The ADL5387 is a broadband quadrature I/Q demodulator that
covers an RF/IF input frequency range from 50 MHz to 2 GHz.
With a NF = 13.2 dB, IP1dB = 12.7 dBm, and IIP3 = 32 dBm @
450 MHz, the ADL5387 demodulator offers outstanding dynamic
range suitable for the demanding infrastructure direct-conversion
requirements. The differential RF/IF inputs provide a wellbehaved broadband input impedance of 50 and are best
driven from a 1:1 balun for optimum performance.
Ultrabroadband operation is achieved with a divide-by-2 method
for local oscillator (LO) quadrature generation. Over a wide
range of LO levels, excellent demodulation accuracy is
achieved with amplitude and phase balances ~0.05 dB and
~0.4°, respectively. The demodulated in-phase (I) and
quadrature (Q) differential outputs are fully buffered and
provide a voltage conversion gain of >4 dB. The buffered
baseband outputs are capable of driving a 2 V p-p differential
signal into 200 .
The fully balanced design minimizes effects from second-order
distortion. The leakage from the LO port to the RF port is
<−70 dBc. Differential dc-offsets at the I and Q outputs are
<10 mV. Both of these factors contribute to the excellent IIP2
specifications > 60 dBm.
The ADL5387 operates off a single 4.75 V to 5.25 V supply. The
supply current is adjustable with an external resistor from the
BIAS pin to ground.
The ADL5387 is fabricated using the Analog Devices, Inc.
advanced silicon-germanium bipolar process and is available in
a 24-lead exposed paddle LFCSP.
Rev. 0
Information furnished by Analog Devices is believed to be accurate and reliable. However, no
responsibility is assumed by Anal og Devices for its use, nor for any infringements of patents or ot her
rights of third parties that may result from its use. Specifications subject to change without notice. No
license is granted by implication or otherwise under any patent or patent rights of Analog Devices.
Trademarks and registered trademarks are the property of their respective owners.
VS = 5 V, TA = 25°C, fRF = 900 MHz, fIF = 4.5 MHz, PLO = 0 dBm, BIAS pin open, ZO = 50 Ω, unless otherwise noted, baseband outputs
differentially loaded with 450 Ω.
Table 1.
Parameter Condition Min Typ Max Unit
OPERATING CONDITIONS
LO Frequency Range External input = 2xLO frequency 0.1 4 GHz
RF Frequency Range 0.05 2 GHz
LO INPUT LOIP, LOIN
Input Return Loss
LO Input Level −6 0 +6 dBm
I/Q BASEBAND OUTPUTS QHI, QLO, IHI, ILO
Voltage Conversion Gain
Demodulation Bandwidth 1 V p-p signal 3 dB bandwidth 240 MHz
Quadrature Phase Error @ 900 MHz 0.4 Degrees
I/Q Amplitude Imbalance 0.1 dB
Output DC Offset (Differential) 0 dBm LO input ±5 mV
Output Common-Mode VPOS − 2.8 V
0.1 dB Gain Flatness 40 MHz
Output Swing Differential 200 Ω load 2 V p-p
Peak Output Current Each pin 12 mA
POWER SUPPLIES VPA, VPL, VPB, VPX
Voltage 4.75 5.25 V
Current BIAS pin open 180 mA
RBIAS = 4 kΩ 157 mA
DYNAMIC PERFORMANCE @ RF = 140 MHz RFIP, RFIN
Conversion Gain 4.7 dB
Input P1dB (IP1dB) 13 dBm
Second-Order Input Intercept (IIP2) −5 dBm each input tone 67 dBm
Third-Order Input Intercept (IIP3) −5 dBm each input tone 31 dBm
LO to RF
RF to LO LOIN, LOIP terminated in 50 Ω −95 dBc
I/Q Magnitude Imbalance 0.05 dB
I/Q Phase Imbalance 0.2 Degrees
LO to I/Q
Noise Figure 12.0 dB
Noise Figure under Blocking Conditions With a −5 dBm interferer 5 MHz away 14.4 dB
AC-coupled into LOIP with LOIN bypassed,
measured at 2 GHz
450 Ω differential load on I and Q outputs
(@ 900 MHz)
200 Ω differential load on I and Q outputs
(@ 900 MHz)
RFIN, RFIP terminated in 50 Ω, 1xLO
appearing at the RF port
RFIN, RFIP terminated in 50 Ω, 1xLO
appearing at the BB port
−10 dB
4.3 dB
3.2 dB
−100 dBm
−39 dBm
Rev. 0 | Page 3 of 28
ADL5387
Parameter Condition Min Typ Max Unit
DYNAMIC PERFORMANCE @ RF = 450 MHz
Conversion Gain 4.4 dB
Input P1dB (IP1dB) 12.7 dBm
Second-Order Input Intercept (IIP2) −5 dBm each input tone 69.2 dBm
Third-Order Input Intercept (IIP3) −5 dBm each input tone 32.8 dBm
LO to RF
RF to LO LOIN, LOIP terminated in 50 Ω −90 dBc
I/Q Magnitude Imbalance 0.05 dB
I/Q Phase Imbalance 0.6 Degrees
LO to I/Q
Noise Figure 13.2 dB
DYNAMIC PERFORMANCE @ RF = 900 MHz
Conversion Gain 4.3 dB
Input P1dB (IP1dB) 12.8 dBm
Second-Order Input Intercept (IIP2) −5 dBm each input tone 61.7 dBm
Third-Order Input Intercept (IIP3) −5 dBm each input tone 31.2 dBm
LO to RF
RF to LO LOIN, LOIP terminated in 50 Ω −88 dBc
I/Q Magnitude Imbalance 0.05 dB
I/Q Phase Imbalance 0.2 Degrees
LO to I/Q
Noise Figure 14.7 dB
Noise Figure under Blocking Conditions With a −5 dBm interferer 5 MHz away 15.8 dB
DYNAMIC PERFORMANCE @ RF = 1900 MHz
RFIN, RFIP terminated in 50 Ω, 1xLO
appearing at the RF port
RFIN, RFIP terminated in 50 Ω, 1xLO
appearing at the BB port
RFIN, RFIP terminated in 50 Ω, 1xLO
appearing at the RF port
RFIN, RFIP terminated in 50 Ω,
1XLO appearing at the BB port
−87 dBm
−38 dBm
−79 dBm
−41 dBm
Conversion Gain 3.8 dB
Input P1dB (IP1dB) 12.8 dBm
Second-Order Input Intercept (IIP2) −5 dBm each input tone 59.8 dBm
Third-Order Input Intercept (IIP3) −5 dBm each input tone 27.4 dBm
LO to RF
RF to LO LOIN, LOIP terminated in 50 Ω −70 dBc
I/Q Magnitude Imbalance 0.05 dB
I/Q Phase Imbalance 0.3 Degrees
LO to I/Q
Noise Figure 16.5 dB
Noise Figure under Blocking Conditions With a −5 dBm interferer 5 MHz away 18.7 dB
RFIN, RFIP terminated in 50 Ω, 1xLO
appearing at the RF port
RFIN, RFIP terminated in 50 Ω, 1xLO
appearing at the BB port
−75 dBm
−43 dBm
Rev. 0 | Page 4 of 28
ADL5387
ABSOLUTE MAXIMUM RATINGS
Table 2.
