FEATURES
Complete RF Detector/Controller Function
Selectable Dual Outputs
49 dB Range at 0.9 GHz (–47.6 dBm to +1.5 dBm re 50 )
Accurate Scaling from 0.1 GHz to 2.5 GHz
Temperature-Stable Linear-in-dB Response
Log Slope of 22 mV/dB
True Integration Function in Control Loop
Low Power: 23 mW at 2.7 V
Power-Down to 11 W
APPLICATIONS
Single-Band, Dual-Band, and Triband Mobile Handsets
(GSM, DCS, PCS, EDGE)
Wireless Terminal Devices
Transmitter Power Control
GENERAL DESCRIPTION
The AD8316 is a complete, low cost subsystem for the precise
control of dual RF power amplifiers (PAs) operating in the
frequency range 0.1 GHz to 2.5 GHz and over a typical dynamic
range of 50 dB. The device is a dual-output version of the AD8315
and intended for use in dual-band or triband cellular handsets
and other battery-operated wireless devices where a separate
FUNCTIONAL BLOCK DIAGRAM
power control signal is required for each band. The logarithmic
amplifier technique provides a much wider measurement range
and better accuracy than is possible using controllers based on
diode detectors. In particular, multiband and multimode cellular designs can benefit from the temperature-stable (–30°C to
+85°C) operation over all cellular telephony frequencies.
Its high sensitivity allows control at low input signal levels, thus
reducing the amount of power that needs to be coupled to the
detector. The selected output, OUT1 or OUT2, has the voltage
range and current drive to directly connect to the gain control
pin of most handset power amplifiers; the deselected output is
pulled low to ensure that the inactive PA remains off. Each
output has a dedicated integrating filter capacitor that allows
separate control loop settings for each PA. OUT1 and OUT2 can
swing from 125 mV above ground to within 100 mV below the
supply voltage. Load currents of up to 12 mA can be supported.
The setpoint control input applied to pin VSET has an operating
range of 0.25 V to 1.4 V. The input resistance of the setpoint
interface is over 100 MΩ, and the bias current is typically 0.5 µA.
The AD8316 is available in 10-lead MSOP and 16-lead LFCSP
packages and consumes 8.5 mA from a 2.7 V to 5.5 V supply.
When it is powered down, the sleep current is 4 µA.
VPOS
ENBL
BSEL
RFIN
COMM
LOW NOISE
GAIN BIAS
DETDETDETDET
10dB
OFFSET
COMPENSATION
LOW NOISE
BAND GAP
REFERENCE
DET
10dB10dB10dB
REV. C
Information furnished by Analog Devices is believed to be accurate and
reliable. However, no responsibility is assumed by Analog Devices for its
use, nor for any infringements of patents or other rights of third parties that
may result from its use. 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.
*Stresses above those listed under Absolute Maximum Ratings may cause perma-
nent 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.
PIN FUNCTION DESCRIPTIONS
Pin No.
MSOP LFCSPMnemonic Function
11 RFINRF Input.
22 ENBLConnect to VPOS for Normal
Operation. Connect pin to
ground for disable mode.
33 VSETSetpoint Input.
44 FLT1Integrator Capacitor for OUT1.
Connect between FLT1 and
COMM.
56 BSELBand Select. LO = OUT2,
HI = OUT1.
67 FLT2Integrator Capacitor for OUT2.
Connect between FLT2 and
COMM.
79 OUT2Band 2 Output.
810, 14COMMDevice Common (Ground).
911OUT1Band 1 Output.
1012VPOSPositive Supply Voltage: 2.7 V
to 5.5 V.
5, 8, 13,NCNo Connection.
15, 16
ORDERING GUIDE
ModelTemperature Range Package DescriptionPackage Option Branding
AD8316ARM–30°C to +85°C10-Lead MSOP, TubeRM-10J8A
AD8316ARM-REEL7 –30°C to +85°CMSOP, 7" Tape and ReelRM-10J8A
AD8316-EVALMSOP Evaluation Board
AD8316ACP-REEL–30°C to +85°C16-Lead LFCSP, 13" Tape and Reel CP-16-3J8A
AD8316ACP-REEL7–30°C to +85°CLFCSP, 7" Tape and ReelCP-16-3J8A
AD8316ACP-EVALLFCSP Evaluation Board
CAUTION
ESD (electrostatic discharge) sensitive device. Electrostatic charges as high as 4000 V readily
accumulate on the human body and test equipment and can discharge without detection. Although the
AD8316 features proprietary ESD protection circuitry, permanent damage may occur on devices
subjected to high energy electrostatic discharges. Therefore, proper ESD precautions are recommended
to avoid performance degradation or loss of functionality.
REV. C
–3–
Page 4
AD8316
–Typical Performance Characteristics
–73–53–43–33–13–23–3–63
1.6
1.4
1.2
1.0
– V
SET
V
0.8
0.6
0.4
0.2
–60
–73–53–43–33–13–23–3–63
1.6
1.4
1.2
1.0
– V
0.8
SET
V
0.6
0.4
TPC 1. V
–30C
INPUT AMPLITUDE – dBV
2.5GHz
0.1GHz
–40–30–200
–50
INPUT AMPLITUDE – dBm
vs. Input Amplitude
SET
INPUT AMPLITUDE – dBV
+85C
+25C
0.9GHz
1.9GHz
–10
–30C
+85C
+25C
10
4
3
2
1
0
–1
–2
ERROR – dB
–73–53–43–33–13–23–3–63
4
3
2
1
0
2.5GHz
ERROR – dB
–1
–2
–3
–4
–60
–50
INPUT AMPLITUDE – dBV
1.9GHz
0.1GHz
–40–30–200
INPUT AMPLITUDE – dBm
–10
1.9GHz
0.9GHz
10
TPC 4. Log Conformance vs. Input Amplitude at
Selected Frequencies
–73–53–43–33–13–23–3–63
1.6
1.4
1.2
1.0
–30C
– V
0.8
SET
V
0.6
+85C
0.4
INPUT AMPLITUDE – dBV
4
3
+25C
2
+85C
1
0
–1
–2
ERROR – dB
0.2
0
–60
TPC 2. V
–40–30–200
–50
