FEATURES
Two Independent Linear-in-dB Channels
Input Noise at Maximum Gain: 1.8 nV/√Hz, 2.7 pA/√Hz
Bandwidth: 40 MHz (–3 dB)
Differential Input
Absolute Gain Range Programmable:
–14 dB to +34 dB (FBK Shorted to OUT), through
0 dB to +48 dB (FBK Open)
Variable Gain Scaling: 20 dB/V through 40 dB/V
Stable Gain with Temperature and Supply Variations
Single-Ended Unipolar Gain Control
Output Common-Mode Independently Set
Power Shutdown at Lower End of Gain Control
Single 5 V Supply
Low Power: 90 mW/Channel
Drives A/D Converters Directly
APPLICATIONS
Ultrasound and Sonar Time-Gain Control
High Performance AGC Systems
Signal Measurement
Variable Gain Amplifier
AD605
FUNCTIONAL BLOCK DIAGRAM
PRODUCT DESCRIPTION
The AD605 is a low noise, accurate, dual channel, linear-in-dB
variable gain amplifier, which is optimized for any application
requiring high performance, wide bandwidth variable gain control. Operating from a single 5 V supply, the AD605 provides
differential inputs and unipolar gain control for ease of use.
Added flexibility is achieved with a user determined gain range
and an external reference input which provides user determined
gain scaling (dB/V).
The high performance linear-in-dB response of the AD605 is
achieved with the differential input, single supply, exponential
amplifier (DSX-AMP) architecture. Each of the DSX-AMPs
comprise a variable attenuator of 0 dB to –48.4 dB followed by
a high speed fixed gain amplifier. The attenuator is based on a
7-stage R-1.5-R ladder network. The attenuation between tap
points is 6.908 dB and 48.360 dB for the entire ladder network.
The DSX-AMP architecture results in 1.8 nV/√Hz input noise
spectral density and will accept a ±2.0 V input signal when
VOCM is biased at VP/2.
Each independent channel of the AD605 provides a gain range
of 48 dB which can be optimized for the application. Gain
REV. B
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.
ranges between –14 dB to +34 dB and 0 dB to +48 dB can be
selected by a single resistor between pins FBK and OUT. The
lower and upper gain range are determined by shorting pin FBK
to OUT, or leaving pin FBK unconnected respectively. The two
channels of the AD605 can be cascaded to provide 96 dB of very
accurate gain range in a monolithic package.
The gain control interface provides an input resistance of approximately 2 MΩ and scale factors from 20 dB/V to 30 dB/V for a
VREF input voltage of 2.5 V to 1.67 V respectively. Note that
scale factors up to 40 dB are achievable with reduced accuracy
for scales above 30 dB. The gain scales linearly with control voltages (VGN) of 0.4 V to 2.4 V for the 20 dB/V scale and 0.20 V
to 1.20 V for the 40 dB/V scale. When VGN is <50 mV the
amplifier is powered-down to draw 1.9 mA. Under normal
operation, the quiescent supply current of each amplifier channel is only 18 mA.
The AD605 is available in a 16-lead plastic DIP and SOIC, and
is guaranteed for operation over the –40°C to +85°C temperature range.
Group Delay Variation1 MHz < f < 10 MHz, Full Gain Range± 2.0±2.0ns
VOCM Input Resistance4545kΩ
ACCURACY
Absolute Gain Error
–14 dB to –11 dB0.25 V < VGN < 0.40 V–1.2 +1.0+3.0–1.2 +0.75+3.0dB
–11 dB to +29 dB0.40 V < VGN < 2.40 V–1.0 ±0.3+1.0–1.0 ± 0.2+1.0dB
+29 dB to +34 dB2.40 V < VGN < 2.65 V–3.5 –1.25+1.2–3.5 –1.25+1.2dB
*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.
