JP2007208779A - Digital orthogonal detector - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To simplify the configuration of a digital orthogonal detector, and to improve its gain also. <P>SOLUTION: The detector comprises an alternative distributing means (6) for inputting a digital IF signal sequence, whose sampling frequency is 8 times of a chip rate, and whose central frequency is four times of the sampling frequency and for distributing to two signal sequences alternately by each sampling, and an alternate inversion means (7, 8) for inputting the distributed signal sequence and for setting one side as I-phase component and the other side as a Q-phase component, after performing code inversion, alternately for each sampling about each signal sequence. The detector further outputs as a baseband the complex signal, grouped with the I-phase component and the Q-phase component from which one sampling time is shifted. As a result, gain is improved by 3 dB, by making a double wave always perform the same phase addition, and deterioration in the gain, caused by the deviation from synchronization point of sampling timing, can be prevented. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a digital quadrature detector that receives a modulated signal and obtains its in-phase component (I-phase component) and quadrature component (Q-phase component) by digital signal processing.

FIG. 6 is a block diagram of a conventional analog quadrature detector. The RF band signal is sequentially frequency-converted into an IF band and a BB band and branched into two. Then, the branched BB band signal is mixed with two local signals (referred to as a sine wave and a cos wave, respectively) having the same frequency but 90 degrees out of phase. The second harmonic component generated at this time is the LPF (Low
After being removed by the Pass Filter), it is output as a quadrature detection signal (complex signal) consisting of I-phase and Q-phase. If the frequency relationship between the input signal and the local signal is deviated, the band is expanded and distortion is caused. In addition, in the case of a modulation method using I-phase and Q-phase such as QPSK, the phase shift and level shift between the sin wave and cos wave of the local signal both cause distortion, and I / Q balance and IQ orthogonality correction, DC offset correction is required.

FIG. 7 is a block diagram of a conventional digital quadrature detector, showing an IF band sampling method for undersampling the IF band. In this case, multiplication with the local signal is inevitably performed by digital processing, but the sine wave and cos wave are distributed using a large-scale table, or a simple +1, -1 distribution cycle. However, because the principle is the same as that of the analog quadrature detection, a double wave is generated, and a procedure for removing it by the subsequent digital LPF is taken. In digital quadrature detection, no synchronization shift or I-phase / Q-phase level shift occurs.

When demodulating (demapping) a signal subjected to quadrature detection, it is desirable to employ a sample timing most suitable for demodulation, that is, a sample timing at which the eye pattern is most opened (hereinafter referred to as a synchronization point).
FIG. 4 is a diagram showing a relationship between a synchronization point timing error and gain deterioration in a conventional quadrature detector. As the synchronization point deviates from the sample timing, the gain of the quadrature detection signal decreases. Since the I and Q phases are naturally at the same timing, the I and Q phases are similarly affected by the timing error. The gain deterioration of the quadrature detection signal is the sum of the gain deterioration of the I-phase and the Q-phase. When 32/256 chips (0.5 samples at 4 times oversampling) are shifted, the deterioration is 0.34 dB.

However, if the sample rate is increased in order to reduce the error in sample timing, the amount of signal processing increases, so the sample timing is determined by a trade-off with the amount of processing.
For example, in a Code Division Multiple Access (CDMA) receiver, the sample rate at the time of quadrature detection is ²ⁿ times (for example, eight times) the chip rate, and the sample rate is reduced to, for example, ½ after quadrature detection to reduce the processing amount. I often sample.

In addition to the above, there is known a conventional technique that performs digital quadrature detection using a clock 4n (n = integer) times the carrier frequency of the received signal (see, for example, Patent Document 1).

JP-A-9-83588

However, with the conventional digital quadrature detector, if the analog method is digitized as it is, half of the energy of the received signal is taken to the double wave, or the influence of the error of the sample timing is exerted and the performance is fully demonstrated. There was a problem that I could not.

The present invention has been made in view of the above circumstances, and an object of the present invention is to provide a digital quadrature detector having a simple configuration and little gain deterioration.

An alternating distribution means for inputting a digital IF signal sequence having a sampling frequency of 8 times the chip rate and a center frequency of 4 times the sampling frequency, and alternately distributing two digital signal sequences for each sample;
The distributed signal sequence is input, and each signal sequence is alternately inverted for each sample, and includes an alternating inversion means having one as an I-phase component and the other as a Q-phase component,
A digital quadrature detector characterized in that a complex signal in which the I-phase component and Q-phase component shifted by one sample time are output as a baseband signal at a sampling frequency four times the chip rate.

According to the present invention, digital quadrature detection can be realized with a simple configuration that does not require an LPF, gain can be improved, and resistance to a deviation of a sampling point from a synchronization point can be provided.

Hereinafter, the embodiments will be described with reference to the drawings. However, all combinations of configurations described in the embodiments are not necessarily essential to the present invention. In addition, any combination of the features of the embodiments and combinations with the cited prior art can be included in the present invention.