INPUT AMPLITUDE – dBm
and Log Conformance vs. Input
SET
Amplitude at 0.1 GHz
–73–53–43–33–13–23–3–63
1.6
1.4
1.2
1.0
–30C
– V
0.8
SET
V
0.6
0.4
0.2
0
–60
TPC 3. V
+85C
+25C
–50
INPUT AMPLITUDE – dBV
+85C
+25C
–30C
–40–30–200
INPUT AMPLITUDE – dBm
and Log Conformance vs. Input
SET
Amplitude at 0.9 GHz
–10
–10
–3
–4
10
4
3
2
1
0
–1
–2
–3
–4
10
ERROR – dB
0.2
0
–60
TPC 5. V
+25C
–30C
–40–30–200
–50
INPUT AMPLITUDE – dBm
and Log Conformance vs. Input
SET
Amplitude at 1.9 GHz
–73–53–43–33–13–23–3–63
1.6
1.4
1.2
1.0
– V
0.8
SET
V
0.6
0.4
0.2
+25C
0
–60
TPC 6. V
–30C
+85C
–50
INPUT AMPLITUDE – dBV
+25C
+85C
–30C
–40–30–200
INPUT AMPLITUDE – dBm
and Log Conformance vs. Input
SET
Amplitude at 2.5 GHz
–10
–10
–3
–4
10
4
3
2
1
0
–1
–2
–3
–4
10
REV. C–4–
ERROR – dB
Page 5
AD8316
–73–53–43–33–13–23–3–63
4
3
2
1
0
ERROR – dB
–1
–2
–3
–4
–60
–50
INPUT AMPLITUDE – dBV
+85C
–30C
–40–30–200
INPUT AMPLITUDE – dBm
–10
10
TPC 7. Distribution of Error at Temperature after Ambient
Normalization vs. Input Amplitude, 3 Sigma to Either Side
of Mean, 0.1 GHz
–73–53–43–33–13–23–3–63
4
3
2
1
0
ERROR – dB
–1
–2
–3
–4
–60
–50
INPUT AMPLITUDE – dBV
+85C
–30C
–40–30–200
INPUT AMPLITUDE – dBm
–10
10
TPC 8. Distribution of Error at Temperature after Ambient
Normalization vs. Input Amplitude, 3 Sigma to Either Side
of Mean, 0.9 GHz
–73–53–43–33–13–23–3–63
4
3
2
1
0
ERROR – dB
–1
–2
–3
–4
–60
–50
INPUT AMPLITUDE – dBV
+85C
–30C
–40–30–200
INPUT AMPLITUDE – dBm
–10
10
TPC 10. Distribution of Error at Temperature after Ambient
Normalization vs. Input Amplitude, 3 Sigma to Either Side
of Mean, 1.9 GHz
–73–53–43–33–13–23–3–63
4
3
2
1
0
ERROR – dB
–1
–2
–3
–4
–60
–50
INPUT AMPLITUDE – dBV
+85C
–30C
–40–30–200
INPUT AMPLITUDE – dBm
–10
10
TPC 11. Distribution of Error at Temperature after Ambient
Normalization vs. Input Amplitude, 3 Sigma to Either Side
of Mean, 2.5 GHz
3100
2800
2500
2200
1900
1600
1300
RESISTANCE –
1000
700
400
100
REV. C
X (MSOP)
FREQ MSOP CHIP-SCALE (LFCSP)
(GHz)
R –
jX
R –
0.1
3100 –
600 –
320 –
110 –
j1220
j194
j134
j86
0.9
1.9
X (LFCSP)
R (MSOP)
0
0.51.52.0
2.5
1.0
FREQUENCY – GHz
2630 –
1000 –
620 –
435 –
R
jX
j1800
j270
j130
j110
R (CSP)
TPC 9. Input Impedance vs. Frequency
0
–200
–400
–600
–800
–1000
X
–1200
REACTANCE –
–1400
–1600
–1800
–2000
2.5
8
6
4
DECREASING
V
ENBL
SUPPLY CURRENT – mA
2
0
0.8
1.0 1.1 1.21.41.31.5
0.9
V
TPC 12. Supply Current vs. V
INCREASING
V
ENBL
1.6 1.7 1.8 1.9 2.0 2.1 2.2 2.3
– V
ENBL
ENBL
–5–
Page 6
AD8316
23
22
+25C
SLOPE – mV/dB
21
20
0
0.51.0
FREQUENCY – GHz
–30C
+85C
1.52.02.5
TPC 13. Slope vs. Frequency at Selected Temperatures
22.5
22.0
21.5
SLOPE – mV/dB
21.0
0.9GHz
0.1GHz
1.9GHz
2.5GHz
–60
–62
–64
INTERCEPT – dBm
–66
–68
0
0.51.02.01.5
+25C
–30C
+85C
2.5
FREQUENCY – GHz
TPC 16. Intercept vs. Frequency at Selected Temperatures
–58
–60
–62
–64
–66
INTERCEPT – dBm
–68
1.9GHz
2.5GHz
0.1GHz
0.9GHz
20.5
2.5
3.03.5
4.04.55.0
VS – V
5.5
TPC 14. Slope vs. Supply Voltage
0
40
30
20
10
0
–10
–20
AMPLITUDE – dB
–30
–40
–50
–60
1
10100501k10k100k1M
C
FLT
= 220pF
C
= 0pF
FLT
FREQUENCY – Hz
10M 100M
–20
–40
–60
–80
–100
–120
–140
–160
–180
–200
–210
TPC 15. AC Response from VSET to OUT1 and OUT2
PHASE – Degrees
–70
2.5
3.03.55.04.5
4.0
V
– V
TPC 17. Intercept vs. Supply Voltage
10000
–50dBm
Hz
NOISE SPECTRAL DENSITY – nV/
1000
100
10
100
–40dBm
–25dBm
0dBm
–20dBm
–10dBm
1k10k10M1M
FREQUENCY – Hz
100k
RF INPUT
28dBm
TPC 18. Output Noise Spectral Density, RL = ,
= 220 pF, by RF Input Amplitude
C
FLT
5.5
100M
REV. C–6–
Page 7
3.5
V
– V
2.7
V
OUT
– V
2.2
2.9
2.8
2.6
2.8
2.7
2.5
2.4
3.0
6mA
2.3
SHADING INDICATES 3 SIGMA
2mA
AD8316
– V
OUT
V
3.3
3.1
2.9
2.7
2.5
2.3
2.1
1.9
2.7
2.82.93.23.1
I
LOAD
0mA
3.0
2mA
4mA
10mA
12 mA
8mA
6mA
3.43.5
V
– V
3.3
TPC 19. Maximum OUT Voltage vs. Supply Voltage by
Load Current, AD8316 Sourcing
AVERAGING = 16 SAMPLES
C
= 68pF
FLT
C
= 220pF
FLT
V
OUT
GND
50mV PER
VERTICAL
DIVISION
TPC 22. Distribution of Maximum OUT Voltage vs. Supply
Voltage with 2 mA and 6 mA Loads, 3 Sigma to Either
Side of Mean, AD8316 Sourcing
AVERAGING = 16 SAMPLES
V
10mV PER VERTICAL DIVISION
OUT2
GND
BSEL INPUT
GND
V
ENBL
2s PER HORIZONTAL DIVISION
1V PER
VERTICAL
DIVISION
TPC 20. ENBL Response Time, Rise/Fall Time = 250 ns
H-P 8110A
PULSE
C
FLT
0.1F
R
1k
GENERATOR
PULSE OUT
TEK P6204
FET PROBE
TEK P6204
FET PROBE
L
TEK 1103
PWR SUPPLY
TEK TDS3054
SCOPE
H-P 8648B
SIGNAL
GENERATOR
52.3
C
FLT
RF OUT
2.7V
0.1GHz
60dBm
AD8316
RFIN
ENBL
VSET
FLT1
BSEL
BSEL HIGH
*
BSEL LOW
*
VPOS
OUT1
COMM
OUT2
FLT2
OUT1;
OUT2
2.7V
TPC 21. Test Setup for ENBL Response Time
1V PER
VERTICAL
DIVISION
GND
2s PER HORIZONTAL DIVISION
TPC 23. BSEL Response Time, ENBL Grounded
0.1F
C
FLT
H-P 8110A
PULSE
GENERATOR
PULSE OUT
TEK P6204
FET PROBE
TEK P6204
FET PROBE
R
L
1k
H-P 8648B
SIGNAL
GENERATOR
RF OUT
52.3
0.1GHz
–60dBm
AD8316
RFIN
ENBL
VSET
FLT1
C
FLT
BSEL*
*BSEL HIGH OUT1;
BSEL LOW
OUT2
2.7V
VPOS
OUT1
COMM
OUT2
FLT2
TPC 24. Test Setup for BSEL Response Time
TEK 1103
PWR SUPPLY
TEK TDS3054
SCOPE
REV. C
–7–
Page 8
AD8316
AVERAGING = 16 SAMPLES