AD605AN–40°C to +85°CPlastic DIPN-1685°C/W
AD605AR–40°C to +85°CSmall Outline IC (SOIC)R-16A100°C/W
AD605BN–40°C to +85°CPlastic DIPN-1685°C/W
AD605BR–40°C to +85°CSmall Outline IC (SOIC)R-16A100°C/W
AD605ACHIPSDIE
AD605AR-REEL13" Reel
AD605AR-REEL77" Reel
AD605BR-REEL13" Reel
AD605BR-REEL77" Reel
AD605-EBEvaluation Board
PIN CONFIGURATION
VGN1
–IN1
+IN1
GND1
GND2
+IN2
–IN2
VGN2
1
2
3
AD605
4
TOP VIEW
5
(Not to Scale)
6
7
8
16
15
14
13
12
11
10
9
VREF
OUT1
FBK1
VPOS
VPOS
FBK2
OUT2
VOCM
JA
PIN FUNCTION DESCRIPTIONS
16-Lead Package for Dual Channel AD605
Pin No.MnemonicDescription
1VGN1CH1 Gain-Control Input and Power-Down Pin. If grounded, device is off, otherwise positive voltage
increases gain.
2–IN1CH1 Negative Input.
3+IN1CH1 Positive Input.
4GND1Ground.
5GND2Ground.
6+IN2CH2 Positive Input.
7–IN2CH2 Negative Input.
8VGN2CH2 Gain-Control Input and Power-Down Pin. If grounded, device is off, otherwise positive voltage
increases gain.
9VOCMInput to this pin defines common-mode voltage for OUT1 and OUT2.
10OUT2CH2 Output.
11FBK2Feedback Pin that Selects Gain Range of CH2.
12VPOSPositive Supply.
13VPOSPositive Supply.
14FBK1Feedback Pin that Selects Gain Range of CH1.
15OUT1CH1 Output.
16VREFInput to this pin sets gain-scaling for both channels: 2.5 V = 20 dB/V, 1.67 V = 30 dB/V.
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 AD605 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.
= 2.5 V (20 dB/V Scaling), f = 1 MHz, RL = 500 ⍀, CL = 5 pF, TA = 25ⴗC, VSS = 5 V)
REF
40
30
20
10
GAIN – dB
0
–10
–20
0.1 0.52.9
–40 C, +25 C, +85 C
0.9 1.3 1.7 2.1 2.5
VGN – Volts
TPC 1. Gain vs. VGN
40.0
37.5
35.0
32.5
30.0
27.5
25.0
GAIN SCALING – dBV
22.5
20.0
1.251.502.50
ACTUAL
THEORETICAL
1.752.002.25
V
REF
– Volts
TPC 4. Gain Scaling vs. V
REF
50
40
GAIN – dB
–10
–20
30
20
10
FBK (OPEN)
FBK (SHORT)
0
0.1 0.52.9
0.9 1.3 1.7 2.1 2.5
VGN – Volts
TPC 2. Gain vs. VGN for Different
Gain Ranges
3.0
2.5
2.0
1.5
1.0
0.5
0.0
–0.5
–1.0
GAIN ERROR – dB
–1.5
–2.0
–2.5
–3.0
0.20.72.7
1.21.72.2
VGN –Volts
–40 C
+85 C
+25 C
TPC 5. Gain Error vs. VGN at
Different Temperatures
40
30
30dB/V
= 1.67V)
(V
REF
20
10
GAIN – dB
0
–10
–20
0.1 0.52.9
0.9 1.3 1.7 2.1 2.5
ACTUAL
20dB/V
(V
REF
VGN – Volts
ACTUAL
= 2.50V)
TPC 3. Gain vs. VGN for Different
Gain Scalings
2.0
1.5
1.0
0.5
0.0
–0.5
GAIN ERROR – dB
–1.0
–1.5
–2.0
0.20.72.7
f = 1MHz
f = 5MHz
1.21.72.2
VGN – Volts
f = 10MHz
TPC 6. Gain Error vs. VGN at
Different Frequencies
2.0
1.5
1.0
0.5
0.0
–0.5
GAIN ERROR – dB
–1.0
–1.5
–2.0
0.20.72.71.21.72.2
30dB/V
V
= 1.67V
REF
VGN – Volts
V
REF
20dB/V
= 2.50V
TPC 7. Gain Error vs. VGN for
Different Gain Scalings
20
18
16
14
12
10
8
PERCENTAGE
6
4
2
0
–0.8 –0.6 –0.4 –0.2 0.0 0.2 0.4
DELTA GAIN – dB
N = 50
⌬G(dB) =
G(CH1) – G(CH2)
0.6 0.8
TPC 8. Gain Match, VGN1 = VGN2 =
1.0 V
–4–
20
18
N = 50
⌬G(dB) =
16
G(CH1) – G(CH2)
14
12
10
8
PERCENTAGE
6
4
2
0
–0.8 –0.6 –0.4 –0.2 0.0 0.2 0.4 0.6 0.8
DELTA GAIN – dB
TPC 9. Gain Match, VGN1 = VGN2 =
2.50 V
REV. B
Page 5
AD605
60
VGN = 2.9V (FBK = OPEN)
VGN = 2.9V (FBK = SHORT)
40
VGN = 1.5V (FBK = OPEN)
20
VGN = 1.5V (FBK = SHORT)
VGN = 0.1V (FBK = OPEN)
0
VGN = 0.1V (FBK = SHORT)
GAIN – dB
–20
–40
–60
100k1M100M
VGN = 0.0V
FREQUENCY – Hz
10M
TPC 10. AC Response
1000
100
10
NOISE – nV/ Hz
1
0.1 0.52.1
0.9 1.3 1.72.5 2.9
VGN – Volts
TPC 13. Input Referred Noise vs.
