![](http://datasheet.mmic.net.cn/Analog-Devices-Inc/AD9225ARS_datasheet_100564/AD9225ARS_19.png)
AD9225
–19–
and detection, among other things. See Application Note AN-302
on using this technique in digital receivers.
In direct IF down conversion applications, one exploits the
inherent sampling process of an ADC in which an IF signal lying
outside the baseband region can be aliased back into the baseband
region in a similar manner that a mixer will down-convert an IF
signal. Similar to the mixer topology, an image rejection filter
is required to limit other potential interferring signals from also
aliasing back into the ADC’s baseband region. A trade-off exists
between the complexity of this image rejection filter and the ADC’s
sample rate as well as dynamic range.
The AD9225 is well suited for various IF sampling applications.
The AD9225’s low distortion input SHA has a full-power band-
width extending beyond 130 MHz thus encompassing many
popular IF frequencies. A DNL of
±0.4 LSB (typ) combined
with low thermal input referred noise allows the AD9225 in the
2 V span to provide 69 dB of SNR for a baseband input sine
wave. Its low aperture jitter of 0.8 ps rms ensures minimum
SNR degradation at higher IF frequencies. In fact, the AD9225
is capable of still maintaining 68 dB of SNR at an IF of 71 MHz
with a 2 V input span. Although the AD9225 can yield a 1 dB
to 2 dB improvement in SNR when configured for the larger
4 V span, the 2 V span achieves the optimum full-scale distor-
tion performance at these higher input frequencies. The 2 V
span reduces only the performance requirements of the input
driver circuitry (i.e., IP3) and thus may also be more attractive
from a system implementation perspective.
Figure 26 shows a simplified schematic of the AD9225 config-
ured in an IF sampling application. To reduce the complexity of
the digital demodulator in many quadrature demodulation
applications, the IF frequency and/or sample rate are strategi-
cally selected such that the band-limited IF signal aliases back
into the center of the ADC’s baseband region (i.e., fS/4). This
demodulation technique typically reduces the complexity of the
post digital demodulator ASIC that follows the ADC.
OPTIONAL
BANDPASS
FILTER
SAW
FILTER
HIGH
LINEARITY
RF
AMPLIFIER
FROM
PREVIOUS
STAGES
MIXER
200
20
0.1 F
MINICIRCUITS
T4-6T
RF2317
RF2312
0.1 F
10 F
VINA
VINB
CML
AD9225
VREF
SENSE
REFCOM
Figure 26. Example of AD9225 IF Sampling Circuit
To maximize its distortion performance, the AD9225 is configured
in the differential mode with a 2 V span using a transformer. The
center tap of the transformer is biased at midsupply via the CML
output of the AD9225. Preceding the AD9225 and transformer is
an optional band-pass filter as well as a gain stage. A low Q passive
band-pass filter can be inserted to reduce the out-of-band distor-
tion and noise that lies within the AD9225’s 130 MHz bandwidth.
A large gain stage(s) is often required to compensate for the high
insertion losses of a SAW filter used for channel selection and
image rejection. The gain stage will also provide adequate isolation
for the SAW filter from the charge kickback currents associated
with the AD9225’s switched capacitor input stage.
The distortion and noise performance of an ADC at the given IF
frequency is of particular concern when evaluating an ADC for a
narrowband IF sampling application. Both single-tone and dual-
tone SFDR versus amplitude are very useful in assessing an
ADC’s dynamic and static nonlinearities. SNR versus amplitude
performance at the given IF is useful in assessing the ADC’s noise
performance and noise contribution due to aperture jitter. In any
application, one is advised to test several units of the same device
under the same conditions to evaluate the given applications sensi-
tivity to that particular device.
Figures 27 to 30 combine the dual-tone SFDR as well as single-
tone SFDR and SNR performances at IF frequencies of 35 MHz,
45 MHz, 70 MHz, and 85 MHz. Note that the SFDR versus
amplitude data is referenced to dBFS while the single-tone SNR
data is referenced to dBc. The performance characteristics in these
figures are representative of the AD9225 without any preceding
gain stage. The AD9225 was operated in the differential mode (via
transformer) with a 2 V span and a sample rate between 28 MSPS
and 36 MSPS. The analog supply (AVDD) and the digital supply
(DRVDD) were set to 5 V and 3.3. V, respectively.
AIN (dBFS)
90
95
–15
0
SNR/SFDR
(dBFS)
–10
–5
85
70
55
50
100
80
75
65
60
SFDR
DUAL-TONE
(dBFS)
SNR
SINGLE-TONE
(dBc)
SFDR
SINGLE-TONE
(dBFS)
Figure 27. IF Undersampling at 35 MHz (F1 = 34.63 MHz,
F2 = 35.43 MHz, CLOCK = 20 MHz)
AIN (dBFS)
90
95
–15
0
SNR/SFDR
(dBFS)
–10
–5
85
70
55
50
100
80
75
65
60
SFDR
DUAL-TONE
(dBFS)
SNR
SINGLE-TONE
(dBc)
SFDR
SINGLE-TONE
(dBFS)
Figure 28. IF Undersampling at 45 MHz (F1 = 44.81 MHz,
F2 = 45.23 MHz, CLOCK = 20 MHz)
Rev. C