&Dgr;&Sgr; modulator, DA converter and AD converter

Details &Dgr;&Sgr; modulator, DA converter and AD converter &Dgr;&Sgr; modulator, DA converter and AD converter

1999-11-10

2001-11-27

JeanPierre, Peguy (Department: 2819)

Coded data generation or conversion

Analog to or from digital conversion

Differential encoder and/or decoder

C341S144000

Reexamination Certificate

active

06323794

ABSTRACT:

BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to a &Dgr;&Sgr; (delta-sigma) modulator, DA converter and AD converter, and, more particularly, relates to an improvement with which a noise-shaping equal to that of a multi-bit higher-order modulator required for improved accuracy is realized on a small-scale circuit while retaining stability.
2. Description of the Background Art
Presently, converters equipped with &Dgr;&Sgr; modulator are widely used as a high-precision DA converter or AD converter having a resolution over 14 bits. In a DA or AD converter having a &Dgr;&Sgr; modulator, a quantization with a small number of bit, such as 1-bit quantization, is performed while a folding noise and quantization noise are reduced by oversampling (for example, when handling voice signals, sampling is performed by an-integral multiple of the inherent sampling frequency, e.g., 44.1 kHz) and noise-shaping (i.e., the frequency characteristic of a quantization noise inherently flat is more enhanced at a higher frequency band).
FIGS. 13 and 14
are block diagrams showing the configurations of conventional DA converter and AD converter, each having a &Dgr;&Sgr; modulator, respectively. In the DA converter
151
shown in
FIG. 13
, an inputted digital signal ID is converted while it passes through an interpolation filter
70
, &Dgr;&Sgr; modulator
71
, internal DA conversion circuit and lowpass filter
73
in this order, and then outputted as an analog signal OA. The &Dgr;&Sgr; modulator
71
has in a feedback loop thereof (i.e., a signal transmission path in which a signal circulates through the input, the output, and the input) an integrator
76
, quantizer
77
and delay circuit
78
. For enabling feedback, an arithmetic operation unit
75
that serves as a substracter is provided at the intersection between the input line and feedback loop. The delay circuit
78
may be placed in anywhere within the feedback loop.
In the AD converter
152
shown in
FIG. 14
, an inputted analog signal IA is converted while it passes through an anti-aliasing filter
80
, &Dgr;&Sgr; modulator
81
and decimation filter
82
in this order, and then outputted as a digital signal OD. The &Dgr;&Sgr; modulator
81
has in a feedback loop thereof an integrator
86
, quantizer
87
, delay circuit
88
and internal DA conversion circuit
89
. For enabling feedback, an arithmetic operation unit
85
is provided at the intersection between the input line and feedback loop. The delay circuit
88
may be placed in anywhere within the feedback loop.
Since in the feedback loop of the &Dgr;&Sgr; modulator
71
, digital signals are regarded as a processing object, all the circuits within the loop are configured as a digital circuit. Whereas in the feedback loop of the &Dgr;&Sgr; modulator
81
, except for a signal transmission path from the output of the quantizer
87
to the input of the internal DA conversion circuit
89
, analog signals are regarded as a processing object. Therefore, all the circuits within the feedback loop, except for the signal transmission path from the output of the quantizer
87
to the input of the internal DA conversion circuit
89
, are configured as an analog circuit.
In the DA converter
151
and AD converter
152
, their signal conversion accuracy as a whole, depends upon the accuracy of signal conversion in the &Dgr;&Sgr; modulators
71
and
81
, respectively. To increase the conversion accuracy of the &Dgr;&Sgr; modulators
71
and
81
, there have been known conventionally a method of increasing the order (order increasing method) and a method of increasing the number of output bit of a quantizer (multi-bit method).
FIG. 15
is a block diagram showing an example where the order increasing method is applied.
FIG. 16
is a block diagram showing an example where the multi-bit method is applied in addition to the order increasing method.
