AN-851

APPLICATION NOTE
One Technology Way • P.O. Box 9106 • Norwood, MA 02062-9106, U.S.A. • Tel: 781.329.4700 • Fax: 781.461.3113 • www.analog.com

A WiMax Double Downconversion IF Sampling Receiver Design

by Eric Newman and Cecile Masse
INTRODUCTION station is not expected to decode a 0 dBm input level success-
fully, the equipment must be able to withstand the large 0 dBm
This application note describes an intermediate frequency (IF)
input without damage.
sampling receiver intended for use in the wireless communica-
tion services (WCS) band from 2.3 GHz to 2.36 GHz and the The 802.16 standard describes an adjacent channel rejection
unlicensed ISM band from 2.4 GHz to 2.48 GHz. The receiver requirement for the receiver. The following adjacent and
is designed for broadband Orthogonal Frequency Division nonadjacent channel interferer to desired channel power ratios
Multiplexing (OFDM) systems, as described according to IEEE must be achieved while maintaining a 10−6 bit error rate (BER)
802.16 standards documentation. The design approach and while the desired signal is no more than 3 dB above the
implementation procedures are presented to allow designers to specified reference sensitivity.
easily modify the receiver chain to address other bands such as
Table 1. Adjacent and Nonadjacent Channel Rejection
Wireless Broadband (WiBro) and other cellular standards such
Requirements as Described in the 802.16-2004 Standard
as TDS-CDMA. The presented design is capable of addressing
Nonadjacent
channel bandwidths as wide as 10 MHz using the depicted Adjacent Channel Channel
surface acoustic wave (SAW) filter; however, larger bandwidths Modulation/Coding Rejection (dB) Rejection (dB)
can be addressed using wider bandwidth channel-select filters 16-QAM-¾ −11 −30
and increased sampling rates. 64-QAM-¾ −4 −23
BACKGROUND
ARCHITECTURE
OFDM is the modulation scheme used in the next emerging
Figure 1 depicts a classic double downconversion IF sampling
WiMax standard. OFDM utilizes multiple subcarriers of various IQ
receiver. IF sampling architectures are quite appropriate when
modulations to achieve a high aggregate data rate that has inherent
dealing with large signal bandwidths as employed in WiMax, or
immunity to multipath propagation. The baseband information is
even multicarrier systems. By using multiple down conversions,
spread among subcarriers so that little information is lost if
it is possible to employ several channel select filters that help to
multiple propagation paths result in destructive interference and
improve the selectivity of the receiver and improve immunity to
heavily attenuate a portion of the transmitted spectrum. The
blocking signals that would otherwise degrade receiver sensitiv-
variability in the subcarrier modulation schemes allows for an
ity. The double downconversion also allows the use of a high
adaptive signaling approach that propagates lower data rates at the
enough first IF where the image frequency band falls out of the
greatest distances, while the highest data rates can be utilized when
passband of the front-end, band-select RF filter.
the received signal-to-noise ratio (SNR) is high.
The receiver architecture described in this application note is
The variety of subcarrier modulation schemes and coding
based on a 14-bit ADC. A 12-bit ADC could be used to address
results in different SNR requirements at the receiver. The
the 802.16 receiver requirements, although it is recommended
reference sensitivity of a WiMax receiver is defined in the IEEE
to use a 14-bit ADC for single-down conversion or for multicar-
Std. 802.16-2004 to be −91 dBm for a 1.5 MHz channel using ½
rier architectures in order to compensate for the less efficient
coding rate QPSK, and −65 dBm for a 20 MHz channel using ¾
selectivity and avoid the saturation of the ADC in presence of
coding rate 64-QAM. This results in a receiver NF requirement
high interferer levels. In order to design a receiver capable of
of 7 dB with 5 dB of implementation margin.
successfully addressing the multiple data rate options available,
The 802.16 standard defines a maximum input power level of it is necessary to select IF center frequencies carefully and
−30 dBm for successful detection, with a maximum tolerated ensure that appropriate SAW filter selections are available.
power level of 0 dBm. While the base station or subscriber

Rev. 0 | Page 1 of 8
AN-851
ADF4153 ADF4360-8

~100Ω ~1kΩ

BAND ATF541M4 IMAGE ATF541M4 SAWTEK LCR AD8370 LCR

LO1 LO2 AD9246 FULL SCALE =
SELECT REJECT 855898 LPF AAF
–3dBV rms
LNA1 LNA2 IF DGA ADC en = 21nV/ Hz
ADL5350 AD8344 IIP3 = 33dBV rms
AV = –1.5dB AV = –2.5dB AV = –8dB AV = –1dB AV = +10dB AV = 0dB
AV = +14.5dB AV = +14.5dB AV = –7.5dB AV = +11.5dB NF = 20.5dB
AV = –15dB TO +34dB