In this example, a description will be given of a case in which a 1-carrier W-CDMA signal is undersampled in the IF band to obtain an 8-times oversampling baseband (BB) signal, but the reception method is not limited.
When this example is outlined, the IF signal includes both the I-phase component and the Q-phase component, but they are in a quadrature relationship, and one of them becomes zero depending on how the phase and period are taken, and the other is the maximum. There is a timing to become. That is, 8-times oversampling data can be extracted in a state where only the I component and only the Q component are included. This is performed by the alternate distribution function of this embodiment. Since the phases of the distributed signals are alternately inverted, the inversion function is used to invert the signals in order to match the phases.

FIG. 1 is a configuration diagram of the digital quadrature detector according to the first embodiment. First, the configuration of this example will be described with reference to FIG.
A BPF (Band Pass filter) 1 inputs an RF band signal received by an antenna or the like, and passes and outputs a frequency band including a modulation signal to be received.
The mixer 2 mixes the RF band signal output from the BPF 1 and the local oscillation signal, and outputs a signal frequency-converted to the IF band.
The BPF 3 passes and outputs only the frequency for a desired one of the signals generated by the mixing in the mixer 2.

An ADC (Analog to Digital Converter) 4 samples and quantizes the IF signal output from the BPF 3 and outputs the sampled signal as a digital IF signal. The ADC 4 sample rate f _s is 30.72 MHz, which is eight times the W-CDMA chip rate f _c (3.84 MHz), and further the USB (Upper Side Band) center frequency f _IFU or LSB (IFB) of the IF signal selected by the BPF 3. Upper Side Band) For center frequency f _IFL
f _IFU = m · f _s (Formula 1)
Or, f _IFL = (m + 1/2) · f _s (Formula 2)
Meet.

The DDC (Digital Down Convert) 5 is necessary only when the ADC 4 performs A / D conversion at a sample rate other than 8 times oversampling, and the sample rate of the digital IF signal output from the ADC 4 is 8 Convert to double oversample rate (30.72MHz). The following description is based on the assumption that the DDC 5 is not provided.
The selector 6 alternately distributes and outputs the digital IF signal output from the ADC 4 for each sample. That is, as shown in FIG. 2, when the output sequence of the ADC 4 is (a), the output sequence is divided into two sequences (b) and (c).

The multiplier 7 alternately multiplies −1 and +1 by one of the two sequences distributed by the selector 6 and outputs the result as an I-phase component.
The multiplier 8 alternately multiplies −1 and +1 by the other of the two sequences distributed by the selector 6 and outputs the result as a Q-phase component.
The selector 6 implements an alternating distribution function, and the multipliers 7 and 8 implement an alternating inversion function, thereby achieving quadrature detection. Each of the multipliers 7 and 8 inverts the sign of the signal alternately, and may be a simple sign inverter.

The relationship between the multipliers 7 and 8 differs depending on whether the original IF signal to be A / D converted is USB or LSB. Since USB is normally used for the BB signal, when the IF signal is USB, the data after quadrature detection is expressed as shown in (d) of FIG. On the other hand, when the original IF signal is LSB, the phase on the orthogonal side is inverted as shown in FIG.

In the complex signal data after quadrature detection, each of the I component and the Q component is four times oversampling. I component and Q component are obtained alternately every sample time, and it is not possible to obtain both components at the same time, but the same I component and one Q component that are shifted by one sample time from each other are the same Output as quadrature oversampling as a quadrature detection signal of time.

In the quadrature detector of this example, complete I / Q separation is possible, but the timing is shifted by 1/8 chip (one sample) between the I component and the Q component. In this example, tolerance for the deviation of the sampling point from the synchronization point is obtained by this deviation.

FIG. 5 is a diagram illustrating the relationship between the timing error at the synchronization point and the gain in the quadrature detector according to the first embodiment. In this figure and FIG. 4, the gain curve is approximated by a sine curve. In the same 4 times oversampling as before, the average value of gain degradation does not change, but the maximum value decreases from 0.34 dB to 0.17 dB. As a result, demodulation errors due to timing errors can be reduced.

There is a concern that this deviation may cause distortion similar to the deviation of the I / Q orthogonality. However, the deviation of the I / Q orthogonality in the analog quadrature modulator is a deviation in a fixed direction independent of the sample timing, whereas the deviation in this example is positive or negative depending on whether the sample is an even number or an odd number in 8-times oversampling. The average value is 0. Therefore, the influence of the deviation can be further suppressed if the CDMA receiver performs RAKE combining.

Next, IF band undersampling in the ADC 4 will be described.
FIG. 3 is a spectrum diagram illustrating IF band undersampling. The bandwidth of the IF signal is limited so as to be smaller than the Nyquist frequency in the sampling theorem. In this example, one carrier of the WCDMA signal is assumed to be about 5 MHz. In the spectrum, image components are repeated at a frequency interval of the ADC 4 sample rate (f _s = 30.72 MHz). Therefore, even if an IF signal having a frequency higher than f _s is A / D converted, it can be extracted as a signal having a center frequency f _IF comparable to that of the BB band.