A quantizer
77
of a &Dgr;&Sgr; modulator
71
a
in
FIG. 15
is an 1-bit quantizer, and a quantizer
79
of a &Dgr;&Sgr; modulator
71
b
in
FIG. 16
is a multi-bit quantizer. In both &Dgr;&Sgr; modulators
71
a
and
71
b
, their feedback loops are branched into four, and one to four integrators
76
are interposed in each branch. That is, the &Dgr;&Sgr; modulator
71
a
is configured as an 1-bit fourth-order modulator, and the &Dgr;&Sgr; modulator
71
b
is configured as a multi-bit fourth-order modulator. The relationship between input signal X and output signal Y in the &Dgr;&Sgr; modulators
71
a
and
71
b
is expressed by using z conversion which is effective in describing a sample data control system, as follows:
Signals A to C, and Y in
FIGS. 15 and 16
are expressed by the following formulas 1 to 4, respectively:
A
=(
X−z
−1
Y
)/(1
−z
−1
)  (1);
B
=(
A−z
−1
Y
)/(1
−z
−1
)  (2);
C
=(
B−z
−1
Y
)/(1
−z
−1
)  (3);
and
Y
=(
C−z
−1
Y
)/(1
−z
−1
)+
e
  (4)
wherein e is a quantizing error in the quantizer
77
or
79
. From these formulas, there is obtained:
Y=X+
(1
−z
−1
)
4
e
  (5)
Formula 5 shows that in the &Dgr;&Sgr; modulators
71
a
and
71
b
, a quantization noise e contributes to an output signal Y in the form proportional to the fourth-order differential coefficient of fourth order of the quantization noise e. Since the frequency characteristic of the quantization noise e is flat, the frequency characteristic of the fourth-order differential coefficient of the quantization noise e is expressed by the fourth-order function of frequency. Specifically, Formula 5 shows that, according to the fourth-order function, the influence of the quantization noise e on the output signal Y increases as frequency increases.
Thus in a higher-order modulator, the quantization noise e is minimized in a low frequency band whereas it is enhanced in a high frequency band, based on the function of higher order of frequency. Namely, noise-shaping is enhanced based on the function of fourth order of frequency. From the interpolation filter
70
, the input signal X obtained by oversampling is inputted to the &Dgr;&Sgr; modulator
71
, thereby the quantization noise e in the inherently required low frequency band (e.g., a band of not more than 44.1 kHz is required in voice signal processing) can be reduced effectively.
In addition, the quantization noise e itself is reduced by multi-bit method. Specifically, the quantization noise e reduces as the number of output bit of a quantizer is higher. In the &Dgr;&Sgr; modulator
71
b
, the number of bit of the quantizer
79
is larger than that of the quantizer
77
in the &Dgr;&Sgr; modulator
71
a
. Thereby, the quantization noise component in Formula 5 is lowered by that amount in the &Dgr;&Sgr; modulator
71
b.
The foregoing is true for the &Dgr;&Sgr; modulator
81
as well.
FIG. 17
is a block diagram illustrating an example where the order increasing method and multi-bit method are applied to the &Dgr;&Sgr; modulator
81
, to obtain a &Dgr;&Sgr; modulator
81
a
. A quantizer
90
of the &Dgr;&Sgr; modulator
81
a
in
FIG. 17
is a multi-bit quantizer. The feedback loop of the &Dgr;&Sgr; modulator
81
a
is branched into two, and one to two integrators
86
are interposed in each branch. That is, the &Dgr;&Sgr; modulator
81
a
is configured as a multi-bit second-order modulator. The relationship between input signal X and output signal Y in the &Dgr;&Sgr; modulator
81
a
is expressed as follows:
Signals D and Y in
FIG. 17
are expressed as follows:
D=
1/(1
−z
−1
)·(
X−z
−1
Y−e
D
)  (6);
and
Y=
1/(1
−z
−1
)·(
D−z
−1
Y−e
D
)+
e
  (7),
respectively,
wherein e is a quantization error in the quantizer
90
, and e
D
is a conversion error in the internal DA conversion circuit
89
.
From Formulas 6 and 7, the signal Y is expressed as follows:
Y=X−e
D
−(1
−z
&minu

Affiliated with

Also associated with

Rating

&Dgr;&Sgr; modulator, DA converter and AD converter does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with &Dgr;&Sgr; modulator, DA converter and AD converter, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and &Dgr;&Sgr; modulator, DA converter and AD converter will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-2575862