Parameter Rating
Supply Voltage VPOS1, VPOS2, VPOS3 5.5 V
LO Input Power 13 dBm (re: 50 Ω)
RF/IF Input Power 15 dBm (re: 50 Ω)
Internal Maximum Power Dissipation 1100 mW
θ
JA
Maximum Junction Temperature 150°C
Operating Temperature Range −40°C to +85°C
Storage Temperature Range −65°C to +125°C
Stresses above those listed under Absolute Maximum Ratings
may cause permanent damage to the device. This is a stress
rating only; functional operation of the device at these or any
other conditions above those indicated in the operational
section of this specification is not implied. Exposure to absolute
maximum rating conditions for extended periods may affect
device reliability.
54°C/W
ESD CAUTION
Rev. 0 | Page 5 of 28
ADL5387
PIN CONFIGURATION AND FUNCTION DESCRIPTIONS
24
2322212019
CMRF CMRF RFIP
1
VPA
RFIN CMRF VPX
VPB
18
2
COM
3
BIAS
VPB
QHI
17
16
ADL5387
4
VPL
5
VPL
6
VPL
CML
789101112
TOP VIEW
(Not to Scale)
LOIP L OIN CML CML COM
QLO
IHI
ILO
15
14
13
06764-002
Figure 2. Pin Configuration
Table 3. Pin Function Descriptions
Pin No. Mnemonic Description
1, 4 to 6,
17 to 19
2, 7, 10 to 12,
VPA, VPL, VPB, VPX
Supply. Positive supply for LO, IF, biasing and baseband sections, respectively. These pins should
be decoupled to board ground using appropriate sized capacitors.
COM, CML, CMRF Ground. Connect to a low impedance ground plane.
20, 23, 24
3 BIAS
Bias Control. A resistor can be connected between BIAS and COM to reduce the mixer core current.
The default setting for this pin is open.
8, 9 LOIP, LOIN
Local Oscillator. External LO input is at 2xLO frequency. A single-ended LO at 0 dBm can be applied
through a 1000 pF capacitor to LOIP. LOIN should be ac-grounded, also using a 1000 pF. These inputs
can also be driven differentially through a balun (recommended balun is M/A-COM ETC1-1-13).
13 to 16 ILO, IHI, QLO, QHI
I-Channel and Q-Channel Mixer Baseband Outputs. These outputs have a 50 Ω differential output
impedance (25 Ω per pin). The bias level on these pins is equal to VPOS − 2.8 V. Each output pair can
swing 2 V p-p (differential) into a load of 200 Ω. Output 3 dB bandwidth is 240 MHz.
21, 22 RFIN, RFIP
RF Input. A single-ended 50 Ω signal can be applied to the RF inputs through a 1:1 balun (recommended
balun is M/A-COM ETC1-1-13). Ground-referenced inductors must also be connected to RFIP and
RFIN (recommended values = 120 nH).
EP Exposed Paddle. Connect to a low impedance ground plane
Rev. 0 | Page 6 of 28
ADL5387
TYPICAL PERFORMANCE CHARACTERISTICS
VS = 5 V, TA = 25°C, LO drive level = 0 dBm, R
20
TA = –40°C
TA = +25°C
TA = +85°C
15
10
GAIN (dB), IP1dB (dBm)
5
INPUT P1dB
GAIN
= open, unless otherwise noted.
BIAS
5
NORMALIZED TO 1MHz
0
–5
–10
–15
BB RESPONSE (dB)
–20
–25
0
0200 400 600 800 1000 1200 1400 1600 1800 2000
RF FREQUENCY ( MHz)
Figure 3. Conversion Gain and Input 1 dB Compression Point (IP1dB) vs.