VGN
2.525
V
= 2.50V
OCM
2.520
2.515
2.510
2.505
2.500
– Volts
OS
2.495
V
2.490
2.485
2.480
2.475
0
–40 C
+25 C
+85 C
1.01.52.02.5
0.53.0
VGN – Volts
TPC 11. Output Offset vs. VGN
2.00
VGN = 2.9V
1.95
1.90
1.85
1.80
1.75
NOISE – nV/ Hz
1.70
1.65
1.60
–200 20406080
–4090
TEMPERATURE – C
TPC 14. Input Referred Noise vs.
Temperature
130
125
120
115
110
105
NOISE – nV/ Hz
100
95
90
0.53.01.01.52.02.5
0
+85 C
+25 C
–40 C
VGN – Volts
TPC 12. Output Referred Noise vs.
VGN
1.90
VGN = 2.9V
1.85
1.80
1.75
1.70
NOISE – nV/ Hz
1.65
1.60
100k1M10M
FREQUENCY – Hz
TPC 15. Input Referred Noise vs.
Frequency
100
VGN = 2.9V
10
1
NOISE – nV/ Hz
0.1
110 1k100
R
SOURCE
FREQUENCY – ⍀
ALONE
TPC 16. Input Referred Noise vs.
R
SOURCE
30
VGN = 2.9V
25
20
15
NOISE FIGURE – dB
10
5
1101k
R
TPC 17. Noise TPC vs. R
SOURCE
– ⍀
100
SOURCE
60
50
40
30
20
NOISE FIGURE – dB
10
0
0.1 0.52.9
0.9 1.3 1.7 2.1 2.5
RS = 50⍀
VGN – Volts
TPC 18. Noise TPC vs. VGN
REV. B
–5–
Page 6
AD605
–30
VO = 1V p-p
–35
VGN = 1.0V
–40
–45
–50
–55
–60
HARMONIC DISTORTION – dBc
–65
–70
100k1M100M
HD2
FREQUENCY – Hz
HD3
10M
TPC 19. Harmonic Distortion vs.