C
= 220pF
FLT
V
OUT
50mV PER
VERTICAL
DIVISION
2V PER
VERTICAL
DIVISION
GND
GND
C
= 68pF
FLT
V
POS/VENBL
250ns
RISE TIME
2s PER HORIZONTAL DIVISION
TPC 25. Power-On and Power-Off Response with
VSET Grounded, Rise/Fall Time = 250 ns
AVERAGING = 16 SAMPLES
C
= 68pF
FLT
50mV PER
VERTICAL
DIVISION
2V PER
VERTICAL
DIVISION
GND
GND
V
OUT
C
= 220pF
FLT
V
POS/VENBL
1s
RISE TIME
2s PER HORIZONTAL DIVISION
AVERAGING = 16 SAMPLES
1V PER
PULSED RF
INPUT
0.1GHz, –3dBm
GND
GND
V
OUT
VERTICAL
DIVISION
100ns PER HORIZONTAL DIVISION
TPC 28. Pulse Response Time, Full-Scale Amplitude
Change, Open Loop, C
V
OUT
1V PER
VERTICAL
GND
GND
DIVISION
2s PER HORIZONTAL DIVISION
= 0 pF
FLT
AVERAGING = 16 SAMPLES
PULSED RF
INPUT
0.1GHz, –3dBm
TPC 26. Power-On and Power-Off Response with
µ
AD811
732
FLT
s
H-P 8110A
PULSE
GENERATOR
TEK P6204
FET PROBE
TEK P6204
FET PROBE
R
L
1k
PWR SUPPLY
TEK TDS3054
PULSE OUT
49.9
TEK 1103
SCOPE
VSET Grounded, Rise/Fall Time = 1
H-P 8648B
SIGNAL
GENERATOR
52.3
0.1GHz
–60dBm
RF OUT
C
FLT
*BSEL HIGH OUT1;
BSEL LOW
AD8316
RFIN
ENBL
VSET
FLT1
BSEL*
COMM
OUT2
VPOS
OUT1
OUT2
FLT2
C
TPC 27. Test Setup for Power-On and Power-Off
Response with VSET Grounded
TPC 29. Pulse Response Time, Full-Scale Amplitude
OUT1;
OUT2
FLT
VPOS
OUT1
COMM
OUT2
FLT2
= 68 pF
PULSE OUT
2.7V
C
EXT TRIG
0.1F
FLT
H-P 8110A
PULSE
GENERATOR
TEK P6204
FET PROBE
TEK P6204
FET PROBE
R
L
TEK 1103
1k
PWR SUPPLY
TEK TDS3054
SCOPE
TRIG OUT
Change, Open Loop, C
H-P 8648B
SIGNAL
GENERATOR
0.1GHz
0dBm
PWR DIVIDER
52.3
C
RFOUT
3dB
FLT
10MHz REF OUT
PULSE MODE IN
2.7V
0.4V
2.7V
*
BSEL HIGH
BSEL LOW
AD8316
RFIN
ENBL
VSET
FLT1
BSEL*
TPC 30. Test Setup for Pulse Response Time
REV. C–8–
Page 9
AD8316
AVERAGING = 16 SAMPLES
V
OUT
V
POS
INPUT
250ns
RISE TIME
VERTICAL DIVISION
2s PER HORIZONTAL DIVISION
RISE TIME
2V PER
1s
10mV PER
VERTICAL
DIVISION
TPC 31. Power-On and Power-Off Response with
VSET and ENBL Grounded
732
H-P 8110A
PULSE
GENERATOR
TEK P6204
FET PROBE
TEK P6204
FET PROBE
R
L
1k
PWR SUPPLY
TEK TDS3054
PULSE OUT
49.9
TEK 1103
SCOPE
H-P 8648B
SIGNAL
GENERATOR
RFOUT
52.3
0.1GHz
–60dBm
AD8316
RFIN
ENBL
VSET
FLT1
C
FLT
BSEL*
*BSEL HIGH OUT1;
BSEL LOW
OUT2
VPOS
OUT1
COMM
OUT2
FLT2
AD811
C
FLT
TPC 32. Test Setup for Power-On and Power-Off
Response with VSET and ENBL Grounded
GENERAL DESCRIPTION AND THEORY
The AD8316 is a wideband logarithmic amplifier (log amp) with
two selectable outputs suitable for dual-band/dual-mode power
amplifier control. It is strictly optimized for power control applications rather than for use as a measurement device. Figure 1
shows its main features in block schematic form. The output
pins, OUT1 and OUT2, are intended to be applied directly to
the automatic power control (APC) pins of two distinct power
amplifiers. When the band select pin, BSEL, directs one of the
controller outputs to servo its amplifier toward the setpoint
indicated by the power control pin VSET, the other output is
forced to ground, disabling the second amplifier. Each output
has a dedicated filter pin, FLT1 and FLT2, that allows the
filtering and loop dynamics for each control loop to be optimized independently.
Basic Theory
Logarithmic amplifiers provide a type of compression in which a
signal with a large range of amplitudes is converted to one of
a smaller range. The use of the logarithmic function uniquely
results in the output representing the decibel value of the input.
The fundamental mathematical form is
V
VV
=log
OUTSLP
IN
V
Z
(1)
Here VIN is the input voltage and VZ is called the intercept (voltage) because when VIN = VZ the argument of the logarithm is
unity, and thus the result is zero; V
is called the slope (volt-
SLP
age), which is the amount by which the output changes for a
certain change in the ratio (V
IN/VZ
).
Because log amps do not respond to power, but only to voltages,
and the calibration of the intercept is waveform dependent and
only quoted for a sine wave signal, the equivalent power response
can be written as
VVPP
=(–)
OUTDBINZ
(2)
where the input power PIN and the equivalent intercept PZ are
both expressed in dBm (thus, the quantity in the parentheses is
simply a number of decibels), and V
is the slope expressed as
DB
so many mV/dB. When base 10 logarithms are used, denoted by
, V
the function log
corresponds to 20 dB, V
represents V/dec, and since a decade
10
SLP
/20 represents the change in V/dB. For
SLP
the AD8316, a nominal (low frequency) slope of 22 mV/dB
(corresponding to a V
was chosen, and the intercept V
of 0.022 mV/dB × 20 dB = 440 mV)
SLP
was placed at the equivalent of
Z
–74 dBV, or 199 µV rms, for a sine wave input. This corre-
sponds to a power level of –61 dBm when the net resistive part
of the input impedance of the log amp is 50 Ω. However, both
the slope and the intercept are dependent on frequency (see for
example, TPC 13 and TPC 16).
For a log amp with a slope V
of +22 mV/dB and an inter-
DB
cept at –61 dBm, the output voltage for an input power of
–30 dBm is 0.022 × (–30 – [–61]) = 0.682 V.