RF Frequency
80
70
60
50
40
IIP2, IIP3 (dBm)
30
20
10
0200 400 600 800 1000 1200 1400 1600 1800 2000
INPUT IP2
INPUT IP3
(I AND Q CHANNELS)
RF FREQUENCY (M Hz)
I CHANNEL
Q CHANNEL
TA = +85°C
TA = +25°C
TA = –40°C
Figure 4. Input Third-Order Intercept (IIP3) and
Input Second-Order Intercept Point (IIP2) vs. RF Frequency
2.0
TA = –40°C
TA = +25°C
1.5
TA = +85°C
1.0
–30
1100010010
06764-003
BB FREQUENCY (MHz)
06764-006
Figure 6. Normalized I/Q Baseband Frequency Response
19
TA = –40°C
TA = +25°C
TA = +85°C
17
15
13
11
NOISE FI GURE (dB)
9
7
0200 400 600 800 1000 1200 1400 1600 1800 2000
06764-004
RF FREQUENCY ( MHz)
06764-007
Figure 7. Noise Figure vs. RF Frequency
4
TA = –40°C
TA = +25°C
3
TA = +85°C
2
0.5
0
–0.5
MAGNITUDE ERRO R (dB)
–1.0
–1.5
–2.0
0200 400 600 800 1000 1200 1400 1600 1800 2000
RF FREQUENCY ( MHz)
Figure 5. I/Q Gain Mismatch vs. RF Frequency
06764-005
Rev. 0 | Page 7 of 28
1
0
–1
–2
–3
QUADRATURE PHASE ERRO R (Degrees)
–4
0200 400 600 800 1000 1200 1400 1600 1800 2000
RF FREQUENCY ( MHz)
Figure 8. I/Q Quadrature Phase Error vs. RF Frequency
06764-008
ADL5387
20
INPUT IP2, Q CHANNEL
80
20
INPUT IP2, I CHANNEL
80
NOISE FI GURE
15
10
GAIN (dB), INPUT P1dB (dBm), NOISE FIGURE (dB)
INPUT IP2, I CHANNEL
INPUT P1dB
NOISE FIGURE
GAIN
5
0
–6–5–4–3–2–10123456
INPUT IP3
LO LEVEL (dBm)
Figure 9. Conversion Gain, Noise Figure, IIP3, IIP2, and IP1dB vs.
LO Level, f
32
TA = –40°C
TA = +25°C
TA = +85°C
28
INPUT IP3
24
20
16
12
IIP3 (dBm) AND NO ISE FI GURE (dB)
= 140 MHz
RF
SUPPLY
CURRENT
NOISE FI GURE
65
50
35
20
195
185
175
165
155
145
15
INPUT IP2, Q CHANNEL
10
GAIN
INPUT IP2, INPUT IP3 (dBm)
06764-009
5
GAIN (dB), INPUT P1dB (dBm), NOISE FIGURE (dB)
0
–6–5–4–3–2–10123456
INPUT IP3
INPUT P1dB
LO LEVEL (dBm)
65
50
35
INPUT IP2, INPUT IP3 (dBm)
20
06764-012
Figure 12. Conversion Gain, Noise Figure, IIP3, IIP2, and IP1dB vs.
LO Level, f
32
TA = –40°C
TA = +25°C
TA = +85°C
28
24
20
SUPPLY CURRENT (mA)
16
12
IIP3 (dBm) AND NO ISE FI GURE (dB)
= 900 MHz
RF
NOISE FIGURE
INPUT IP3
8
110100
R
(kΩ)
BIAS
Figure 10. Noise Figure, IIP3, and Supply Current vs. R
25
20
R
= 100kΩ
15
10
NOISE FIGURE (dB)
5
0
–3050–5–10–15–20–25
BIAS
R
= 10kΩ
BIAS
R
= 4kΩ
BIAS
R
= 1.4kΩ
BIAS
RF BLOCKER INP UT POWER ( dBm)
Figure 11. Noise Figure vs. Input Blocker Level, f
(RF Blocker 5 MHz Offset)
, fRF = 140 MHz
BIAS
= 900 MHz
RF
135
8
110
06764-010
Figure 13. IIP3 and Noise Figure vs. R
R
(kΩ)
BIAS
, fRF = 900 MHz
BIAS
100
06764-013
80
70
60
50
40
30
20
10
GAIN (dB), IP1dB, IIP2, I AND Q CHANNELS (d Bm)
0
110
06764-011
Figure 14. Conversion Gain, IP1dB, IIP2 I Channel, and IIP2 Q Channel vs. R
140MHz: GAIN
140MHz: IP1d B
140MHz: IIP2, I CHANNEL
140MHz: IIP2, Q CHANNEL
450MHz: GAIN
450MHz: IP1d B
450MHz: IIP2, I CHANNEL
450MHz: IIP2, Q CHANNEL
R
(kΩ)
BIAS
100
06764-014
BIAS
Rev. 0 | Page 8 of 28
ADL5387
–
–
35
80
20
30
25
20
IP1dB, IIP3 (dBm)
15
10
TA = –40°C
TA = +25°C
TA = +85°C
5
05
IIP3
IP1dB
BB FREQUENCY (MHz)
INPUT IP2,
I CHANNEL
INPUT IP2,
Q CHANNEL
75
70
65
60
55
INPUT IP2, I AND Q CHANNELS (dBm)
50
045403530252015105
06764-015
Figure 15. IIIP3, IIP2, IP1dB vs. Baseband Frequency
–30
–40
–50
–60
–70
LO LEAKAGE (dBm)
–80
–90
–100
0200018001600140012001000800600400200
Figure 18. LO-to-RF Leakage vs. Internal 1xLO Frequency
0
–10
–20
–30
–40
–50
FEEDTHROUG H (dBm)
–60
–70
1xLO (INT ERNAL)
2xLO (EXT ERNAL)
LEAKAGE (dBc)
–100
20
–40
–60
–80
1xLO
2xLO
INTERNAL 1xL O FREQUENCY (MHz)
06764-018
–80
0200018001600140012001000800600400200
INTERNAL 1xL O FREQUE NCY (MHz)
06764-016
Figure 16. LO-to-BB Feedthrough vs. 1xLO Frequency (Internal LO Frequency)
0
–5
–10
–15
RETURN LOSS ( dB)
–20
–25
0200018001600140012001000800600400200
RF FREQUENCY ( MHz)
06764-017
Figure 17. RF Port Return Loss vs. RF Frequency, Measured on
Characterization Board through ETC1-1-13 Balun with 120 nH Bias Inductors
–120
RETURN LOSS ( dB)
–10
–15
–20
–25
–30
0200018001600140012001000800600400200
RF FREQUENCY ( MHz)
Figure 19. RF-to-LO Leakage vs. RF Frequency
0
–5
04000350030002500200015001000500
FREQUENCY (MHz )
Figure 20. Single-Ended LO Port Return Loss vs.