Frequency
15
10
INPUT
5
GENERATOR
LIMIT
0
– dBm
–5
IN
P
–10
–15
–20
0.1 0.52.9
FREQ = 10MHz
FREQ = 1MHz
0.9 1.3 1.7 2.1 2.5
VGN – Volts
TPC 22. 1 dB Compression vs. VGN
–35
–40
–45
HD2
–50
(1MHz)
–55
–60
–65
HARMONIC DISTORTION – dBc
–70
HD3
(1MHz)
–75
0.5 0.82.91.1 1.4 1.7 2.0 2.3 2.6
HD2
(10MHz)
VGN – Volts
HD3
(10MHz)
TPC 20. Harmonic Distortion vs. VGN
35
30
25
20
15
10
INTERCEPT – dBm
5
0
–5
0.613
f = 10MHz
1.41.82.22.6
VO = 1V p-p
f = 1MHz
VGN – Volts
TPC 23. Third Order Intercept vs. VGN
–20
–30
–40
–50
–60
–70
– dBm
OUT
–80
P
–90
–100
–110
–120
9.92
9.961010.02 10.04
FREQUENCY – MHz
f = 10MHz
= 1V p-p
V
O
VGN = 1.0V
TPC 21. Intermodulation Distortion
2V
VO = 2V p-p
VGN = 1.5V
400mV / DIV
–2V
253ns
100ns / DIV
1.253s
TPC 24. Large Signal Pulse Response
200
VO = 200mV p-p
VGN = 1.5V
40mV / DIV
TRIG'D
–200
253ns
100ns / DIV
1.253s
TPC 25. Small Signal Pulse Response
2.9V
VGN – Volts
0.0V
500mV
100
90
10
0%
500mV
200ns
TPC 26. Power-Up/Down Response
2.9V
VGN – Volts
0.1V
500mV
100
90
10
0%
500mV
TPC 27. Gain Response
100ns
–6–
REV. B
Page 7
AD605
–30
VGN1 = 1V
= 1V p-p
V
OUT1
–40
V
= GND
IN2
–50
–60
–70
CROSSTALK – dB
–80
–90
100k1M100M
VGN2 = 2.9V
VGN2 = 2.5V
FREQUENCY – Hz
VGN2 = 2.0V
VGN2 = 0.1V
10M
TPC 28. Crosstalk (CH1 to CH2) vs.
Frequency
25
+IS (AD605)
20
15
10
SUPPLY CURRENT – mA
5
0
–4090
+IS (VGN = 0)
–20020406080
TEMPERATURE – C
0
VIN = 0dBm
–10
–20
–30
CMRR – dB
–40
–50
–60
100k1M100M10M
VGN = 2.9V
VGN = 2.5V
VGN = 2.0V
VGN = 0.1V
FREQUENCY – Hz
TPC 29. Common-Mode Rejection vs.
Frequency
16
14
12
10
DELAY – ns
8
6
4
100k1M100M10M
180
175
170
165
160
155
150
INPUT IMPEDANCE – ⍀
145
140
100k1M100M
TPC 30. Input Impedance vs.
Frequency
VGN = 0.1V
VGN = 2.9V
FREQUENCY – Hz
VGN = 2.9V
FREQUENCY – Hz
10M
TPC 31. Supply Current (One
Channel) vs. Temperature
TPC 32. Group Delay vs. Frequency
REV. B
–7–
Page 8
AD605
VREF
VGN
+IN
–IN
VOCM
EXT
C1
EXT
C2
VPOS
R3
200k⍀
C3
R4
200k⍀
175⍀
175⍀
CONTROL
DIFFERENTIAL
ATTENUATOR
Figure 1. Simplified Block Diagram of a Single Channel of the AD605
THEORY OF OPERATION
The AD605 is a dual channel, low noise variable gain amplifier.
Figure 1 shows the simplified block diagram of one channel.
Each channel consists of a single-supply X-AMP (hereafter
called DSX, Differential Single-supply X-AMP) made up of:
(a) a precision passive attenuator (differential ladder)
(b) a gain control block
(c) a VOCM buffer with supply splitting resistors R3 and R4
(d) an Active Feedback Amplifier
1
(AFA) with gain setting
resistors R1 and R2.
The linear-in-dB gain response of the AD605 can generally be
described by Equation 1:
G (dB) = (Gain Scaling (dB/V)) × (Gain Control (V)) –
(19 dB – (14 dB) × (FB))(1)
where FB = 0 if FBK-to-OUT are shorted,
FB = 1 if FBK-to-OUT is open.
Each channel provides between –14 dB to +34.4 dB through
0 dB to +48.4 dB of gain depending on the value of the resistance
connected between pin FBK and OUT. The center 40 dB of
gain is exactly linear-in-dB while the gain error increases at the
top and bottom of the range. The gain is set by the gain control
voltage (VGN). The VREF input establishes the gain scaling—
the useful gain scaling range is between 20 dB/V and 40 dB/V for
a VREF voltage of 2.5 V and 1.25 V respectively. For example,
if FBK to OUT were shorted and VREF were set to 2.50 V (to
establish a gain scaling of 20 dB/V), the gain equation would
simplify to:
G (dB) = (20 (dB/V )) × (VGN (V )) – 19 dB(2)
The desired gain can then be achieved by setting the unipolar
gain control (VGN) to a voltage within its nominal operating
range of 0.25 V to 2.65 V (for 20 dB/V gain scaling). The gain is
monotonic for a complete gain control range of 0.1 V to 2.9 V.