REV. C
VPOS
ENBL
BSEL
RFIN
COMM
LOW NOISE
GAIN BIAS
10dB
OFFSET
COMPENSATION
LOW NOISE
BAND GAP
REFERENCE
DETDETDETDETDET
10dB10dB10dB
OUTPUT
ENABLE
DELAY
INTERCEPT
POSITIONING
Figure 1. Block Schematic of the AD8316
–9–
HI-Z
HI-Z
1.35
V–I
1.35
FLT1
OUT1
LOW NOISE
RAIL-TO-RAIL
BUFFERS
OUT2
FLT2
VSET
325mV TO
1.4V = 49dB
Page 10
AD8316
Further details about the structure and function of log amps are
provided in data sheets for other log amps produced by Analog
Devices. The AD640 and AD8307 include detailed discussions
of the basic principles of operation and explain why the intercept
depends on waveform, an important consideration when complex
modulation is imposed on an RF carrier.
The intercept need not correspond to a physically realizable part
of the signal range for the log amp. Thus, for the AD8316, the
specified intercept is –62 dBm at 0.9 GHz, whereas the lowest
acceptable input for accurate measurement (+1 dB error) is
–48 dBm. At 2.5 GHz, the +1 dB error point shifts to –52 dBm.
This positioning of the intercept is deliberate and ensures that
the VSET voltage is within the capabilities of certain DACs,
whose outputs cannot swing below 200 mV. Figure 2 shows the
0.9 GHz response of the AD8316; the vertical axis represents
the value required at the power control pin VSET to null the
control loop rather than the voltage at the OUT1 or OUT2
pins.
V
, dBV
IN
100V
80dBV
1.5V
1.0V
SET
V
0.5V
0
–67dBm
–62dBm
1mV
60dBV
SLOPE = 22mV/dB
0.308V AT –48dBm
IDEAL
–47dBm–27dBm–7dBm+13dBm
IN
10mV
40dBV
1.408V AT +2dBm
ACTUAL
P
IN
100mV
20dBV
1V (rms)
0dBV
Figure 2. Basic Calibration of the AD8316 at 0.9 GHz
Controller-Mode Log Amps
The AD8316 combines the two key functions required for the
measurement and control of the power level over a moderately wide dynamic range. First, it provides the amplification
needed to respond to small signals with a chain of four amplifier/limiter cells, each with a small signal gain of 10 dB and a
bandwidth of approximately 4 GHz (see Figure 1). At the
output of each of these amplifier stages is a full-wave rectifier, essentially a square-law detector cell that converts the
RF signal voltages to a fluctuating current having an average
value that increases with signal level. A passive detector stage is
added ahead of the first stage. These five detectors are separated
by 10 dB, spanning 50 dB of dynamic range. Their outputs are
in the form of a differential current, making summation a
simple matter. It is readily shown that the summed output can
closely approximate a logarithmic function. The overall accuracy at the extremes of the total range, viewed as the deviation
from an ideal logarithmic response, that is, the law-conformance
error, can be judged by referring to TPC 4, which shows that
errors across the central 40 dB are moderate. Other performance curves show how conformance to an ideal logarithmic
function varies with supply voltage, temperature, and frequency.
In a device intended for measurement applications, this current
would be converted to an equivalent voltage to provide the
) function shown in Equation 1. However, the design of
log(V
IN
the AD8316 differs from standard practice in that its output
needs to be a low noise control voltage for an RF power amplifier, not a direct measure of the input level. Further, it is highly
desirable that this voltage be proportional to the time integral of
the error between the actual input V
and a dc voltage V
IN
SET
(applied to Pin 3, VSET) that defines the setpoint, that is, a
target value for the power level, typically generated by a DAC.
This is achieved by converting the difference between the sum
of the detector outputs (still in current form) and an internally
generated current proportional to VSET to a single-sided
current-mode signal. This, in turn, is converted to a voltage
(at FLT1 or FLT2, the low-pass filter capacitor nodes) to
provide a close approximation to an exact integration of the
error between the power present in the termination at the input
of the AD8316 and the setpoint voltage. Finally, the voltages
developed across the ground referenced filter capacitors C
FLT
are buffered by a special low noise amplifier of low voltage
gain (×1.35) and presented at OUT2 or OUT1 for use as the
control voltage for the appropriate RF power amplifier. This
buffer can provide rail-to-rail swings and can drive a substantial load current, including large capacitors. Note: The RF
power delivered by the power amplifier is assumed to increase monotonically with an increasingly positive voltage on its APC control pin.
Band selection in the AD8316 relies on the fact that dual-band/
dual-mode amplifier systems require only one active amplifier at
a time. This allows both amplifier outputs to share the RF input
of the AD8316 (Pin 1, RFIN) as long as the inactive amplifier is
disabled, i.e., it is not delivering RF power. In this case, power
control is directed solely through the selected amplifier. The
AD8316 ensures that the output control pin associated with the
unselected amplifier pulls its APC pin to ground. It is assumed
that the amplifier is essentially disabled when its APC pin is
grounded.
Control Loop Dynamics
To understand how the AD8316 behaves in a complete control
loop, it is necessary to develop an expression for the current in
1.35
and the
IN
VOUT1
9
the integration capacitor as a function of the input V
setpoint voltage V
VSET
3
V
SET
RFIN
1
V
IN
I
. Refer to Figure 3.
SET
SETPOINT
INTERFACE
LOGARITHMIC
RF DETECTION
SUBSYSTEM
= I
log10 (VIN/VZ)
DET
SLP
I
SET
I
DET
= V
SET
I
/ 4.15k
ERR
FLT1
4
C
FLT
Figure 3. Behavioral Model for the AD8316 with
OUT1 Selected
First, write the summed detector currents as a function of the
input:
IIVV
=log (/)
DETSLPINZ
where I
is the partially filtered demodulated signal, whose
DET
10
(3)
exact average value will be extracted through the subsequent
integration step; I
of 106 mA per decade (that is, 5.3 mA/dB); V
is the current-mode slope, and has a value
SLP
is the input in
IN
REV. C–10–
Page 11
AD8316
volts rms; and VZ is the effective intercept voltage, which, as
previously noted, is dependent on waveform but is 199 µV rms
for a sine wave input. Now, the current generated by the setpoint
interface is simply
IVk
=Ω/.415
SETSET
I
, the difference between this current and I
ERR
the loop filter capacitor C
on this capacitor, V
. It follows that the voltage appearing
FLT
, is the time integral of the difference
FLT
, is applied to
DET
(4)
current
Vs I I sC
() (–)/=
FLTSETDETFLT
ΩVkIVV
/.–log (/ )415
SETSLPIN Z
=
sC
The control output V
10
FLT
is slightly greater than this, since the
OUT
(5)
(6)
gain of the output buffer is ×1.35. Also, an offset voltage is deliberately introduced in this stage, but this is inconsequential, since
the integration function implicitly allows for an arbitrary constant
to be added to the form of Equation 6. The polarity is such that
will rise to its maximum value for any value of V
V
OUT
than the equivalent value of V
. In practice, the output will rail
IN
SET
greater
to the positive supply under this condition unless the control
point for understanding the more complex situation that arises
when the gain control law is less than ideal.