LO Frequency, LOIN AC-Coupled to Ground
06764-019
06764-020
Rev. 0 | Page 9 of 28
ADL5387
DISTRIBUTIONS FOR fRF = 140 MHz
100
TA = –40°C
TA = +25°C
TA = +85°C
80
100
TA = –40°C
TA = +25°C
TA = +85°C
80
I CHANNEL
Q CHANNEL
60
40
PERCENTAGE (%)
20
0
283330293231
INPUT IP3 (dBm)
Figure 21. IIP3 Distributions
100
TA = –40°C
TA = +25°C
TA = +85°C
80
60
40
PERCENTAGE (%)
20
0
101512111413
INPUT P1dB (dBm)
Figure 22. IP1dB Distributions
100
TA = –40°C
TA = +25°C
TA = +85°C
80
60
40
PERCENTAGE (%)
20
0
60756570
06764-021
INPUT IP2 (dBm)
06764-024
Figure 24. IIP2 Distributions for I Channel and Q Channel
100
TA = –40°C
TA = +25°C
TA = +85°C
80
60
40
PERCENTAGE (%)
20
0
10.513.513.012.512.011.511.0
06764-022
NOISE FI GURE (dB)
06764-025
Figure 25. Noise Figure Distributions
100
TA = –40°C
TA = +25°C
TA = +85°C
80
60
40
PERCENTAGE (%)
20
0
–0.20.20–0.10.1
I/Q GAI N MISMATCH (dB)
Figure 23. I/Q Gain Mismatch Distributions
60
40
PERCENTAGE (%)
20
0
–1.01.00.50–0.5
06764-023
QUADRATURE PHASE ERRO R (Degrees)
06764-026
Figure 26. I/Q Quadrature Error Distributions
Rev. 0 | Page 10 of 28
ADL5387
DISTRIBUTIONS FOR fRF = 450 MHz
100
TA = –40°C
TA = +25°C
TA = +85°C
80
60
100
TA = –40°C
TA = +25°C
TA = +85°C
80
60
I CHANNEL
Q CHANNEL
40
PERCENTAGE (%)
20
0
303534333231
100
80
60
40
PERCENTAGE (%)
20
0
101514131211
100
80
INPUT IP3 (dBm)
Figure 27. IIP3 Distributions
TA = –40°C
TA = +25°C
TA = +85°C
INPUT P1dB (dBm)
Figure 28. IP1dB Distributions
TA = –40°C
TA = +25°C
TA = +85°C
40
PERCENTAGE (%)
20
0
60756570
06764-027
INPUT IP2 (dBm)
06764-030
Figure 30. IIP2 Distributions for I Channel and Q Channel
100
80
60
40
PERCENTAGE (%)
20
0
12.015.014.514.013.513.012.5
06764-028
NOISE FI GURE (dB)
TA = –40°C
TA = +25°C
TA = +85°C
06764-031
Figure 31. Noise Figure Distributions
100
TA = –40°C
TA = +25°C
TA = +85°C
80
PERCENTAGE (%)
60
40
20
0
–0.20.20.10–0.1
I/Q GAI N MISMATCH (dB)
Figure 29. I/Q Gain Mismatch Distributions
06764-029
Rev. 0 | Page 11 of 28
60
40
PERCENTAGE (%)
20
0
–1.0–0.500.51. 0
QUADRATURE PHASE ERRO R (Degrees)
Figure 32. I/Q Quadrature Error Distributions
06764-032
ADL5387
DISTRIBUTIONS FOR fRF = 900 MHz
100
TA = –40°C
TA = +25°C
TA = +85°C
80
60
100
TA = –40°C
TA = +25°C
TA = +85°C
80
60
I CHANNEL
Q CHANNEL
40
PERCENTAGE (%)
20
0
303133323435
INPUT IP3 (dBm)
Figure 33. IIP3 Distributions
100
TA = –40°C
TA = +25°C
TA = +85°C
80
60
40
PERCENTAGE (%)
20
0
101113121415
INPUT P1dB (dBm)
Figure 34. IP1dB Distributions
100
TA = –40°C
TA = +25°C
TA = +85°C
80
40
PERCENTAGE (%)
20
0
5575656070
06764-033
INPUT IP2 (dBm)
06764-036
Figure 36. IIP2 Distributions for I Channel and Q Channel
100
TA = –40°C
TA = +25°C
TA = +85°C
80
60
40
PERCENTAGE (%)
20
0
13.013.514.014.515.015.516.0
06764-034
NOISE FI GURE (dB)
06764-037
Figure 37. Noise Figure Distributions
100
TA = –40°C
TA = +25°C
TA = +85°C
80
60
40
PERCENTAGE (%)
20
0
–0.2–0.100.10. 2
I/Q GAI N MISMATCH (dB)
Figure 35. I/Q Gain Mismatch Distributions
06764-035
Rev. 0 | Page 12 of 28
60
40
PERCENTAGE (%)
20
0
–1.01.00.50–0.5
Figure 38. I/Q Quadrature Error Distributions
QUADRATURE PHASE ERRO R (Degrees)
06764-038
ADL5387
DISTRIBUTIONS FOR fRF = 1900 MHz
100
TA = –40°C
TA = +25°C
TA = +85°C
80
100
80
60
40
PERCENTAGE (%)
20
0
263129302827
100
80
60
40
PERCENTAGE (%)
20
0
101513141211
100
80
INPUT IP3 (dBm)
Figure 39. IIP3 Distributions
TA = –40°C
TA = +25°C
TA = +85°C
INPUT P1dB (dBm)
Figure 40. IP1dB Distributions
TA = –40°C
TA = +25°C
TA = +85°C
60
40
PERCENTAGE (%)
20
0
526866646260585654
06764-039
INPUT IP2 (dBm)
TA = –40°C
TA = +25°C
TA = +85°C
I CHANNEL
Q CHANNEL
06764-042
Figure 42. IIP2 Distributions for I Channel and Q Channel
100
80
60
40
PERCENTAGE (%)
20
0
15.018.017.517.016.516.015.5
06764-040
TA = –40°C
TA = +25°C
TA = +85°C
NOISE FI GURE (dB)
06764-043
Figure 43. Noise Figure Distributions
100
TA = –40°C
TA = +25°C
TA = +85°C
80
PERCENTAGE (%)
60
40
20
0
–0.20.20.10–0.1
I/Q GAI N MISMATCH (dB)
Figure 41. I/Q Gain Mismatch Distributions
06764-041
Rev. 0 | Page 13 of 28
60
40
PERCENTAGE (%)
20
0
–1.01.00.50–0.5
Figure 44. I/Q Quadrature Error Distributions
QUADRATURE PHASE ERRO R (Degrees)
06764-044
ADL5387
CIRCUIT DESCRIPTION
The ADL5387 can be divided into five sections: the local
oscillator (LO) interface, the RF voltage-to-current (V-to-I)
converter, the mixers, the differential emitter follower outputs,
and the bias circuit. A detailed block diagram of the device is
shown in
Figure 45.
