CS5101A CS5102A
Count
8192
Count
8192
6144
4096
2048
6144
4096
2048
Noiseless
Converter
Noiseless
Converter
CS5101A
CS5102A
7FFB 7FFC 7FFD 7FFE 7FFF 8000 8001
Code (Hexadecimal)
989 6359 844
7FFD 7FFE 7FFF 8000(H) 8001
Code (Hexadecimal)
8002 8003
Counts:
0
0
0
0
Counts:
0
5
1727 4988 1467
5
0
Figure 21. 5101A Histogram Plot of 8192 Conversion
Inputs
Figure 22. 5102A Histogram Plot of 8192 Conversion
Inputs
shows a histogram plot of output code occur-
rences obtained from 8192 samples taken from a
CS5101A in the bipolar mode. Hexadecimal code
7FFE was arbitrarily selected and the analog in-
put was set close to code center. With a noiseless
converter, code 7FFE would always appear. The
histogram plot of the device has a "bell" shape
with all codes other than 7FFE due to internal
noise. Figure 22 illustrates the noise histogram of
the CS5102A.
averaging multiple samples for each word. Over-
sampling spreads the device’s noise over a wider
band (for lower noise density), and averaging ap-
plies a low-pass response which filters noise
above the desired signal bandwidth. In general,
the device’s noise performance can be maximized
in any application by always sampling at the
maximum specified rate of 100 kHz (CS5101A)
or 20 kHz (CS5102A) (for lowest noise density)
and digitally filtering to the desired signal band-
width.
In a sampled data system all information about
the analog input applied to the sample/hold ap-
pears in the baseband from dc to one-half the
sampling rate. This includes high-frequency com-
ponents which alias into the baseband. Low-pass
(anti-alias) filters are therefore used to remove
frequency components in the input signal which
are above one-half the sample rate. However, all
wideband noise introduced by the CS5101A and
CS5102A still aliases into the baseband. This
"white" noise is evenly spread from dc to one-
half the sampling rate and integrates to 35 µV rms
in unipolar mode.
Aperture Jitter
Track-and-hold amplifiers commonly exhibit two
types of aperture jitter. The first, more appropri-
ately termed "aperture window", is an input
voltage dependent variation in the aperture delay.
Its signal-dependency causes distortion at high
frequencies. The proprietary architecture of the
CS5101A and CS5102A avoids applying the in-
put voltage across a sampling switch, thus
avoiding any "aperture window" effects. The sec-
ond type of aperture jitter, due to component
noise, assumes a random nature. With only
100 ps peak-to-peak aperture jitter, the CS5101A
and CS5102A can process full-scale signals up to
Noise in the digital domain can be reduced by
sampling at higher than the desired word rate and
28
DS45F2
CIRRUS [ CIRRUS LOGIC ]