Maximum gain can be achieved at a VGN of 2.9 V.
Since the two channels are identical, only Channel 1 will be
used to describe their operation. VREF and VOCM are the only
inputs that are shared by the two channels, and since they are
normally ac grounds, crosstalk between the two channels is
minimized. For highest gain scaling accuracy, VREF should
have an external low impedance voltage source. For low accuracy 20 dB/V applications, the VREF input can be decoupled with
a capacitor to ground. In this mode the gain scaling will be
GAIN
R2
20⍀
DISTRIBUTED G
G1
G2
R1
820⍀
M
Ao
OUT
3.36k⍀
FBK
determined by the midpoint between +VCC and GND, so care
should be taken to control the supply voltage to 5 V. The input
resistance looking into the VREF pin is 10 kΩ ± 20%.
The AD605 is a single-supply circuit and the VOCM pin is used
to establish the dc level of the midpoint of this portion of the
circuit. VOCM needs only an external decoupling capacitor to
ground to center the midpoint between the supply voltages (5 V,
GND); however if the dc level of the output is important to the
user (see Applications section for AD9050 example), then VOCM
can be specifically set. The input resistance looking into the
VOCM pin is 45 kΩ ± 20%.
Differential Ladder (Attenuator)
The attenuator before the fixed gain amplifier is realized by a
differential seven-stage R–1.5R resistive ladder network with an
untrimmed input resistance of 175 Ω single-ended or 350 Ω
differentially. The signal applied at the input of the ladder
network (Figure 2) is attenuated by 6.908 dB per tap; thus, the
attenuation at the first tap is 6.908 dB, at the second, 13.816 dB,
and so on all the way to the last tap where the attenuation is
48.356 dB. A unique circuit technique is used to interpolate
continuously between the tap points, thereby providing continuous attenuation from 0 dB to –48.36 dB. One can think of the
ladder network together with the interpolation mechanism as a
voltage-controlled potentiometer.
Since the DSX is a single-supply circuit, some means of biasing
its inputs must be provided. Node MID together with the
VOCM buffer performs this function. Without internal biasing,
external biasing would be required. If not done carefully, the
biasing network can introduce additional noise and offsets. By
providing internal biasing, the user is relieved of this task and
only needs to ac couple the signal into the DSX. It should be
made clear again that the input to the DSX is still fully differential if driven differentially, i.e., pins +IN and –IN see the same
signal but with opposite polarity. What changes is the load as
seen by the driver; it is 175 Ω when each input is driven singleended, but 350 Ω when driven differentially. This can be easily
explained when thinking of the ladder network as just two 175 Ω
resistors connected back-to-back with the middle node, MID,
being biased by the VOCM buffer. A differential signal applied
between nodes +IN and –IN will result in zero current into
node MID, but a single-ended signal applied to either input
+IN or –IN while the other input is ac grounded will cause the
current delivered by the source to flow into the VOCM buffer
via node MID.
1
To understand the active-feedback amplifier topology, refer to the AD830 data
sheet. The AD830 is a practical implementation of the idea.
–8–
REV. B
Page 9
AD605
1.5R
1.5R
R
R
–13.82dB
1.5R
1.5R
R
R
–20.72dB
1.5R
1.5R
R
R
–27.63dB
R
+IN
MID
R
–IN
NOTE: R = 96⍀
1.5R = 144⍀
–6.908dB
Figure 2. R–1.5R Dual Ladder Network
One feature of the X-AMP architecture is that the output referred
noise is constant versus gain over most of the gain range. This
can be easily explained by looking at Figure 2 and observing
that the tap resistance is equal for all taps after only a few taps
away from the inputs. The resistance seen looking into each tap is
54.4 Ω which makes 0.95 nV/√Hz of Johnson noise spectral
density. Since there are two attenuators, the overall noise
contribution of the ladder network is √2 times 0.95 nV/√Hz
or 1.34 nV/√Hz, a large fraction of the total DSX noise. The rest
of the DSX circuit components contribute another 1.20 nV/√Hz
which together with the attenuator produces 1.8 nV/√Hz of
total DSX input referred noise.