RFIP
RFIN
BIAS
DIVIDE-BY-T WO
QUADRATURE
PHASE SPLITTER
Figure 45. Block Diagram
IHI
ILO
LOIP
LOIN
QHI
QLO
6764-045
The LO interface generates two LO signals at 90° of phase
difference to drive two mixers in quadrature. RF signals are
converted into currents by the V-to-I converters that feed into
the two mixers. The differential I and Q outputs of the mixers
are buffered via emitter followers. Reference currents to each
section are generated by the bias circuit. A detailed description
of each section follows.
LO INTERFACE
The LO interface consists of a buffer amplifier followed by a
frequency divider that generate two carriers at half the input
frequency and in quadrature with each other. Each carrier is
then amplified and amplitude-limited to drive the doublebalanced mixers.
V-TO-I CONVERTER
The differential RF input signal is applied to a resistively
degenerated common base stage, which converts the differential
input voltage to output currents. The output currents then
modulate the two half-frequency LO carriers in the mixer stage.
MIXERS
The ADL5387 has two double-balanced mixers: one for the
in-phase channel (I channel) and one for the quadrature channel
(Q channel). These mixers are based on the Gilbert cell design
of four cross-connected transistors. The output currents from
the two mixers are summed together in the resistive loads that
then feed into the subsequent emitter follower buffers.
EMITTER FOLLOWER BUFFERS
The output emitter followers drive the differential I and Q
signals off-chip. The output impedance is set by on-chip 25 Ω
series resistors that yield a 50 Ω differential output impedance
for each baseband port. The fixed output impedance forms a
voltage divider with the load impedance that reduces the effective
gain. For example, a 500 Ω differential load has 1 dB lower
effective gain than a high (10 kΩ) differential load impedance.
BIAS CIRCUIT
A band gap reference circuit generates the proportional-toabsolute temperature (PTAT) as well as temperature-independent
reference currents used by different sections. The mixer current
can be reduced via an external resistor between the BIAS pin
and ground. When the BIAS pin is open, the mixer runs at
maximum current and hence the greatest dynamic range. The
mixer current can be reduced by placing a resistance to ground;
therefore, reducing overall power consumption, noise figure,
and IIP3. The effect on each of these parameters is shown in
Figure 10, Figure 13, and Figure 14.
Rev. 0 | Page 14 of 28
ADL5387
APPLICATIONS INFORMATION
BASIC CONNECTIONS
Figure 47 shows the basic connections schematic for the ADL5387.
POWER SUPPLY
The nominal voltage supply for the ADL5387 is 5 V and is
applied to the VPA, VPB, VPL, and VPX pins. Ground should
be connected to the COM, CML, and CMRF pins. Each of
the supply pins should be decoupled using two capacitors;
recommended capacitor values are 100 pF and 0.1 µF.
LOCAL OSCILLATOR (LO) INPUT
The LO port is driven in a single-ended manner. The LO signal
must be ac-coupled via a 1000 pF capacitor directly into LOIP,
and LOIN is ac-coupled to ground also using a 1000 pF capacitor.
The LO port is designed for a broadband 50 Ω match and
therefore exhibits excellent return loss from 100 MHz to 4 GHz.
The LO return loss can be seen in
the LO input configuration.
Figure 20. Figure 46 shows
RFC
LO INPUT
1000pF
Figure 46. Single-Ended LO Drive
The recommended LO drive level is between −6 dBm and
+6 dBm. The LO frequency at the input to the device should be
twice that of the desired LO frequency at the mixer core. The
applied LO frequency range is between 100 MHz and 4 GHz.
ETC1-1-13
1000pF
8
LOIP
9
LOIN
06764-047
120nH120nH
1000pF 1000pF
V
POS
100pF100pF0.1µF
V
POS
0.1µF
100pF
Figure 47. Basic Connections Schematic for ADL5387
242322212019
RFIP
ADL5387
LOIN
1000pF
RFIN
CML
CMRF
1
2
3
4
5
6
VPA
COM
BIAS
VPL
VPL
VPL
CMRF
CML
LOIP
789101112
1000pF
LO
VPX
CMRF
18
VPB
17
VPB
16
QHI
15
QLO
14
IHI
13
ILO
CML
COM
IHI
ILO
QLO
0.1µF
QHI
V
POS
06764-046
Rev. 0 | Page 15 of 28
ADL5387
–
RF INPUT
The RF inputs have a differential input impedance of
approximately 50 Ω. For optimum performance, the RF port
should be driven differentially through a balun. The recommended
balun is M/A-COM ETC1-1-13. The RF inputs to the device
should be ac-coupled with 1000 pF capacitors. Ground-referenced
choke inductors must also be connected to RFIP and RFIN
(recommended value = 120 nH, Coilcraft 0402CS-R12XJL) for
appropriate biasing. Several important aspects must be taken
into account when selecting an appropriate choke inductor for
this application. First, the inductor must be able to handle the
approximately 40 mA of standing dc current being delivered
from each of the RF input pins (RFIP, RFIN). (The suggested
0402 inductor has a 50 mA current rating). The purpose of the
choke inductors is to provide a very low resistance dc path to
ground and high ac impedance at the RF frequency so as not to
affect the RF input impedance. A choke inductor that has a selfresonant frequency greater than the RF input frequency ensures
that the choke is still looking inductive and therefore has a more
predictable ac impedance (jωL) at the RF frequency.
shows the RF input configuration.