This idealized control loop is shown in Figure 4. With some
manipulation, it is found that the characteristic equation of this
system is
Vs
()
=
OUT
VV V VkGV V
()/log (/)
SET GSCSLPGSCO CWZ
−
1
+
where k is the voltage coupling factor from the output of the
power amplifier to the input of the AD8316 (e.g., ×0.1 for a 20 dB
coupler) and T
is a modified time constant (V
O
This is quite easy to interpret. First, it shows that a system of
this sort will exhibit a simple single-pole response, for any power
level, with the customary exponential time domain form for
either increasing or decreasing step polarities in the demand
level V
or the carrier input VCW. Second, it reveals that the
SET
final value of the control voltage V
several fixed factors
VtVV V VkGVV
=∞
=−()/log(/)
()
OUTSET GSCSLPGSCO CWZ
loop through the power amplifier is present. In other words, the
AD8316 seeks to drive the RF power to its maximum value whenever it falls below the setpoint. The use of exact integration results
in a final error that is theoretically zero, and the logarithmic
detection law would ideally result in a constant response time
following a step change of either the setpoint or the power level, if
the power amplifier control function were likewise “linear-in-dB.”
This latter condition is rarely true, however, and it follows that
the loop response time will, in practice, depend on the power level,
and this effect can strongly influence the design of the control loop.
Equation 6 can be clarified by noting that it can be restated in
the following way
Figure 4. Idealized Control Loop for Dynamic
Analysis, OUT1 Selected
V
= kV
IN
RF
V
SET
RESPONSE-SHAPING
OF OVERALL CONTROL
LOOP (EXTERNAL CAP)
Example
VVVV
–log (/)
Vs
where V
OUT
SLP
SETSLPINZ
()
=
is the volts-per-decade slope from Equation 1, having a
value of 440 mV/dec, and T is an effective time constant for
the integration, being equal to (4.15 kΩ × C
tor value comes from the setpoint interface scaling Equation 4
and the factor 1.35 arises as a result of the voltage gain of the
buffer. So the integration time constant can be written as
TC
=×
307.
in s whenCisressed in nF
µexp
()
FLT
FLT
To simplify understanding of the control loop dynamics, begin
by assuming that the power amplifier gain function actually is
linear-in-dB; for now, we will also use voltages to express the
signals at the power amplifier input and output. Let the RF output
voltage be V
and its input be VCW; further, to characterize the
PA
gain control function, this form is used
VGV
=10
PAO CW
(/)VV
where GO is the gain of the power amplifier when V
is the gain scaling. While few amplifiers will conform so
V
GSC
conveniently to this law, it nevertheless provides a clearer starting
OUT GSC
sT
10
)/1.35; the resis-
FLT
(7)
Assume that the gain magnitude of the power amplifier runs from
a minimum value of ×0.316 (–10 dB) at V
(40 dB) at V
0.316 and V
= 2.5 V. Applying Equation 9, we find GO =
OUT
= 1 V. Using a coupling factor of k = 0.0316
GSC
(that is, a 30 dB directional coupler) and recalling that the nominal
value of V
is 440 mV and VZ = 199 µV for the AD8316, we will
SLP
first calculate the range of values needed for V
output range of +32 dBm to –17 dBm. Note that, in the steady
state, the numerator of Equation 7 must be zero, that is
(8)
VVkVV
=
log
SETSLPPAZ
()
10
when VIN is expanded to kVPA, the fractional voltage sample of
the power amplifier output. Now, for +32 dBm, V
this evaluates to
= 0 and
OUT
(9)
Vmax mV/ V
()=()
SET
V
139
.
For a delivered power of –17 dBm, VPA = 31.6 mV rms,
Vmin. mV/ V
()
SET
V
0 310
.
=
.log
.log
10
10
Note: The power range is 49 dB, which corresponds to a voltage
change of 49 dB × 22 mV/dB = 1.08 V in V
(10)
)T.
(11)
sT
O
V
AD8316
10
GSC/VSLP
will be determined by
OUT
10
V
RF DRIVE: UP
TO 2.5GHz
V
OUT1
OUT
SET
CW
= 0 to ×100
to control an
RFDIRECTIONAL COUPLER
RF PA
C
FLT
(12)
= 8.9 V rms,
PA
=044281199
µ
=04410199
()
µ
SET
.
(13)
(14)
REV. C
–11–
Page 12
AD8316
The value of V
is of interest, although it is a dependent param-
OUT
eter inside the loop. It depends on the characteristics of the
power amplifier, and the value of the carrier amplitude V
Using the control values derived above, that is, G
V
= 1 V, and assuming that the applied power is fixed at
GSC
–7 dBm (so that V
VmaxVVVkG VV
OUTSET GSCSLPO CWZ
=
()
=×
VminVVVkG VV
OUTSET GSCSLPO CWZ
=−=
=
()
=×
=−=
= 100 mV rms), Equation 11 shows
CW
log
()
..log
139 1 044
()
...
32 07 25
()−()
031 1 044
..log
()
07 07 0
..
−
()
10
..
−
0 03160 316
10
./
01 199
V
log
10
0 0316 0 316
..
−
10
01 199
./
= 0.316 and
O
××
V
µ
××
V
µ
.
CW
(15)
(16)
Both results are consistent with the assumptions made about the
amplifier control function. Note that the second term is independent of the delivered power and is a fixed function of the
drive power.
Finally, the loop time constant for these parameters, using an
illustrative value of 2 nF for the filter capacitor C
TV VT
=
OGSCSLP
=
/
()
××
/...1044 3072 1395µµ
()
snFs
=
()
, evaluates to
FLT
(17)
Practical Loop
At the present time, power amplifiers, or VGAs preceding such
amplifiers, do not provide an exponential gain characteristic. It
follows that the loop dynamics (the effective time constant) will
vary with the setpoint, since the exponential function is unique
in providing constant dynamics. The procedure must therefore
be as follows. Beginning with the curve usually provided for the
power output versus APC voltage, draw a tangent at the point
on this curve where the slope is highest (see Figure 5). Using
this line, calculate the effective minimum value of the variable
, and use it in Equation 17 to determine the time constant.
V
GSC
(Note that the minimum in V
rate of change in the output power versus V
corresponds to the maximum
GSC
OUT
.)
For example, suppose it is found that, for a given drive power,
the amplifier generates an output power of P
P
at V
2
= V2. Then, it is readily shown that
OUT
VVVPP
=
20
–/–
GSC
()()
21 21
at V
1
= V1, and
OUT
(18)
This should be used to calculate the filter capacitance. The
response time at high and low power levels (on the “shoulders”
of the curve shown in Figure 5) will be slower. Note also that it
is sometimes useful to add a zero in the closed-loop response by
placing a resistor in series with C
FLT
.
A Note About Power Equivalency
Users of the AD8316 must understand that log amps fundamentally do not respond to power. For this reason, dBV
(decibels above 1 V rms) are included in addition to the commonly used metric dBm. The dBV scaling is fixed, independent
of termination impedance, while the corresponding power level
is not. For example, 224 mV rms is always –13 dBV, with one
further condition of an assumed sinusoidal waveform; see the
AD640 data sheet for more information about the effect of waveform on logarithmic intercept. This corresponds to a power
of 0 dBm when the net impedance at the input is 50 Ω. When
this impedance is altered to 200 Ω, however, the same voltage
corresponds to a power level that is four times smaller (P = V
2
/R),
or –6 dBm. A dBV level may be converted to dBm in the special
case of a 50 Ω system and a sinusoidal signal simply by adding
13 dB. 0 dBV is then, and only then, equivalent to 13 dBm.