06147-001
NF = 0.9dB NF = 0.9dB NF = 7.5dB NF = 8dB IIP3 = 33dBm
IIP3 = 18dBm IIP3 = 18dBm IIP3 = 24dBm IIP3 = 12dBm NF = 8dB (AT MAX GAIN)
OIP3 = 35dBm

Figure 1. Double Downconversion IF Sampling Receiver Chain

Figure 2 shows the spurious trajectories for a downconverting Next it is necessary to consider the final IF. The second down
mixer using low-side LO injection. In order to cover the converting mixer has the benefit of a narrower input band-
2.3 GHz to 2.4 GHz bands while minimizing the number of width, typically no more than 20 MHz. This allows several
spurious mixer components that can interfere with the desired possible second IFs to be considered, with ranges from 10 MHz
signal, inspection of the spurious trajectories indicates that the to 70 MHz and higher IF frequency bands centered at 107 MHz
first IF should be within a range of 210 MHz to 400 MHz. The and 140 MHz. A 70 MHz IF is selected in order to avoid higher
grey shaded area indicates the fractional bandwidth bounded by order LO harmonics that can leak back to the antenna receive
fRF/fLO and fIF/fLO for a 374 MHz IF using a low-side LO (LO1 = port and fall in the desired receive band. Additionally, the
1926 MHz to 2026 MHz). It is possible to substitute other IF 70 MHz IF allows for a wide selection of commonplace SAW
center frequencies, but caution must be taken to ensure that the filters to be applied, or alternatively, practically lumped element
spurious responses of the first mixer do not cause in-channel LC passive filters.
interference. For this design, a first IF of 374 MHz is selected.
There are several commercially available SAW filters of various After establishing the target IF frequencies it is possible to
bandwidths from Sawtek, as well as other manufacturers, that consider the cascaded performance. Table 2 depicts the
address this center frequency. anticipated dynamic performance of the double downconver-
sion receiver from the output of the band-select filter down to
1.0
the ADC interface.
L

-2L
4R-4

0.9
3L

2R

Table 2. Anticipated Dynamic Performance of Receiver

3R-

L
0.8 R-
Signal Chain depicted in Figure 1
5L-

-4L

-5L 3L

0.7 Gain (dB) IIP3 (dBm) NF (dB)

-
3R

3R

2R
4L

0.6 65.5 −32.8 2.6

-2

3R
R

58.5 −26 2.7

fLO

0.5
fIF

2L
3L

- R 48.5 −14.5 3.5

-2R

0.4
37 −7.7 4.6
0.3 1.0
23.5 −5.6 9.7
4L-

0.2
3R

0.1

0
1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0
fRF
06147-002

fLO

Figure 2. Spurious Trajectories for a Down Converting Mixer Using Low-Side

LO Injection (Note that the grey box indicates the fractional bandwidth
boundaries for a 374 MHz IF using low-side LO injection for the 2.3 GHz to
2.4 GHz WiBro frequency band.)

Rev. 0 | Page 2 of 8
 AN-851

IMPLEMENTATION
25 5
The first component in the receiver chain is a band-select filter.
Several manufacturers, such as Anatech Electronics, Inc., K & L
Microwave Inc., and Digital Communications Inc., provide 20 4

cavity- and ceramic-based filters that address the WCS and

unlicensed ISM frequency bands utilized for 2.3 GHz to 15 3

GAIN (dB)
2.5 GHz WiMax deployments. The cavity-based filters are

NF (dB)
capable of low insertion loss of less than 2 dB with up to 60 dB
10 2
of stop-band rejection at only 25 MHz offset from the center of
the pass band. The lower cost ceramic-based filters offer closer
to ~50 dB of stop-band rejection at 150 MHz offsets. The choice 5 1
of band select filter strongly depends on the image rejection
performance of the receiver and the expected magnitude of the
0 0