AC Coupling
As already mentioned, the DSX is a single single-supply circuit
and therefore its inputs need to be ac coupled to accommodate
ground-based signals. External capacitors C1 and C2 in
Figure 1 level shift the input signal from ground to the dc
value established by VOCM (nominal 2.5 V). C1 and C2,
together with the 175 Ω looking into each of DSX inputs (+IN
and –IN), will act as high-pass filters with corner frequencies
depending on the values chosen for C1 and C2. For example,
if C1 and C2 are 0.1 µF, then together with the 175 Ω input
resistance seen into each side of the differential ladder of the
DSX, a –3 dB high-pass corner at 9.1 kHz is formed.
If the DSX output needs to be ground referenced, then another
ac coupling capacitor will be required for level shifting. This
capacitor will also eliminate any dc offsets contributed by the
DSX. With a nominal load of 500 Ω and a 0.1 µF coupling
capacitor, this adds a high-pass filter with –3 dB corner frequency at about 3.2 kHz.
The choice for all three of these coupling capacitors depends on
the application. They should allow the signals of interest to pass
unattenuated, while at the same time they can be used to limit
the low frequency noise in the system.
Gain Control Interface
The gain-control interface provides an input resistance of
approximately 2 MΩ at pin VGN1 and gain scaling factors
from 20 dB/V to 40 dB/V for VREF input voltages of 2.5 V to
1.25 V respectively. The gain varies linearly-in-dB for the center
40 dB of gain range, that is for VGN equal to 0.4 V to 2.4 V
for the 20 dB/V scale, and 0.25 V to 1.25 V for the 40 dB/V
scale. Figure 3 shows the ideal gain curves when the FBK to
OUT connection is shorted which are described by the following equations:
From these equations one can see that all gain curves intercept
at the same –19 dB point; this intercept will be 14 dB higher
(–5 dB) if the FBK to OUT connection is left open. Outside of
the central linear range, the gain starts to deviate from the ideal
control law but still provides another 8.4 dB of range. For a
given gain scaling one can calculate V
as shown in Equa-
REF
tion (6).
2.500V ×20 dB /V
V
GAIN – dB
–10
–15
–20
=
REF
35
30
25
20
15
10
5
0
–5
Gain Scale
40dB/V 30dB/V20dB/V
1.00.51.52.02.53.0
GAIN CONTROL VOLTAGE
Figure 3. Ideal Gain Curves vs. V
LINEAR-IN-dB RANGE
OF AD605
REF
(6)
Usable gain control voltage ranges are 0.1 V to 2.9 V for 20 dB/V
scale and 0.1 V to 1.45 V for the 40 dB/V scale. VGN voltages
of less than 0.1 V are not used for gain-control since below
50 mV the channel is powered down. This can be used to conserve power and at the same time gate-off the signal. The supply
current for a powered-down channel is 1.9 mA, the response
time to power the device on-or-off, is less than 1 µs.
Active Feedback Amplifier (Fixed Gain Amp)
To achieve single supply operation and a fully differential input
to the DSX, an active-feedback amplifier (AFA) was utilized.
The AFA is basically an op amp with two g
stages; one of the
m
active stages is used in the feedback path (therefore the name),
while the other is used as a differential input. Note that the
differential input is an open-loop g
stage which requires that it
m
be highly linear over the expected input signal range. In this
design, the g
a distributed one, for example, there are as many g
stage that senses the voltages on the attenuator is
m
stages as
m
there are taps on the ladder network. Only a few of them are on
at any one time, depending on the gain-control voltage.