P
RF
33dBm
23dBm
13dBm
3dBm
–7dBm
–67dBm–47dBm–27dBm–7dBm
V
1, P1
V
2, P2
V
OUT1
+13dBm
Figure 5. Typical Power Control Curve
Therefore, the external termination added ahead of the AD8316
determines the effective power scaling. This often takes the
form of a simple resistor (52.3 Ω will provide a net 50 Ω input),
but more elaborate matching networks may be used. The choice
of impedance determines the logarithmic intercept, that is, the
input power for which the V
versus PIN function would cross
SET
the baseline if that relationship were continuous for all values of
. This is never the case for a practical log amp; the intercept
V
IN
(so many dBV) refers to the value obtained by the minimum-error
straight-line fit to the actual graph of V
generally, V
). Where the modulation is complex, as in CDMA,
IN
versus PIN (more
SET
the calibration of the power response needs to be adjusted; the
intercept will remain stable for any given arbitrary waveform.
When a true power (waveform independent) response is needed,
a mean-responding detector, such as the AD8361, should be
considered.
The logarithmic slope, V
in Equation 1, which is the amount by
SLP
which the setpoint voltage needs to be changed for each decade
of input change (voltage or power) is, in principle, independent
of waveform or termination impedance. In practice, it usually
falls off somewhat at higher frequencies, because of the declining
gain of the amplifier stages and other effects in the detector
cells (see TPC 13).
Basic Connections
Figure 6 shows the basic connections for operating the AD8316
and Figure 7 shows a block diagram of a typical application.
The AD8316 is typically used in the RF power control loop of
dual mode and trimode mobile handsets where there is more
than one RF power control line.
REV. C–12–
Page 13
AD8316
FLT2
10
9
8
7
6
0.1F
C1
+V
S
2.7 TO 5.5V
V
OUT1
V
OUT2
C
FLT2
R1
52.3
RFIN
+V
S
V
SET
C
FLT1
V
BSEL
1
2
3
4
5
RFIN
ENBL
VSET
FLT1
BSEL
AD8316
VPOS
OUT1
COMM
OUT2
Figure 6. Basic Connections (Shown with MSOP Pinout)
RX1
RX2
DIRECTIONAL
COUPLER
ANT
ATTN
TX1
TX2
R1
52.3
PWR
AMP
OUT1
RFIN
FLT1
C
FLT1CFLT2
OUT2
VSET
BSEL
FLT2
GAIN
CONTROL
VOLTAGES
DAC
RFIN2
RFIN1
BAND
SELECT
Figure 7. Block Diagram of Typical Application
A supply voltage of 2.7 V to 5.5 V is required for the AD8316.
The supply to the VPOS pin should be decoupled with a low
inductance 0.1 µF surface-mount ceramic capacitor close to the
device. The AD8316 has an internal input coupling capacitor,
which negates the need for external ac coupling. This capacitor,
along with the device’s low frequency input impedance of approximately 3.0 kΩ, sets the minimum usable input frequency to around
20 MHz. A broadband 50 Ω input match is achieved in this
example by connecting a 52.3 Ω resistor between RFIN and
ground (COMM). A plot of input impedance versus frequency
is shown TPC 9. Other matching methods are also possible
(see the Input Coupling Options section).
In a power control loop, the AD8316 provides both the detector
and controller functions.
A number of options exist for coupling the RF signal from the
power amplifiers (PA) to the AD8316 input. Because only one
PA output is active at any time, a single RF input on the
AD8316 is sufficient in all cases.
Two directional couplers can be used directly at the PA outputs.
The outputs of these couplers would be passively combined
before being applied to the AD8316 RF input (in general,
some additional attenuation will be required between the coupler
and the AD8316). Another option involves using a dual-directional coupler between the PA and T/R switch. This device has
two inputs/outputs and a single-coupled output so that no external combiner is required.
A third option is to use a single broadband directional coupler
at the output of the transmit/receive (T/R) switch (the outputs
from the two PAs are combined in the T/R switch). This is
shown in Figure 7. This provides the advantage of enabling the
power at the output of the T/R switch to be precisely set, eliminating any errors due to insertion loss and insertion loss
variations of the T/R switch.
A setpoint voltage is applied to VSET from the controlling
source, generally a DAC. Any imbalance between the RF input
REV. C
–13–
level and the level corresponding to the setpoint voltage will be
corrected by the selected output, OUT1 or OUT2, which drives
the gain control terminal of the PAs. This restores a balance
between the actual power level sensed at the input of the AD8316
and the demanded value determined by the setpoint. This assumes
that the gain control sense of the variable gain element is positive; that is, an increasing voltage from OUT1 or OUT2 will
tend to increase gain. The outputs can swing from 100 mV
above ground to within 100 mV of the supply rail and can source
up to 12 mA. (A plot of maximum output voltage versus output
current is shown in TPC 19.) OUT1/OUT2 are capable of
sinking more than 200 µA.
Range on VSET and RF Input
The relationship between RF input level and the setpoint voltage follows from the nominal transfer function of the device (see
TPCs 2, 3, 5, and 6). At 0.9 GHz, for example, a voltage of 1 V
on VSET indicates a demand for –17 dBm (–30 dBV) at RFIN.
The corresponding power level at the output of the power amplifier will be greater than this amount due to the attenuation
through the directional coupler. For setpoint voltages of less
than approximately 200 mV and RF input amplitudes greater
than approximately –50 dBm, V
will remain unconditionally
OUT
at its minimum level of approximately 250 mV. This feature can
be used to prevent any spurious emissions during power-up and
power-down phases. Above 250 mV, VSET will have a linear
control range up to 1.4 V, corresponding to a dynamic range of
49 dB. This results in a slope of 22.2 mV/dB or approximately
45.5 dB/V.
Transient Response
The time domain response of power amplifier control loops,
using any kind of controller, is only partially determined by the
choice of filter which, in the case of the AD8316, has a true
integrator form 1/sT, as shown in Equation 7, with a time constant given by Equation 8. The large signal step response is also
strongly dependent on the form of the gain control law. Nevertheless, some simple rules can be applied. When the filter capacitor
is very large, it will dominate the time domain response,
C
FLT
but the incremental bandwidth of this loop will still vary as
traverses the nonlinear gain control function of the PA, as
V
OUT
shown in Figure 5. This bandwidth will be highest at the
point where the slope of the tangent drawn on this curve is
greatest—that is, for power outputs near the center of the PA’s
range—and will be much reduced at both the minimum and the
maximum power levels, where the slope of the gain control
curve is lowest, due to its S-shaped form. Using smaller values
, the loop bandwidth will generally increase, in inverse
of C
FLT
proportion to its value. Eventually, however, a secondary effect
will appear, due to the inherent phase lag in the power amplifier’s
control path, some of which may be due to parasitic or deliberately added capacitance at the OUT1 and OUT2 pins. This results
in the characteristic poles in the ac loop equation moving off the
real axis and thus becoming complex (and somewhat resonant).
This is a classic aspect of control loop design.
The lowest permissible value of C
needs to be determined
FLT
experimentally for a particular amplifier and circuit board layout. For GSM and DCS power amplifiers, C
will typically
FLT
range from 150 pF to 300 pF.
