Coded data generation or conversion – Analog to or from digital conversion – Analog to digital conversion
Reexamination Certificate
2002-07-29
2003-11-25
JeanGlaude, Jean Bruner (Department: 2819)
Coded data generation or conversion
Analog to or from digital conversion
Analog to digital conversion
C341S144000
Reexamination Certificate
active
06653964
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to a filtering method and an A/D conversion apparatus having a filtering function, for generating digital data from an analog input signal after removing unnecessary high frequency signal components therefrom.
2. Description of the Related Art
Conventionally, in a control unit consisting of a microcomputer or the like, there has been used an A/D conversion apparatus constructed of an analog CR filter
3
, an A/D converter
4
, and a digital moving average filter
6
as shown in
FIG. 12
a
, for example, in order to fetch detection signals (analog input signals Vin) from various sensors that detect an operational status of an object to be controlled.
The digital moving average filter
6
is a unit that removes unnecessary noise components (high-frequency signal components) from digital data Dad that occur as a result of A/D conversion by the A/D converter
4
. For example, as shown in
FIG. 12
c
, the digital moving average filter
6
is structured such that latch circuits LT at a plurality of stages operate in synchronism with a clock CKSD at a constant period to sequentially latch the digital data Dad from the A/D converter
4
. Adder circuits ADD add outputs from these latch circuits LT, thereby to carry out moving average processing of the digital data Dad.
In other words, in the digital moving average filter
6
, the latch circuits LT sequentially sample the digital data Dad at a plurality of stages in synchronism with the clock CKSD. The adder circuits ADD add the digital data Dad that have been sampled a plurality of times in the past, and average the digital data Dad. The averaged result is output as digital data DT that expresses a true A/D conversion result.
The digital moving average filter
6
may be realized by arithmetic processing (what is called a smoothing processing) of a microcomputer that constitutes the control unit, instead of the digital circuit as shown in
FIG. 12
c.
On the other hand, when the A/D converter
4
carries out A/D conversion of the analog input signal Vin, an aliasing phenomenon of high-frequency signal components occurs, when the frequency of the analog input signal Vin becomes equal to or above one half of a sampling frequency fad of the A/D converter
4
. Therefore, the analog CR filter
3
is provided as what is called a pre-filter at a pre-stage of the A/D converter
4
, in order to remove the frequency components of not less than one half of the sampling frequency fad of the A/D converter
4
from the analog input signal Vin.
The analog CR filter
3
is structured as shown in
FIG. 12
b
. The analog CR filter
3
has a non-inversion input terminal (+) of an operation amplifier OP
1
grounded to the earth, has a capacitor C
1
and a resistor RI connected in parallel between an inversion input terminal (−) of the operation amplifier OP
1
and an output terminal, and has an input signal (analog) input to the inversion input terminal (−) of the operation amplifier OP
1
via a resistor R
2
.
In other words, the analog CR filter
3
integrates the analog input signal Vin based on the resistances of resistors R
1
and R
2
, the capacity of a capacitor C
1
, and the determined time constant. Based on this integration, the analog CR filter
3
limits the frequency of the analog input signal Vin input to the A/D converter
4
to less than one half of the sampling frequency fad of the A/D converter
4
according to a known “sampling theorem”.
According to the A/D conversion apparatus that has the conventional filtering function having the above structure, the attenuation of a signal based on the digital moving average filter
6
becomes extremely small (approximately zero) at each of the frequencies of n times (where n is a positive integer including 1) the sampling frequency fsd (the frequency of the clock CKSD) (hereinafter referred to as “frequency range”). Unnecessary signal components pass through in this frequency range. Therefore, it is necessary to set the frequency versus attenuation characteristics of the analog CR filter
3
to be used as the pre-filter, such that the attenuation changes as steeply as possible in the frequency range in which the frequency exceeds the cutoff frequency. For this purpose, it has been necessary to increase the order of the analog CR filter
3
, and lower the cutoff frequency.
FIG. 13
a
shows the frequency versus attenuation characteristics of the digital moving average filter
6
that carries out the moving average processing twice to average the digital data for the past sixteen times, with the sampling frequency fsd set to 100 kHz. As is clear from this diagram, according to the digital moving average filter
6
that uses the sampling frequency fsd as 100 kHz, the attenuation becomes approximately zero in the frequency range in which the frequency is n times the sampling frequency fsd like 100 kHz, 200 kHz, 300 kHz, etc. Therefore, it is not possible to remove unnecessary signal components in this frequency area.
For the above reason, when a low-order filter (specifically, a CR linear filter) is used for the analog CR filter
3
, for example, it is not possible to make the signal attenuation sufficiently large in a high-frequency signal passing range (a frequency range of n×fsd) that is generated by the digital moving average filter
6
as shown by a one-point chain line in
FIG. 13
b
. As a result, unnecessary high-frequency signal components pass through this frequency range.
Therefore, in order to prevent the above problem, it is necessary to increase the order of the analog CR filter
3
to as high a level as possible, and the analog CR filter
3
attenuates the high-frequency signal components that cannot be attenuated by the digital moving average filter
6
. For this purpose, it is necessary to connect filters, as shown in
FIG. 12
b
, at many stages. This leads to an increase in the size of the analog CR filter
3
and the size of the A/D conversion apparatus. Further, there occurs a problem of an increase in the cost of the A/D conversion apparatus.
Further, in order to make the analog CR filter
3
attenuate the high-frequency signal components that cannot be attenuated by the digital moving average filter
6
, there is a method of lowering the cutoff frequency of the analog CR filter
3
. However, in order to lower the cutoff frequency, it is necessary to increase the capacity of the capacitor C and the resistance of the resistor R that constitute the analog CR filter
3
. This measure also has a problem of increasing the sizes of the analog CR filter
3
and the A/D conversion apparatus. Further there occurs a problem of an increase in the cost of the A/D conversion apparatus.
In the frequency versus attenuation characteristics of the digital moving average filter shown in
FIGS. 13
a
and
13
b
, the attenuation becomes a maximum (infinite) at predetermined frequency intervals at a lower frequency side than the sampling frequency fsd (100 kHz) of the digital moving average filter
6
. This is because the digital moving average filter
6
executes the moving average processing, and the number of peaks of the attenuation corresponds to the number of digital data to be averaged. Specifically, when the number of digital data to be averaged by the digital moving average filter
6
is “2” (that is, the averaging of the data for the past two times), the number of the peaks of the attenuation becomes “1”. When the number of digital data to be averaged is “4” (that is, the averaging of the data for the past four times), the number of the peaks of the attenuation becomes “3”.
In
FIGS. 13
a
and
13
b
, the abscissa shows frequency as a logarithm, and the ordinate linearly displays attenuation. Therefore, it is difficult to know from these diagrams the number of peaks of attenuation generated in the frequency area of the sampling frequency 100 kHz and below. Actually, the number of digital data to be averaged by the digital moving average filter
6
is “16”. Therefore, the peak number of att
Mizuno Tamotsu
Watanabe Takamoto
JeanGlaude Jean Bruner
Lauture Joseph
LandOfFree
Filtering method and A/D conversion apparatus having... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Filtering method and A/D conversion apparatus having..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Filtering method and A/D conversion apparatus having... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3144505