120nH
21
1000pF
ETC1-1-13
1000pF
RF INPUT
120nH
Figure 48. RF Input
RFIN
22
RFIP
06764-048
Figure 48
The differential RF port return loss has been characterized as
shown in
Figure 49.
10
–12
–14
–16
–18
–20
S(1, 1) (dB)
–22
–24
–26
–28
00.20.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
Figure 49. Differential RF Port Return Loss
FREQUENCY (GHz)
06764-049
BASEBAND OUTPUTS
The baseband outputs QHI, QLO, IHI, and ILO are fixed
impedance ports. Each baseband pair has a 50 Ω differential
output impedance. The outputs can be presented with differential
loads as low as 200 Ω (with some degradation in linearity and
gain) or high impedance differential loads (500 Ω or greater
impedance yields the same excellent linearity) that is typical of
an ADC. The TCM9-1 9:1 balun converts the differential IF
output to single-ended. When loaded with 50 Ω, this balun
presents a 450 Ω load to the device. The typical maximum
linear voltage swing for these outputs is 2 V p-p differential.
The bias level on these pins is equal to VPOS − 2.8 V. The
output 3 dB bandwidth is 240 MHz.
baseband output configuration.
16
QHI
Figure 50 shows the
QHI
QLO
ILO
15
14
IHI
13
QLO
IHI
ILO
06764-050
Figure 50. Baseband Output Configuration
Rev. 0 | Page 16 of 28
ADL5387
ERROR VECTOR MAGNITUDE (EVM)
PERFORMANCE
EVM is a measure used to quantify the performance of a digital
radio transmitter or receiver. A signal received by a receiver
would have all
various imperfections in the implementation (such as
leakage
constellation points to deviate from the ideal locations.
The ADL5387 shows excellent EVM performance for various
modulation schemes.
over input power range for a point-to-point application with
16 QAM modulation schemes and zero-IF baseband. The
differential dc offsets on the ADL5387 are in the order of a
few mV. However, ac coupling the baseband outputs with 10 µF
capacitors helps to eliminate dc offsets and enhances EVM
performance. With a 10 MHz BW signal, 10 µF ac coupling
capacitors with the 500 Ω differential load results in a high-pass
corner frequency of ~64 Hz which absorbs an insignificant
amount of modulated signal energy from the baseband signal.
By using ac coupling capacitors at the baseband outputs, the dc
offset effects, which can limit dynamic range at low input power
levels, can be eliminated.
–10
–15
–20
–25
EVM (dB)
–30
–35
–40
–45
–50
Figure 51. RF = 140 MHz, IF = 0 Hz, EVM vs. Input Power for a 16 QAM
constellation points at the ideal locations; however,
carrier
, phase noise, and quadrature error) cause the actual
Figure 51 shows typical EVM performance
0
–5
–70100–10–20–30–40–50–60
INPUT POWER (dBm)
10 Msym/s Signal (AC-Coupled Baseband Outputs)
06764-051
Figure 52 shows the EVM performance of the ADL5387 when
ac-coupled, with an IEEE 802.16e WiMAX signal.
0
–5
–10
–15
–20
–25
EVM (dB)
–30
–35
–40
–45
–50
–5020100–10–20–30–40
INPUT POW ER (dBm)
06764-052
Figure 52. RF = 750MHz MHz, IF = 0 Hz, EVM vs. Input Power for a 16 QAM
10 MHz Bandwidth Mobile WiMAX Signal (AC-Coupled Baseband Outputs)
Figure 53 exhibits the zero IF EVM performance of a WCDMA
signal over a wide RF input power range.
0
–5
–10
–15
–20
–25
EVM (dB)
–30
–35
–40
–45
–70–60100–10–20–30–40–50
INPUT POW ER (dBm)
06764-053
Figure 53. RF = 1950 MHz, IF = 0 Hz, EVM vs. Input Power for a WCDMA
(AC-Coupled Baseband Outputs)
Rev. 0 | Page 17 of 28
ADL5387
COS
ωLOt
ω
ω
IF
IF
ω
LSB
ω
ω
USB
LO
SIN
ωLOt
Figure 54. Illustration of the Image Problem
LOW IF IMAGE REJECTION
The image rejection ratio is the ratio of the intermediate
frequency (IF) signal level produced by the desired input
frequency to that produced by the
rejection ratio is expressed in
rejection is critical because the image power can be much
higher than that of the desired signal, thereby plaguing the
down conversion process.
Figure 54 illustrates the image
problem. If the upper sideband (lower sideband) is the desired
band, a 90° shift to the Q channel (I channel) cancels the image
at the lower sideband (upper sideband).
Figure 55 shows the excellent image rejection capabilities of the
ADL5387 for low IF applications, such as CDMA2000. The
ADL5387 exhibits image rejection greater than 45 dB over the
broad frequency range for an IF = 1.23 MHz.
0
–10
–20
–30
–40
–50
IMAGE REJECTION AT 1.23MHz (dB)
–60
–70
50 250 450 650 850 1050 1250 1450 1650 1850
RF Input Frequency for a CDMA2000 Signal, IF = 1.23 MHz
RF INPUT FREQUENCY (MHz)
Figure 55. Image Rejection vs.
image frequency. The image
decibels. Appropriate image
06764-055
0°
0+
ω
–
–
ω
ω
IF
IF
0+
ω
IF
0+
ω
IF
–90°
+90°
0°
IF
0+
ω
IF
06764-054
EXAMPLE BASEBAND INTERFACE
In most direct conversion receiver designs, it is desirable to
select a wanted carrier within a specified band. The desired
channel can be demodulated by tuning the LO to the appropriate
carrier frequency. If the desired RF band contains multiple
carriers of interest, the adjacent carriers would also be down
converted to a lower IF frequency. These adjacent carriers can
be problematic if they are large relative to the wanted carrier as
they can overdrive the baseband signal detection circuitry. As a
result, it is often necessary to insert a filter to provide sufficient
rejection of the adjacent carriers.