06147-004
interfering signals present in the vicinity of the band of 2100 2200 2300 2400 2500 2600
FREQUENCY (MHz)
operation. For this demonstration, a Digital Communications
Inc. 10-section band-pass cavity filter is selected at a center Figure 4. ATF541M4 Gain and NF, Input Return Loss is Measured to be −8 dB
with an IIP3 of 18 dBm
frequency of 2350 MHz.
In order to help improve the image rejection of the front-end
The first LNA stage follows the front-end, band-select filter. The
design and to minimize the presented broadband noise to the
insertion loss of the filter and the noise figure of the first stage
first mixer, a simple high-pass filter was used between the first
dominate the overall cascaded sensitivity of the receiver. As a
and second LNA. The rejection requirement is not very critical
result, it is critical to achieve very low noise figure in the first
considering the first band-select filter provides >60 dB rejection
LNA stage. The ATF541M4 GaAs E-mode pHEMT was selected
of unwanted signals appearing at the image frequency. Using a
from Agilent Technologies for its low noise and high OIP3. The
374 MHz IF with low-side LO injection, the fundamental image
circuit implementation for the first and second stage LNAs is
band is from 1552 MHz to 1652 MHz for a 2.3 GHz to 2.4 GHz
presented in Figure 3, along with the measured gain and NF
input frequency range. A simple 3-pole lumped element filter
performance in Figure 4.
was designed between the first and second LNA stages. The
5V
filter provides better than 20 dB of image band rejection and
10nF less than 2.5 dB of insertion loss in the desired pass band. The
1nF
measured frequency response and circuit implementation is
32.4Ω depicted in Figure 5.
294Ω 1.21kΩ

8.2pF 0
50Ω 8.2pF
–5
5.6nH
1.5nH –10
1.5nH
–15
8.2pF
3.3pF
06147-003

0.6pF –20
S21 (dB)

–25
Figure 3. Basic LNA Implementation Schematic
–30 4.3nH

–35

–40 2.2pF
2.2nH 2.2nH
–45

–50
06147-005

0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000
FREQUENCY (MHz)

Figure 5. Simple Lumped Element Image Reject Filter Implementation and

Measured Response

Rev. 0 | Page 3 of 8
AN-851
30 12
After the second LNA stage, the signal is downconverted to a
IIP3
fixed high IF of 374 MHz using the ADL5350 mixer. The 25 9
ADL5350 is a single-ended passive mixer with an onboard LO
20 6
buffer amplifier. The mixer relies on off-chip filtering networks to

CONVERSION GAIN (dB)

IIP3, IP1dB (dBm)
isolate the RF and IF ports. The subcircuit depicted in Figure 7 15 3
includes the ADL5350 mixer, external filter networks, and an IP1dB
10 0
external LO buffer to provide sufficient drive into the LOIN pin.
5 –3
The additional AD8353 gain stage is needed to amplify the
output power of the PLL/VCO synthesizer circuit to ~4 dBm 0 GAIN –6

drive level at the LOIN pin. There is additional filtering applied –5 –9

to the LO input to help reduce harmonic content coming from
the LO synthesizer. The filter is a simple 3-pole, low-pass filter –10 –12

06147-007
2200 2250 2300 2350 2400 2450 2500
constructed using lumped element components. Without the RF FREQUENCY (MHz)
filter the second harmonic of the LO would cause some addi-
Figure 6. Measured Performance of ADL5350 Mixer
tional high frequency noise power to be downconverted to the
desired IF, degrading the reference sensitivity of the receiver. During the downconversion, the phase noise of the PLL is
The measured dynamic performance of the ADL5350, includ- imposed onto each subcarrier of the OFDM modulated signal
ing the external LO buffer and filter networks, is presented in through convolution. To minimize the impact on the receiver
Figure 6. The mixer provides better than 24 dBm input IIP3 sensitivity level, the first agile PLL for the generation of LO1 is a
with ~8 dB conversion loss and single-sideband NF. fractional-N PLL. It is designed with the ADF4153 synthesizer.
The closed loop bandwidth is about 30 kHz and the estimated
phase jitter is 0.3 degrees.

