
Coefficient 2, coefficient 3, and coefficient 4 can be determined. Therefore, the coefficient can be determined by performing measurement at the time of manufacture.

As described above, according to the fourth embodiment of the present invention, the path linearity compensation circuit of the Cartesian feedback circuit is corrected by the amplitude angle error correction circuit according to the in-phase input and the quadrature input, and the analog conversion is performed. The in-phase signal and the quadrature signal are then subtracted, and the in-phase error signal and the quadrature error signal are calculated by subtracting the local in-phase demodulation signal and the local quadrature demodulation signal obtained by locally demodulating a part of the transmission output from the in-phase signal and the quadrature signal, respectively. Therefore, since the distortion is corrected by the baseband orthogonal coordinates, the orthogonal modulator 7 and the orthogonal demodulator 9
The amplitude error and the angle error of the entire Cartesian feedback caused by the amplitude error and the angle error can be corrected.

[0056]

As is apparent from the above description, according to the present invention, the linear compensation circuit of the Cartesian feedback circuit for correcting distortion in the orthogonal coordinates of the baseband includes the local demodulated local in-phase demodulated signal I4 and the local quadrature demodulated signal. An analog-to-digital conversion circuit 10 for converting the in-phase error signal I3 and the quadrature error signal Q3 obtained by subtracting the demodulated signal Q4 from the in-phase signal I2 and the quadrature signal Q2, respectively, into digital signals.
And an analog / digital conversion circuit 11, a correction signal generation circuit 12 which receives the output thereof, generates a correction signal, and inputs the correction signal to a digital / analog conversion circuit 13 and a digital / analog conversion circuit 14, and a digital / analog conversion circuit
An adding circuit 3 for adding the common mode error signal I3 to the output of
Since the configuration is provided with the addition circuit 4 for adding the quadrature error signal Q3 to the output of the digital / analog conversion circuit 14, a loop filter, a quadrature modulator, a high-frequency amplifier, and a quadrature demodulator, the loop gain is reduced. This makes it possible to expand the distortion improvement bandwidth.

When the distortion improvement bandwidth and the loop gain are fixed, distortion improvement is improved by adding a distortion correction signal in addition to the distortion improvement for the loop gain. Distortion due to transmitter noise, characteristic changes due to temperature and power supply voltage changes, etc. can also be improved.

Furthermore, there is no need to improve the distortion characteristics of the high-frequency amplifier itself, and no need to control the gain of the high-frequency amplifier.
All of the processing is performed in the baseband unit, and the power newly required for performing the linear compensation can be reduced. Therefore, the processing can be used for the linear compensation of a high-frequency amplifier having a small transmission power.

[Brief description of the drawings]

FIG. 1 is a block diagram of a linear compensation circuit according to a first embodiment of the present invention;

FIG. 2 is a block diagram of a correction signal generation circuit of the linear compensation circuit according to the first embodiment of the present invention;

FIG. 3 illustrates the principle of the present invention;

FIG. 4 is a diagram of a Cartesian feedback circuit that is a conventional linear compensation circuit;

FIG. 5 is a diagram showing the principle of a Cartesian feedback circuit,

FIG. 6 is a diagram showing input / output characteristics of a high-frequency amplifier.

FIG. 7 is a diagram showing input / output reverse characteristics of a high-frequency amplifier.

FIG. 8 is a diagram illustrating distortion correction characteristics of a high-frequency amplifier.

FIG. 9 is a diagram of a pre-distortion circuit which is a conventional linear compensation circuit;

FIG. 10 is a block diagram of a correction circuit in a conventional predistortion circuit,

FIG. 11 is a diagram showing phase characteristics of the circuit of the principle diagram of Cartesian feedback,

FIG. 12 is a block diagram showing a configuration of a linear compensation circuit according to a second embodiment of the present invention;

FIG. 13 is a block diagram showing a configuration of a linear compensation circuit according to a third embodiment of the present invention;

FIG. 14 is a block diagram showing a configuration of a linear compensation circuit according to a fourth embodiment of the present invention;

FIG. 15 is a block diagram illustrating a configuration of an amplitude angle error correction circuit of a linear compensation circuit according to a fourth embodiment of the present invention.

[Explanation of symbols]

 1, 2 Subtraction circuit 3, 4, 21, 23, 24 Addition circuit 5, 6 Loop filter 7 Quadrature modulator 8 High frequency amplifier 9 Quadrature demodulator 10, 11 Analog / digital conversion circuit 12 Correction signal generation circuit 13, 14, 25, 26 Digital / analog conversion circuit 15, 16, 17, 18, 19, 20 Multiplication circuit 22 Correction signal generation table 27 Correction table 28 Amplitude angle error correction circuit

 ──────────────────────────────────────────────────続 き Continued on the front page F-term (reference) 5J090 AA04 AA41 CA25 CA26 CA62 FA08 FA17 GN03 GN05 GN06 HA25 HN03 HN04 HN07 HN08 HN17 KA17 KA26 KA34 KA42 KA53 KA55 MA13 NN16 SA14 TA01 TA02 5K004 JA07 FF07J04 FF06 HH01 HH09 HH11 KK03 KK04 KK06 LL24

Claims (4)

