The present invention relates to a signal processing device, a signal processing method, a delta-sigma modulation type fractional frequency division PLL frequency synthesizer, a radio communication device and a delta-sigma modulation type D/A converter.
A delta-sigma (ΔΣ) modulator has a circuit configuration which performs feedback of quantization noise generated in an output to an input via a delayer and sometimes is called “sigma-delta (ΣΔ) modulator” or “noise shaper” because of its function of biasing quantization noise to the high frequency band.
When a frequency synthesizer including a phase locked loop (PLL) is used for a radio communication device such as a cellular phone, in order to ensure many available bands, it is required to change an output frequency with a smaller step size than the frequency of a reference signal. As a frequency synthesizer to meet this requirement, a ΔΣ modulation fractional frequency division PLL frequency synthesizer has been known. An exemplary ΔΣ modulation fractional frequency division PLL frequency synthesizer is described in U.S. Pat. No. 5,070,310. In the PLL frequency synthesizer, a fractional frequency divider for frequency-dividing an output of a voltage control oscillator to feedback the output to a phase comparator includes a ΔΣ modulator and a digital value F representing fraction part (non-integer part) of frequency division data is given to the ΔΣ modulator.
Moreover, a high accuracy digital/analog (D/A) converter including a ΔΣ modulator, i.e., a ΔΣ modulation D/A converter is used for an audio device and the like.
With the ΔΣ modulation fractional frequency division PLL frequency synthesizer, assuming that the frequency of the reference signal given to the phase comparator is Fref and the digital value F representing fraction part of the frequency divided data is binary data of n (n is an integer), an output frequency step size equal to Fref×(F/2n) can be achieved. However, it has been pointed out as a problem that as a result of concentration of quantization noise at a particular frequency when the ΔΣ modulator receives a particular F value (e.g., F=2n−1), a spurious signal is generated. Then, in a known manner, n takes a large value (Fref=26 MHz, n=24 in the above-described United State Patent) and F, which may be a problem, is substituted by F+1 or F−1. Accordingly, two problems, i.e., (1) a problem in which circuit scale is increased and (2) a problem in which an output frequency is slightly shifted from a desired frequency, arise.
With a ΔΣ modulation D/A converter, spurious problems arise such as those described above, which depends on a digital input of ΔΣ modulator.
An object of the present invention is to suppress concentration of quantization noise at a particular frequency.
To achieve the above-described object, the present invention uses a signal processing device which has a configuration including, in addition to a delta-sigma modulator, a dither circuit, located between a digital input and the delta-sigma modulator, for selectively supplying a digital signal which has been discretely changed from the digital input and of which a time average corresponds to the digital input. Thus, even if a bit width of the digital input is not increased, concentration of quantization noise at a particular frequency can be suppressed.
The signal processing device can be applied to a fractional frequency division PLL frequency synthesizer, an A/D converter, a radio communication device and the like.
According to the present invention, a dither circuit, located between a digital input and the delta-sigma modulator, for selectively supplying a digital signal which has been discretely changed from the digital input and of which a time average corresponds to the digital input is adopted. Thus, even if a bit width of the digital input is not increased, concentration of quantization noise at a particular frequency can be suppressed. Therefore, a known spurious problem can be dissolved and also a desired output frequency can be obtained.
FIGS. 4(a) and 4(b) are timing charts describing the operation of the digital dither circuit shown in
In the ΔΣ modulation fractional frequency division PLL frequency synthesizer 2 of
Next, the operation of the fractional frequency divider 28 of
The frequency division number of the output signal Fo of the voltage control oscillator 27 is (P+1)×A until the A counter 36 outputs an output and is P×(N−A) until the N counter 37 outputs an output. Therefore, if respective frequencies of an output signal and a reference signal are assumed to be Fo and Fref, the following equation holds.
Even when P=2n (n is an integer), the number of available bands can be increased by changing A in Equation 1.
To further increase the number of available bands, the ΔΣ modulator 33 is provided. Moreover, to dissolve the spurious problem, the digital dither circuit 32 is provided between the latch 31 and the ΔΣ modulator 33. The latch 31 holds given frequency division data DATA. Note that in
Fo=((P×N+A)+F/2n)×Fref [Equation 2]
Thus, an output frequency step size equal to Fref×(F/2n) is achieved. That is, in a normal operation state, an average frequency Fo of an output signal can be changed with a smaller step size than the frequency Fref of the reference signal, so that the reference frequency Fref can be set at a large level. Thus, a PLL frequency synthesizer having excellent lockup characteristics can be obtained.
The selection circuit 44 outputs Fdiv as EFdiv when the SELECT signal is high, and outputs a fixed value as EFdiv when the SELECT signal is low.
The ½ frequency divider 41 generates a clock signal DFdiv having a half frequency of a frequency of the comparative signal EFdiv from the selection circuit 44.
The selector 42 receives a clock signal DFdiv from the ½ frequency divider 41 at an S input, and selects a positive constant value or a negative constant value alternately in a manner in which a positive cohstant value [+k (A input)] is selected when the logic level of the S input is low, and a negative constant value [−k (13 input)] is selected when the logic level of the S input is high.
The adder 43 receives the F value from the latch 31 at the A input and the constant value [±k] from the selector 42 at the B input, and performs addition operation A+B when a rise pulse of the comparator signal EFdiv is given as a CK input to periodically change a Y output to F+k or F−k.
The selector 45 receives the F value from the latch 31 at the A input, the Y output from the adder 43 at the B input, and the SELECT signal at the S input. The selector 45 selects the F value (A input) when the logic level of the S signal, i.e., the SELECT signal is low, and selects as the Y output the Y output (B input) of the adder 43 when the logic level of the SELECT signal is high.
As has been described, the Y output of the selector 45, i.e., F+k or F is finally supplied to the ΔΣ modulator 33 by the SELECT signal. Change of the SELECT signal will be described later. FIGS. 4(a) and 4(b) illustrate the operation of the digital dither circuit 32.
From
Note that the configuration of the digital dither circuit 32 is not limited to that of
SELECT signal change is performed so as to randomly obtain F+k and F−k in unspecific periodical intervals and output a digital signal of which the time average corresponds to the F value to the ΔΣ modulator 33 only when the F value is a particular value (e.g., F=2n−1, 2n−2 and so on), and the F value itself to the ΔΣ modulator 33 when the F value is a value other than the particular value in order to suppress spurious which occurs as a result of concentration of quantization noise in a particular frequency. Specifically, the SELECT signal is changed to be high when the given F value is a particular value and the SELECT signal is changed to low when the given F value is a value other than the particular value. As a result, the digital dither circuit 32 outputs a digital value which has been discretely changed from the F value and of which the time average corresponds to the F value to the ΔΣ modulator 33 only when the given F value is the particular value (e.g., F=2n−1, 2n−2 and so on), and outputs the F value, as it is, to the ΔΣ modulator 33 when the given F value is a value other than the particular value. Thus, the generation of spurious to an output of the voltage control oscillator 27 in a particular frequency division ration is suppressed, so that the same characteristics as those of the known ΔΣ modulation D/A converter can be achieved in a frequency division ratio other than the particular frequency division ratio.
With the ΔΣ modulation D/A converter 50 of
Number | Date | Country | Kind |
---|---|---|---|
2002-260088 | Sep 2002 | JP | national |
Filing Document | Filing Date | Country | Kind |
---|---|---|---|
PCT/JP03/10885 | 8/27/2003 | WO |