ADL5350 EVALUATION BOARD

SAWTEK
1.5nH 855898
1pF ADL5350 1nF 18pF 6pF
0.67nH
RF INPUT 374MHz
FROM LNA2
1 RFIF RFIF 8 0.7pF IF OUT
2 GND2 GC 7 3V 9.5nH 15nH
3nH 2.1nH
3 LOIN VPOS 6
100pF
4 GND1 GND1 5 100pF 4.7µF
2.2pF

2.7nH 3V 3V
VP VDD
100pF
0.47µF
2.2pF
100nF 3dB PAD VCO190-1960T
1nF 1nF 18Ω 18Ω 1.5kΩ 10MHz
CP REFIN XTAL
100pF 100pF REF
600Ω ADF4153
MUXOUT
18Ω CLK SPI
AD8353 100pF DATA CONTROL
820pF 40nF 2.7nF
EVALUATION BOARD RFINA LE BUS
51Ω
RFINB
RSET
100pF
06147-006

ADF4153 EVALUATION BOARD

Figure 7. First Mix-Down Stage

Rev. 0 | Page 4 of 8
 AN-851
The first IF is then passed through a 374 MHz SAW filter for The first filtered IF signal is then downconverted to the second
channel selection. The excellent stop-band rejection of the IF of 70 MHz using the AD8344 active mixer. A low-side LO
Sawtek 855898 SAW filter allows the receiver to have injection scheme was used to ensure optimal spurious response
exceptional selectivity, improving immunity to adjacent rejection and to achieve greater gain in the mixer. In general,
interfering signals. The filter is matched to 50 Ω using the the AD8344 offers slightly higher conversion gain when using a
external LC component indicated in Figure 7. The filters low-side LO. The 304 MHz LO is provided by the ADF4360-8.
frequency response characteristic is presented in Figure 8. This is an integrated PLL + VCO providing good cost and
0
board space savings over discrete solutions. Two external
inductors that are the tank inductors of the on-chip VCO set
–10
the center frequency. The differential outputs of the ADF4360-8
–20
are combined in a balun. Using a 304 MHz LO the AD8344
–30 offers a conversion gain of ~11 dB, with 12 dBm IIP3 and 8 dB
RESPONSE (dB)

–40 SSB NF. The 70 MHz output is then passed through a fourth
–50 order low-pass filter to help reject LO feedthrough and higher
–60
frequency mixer spurii.
–70

–80

–90

–100
06147-008

300 325 350 375 400 425 450

FREQUENCY (MHz)

Figure 8. Sawtek 855898 Frequency Response

AD8344 1nF
EVALUATION BOARD
5V

2.43kΩ
1nF

5V
12 11 10 9

VPDC PWDN EXRB COMM 1nF

13 COMM COMM 8
100pF 4:1 1nF 220nH 240nH
70MHz
14 RFCM IFOP 7 IF OUT
AD8344 39pF 18pF
374MHz 15 RFIN IFOM 6
INPUT
100pF
5V 16 VPMX COMM 5
1nF
VPLO LOCM LOIN COMM
VDD
1 2 3 4 ADF4360-8
EVALUATION BOARD
1nF 1nF 1nF
ADF4360-8
10MHz
VTUNE REFIN XTAL
REF
4.3kΩ
CP CLK SPI
1.5kΩ DATA CONTROL
LE BUS
150pF 680nF 390pF
43.8nH
VDD L1

240nH L2
1:4
RFOUTA 470Ω
RSET
100pF 4.7kΩ
RFOUTB
06147-009

Figure 9. The AD8344 Active Mixer Driven by the ADF4360-8 Through an External 1:4 Impedance Ratio Balun

Rev. 0 | Page 5 of 8
AN-851
The final 70 MHz IF signal is then passed through the AD8370
digitally controlled variable gain amplifier before being IF
sampled by the AD9246. The AD8370 provides a high output
IIP3 and greater than 40 dB gain adjustment range. This allows
the overall receiver conversion gain to adjust and tailor the
cascaded input dynamic range to accommodate widely varying
input signal powers.
The selected ADC provides excellent spurious free dynamic
range out to greater than 200 MHz IF frequencies while only
consuming ~250 mW. The high analog input bandwidth of
650 MHz allows the AD9246 to be applied to higher IF
frequencies. In this demonstration a 70 MHz IF was selected
using an 80 MSPS sampling clock, placing the IF in the second
Nyquist zone of the ADC.
to raise the presented loading impedance at the desired IF
frequency. The filter network provides an impedance
transformation from 100 Ω to 600 Ω and has the effect of
stepping up the voltage by ~8 dB. The step-up transformation
needs to be accounted for when analyzing the receiver line-up.
For more information regarding the design approach used to
yield the driver/ADC interface, please refer to AN-827, A
Resonant Approach to Interfacing Amplifiers to Switched-
Capacitor ADCs. The key details of the AD8370 and AD9246
network interface are captured in Figure 11. The simulated filter
response is plotted against the measured response through the
ADC in Figure 10.
0
–5

