English
Language : 

DDC101 Datasheet, PDF (20/29 Pages) Burr-Brown (TI) – 20-BIT ANALOG-TO-DIGITAL CONVERTER
0
–5
–10
–15
–20
–25
–30
–35
–40
1k
fOS 10k
100k
1M
Frequency (Hz)
0
–5
–10
–15
–20
–25
–30
–35
–40
1
N = 1000
L = 64
fCONV = 31Hz
10
fCONV
100
1k
Frequency (Hz)
FIGURE 17. Oversampling Frequency Response for M = 256
(fCLK = 2MHz).
FIGURE 18. A Multiple Integration Frequency Response
Example.
Multiple Integration Frequency Response
If the DDC101 is operated in the multiple integrations per
conversion mode of operation, an additional sin(x)/x type
low pass filter is created. The filter creates an initial null
frequency at the conversion frequency, fCONV of the DDC101
and at multiples of fCONV. The –3dB point for this filter is also
at fCONV/2.26. The conversion time, TCONV, is the sum of the
integration times for multiple integrations that are averaged
together by the DDC101. TCONV = LN/fCLK. fCONV = l/TCONV.
If multiple integrations per conversion are used, this filter
will be the dominant low frequency filter of the DDC101.
Table VIII shows examples of the conversion time and
frequency for different parameter selections. Figure 18 shows
an example of the frequency response due to Multiple
Integrations. In the case of Figure 18, the integration time is
500µs (N = 1000 clock periods) and L = 64 integrations per
conversion.
INTEGRATION
TIME
1ms
1ms
1ms
1ms
1ms
10ms
10ms
10ms
10ms
10ms
CONVERSION –3dB
L
TIME FREQUENCY
fCONV
2
2ms
221Hz
500Hz
8
8ms
55Hz
125Hz
16
16ms
27.5Hz
62.5Hz
64
64ms
6.9Hz
15.6Hz
256
256ms
1.73Hz
3.91Hz
2
20ms
22.1Hz
50.0Hz
8
80ms
5.5Hz
12.5Hz
16
160ms
2.75Hz
6.25Hz
64
640ms
0.69Hz
1.56Hz
256
2560ms
0.173Hz
0.39Hz
TABLE VIII. Multiple Integration Time Examples.
System Noise implications
The noise at the digital output of the DDC101 consists of
system noise that is included in the analog input signal and
noise from the DDC101.
DDC101 Noise—The noise of the DDC101 includes low
frequency and broadband noise. The low frequency noise is
reduced by the integrating function and the CDS function of
the DDC101. This is reflected in the basic integration
frequency response and in the multiple integration frequency
response. The broadband electronic noise is reduced prima-
rily by the oversampling function of the DDC101
Signal Noise—The noise of the input signal is filtered and
reduced in a manner similar to the DDC101 noise reduction
through the integrating and oversampling functions of the
DDC101.
Figures 19 and 20 show the frequency response of the
DDC101 for the product of the basic integration and
oversampling frequency response for two different values of
M. In both examples, the integration time is 1ms, the only
difference is in the number of oversamples, M; for Figure
19, M = 256 oversamples was used; for Figure 20, M = 32
oversamples was used. The first null frequency is fMEAS and
subsequent nulls are at multiples of fMEAS. The first example
with the larger number of oversamples (M = 256) clearly
reduces high frequency noise more than the second example
with M = 32.
For M = 256, fOS is 7.8kHz, fMEAS is 1.16kHz, and the –3dB
frequency is 507Hz. For M = 32, fOS is 62.4kHz, fMEAS is
1.02kHz and the –3dB frequency is 453Hz.
0
–5
–10
–15
–20
–25
–30
–35
–40
100
N = 2000
M = 256
K = 16
1k
10k
Frequency (Hz)
100k
FIGURE 19. Product of Frequency Response of Basic Inte-
gration and Oversampling: 1ms Integration
Time, 256 Oversamples.
®
DDC101
20