REV. B
–9–
Page 10
AD605
The AFA makes a differential input structure possible since one
of its inputs (G1) is fully differential; this input is made up of a
distributed gm stage. The second input (G2) is used for feedback. The output of G1 will be some function of the voltages
sensed on the attenuator taps which is applied to a high-gain
amplifier (A0). Because of negative feedback, the differential
input to the high-gain amplifier has to be zero, this in turn implies
that the differential input voltage to G2 times g
(the transcon-
m2
ductance of G2) has to be equal to the differential input voltage
to G1 times g
(the transconductance of G1). Therefore the
m1
overall gain function of the AFA is
gm1
gm2
R1× R2
×
R2
is the effective voltage
ATTEN
m1/gm2
(7)
=
V
OUT
ATTEN
=
where V
V
is the output voltage, V
OUT
sensed on the attenuator, (R1 + R2)/R2 = 42, and g
1.25; the overall gain is thus 52.5 (34.4 dB).
The AFA has additional features: (1) inverting the output signal
by switching the positive and negative input to the ladder network,
(2) the possibility of using the –IN input as a second signal input,
and (3) independent control of the DSX common-mode voltage.
Under normal operating conditions it is best to just connect a
decoupling capacitor to pin VOCM in which case the commonmode voltage of the DSX is half the supply voltage; this allows for
maximum signal swing. Nevertheless, the common- mode voltage
can be shifted up or down by directly applying a voltage to VOCM.
It can also be used as another signal input, the only limitation
being the rather low slew rate of the VOCM buffer.
If the dc level of the output signal is not critical, another coupling
capacitor is normally used at the output of the DSX; again this
is done for level shifting and to eliminate any dc offsets contributed by the DSX (see AC Coupling section).
The gain range of the DSX is programmable by a resistor connected between pins FBK and OUT. The possible ranges go from
–14 dB to +34.4 dB when the pins are shorted together, to 0 dB
to +48.4 dB when FBK is left open. Note that for the higher
gain range, the bandwidth of the amplifier is reduced by a factor
of five to about 8 MHz since the gain increased by 14 dB. This
is the case for any constant gain-bandwidth-product amplifier of
which the Active Feedback Amplifier is one.
APPLICATIONS
The most basic circuit in Figure 4 shows the connections for one
channel of the AD605 with a gain range of –14 dB to +34.4 dB.
The signal is applied at Pin 3. The ac coupling capacitors before
pins –IN1 and +IN1 should be selected according to the required
lower cutoff frequency. In this example the 0.1 µF capacitors
together with the 175 Ω seen into each of the DSX input pins
provides a –3 dB high pass corner of about 9.1 kHz. The upper
cutoff frequency is determined by the amplifier and is 40 MHz.
VGN
V
1
0.1F
IN
0.1F
VGN1
2
–IN1
AD605
3
+IN1
4
GND1
5
GND2
6
+IN2
7
–IN2
8
VGN2
VREF
OUT1
FBK1
VPOS
VPOS
FBK2
OUT2
VOCM
16
15
0.1F
14
13
12
11
10
9
0.1F
2.500V
OUT
5V
Figure 4. Basic Connections for a Single Channel
As shown here, the output is ac coupled for optimum performance. In the case of connecting to the 10-bit 40 MSPS A/D
converter AD9050, ac coupling can be eliminated as long as
pin VOCM is biased by the same 3.3 V common-mode voltage
as the AD9050.
Pin VREF requires a voltage of 1.25 V to 2.5 V, with between
40 dB/V and 20 dB/V gain scaling respectively. Voltage VGN
controls the gain; its nominal operating range is from 0.25 V to
2.65 V for 20 dB/V gain scaling, and 0.125 V to 1.325 V for
40 dB/V scaling. When this pin is taken to ground, the channel
will power down and disable its output.
Connecting Two Amplifiers to Double the Gain Range
Figure 5 shows the two channels of the AD605 connected in
series to provide a total gain range of 96.8 dB. When R1 and R2
are shorts, the gain range will be from –28 dB to +68.8 dB with
a slightly reduced bandwidth of about 30 MHz. The reduction
in bandwidth is due to two identical low-pass circuits being
connected in series; in the case of two identical single-pole lowpass filters, the bandwidth would be reduced by exactly √2. If
R1 and R2 are replaced by open circuits, i.e., Pins FBK1 and
FBK2 are left unconnected, then the gain range will shift up by
28 dB to 0 dB to +96.8 dB. As already pointed out earlier, the
bandwidth of each individual channel will be reduced by a
factor of 5 to about 8 MHz since the gain increased by 14 dB.