When USB is used for the original IF signal, if an IF signal having a center frequency (carrier frequency) of 222.72 MHz satisfying (Equation 1) is used, an IF signal having a center frequency f _IF = 7.68 MHz is obtained by sampling. Further, when LSB is used for the original IF signal, if an IF signal of 238.08 MHz satisfying (Equation 2) is used, an IF signal of f _IF = 7.68 MHz is also obtained.

Next, the principle that the present embodiment can use the second harmonic component will be described.
In the conventional normal quadrature detection, a frequency component of the sum of the frequency of the IF signal and the local signal (because the frequency is almost doubled, henceforth referred to as a double wave component) is generated, which is an LPF (Low Pass Filter). Used to remove. In this example, quadrature detection is achieved without removing the second harmonic component. Here, the symbols are defined as follows.
Modulated (spread) signal: a + jb = exp (jφ)
Quadrature modulation angular frequency: ω
Initial phase: Δ

At this time, the signal after quadrature modulation is as follows.
Re [(a + jb) exp (jωt)] = {a · cos (ωt + Δ) −b · sin (ωt + Δ)} (Formula 3)
However, Re [] indicates a real part. Since ω corresponds to the center frequency f _IF of the sampled digital IF signal, and the BB signal is a signal in which ω is zero, when the quadrature detection is performed directly on the BB signal, the complex signal after the quadrature detection is as follows: Become.
{a ・ cos (ωt + Δ) −b ・ sin (ωt + Δ)} exp (-jωt)
= (1/2) exp (jΔ) exp (jφ) + (1/2) exp (-j (2ωt + Δ) exp (-jφ) (Formula 4)

Here, when the frequency component to be obtained is x and the second harmonic component is y, the following is obtained.
x = (1/2) exp (jΔ) exp (jφ)
Re [x] = (1/2) cos (φ + Δ)
Im [x] =-(1/2) sin (φ + Δ)

y = (1/2) exp (-jΔ) exp (-jφ) exp (-j2ω)
Re [y] = (1/2) cos (2ωt + φ + Δ)
Im [y] =-(1/2) sin (2ωt + φ + Δ)

Taking the extraction timing at every odd multiple of π / 2 with respect to the local signal exp (−jωt), for example, sampling at f _s = (2 / π) ω, the following equation is established.
When ωt = nπ
Re [y] = Re [x]
Im [y] =-Im [x]
When ωt = (n + 0.5) π
Re [y] =-Re [x]
Im [y] = Im [x]

That is, since the second harmonic component that has been discarded by the conventional LPF can always be effectively in-phase synthesized with the frequency component to be obtained in this example, the LPF is not required, and the gain can be improved by 3 dB.
Quadrature detection I phase data when ω = nπ: 2 × Re [x] = cos (φ + Δ) (Formula 5)
Quadrature detection Q phase data when ω = (n + 0.5) π: 2 × Im [x] = − sin (φ + Δ) (Formula 6)

(Formula 5) The quadrature detection as shown in (Formula 6) is possible only when f _s = (2 / π) ω = 4f _IF (the other f _IF = (n / 2 + 1/4) f _s is , Image frequency.) In addition, when (the n 1 or more integer) f _s = 2 ⁿ f _c to choose to usually, when the sampling rate is halved by quadrature detection in f _s is less 4f _c, band limitation and paths since the signal processing such as separation can not be performed sufficiently, f _s = 8f _c clogging 8 times oversampling is most suitable in the processing amount.

According to this embodiment, the advantage of digital quadrature detection that does not generate a perfect I / Q balance or DC component is maintained, and 3 dB gain can be improved over the normal method by preventing energy loss to the second harmonic. An LPF for the purpose of removing the second harmonic is not necessary, but there is no problem even if it exists. Further, a digital filter that performs roll-off filtering on the A / D converted IF or BB signal may be provided.

Configuration diagram of quadrature detector of embodiment 1 Operation timing chart of the quadrature detector according to the first embodiment. The figure explaining IF band undersampling of ADC4 The figure which shows the relationship between the timing error and gain in the quadrature detector of Example 1. The figure which shows the relationship between the timing error and the gain in the conventional quadrature detector Configuration of conventional analog quadrature detector Configuration of conventional digital quadrature detector

Explanation of symbols

1, 3 ... BPF, 2 ... Mixer, 4 ... ADC, 5 ... DDC, 6 ... Selector,
7,8 ... multiplier

Claims (1)

An alternating distribution means for inputting a digital signal sequence having a sampling frequency of 8 times the chip rate and a center frequency of 4 times the sampling frequency and distributing the digital signal sequence alternately to two signal sequences for each sample;
The distributed signal sequence is input, and each signal sequence is alternately inverted for each sample, and includes an alternating inversion means having one as an I-phase component and the other as a Q-phase component,
A digital quadrature detector characterized in that a complex signal in which the I-phase component and Q-phase component shifted by one sample time are output as a baseband signal at a sampling frequency four times the chip rate.

Images