–10
In order to prevent degradation of the ADC’s sensitivity level it MEASURED
was necessary to employ an anti-aliasing filter. The anti-aliasing –15

RESPONSE (dB)
filter helps to reject higher frequency spurious signals such as –20
LO leakage from degrading the perceived noise floor of the SIMULATED
–25
ADC. Additionally the anti-aliasing filter helps to reject
wideband noise generated by the driving amplifier stages that –30

would otherwise alias additional noise into the desired Nyquist –35
band. Elliptical low-pass architecture is selected to help provide
–40
better rejection of strong spurious components in the higher
Nyquist zones. A resonant parallel tank network is formed by –45

06147-011
20 40 60 80 100 120 140 160
the 72 nH bias inductors in combination with the total FREQUENCY (MHz)
presented input capacitance. The resonant tank network helps Figure 10. Simulated and Measured Response of the Anti-Aliasing Filter into
the AD9246 ADC

AD9246
VCOM
SERIAL CONTROL INTERFACE 39pF
1nF
1nF 1nF 47nH 56nH

16 15 14 13 12 11 10 9

INLO ICOM DATA CLCK LTCH VCCO OCOM OPLO 72nH

70MHz 1:4
INPUT
AD8370 15pF 18pF 22pF AD9246-80
72nH
INHI ICOM VCCI PWUP VOCM VCCO OCOM OPHI
1 2 3 4 5 6 7 8

1nF
1nF 1nF 47nH 56nH
06147-010

0.1µF 0.1µF
39pF
5V

Figure 11. Circuit Interface Between the VGA and ADC

Rev. 0 | Page 6 of 8
 AN-851

SUMMARY OF COMPLETE RECEIVER PERFORMANCE

While attempting to verify the overall cascaded performance of Suitable BER test equipment for an 802.16 OFDM waveform
the full receiver, it is necessary to capture the sampled data from was not available at the time of the evaluation. However, error
the AD9246 ADC using the Analog Devices high speed ADC vector magnitude (EVM) analysis allows for a fair estimate of
FIFO USB evaluation kit. The additional first-in-first-out the receiver performance and available dynamic range. In order
(FIFO) daughtercard serves as a data buffer to allow capture of to perform EVM analysis on OFDM 802.16 test signals it was
long data output strings generated by the ADC at the sampling necessary to use a FIFO card with an extended memory depth
rate. The FIFO can then transmit the captured data to the PC at of at least 65 kb. This allows capture of a complete RF burst.
a lower data rate that can be supported via a standard USB
interface. The ADC evaluation hardware is controlled by The captured integer vectors range from 0 to 65,536 (216) with a
midscale value of 32,768. Note that the AD9246 is a 14-bit
Analog Devices ADC Analyzer™ software. The ADC Analyzer
ADC, and that the data output words are 16-bit representation.
allows for time domain and frequency domain analysis. The
The captured integer vectors were post-processed in the
single-tone and two-tone distortion performance is readily
Advanced Design System (ADS) 2006A from Agilent. The ADS
captured using ADC Analyzer, and the results are presented in
program divides the 16-bit represented IF vector by four to
Figure 12 and Figure 13. Note the discontinuity in Figure 13
yield a 214 integer representation. The vector is then shifted and
around 48 dB of signal gain. This is where the AD8370 VGA
scaled in magnitude to yield a zero-mean value waveform with
transitions from a low gain mode to a high gain mode. The
±1 V peak magnitude. The signal is then demodulated using an
discontinuity does not result in any notable degradation of the
ideal IQ demodulator. The IQ vectors are then decimated in
BER performance of the full receiver.
time to yield baseband vectors at ¼ the original sampling rate.
–50
The processed IQ vectors are then fed into the 89600 Vector
–55
Signal Analyzer software from Agilent to reveal the EVM
HARMONIC DISTORTION (dBc)

–60 performance. The measured EVM performance vs. input power

is depicted in Figure 14 for a 64-QAM OFDM signal with a
–65
10 MHz modulation bandwidth and ¾ coding rate. The EVM is
–70 plotted with and without the ADF4153 and ADF4360 PLLs to
HD2
–75
provide a comparison of the performance degradation due to
HD3 phase noise.
–80
0

–85 –5

–90 –10
06147-012

30 35 40 45 50 55 60 65
–15
SIGNAL GAIN (dB)

Figure 12. Second and Third Harmonic Distortion for the Full Receiver –20
EVM (dB)