It is necessary to consider the overall source and load impedance
presented by the ADL5387 and ADC input to design the filter
network. The differential baseband output impedance of the
ADL5387 is 50 Ω. The ADL5387 is designed to drive a high
impedance ADC input. It may be desirable to terminate the
ADC input down to lower impedance by using a terminating
resistor, such as 500 Ω. The terminating resistor helps to better
define the input impedance at the ADC input. The order and
type of filter network depends on the desired high frequency
rejection required, pass-band ripple, and group delay. Filter
design tables provide outlines for various filter types and orders,
illustrating the normalized inductor and capacitor values for a
1 Hz cutoff frequency and 1 Ω load. After scaling the normalized
prototype element values by the actual desired cut-off frequency
and load impedance, the series reactance elements are halved to
realize the final balanced filter network component values.
Rev. 0 | Page 18 of 28
ADL5387
Ω
As an example, a second-order, Butterworth, low-pass filter
design is shown in
Figure 56 where the differential load impedance
is 500 Ω, and the source impedance of the ADL5387 is 50 Ω.
The normalized series inductor value for the 10-to-1, load-tosource impedance ratio is 0.074 H, and the normalized shunt
capacitor is 14.814 F. For a 10.9 MHz cutoff frequency, the
single-ended equivalent circuit consists of a 0.54 µH series
inductor followed by a 433 pF shunt capacitor.
The balanced configuration is realized as the 0.54 µH inductor
is split in half to realize the network shown in
RS = 50
V
S
R
S
= 0.1
R
L
RS = 50Ω
V
S
R
S
= 25Ω
2
V
S
R
S
= 25Ω
2
Figure 56. Second-Order, Butterworth, Low-Pass Filter Design Example
L
= 0.074H
N
NORMALIZE D
SINGLE- ENDED
CONFIGURAT ION
0.54µH
DENORMALIZ ED
SINGLE- ENDED
EQUIVALENT
0.27µH
BALANCED
CONFIGURAT ION
0.27µH
C
N
Figure 56.
14.814F
433pF
433pF
R
= 500Ω
L
f
= 1Hz
C
= 500Ω
R
L
f
= 10.9MHz
C
R
L
= 250Ω
2
R
L
= 250Ω
2
06764-056
A complete design example is shown in Figure 59. A sixth-order
Butterworth differential filter having a 1.9 MHz corner frequency
interfaces the output of the ADL5387 to that of an ADC input.
The 500 Ω load resistor defines the input impedance of the
ADC. The filter adheres to typical direct conversion WCDMA
applications, where 1.92 MHz away from the carrier IF frequency,
1 dB of rejection is desired and 2.7 MHz away 10 dB of rejection
is desired.
Figure 57 and Figure 58 show the measured frequency response
and group delay of the filter.
10
5
0
–5
–10
MAGNITUDE RESPONSE (dB)
–15
–20
033.02.52.01.51.00.5
FREQUENCY (MHz )
.5
06764-157
Figure 57. Baseband Filter Response
900
800
700
600
500
DELAY (ns)
400
300
200
100
00. 20.40.6 0. 81.01.21.41.61.8
FREQUENCY (MHz )
06764-158
Figure 58. Baseband Filter Group Delay
Rev. 0 | Page 19 of 28
ADL5387
V
RFC
120nH120nH
V
POS
100pF0.1µF
POS
0.1µF100pF
24232221 2 019
1
VPA
2
COM
3
BIAS
4
VPL
5
VPL
6
VPL
ETC1-1-13
1000pF
1000pF
RFIP
CMRF
CMRF
RFIN
CMRF
ADL5387
CML
LOIP
LOIN
CML
789101112
1000pF1000pF
LO
CML
VPX
VPB
VPB
QHI
QLO
ILO
COM
C
AC
27µH
27µH
27µH
27µH
27µH
10µH
91pF
68pF
500Ω
ADC INPUTADC INPUT
10µH
10µH
91pF
68pF
500Ω
10µH
10µF
C
10µF
18
100pF0.1µF
17
16
15
14
IHI
13
V
POS
C
10µF
C
10µF
270pF
AC
27µH
AC
27µH
270pF
AC
27µH
06764-159
Figure 59. Sixth Order Low-Pass Butterworth Baseband Filter Schematic
Rev. 0 | Page 20 of 28
ADL5387
CHARACTERIZATION SETUPS
Figure 60 to Figure 62 show the general characterization bench
setups used extensively for the ADL5387. The setup shown in
Figure 62 was used to do the bulk of the testing and used sinusoidal
signals on both the LO and RF inputs. An automated AgilentVEE program was used to control the equipment over the IEEE
bus. This setup was used to measure gain, IP1dB, IIP2, IIP3, I/Q
gain match, and quadrature error. The ADL5387 characterization
board had a 9-to-1 impedance transformer on each of the
differential baseband ports to do the differential-to-singleended conversion.
The two setups shown in
for making NF measurements.
measuring NF with no blocker signal applied while
Figure 60 and Figure 61 were used
Figure 60 shows the setup for
Figure 61
was used to measure NF in the presence of a blocker. For both
setups, the noise was measured at a baseband frequency of
SNS
OUTPUT
CONTROL
10 MHz. For the case where a blocker was applied, the output
blocker was at 15 MHz baseband frequency. Note that great care
must be taken when measuring NF in the presence of a blocker.