In many cases, some improvement in the worst-case response
time can be achieved by including a small resistance in series with
C
; this generates an additional zero in the closed-loop trans-
FLT
fer function, which will serve to cancel some of the higher-order
Page 14
AD8316
)
3.5V
4.7F
1F
T/R SWITCH
GSM/DCS
16.5dBm/19dBm
8-BIT
RAMP DAC
0V2.55V
(R2, R3 OPTIONAL)
(SEE TEXT)
TO
21.5dB ATTENUATOR
R5
54.9
R2
600
1k
BAND SELECT
0V/+V
GSM/(DCS/PCS
LDC15D190A0007A
R4
576
ENABLE
0V/+V
R3
C
220pF
S
7
5
FLT1
1
R7
49.9
48
3
2
6
R1
52.3
RFIN
1
ENBL
S
2
VSET
3
FLT1
4
BSEL
56
AD8316
VPOS
COMM
RF OUTPUT
+35.5dBm MAX
PCS/DCS
RF OUTPUT
+33dBm MAX
0.1F
OUT1
OUT2
FLT2
Figure 8. Dual-Mode (GSM/DCS) PA Control Example (Shown with AD8316 MSOP Pinout)
poles in the overall loop. A combination of main capacitor C
FLT
shunted by a second capacitor and resistor in series will also be
useful in minimizing the settling time of the loop.
Mobile Handset Power Control Example
Figure 8 shows a complete power amplifier control circuit for a
dual-mode handset. The RF3108 (RF Micro Devices), dualinput, trimode (GSM, DCS, PCS) PA is driven by a nominal
power level of 6 dBm at both inputs and has two gain control
lines. Some of the output power from the PA is coupled off
using a dual-band directional coupler (Murata part number
LDC15D190A0007A). This has a coupling factor of approximately 20 dB for the GSM band and 15 dB for DCS and an
insertion loss of 0.38 dB and 0.45 dB, respectively. Because the
RF3108 transmits a maximum power level of approximately
35 dBm for GSM and 32 dBm for DCS/PCS, additional attenuation of 20 dB is required before the coupled signal is applied to
the AD8316. This results in peak input levels of –5 dBm (GSM)
and –3 dBm (DCS). While the AD8316 gives a linear response
for input levels up to +3 dBm, for highly temperature-stable
performance at maximum PA output power, the maximum
input level should be limited to approximately –3 dBm (see
TPC 3 and TPC 5). This does, however, reduce the sensitivity
of the circuit at the low end.
The operational setpoint voltage, in the range 250 mV to 1.4 V,
is applied to the VSET pin of the AD8316. This will typically be
supplied by a DAC. The desired output is selected by applying
a high or low signal to the BSEL pin (HI = OUT1, LO = OUT2).
The selected output directly drives the level control pin of the
power amplifier. In this case a minimum supply voltage of 2.9 V
is required and V
2.6 V while delivering about 5 mA to the PA’s V
power amplifiers with lower V
reaches a maximum value of approximately
OUT
input ranges, a corresponding
APC
input. For
APC
low power supply to the AD8316 can be used. For example, on
GSM
GSM RF IN
+6dBm
C1
10
9
8
7
+V
S
2.7V TO 5.5V
C
FLT2
220pF
DCS/PCS V
R8
R6
RF3108
APC
APC
GSM V
(R5, R6 OPTIONAL)
(SEE TEXT)
GND
DCS/PCS RF IN
+6dBm
a 2.7 V supply, the voltage on OUT1/OUT2 can come to within
approximately 100 mV of the supply rail. This will depend, however, on the current draw (see TPC 19).
During initialization and completion of the transmit sequence,
V
should be held at its minimum level of 250 mV by keeping
OUT
V
below 200 mV. In this example, V
SET
is supplied by an 8-bit
SET
DAC that has an output range from 0 V to 2.55 V or 10 mV
per bit. This sets the control resolution of V
to 0.4 dB/bit
SET
(0.04 dB/mV 10 mV). If finer resolution is required, the
DAC’s output voltage can be scaled using two resistors as
shown. This converts the DAC’s maximum voltage of 2.55 V
down to 1.6 V and increases the control resolution to 0.25 dB/bit.
Two filter capacitors (C
FLT1/CFLT2
the loop for each band. The choice of C
) must be used to stabilize
will depend to a
FLT
large degree on the gain control dynamics of the power amplifier, something that is frequently poorly characterized, so some
trial and error may be necessary. In this example, a 220 pF
capacitor is used. The user may want to add a resistor in series
with the filter capacitor. The resistor adds a zero to the control
loop and increases the phase margin, which helps to make the
step response of the circuit more stable when the slope of the
PA’s power control function is the steepest. In this example,
the two filter capacitors are equal values; however, this is not
a requirement.
A smaller filter capacitor can be used by inserting a series resistor between V
and the control input of the PA. A series
OUT
resistor will work with the input impedance of the PA to create a
resistor divider and will reduce the loop gain. The size of the
resistor divider ratio depends upon the available output swing of
V
and the required control voltage on the PA. This tech-
OUT
nique can also be used to limit the control voltage in situations
where the PA cannot deliver the power level demanded by V
OUT
. Over-
drive of the control input of some PAs causes increased distortion.
REV. C–14–
Page 15
AD8316
Enable and Power-On
The AD8316 may be disabled by pulling the ENBL pin to
ground. This reduces the supply current from its nominal level
of 8.5 mA to 3 µA at 2.7 V. The logic threshold for turning on
the device is at 1.8 V at 2.7 V. A plot of the enable glitch is
shown in TPC 20. Alternatively, the device can be completely
disabled by pulling the supply voltage to ground; ENBL would be
connected to VPOS. The glitch in this mode of operation is
shown on TPC 25 and TPC 26. If VPOS is applied before the
device is enabled, a narrow glitch of less than 50 mV will result.
This is shown in TPC 31.
In both situations, the voltage on V
should be kept below
SET
250 mV during power-on and power-off, preventing any unwanted
transients on V
Input Coupling Options
OUT
.
The internal 5 pF coupling capacitor of the AD8316, along with
the low frequency input impedance of 3 kΩ, result in a high-pass
input corner frequency of approximately 20 MHz. This sets the
minimum operating frequency. Figure 9 shows three options for
input coupling. A broadband resistive match can be implemented
by connecting a shunt resistor to ground at RFIN. This 52.3 Ω
resistor (other values can also be used to select different overall
input impedances) combines with the input impedance of the
AD8316 (3 kΩ1 pF) to give a broadband input impedance of
50 Ω. While the input resistance and capacitance (C
and RIN)
IN
will vary by approximately ±20% from device to device, the
dominance of the external shunt resistor means that the variation in the overall input impedance will be close to the tolerance
of the external resistor. This method of matching is most useful
in wideband applications or in multimode systems where there
is more than one operating frequency and those frequencies are
quite far apart.
A reactive match can also be implemented as shown in Figure 9b.
This is not recommended at low frequencies because device
tolerances will vary the quality of the match dramatically because
of the large input resistance. For low frequencies, Option 9a or
Option 9c is recommended.
In Figure 9b, the matching components are drawn as generic
reactances. Depending on the frequency, the input impedance
at that frequency, and the availability of standard value components, either a capacitor or an inductor will be used. As in the
previous case, the input impedance at a particular frequency is
plotted on a Smith chart and matching components are chosen
(shunt or Series L, shunt or Series C) to move the impedance to
the center of the chart.