Cascade Measured at −1 dBFS Input Levels at the AD9246 Input –25

0 7 –30

–35
–5 6 WITH ADF4153 AND
ADF4360-8 PLLs
–40
–10 5
–45
USING EXTERNAL GENERATORS FOR LO SIGNALS
SSB NF (dB)

–50
IIP3 (dBm)

–15 4
06147-014

–80 –70 –60 –50 –40 –30 –20 –10

INPUT POWER (dBm)
–20 3
Figure 14. EVM Performance Measured Through the Full Receiver for 64-QAM
¾ Rate Coding, 10 MHz OFDM Signaling
–25 2

–30 1

–35 0
06147-013

30 35 40 45 50 55 60 65 70
SIGNAL GAIN (dB)

Figure 13. Two-Tone IIP3 and Single-Sideband NF for the Full Receiver
Cascade Measured vs. Receiver Conversion Gain

Rev. 0 | Page 7 of 8
AN-851
In order to access the selectivity of the receiver, it is necessary to The typical spectral characteristics of the full receiver is
measure the relative response for single-tone inputs swept over depicted in Figure 16. By using the ADC Analyzer software, it
a reasonable input frequency range. Figure 15 presents the is easy to see the instantaneous spectrum and capture time
selectivity of the full receiver in the absence of the front-end domain data for later signal processing. Whereas the IF is at
band-select filter. The frequency selective nature of the double 70 MHz with an 80 MSPS sampling clock, the signal appears as
downconversion receiver provides very high immunity to a 10 MHz signal in the first Nyquist zone. The ADC Analyzer
nearby interferers. With the addition of the front-end band- software makes it possible to quickly analyze and debug
select filter, greater than 60 dB rejection can be achieved at spurious clock and LO signal components that can otherwise
adjacent RF frequency bands. compromise in-band receiver performance. The multiple IF
10
filter stages and LO filters helped to minimize spurious clutter
to negligible levels. Note the weak spur at 32 MHz. This is the
0
LO to IF leakage of the first mixer.
NORMALIZED RESPONSE (dB)

–10
A complete summary of the receiver’s performance is provided
–20 in Table 3. The design provides better than −25 dB EVM per-
formance for input signals greater than −74 dBm, with more
–30
than 60 dB image rejection and excellent adjacent and nonadja-
–40 cent channel rejection.
–50 Table 3. Performance Summary (measured with 64-QAM ¾
rate coding OFDM 10 MHz bandwidth input signal)
–60
Measured
–70 Parameter Performance
06147-015

2280 2300 2320 2340 2360 2380 2400 2420 2440

Gain Range 30 dB to 67 dB
RF INPUT FREQUENCY (MHz)
EVM from −74 dBm to −20 dBm Input Power <−25 dB
Figure 15. Receiver Selectivity vs. Input Frequency, Local Oscillators Tuned for
2350 MHz Input Signal (fLO1 = 1976 MHz, fLO2 = 304 MHz) IIP3 at Mid-Gain (Av = 48 dB) >−8 dBm
NF at Mid-Gain (Av = 48 dB) <4 dB
–20
Image Rejection >60 dB
–30
Adjacent Channel Rejection >58 dB
–40
Power Dissipation (from LNA to ADC, ~2.2 Watts
SPECTRAL MAGNITUDE (dBm)

–50 including PLLs)

–60

–70
REFERENCES
–80 Newman, Eric and Reeder, Rob. 2006. “A Resonant Approach to
–90
Interfacing Amplifiers to Switched-Capacitor ADCs.”
–100
Application Note AN-827. Analog Devices, Inc. (January)
–110
IEEE Std 802.16-2004. IEEE Standard for Local and
–120 Metropolitan Area Networks - Part 16: Air Interface for Fixed
–130 Broadband Wireless Access Systems. Institute of Electrical and
06147-016

0 5 10 15 20 25 30 35 40
FREQUENCY (MHz) Electronics Engineers, Inc. (June).
Figure 16. FFT Plot of ADC Spectrum at Mid-Gain for a 64-QAM ¾ Rate
Coded, 10 MHz Bandwidth OFDM Input at 2.342 GHz (Note that the signal
appears as a 10 MHz IF but is actually a 70 MHz input in the second Nyquist
zone of the converter.)

©2006 Analog Devices, Inc. All rights reserved. Trademarks and

registered trademarks are the property of their respective owners.
AN06147-0-6/06(0)

Rev. 0 | Page 8 of 8