The RF blocker generator must be filtered to prevent its noise
(which increases with increasing generator output power) from
swamping the noise contribution of the ADL5387. At least
30 dB of attention at the RF and image frequencies is desired.
For example, with a 2xLO of 1848 MHz applied to the ADL5387,
the internal 1xLO is 924 MHz. To obtain a 15 MHz output
blocker signal, the RF blocker generator is set to 939 MHz and
the filters tuned such that there is at least 30 dB of attenuation
from the generator at both the desired RF frequency (934 MHz)
and the image RF frequency (914 MHz). Finally, the blocker
must be removed from the output (by the 10 MHz low-pass
filter) to prevent the blocker from swamping the analyzer.
AGILENT N8974A
NOISE FI GURE ANALYZER
HP 6235A
POWER SUPPLY
AGILENT 8665B
SIGNAL GENERATOR
RF
GND
ADL5387
CHAR BOARD
V
POS
LO
6dB PAD
Figure 60. General Noise Figure Measurement Setup
Q
I
IEEE
R1
50Ω
FROM SNS PORT
LOW-PASS
FILTER
INPUT
PC CONTROLLER
IEEE
06764-057
Rev. 0 | Page 21 of 28
ADL5387
R&S SMT03
SIGNAL GE NERATOR
HP 6235A
POWER SUPPLY
AGILENT 8665B
SIGNAL GE NERATOR
BAND-PASS
TUNABLE FILTER
6dB PAD
RF
GND
ADL5387
CHAR BOARD
V
POS
LO
6dB PAD
BAND-PASS
CAVITY FI LTER
BAND-REJECT
TUNABLE FILTER
Q
50Ω
6dB PAD
I
R1
LOW-PASS
FILTER
HP87405
LOW NOISE
PREAMP
R&S FSEA30
SPECTRUM ANALYZ ER
06764-058
Figure 61. Measurement Setup for Noise Figure in the Presence of a Blocker
3dB PAD
RF
R&S SMT-06
RF
3dB PAD
3dB PAD
AGILENT
11636A
AMPLIFIER
VP GND
OUTIN
3dB PAD
RF
R&S SMT-06
GND
V
POS
AGILENT E3631
PWER SUPPLY
IEEEIEEEIEEEIEEE
AGILENT E8257D
SIGNAL G ENERATOR
IEEE
PC CONT ROLL ER
6dB PAD
RF
ADL5387
CHAR BOARD
LO
6dB PAD
R&S FSEA30
SPECTRUM ANALYZ ER
Q
I
6dB PAD
6dB PAD
IEEE
RF
INPUT
VECTOR VOLTMETER
SWITCH
MATRIX
HP 8508A
IEEE
A AND B
INPUT CHANNELS
06764-059
Figure 62. General ADL5387 Characterization Setup
Rev. 0 | Page 22 of 28
ADL5387
V
EVALUATION BOARD
The ADL5387 evaluation board is available. The board can be
used for single-ended or differential baseband analysis. The default
configuration of the board is for single-ended baseband analysis.
T1
C11
RFIP
CMRF
ADL5387
LOIP
LOIN
C7C6
T4
C10
V
POS
R1
POS
R3
C3C4
RFC
L2L1
R8R7
242322212019
CMRF
1
C2C1
R2
VPA
2
COM
3
BIAS
4
VPL
5
VPL
6
VPL
CML
7 8 9 101112
R17
RFIN
CML
C5
VPX
CMRF
18
VPB
17
VPB
16
QHI
15
QLO
14
IHI
13
ILO
CML
COM
C12
C13
R6
C8
R14
R16
R4
R13
R9
R10
R11
R12
V
POS
C9
Q OUTPUT OR QHI
R15
T2
QLO
I OUTPUT OR IHI
R5
T3
ILO
LO
Figure 63. Evaluation Board Schematic
Rev. 0 | Page 23 of 28
6764-060
ADL5387
Table 4. Evaluation Board Configuration Options
Component Function Default Condition
VPOS, GND Power Supply and Ground Vector Pins. Not Applicable
R1, R3, R6 Power Supply Decoupling. Shorts or power supply decoupling resistors. R1, R3, R6 = 0 Ω (0805)
C1, C2, C3,
C4, C8, C9
C5, C6, C7,
C10, C11
R4, R5,
R9 to R16
L1, L2,
R7, R8
T2, T3
C12, C13
R17
T1
R2 R
The capacitors provide the required dc coupling up to 2 GHz.
AC Coupling Capacitors. These capacitors provide the required ac coupling from
50 MHz to 2 GHz.
Single-Ended Baseband Output Path. This is the default configuration of the evaluation
board. R14 to R16 and R4, R5, and R13 are populated for appropriate balun interface.
R9, R10 and R11, R12 are not populated. Baseband outputs are taken from QHI and IHI.
The user can reconfigure the board to use full differential baseband outputs. R9 to R12
provide a means to bypass the 9:1 TCM9-1 transformer to allow for differential baseband
outputs. Access the differential baseband signals by populating R9 to R12 with 0 Ω and
not populating R4, R5, R13 to R16. This way the transformer does not need to be removed.
The baseband outputs are taken from the SMAs of Q_HI, Q_LO, I_HI, and I_LO.
Input Biasing. Inductance and resistance sets the input biasing of the common base
input stage. Default value is 120 nH.
IF Output Interface. TCM9-1 converts a differential high impedance IF output to a singleended output. When loaded with 50 Ω, this balun presents a 450 Ω load to the device.
The center tap can be decoupled through a capacitor to ground.
Decoupling Capacitors. C12 and C13 are the decoupling capacitors used to reject noise
on the center tap of the TCM9-1.
LO Input Interface. The LO is driven as a single-ended signal. Although, there is no
performance change for a differential signal drive, the option is available by placing a
transformer (T4, ETC1-1-13) on the LO input path.
RF Input Interface. ETC1-1-13 is a 1:1 RF balun that converts the single-ended RF input
to differential signal.
. Optional bias setting resistor. See the Bias Circuit section to see how to use this feature. R2 = Open