Figure 9c shows a third method for coupling the input signal
into the AD8316, applicable where the input signal is larger
than the input range of the log amp. A series resistor, connected
to the RF source, combines with the input impedance of the
AD8316 to resistively divide the input signal being applied to
the input. This has the advantage of very little power being
tapped off in RF power transmission applications.
Using the Chip Scale Package
On the underside of the chip scale package, there is an exposed
paddle. This paddle is internally connected to the chip’s ground.
For better electrical performance, this paddle should be soldered
down to the printed circuit board’s ground plane, even though
there is no thermal requirement to do so.
EVALUATION BOARD
Figures 10 and 11 show the schematics of the AD8316 MSOP and
LFCSP evaluation boards. Note that uninstalled components are
marked as open. The layout and silkscreen of the MSOP evaluation board are shown in Figures 12 and 13. Apart from the slightly
smaller device footprint and number of pins, the LFCSP evaluation board is identical to the MSOP board. The boards are
powered by a single supply in the 2.7 V to 5.5 V range. The power
supply is decoupled by a single 0.1 µF capacitor. Table II details
the various configuration options of the evaluation boards.
For operation in controller mode, both jumpers, LK1 and LK2,
should be removed. OUT1 and OUT2 can be selected with SW3
in Position A and Position B, respectively. The setpoint voltage
is applied to VSET, RFIN is connected to the RF source (PA
output or directional coupler), and OUT1 or OUT2 is connected
to the gain control pins of each PA. When the AD8316 is used
in controller mode, a capacitor and a resistor must be installed
in C4, C6, and R10, R11 for loop stability. For GSM/DCS
handset power amplifiers, this capacitor should typically range
from 150 pF to 300 pF. The series resistor improves the system
phase margin at low power levels, which in turn improves the
step response in the circuit. Typically, this resistor value should
be about 1.5 kΩ.
A quasi-measurement mode (in which the AD8316 delivers an
output voltage that is proportional to the log of the input signal)
can be implemented to establish the relationship between V
SET
and RFIN with the installation of two jumpers, LK1 and LK2.
This mimics an AGC loop. To establish the transfer function of
the log amp, the RF input should be swept while the voltage on
VSET is measured, that is, the SMA connector labeled VSET
acts as an output. This is the simplest method for validating
operation of the evaluation board. When operated in this
mode, a large capacitor (0.01 µF or greater) must be installed in
C4 or C6 (set R10/R11 to 0 Ω) to ensure loop stability.
REV. C
–15–
Page 16
AD8316
R
SHUNT
52.3
RFIN
ANTENNA
AD8316
C
C
C
IN
R
IN
X1
RFIN
X2
AD8316
C
C
C
IN
R
IN
STRIPLINE
PA
R
RFIN
ATTN
AD8316
C
C
C
IN
R
IN
a. Broadband Resistive
J1
INPUT
VSET
VPOS
A
SW1
B
J3
LK1
b. Narrow-Band Reactive
Figure 9. Input Coupling Options
OUT1
OUT2
FLT2
0.1F
10
9
8
7
6
R7
16.2k
C1
C6
(OPEN)
R11
(OPEN)
R6
17.8k
R12
0
VPOS
C3
0.1F
(OPEN)
(OPEN)
R10
(OPEN)
R2
VPOS
52.3
SW2
0.1F
AD8316
1
2
3
4
5
C5
RFIN
ENBL
VSET
FLT1
BSEL
VPOS
COMM
VPOS
R1
0
C4
OUT1(A)
OUT2(B)
AD8031
R5
10k
Figure 10. Schematic of Evaluation Board (MSOP)
R9
R8
10k
R3
0
(OPEN)
LK2
C2
C7
(OPEN)
c. Series Attenuation
J2
R4
(OPEN)
OUT1 (A)
OUT2 (B)
SW3
OUT1
J4
OUT2
INPUT
VSET
13141516
NCNC
R7
16.2k
12
11
10
9
C1
0.1F
R6
17.8k
R12
0
C3
0.1F
VPOS
(OPEN)
R8
10k
R3
0
C2
C7
(OPEN)
R4
(OPEN)
OUT1 (A)
OUT2 (B)
(OPEN)
R9
LK2
J2
OUT1
J4
OUT2
SW3
J1
VPOS
A
B
J3
SW1
LK1
(OPEN)
R10
(OPEN)
R1
0
C4
OUT1 (A)
OUT2 (B)
R2
52.3
NC = NO CONNECT
VPOS
SW2
0.1F
NC NCNC
1
RFIN
ENBL
2
VSET
3
FLT1
4
C5
VPOS
COMM
OUT1
AD8316
COMM
OUT2
FLT2
BSEL
5678
C6
(OPEN)
R11
(OPEN)
VPOS
AD8031
R5
10k
Figure 11. Schematic of Evaluation Board (LFCSP)
REV. C–16–
Page 17
AD8316
Table II. Evaluation Board Configuration Options
ComponentFunctionDefault Condition
TP1, TP2Supply and Ground Vector Pins.Not Applicable
SW1Device Enable. When in Position A, the ENBL pin is connected to VPOS andSW1 = A
the AD8316 is in operating mode. In Position B, the ENBL pin is grounded,
putting the device into power-down mode.
SW2Band Select. When in Position A (OUT1), the BSEL pin is connected to VPOSSW2 = OUT1
and the AD8316 OUT1 is in operation mode. In Position B (OUT2), the
BSEL pin is grounded and the AD8316 OUT2 is in operation while OUT1 pin
is shut down.
R1, R2Input Interface. The 52.3 Ω resistor in Position R2 combines with the AD8316’sR2 = 52.3 Ω (Size 0603)
internal input impedance to provide a broadband input impedance of aroundR1 = 0 Ω (Size 0402)
50 Ω. A reactive match can be implemented by replacing R2 with an inductor
and R1 (0 Ω) with a capacitor. In addition, the RF microstrip line has been
provided with a clean mask ground plane to provide additional matching. Note
that the AD8316’s RF input is internally ac-coupled.
R3, R4, R12,Output Interface. R4 and C2, R9 and C7 can be used to check the responseR4 = C2 = Open (Size 0603)
R9, C2, C7capacitive and resistive loading, respectively. R3/R4 and R12/R9 can be used toR9 = C7 = Open (Size 0603)
reduce the slope of OUT1 and OUT2.R3 = R12 = 0 Ω (Size 0603)
C4, C6, R10,Filter Capacitors/Resistors. The response time of OUT1, OUT2 can be modifiedC4 = C6 = Open (Size 0603)
R11by placing the capacitors between FLT1, FLT2 and resistors R10, R11R10 = R11 = Open (Size 0603)
to ground.
LK1, LK2Measurement Mode. A quasi-measurement mode can be implemented byLK1, LK2 = Installed
installing LK1 and LK2 (connecting an inverted OUT1 or OUT2 to VSET) to
yield the nominal relationship between RFIN and VSET. In this mode, a large
capacitor (0.01 µF or greater) must be installed in C4 and C6 and a 0 Ω
resistors to ground in R10 and R11. To select OUT1 or OUT2, SW3 must
be in the OUT1 position or the OUT2 position, respectively.
SW3Measurement Mode Output Select. When in measurement mode, output 1SW3 = OUT1
or output 2 can be selected by positioning SW3 to the OUT1 position or the
OUT2 position, respectively.
REV. C
–17–
Page 18
AD8316
Figure 12. Silkscreen of Component Side (MSOP)Figure 13. Layout of Component Side (MSOP)