[Claims] 1. A locally in-phase demodulated signal (I4) locally demodulated.
And the local quadrature demodulation signal (Q4) is subtracted from the in-phase signal (I2) and the quadrature signal (Q2), respectively.
3) and a subtraction circuit (1,
2), an addition circuit (3) for adding the in-phase error signal (I3) to the output of the digital / analog conversion circuit (13), and the quadrature error signal (Q3) to the output of the digital / analog conversion circuit (14). Addition circuit (4) for adding, and addition circuit (3)
And a loop filter (5) and a loop filter (6) for band-limiting the output of the adder circuit (4), and a quadrature modulator (7) for quadrature modulating the outputs of the loop filter (5) and the loop filter (6) with a carrier. ), A high-frequency amplifier (8) for amplifying the quadrature-modulated signal, a local in-phase demodulation signal (I4) and a local Quadrature demodulated signal (Q
A quadrature demodulator (9) which demodulates the signal to 4) and outputs it to the subtraction circuit
In a linear compensation circuit using a Cartesian feedback circuit for correcting distortion in the baseband rectangular coordinates, the in-phase error signal (I3) and the quadrature error signal (Q3) output from the subtraction circuits (1, 2) are provided. ) Into a digital signal, an analog-to-digital converter (10) and an analog-to-digital converter (11), and the outputs of the analog-to-digital converter (10) and the analog-to-digital converter (11). A linear compensation circuit, comprising: a correction signal generation circuit (12) for generating a correction signal and inputting the correction signal to the digital / analog conversion circuit (13) and the digital / analog conversion circuit (14). 2. A locally in-phase demodulated signal (I4) locally demodulated.
And the local quadrature demodulation signal (Q4) is subtracted from the in-phase signal (I2) and the quadrature signal (Q2), respectively.
3) and a subtraction circuit (1,
2), an adding circuit (3) for adding the in-phase error signal (I3) to the output of the digital / analog conversion circuit (13), and the quadrature error signal (Q3) to the output of the digital / analog conversion circuit (14). An addition circuit (4) for adding, a loop filter (5) and a loop filter (6) for band-limiting the outputs of the addition circuit (3) and the addition circuit (4), respectively, and the loop filter (5) and the loop filter ( 6) a quadrature modulator (7) for orthogonally modulating the output with a carrier, a high-frequency amplifier (8) for amplifying the quadrature-modulated signal, A quadrature demodulator (9) that demodulates the adjusted carrier into the local in-phase demodulation signal (I4) and the local quadrature demodulation signal (Q4) and outputs the demodulated signals to the subtraction circuits (1, 2). Cartesian F In a linear compensation circuit using a feedback circuit, a correction signal generation circuit (12) for generating a correction signal according to an in-phase input (I1) and a quadrature input (Q1), and an output of the correction signal generation circuit (12) is converted into an analog signal. Digital-to-analog conversion circuit (1
3) and a digital / analog conversion circuit (14), and the in-phase input (I1) and the quadrature input (Q1) are connected to the in-phase signal
A linear compensation circuit comprising: a digital-to-analog conversion circuit (25) and a digital-to-analog conversion circuit (26) for converting (I2) and a quadrature signal (Q2), respectively. 3. A locally demodulated local in-phase demodulated signal (I4).
And the local quadrature demodulation signal (Q4) is subtracted from the in-phase signal (I2) and the quadrature signal (Q2), respectively.
3) and a subtraction circuit (1,
2), the in-phase error signal (I3) and the quadrature error signal (Q
3) a loop filter (5) and a loop filter (6) for band-limiting each of them; a quadrature modulator (7) for quadrature modulating the outputs of the loop filter (5) and the loop filter (6) with a carrier; A high-frequency amplifier (8) for amplifying the amplified signal, and the local in-phase demodulated signal (I4) and the local quadrature demodulated signal (I4) using the amplified signal as a transmission output and using a carrier wave that has been partially branched and phase adjusted. A quadrature demodulator (9) for demodulating the signal to Q4) and outputting it to the subtraction circuit (1, 2), and using a Cartesian feedback circuit for correcting distortion in baseband rectangular coordinates. A correction circuit (7) for correcting a signal according to the in-phase input (I1) and the quadrature input (Q1), and a digital signal for converting the output of the correction circuit (7) into the in-phase signal (I2) and the quadrature signal (Q2).・ Analog conversion Linear compensation circuit which is characterized in that a road (25) and digital-to-analog converter (26). 4. A locally in-phase demodulated signal (I4) locally demodulated.
And the local quadrature demodulation signal (Q4) is subtracted from the in-phase signal (I2) and the quadrature signal (Q2), respectively.
3) and a subtraction circuit (1,
2) and the in-phase error signal (I3) and the quadrature error signal
A loop filter (5) and a loop filter (6) for limiting the band of (Q3), an orthogonal modulator (7) for orthogonally modulating the outputs of the loop filter (5) and the loop filter (6) with a carrier, A high-frequency amplifier (8) for amplifying the modulated signal; and a local in-phase demodulation signal (I4) and a local quadrature demodulation signal using the amplified signal as a transmission output and using a carrier wave that has been partially branched and phase adjusted. A quadrature demodulator (9) which demodulates the signal to (Q4) and outputs it to the subtraction circuits (1, 2)
And a linear compensation circuit using a Cartesian feedback circuit for correcting distortion in the baseband rectangular coordinates, the amplitude angle error correction circuit (28) for correcting a signal according to the in-phase input (I1) and the quadrature input (Q1). ) And the output of the amplitude angle error correction circuit (28)
(I2) and a digital / analog conversion circuit (25) and a digital / analog conversion circuit (26) for converting the signal into a quadrature signal (Q2).