In addition, there is still the √2 reduction because of the series
connection of the two channels which results in a final bandwidth of the higher gain version of about 6 MHz.
VGN
C1
0.1F
V
IN
C2
0.1F
C3
0.1F
C4
0.1F
1
2
3
4
5
6
7
8
VGN1
–IN1
+IN1
GND1
GND2
+IN2
–IN2
VGN2
AD605
VREF
OUT1
FBK1
VPOS
VPOS
FBK2
OUT2
VOCM
16
15
R1
14
13
12
11
R2
10
9
C6
0.1F
5V
C5
0.1F
2.500V
OUT
Figure 5. Doubling the Gain Range with Two Amplifiers
–10–
REV. B
Page 11
AD605
Two other easy combinations are possible to provide a gain
range of –14 dB to +82.8 dB: (1) make R1 a short and R2 an
open, or (2) make R1 an open and R2 a short. The bandwidth
for both of these cases will be dominated by the channel that is
set to the higher gain and will be about 8 MHz. From a noise
standpoint, choice (2) is the better one since by increasing the
gain of the first amplifier, the second amplifier’s noise will have
less of an impact on the total output noise. One further observation regarding noise is that by increasing the gain the output
noise will increase proportionally; therefore, there is no increase
in signal-to-noise ratio. It will actually stay fixed.
It should be noted that by selecting the appropriate values of R1
and R2, any gain range between –28 dB to +68.8 dB and 0 dB
to +96.8 dB can be achieved with the circuit in Figure 5. When
using any value other than shorts and opens for R1 and R2, the
final value of the gain range will depend on external resistors
matching on-chip resistors. Since the internal resistors can vary
by as much as ±20%, the actual values for a particular gain have
to be determined empirically. Note that the two channels within
one part will match quite well; therefore, R1 will track R2 in
Figure 5.
C3 is not required since the common-mode voltage at Pin OUT1
should be identical to the one at Pins +IN2 and –IN2, but since
only 1 mV of offset at the output of the first DSX will introduce
an offset of 53 mV when the second DSX is set to the maximum
gain of the lowest gain range (34.4 dB), and 263 mV when set to
the maximum gain of the highest gain range (48.4 dB), it is
important to include ac-coupling to get the maximum dynamic
range at the output of the cascaded amplifiers. C5 is necessary if
the output signal needs to be referenced to any common-mode
level other than half of the supply as is provided by Pin OUT2.
Figure 6 shows the gain verses VGN for the circuit in Figure 5
at 1 MHz and the lowest gain range (–14 dB to +34.4 dB). Note
that the gain scaling is 40 dB/V, double the 20 dB/V of an individual DSX; this is the result of the parallel connection of the
gain control inputs, VGN1 and VGN2. One could of course
also sequentially increase the gain by first increasing the gain of
Channel 1 and then Channel 2. In that case VGN1 and VGN2
will have to be driven from separate voltage sources, for instance
two separate DACs. Figure 7 shows the gain error of Figure 6.
80
70
f = 1MHz
60
50
40
30
20
10
GAIN – dB
0
–10
–20
–30
–40
0.10.52.9
0.91.31.72.12.5
THEORETICAL
ACTUAL
VGN – Volts
Figure 6. Gain vs. VGN for the Circuit in Figure 5
4
3
2
1
0
–1
GAIN ERROR – dB
–2
–3
–4
0.20.72.7
1.21.72.2
VGN – Volts
f = 1MHz
Figure 7. Gain Error vs. VGN for the Circuit in Figure 5
REV. B
–11–
Page 12
AD605
0.210 (5.33)
MAX
0.160 (4.06)
0.115 (2.93)
0.1574 (4.00)
0.1497 (3.80)
OUTLINE DIMENSIONS
Dimensions shown in inches and (mm).
16-Lead Plastic DIP
(N-16)
0.840 (21.34)
0.745 (18.92)
16
18
PIN 1
0.022 (0.558)
0.014 (0.356)
0.100
(2.54)
BSC
9
0.280 (7.11)
0.240 (6.10)
0.060 (1.52)
0.015 (0.38)
0.070 (1.77)
0.045 (1.15)
0.130
(3.30)
MIN
SEATING
PLANE
16-Lead Small Outline IC
(R-16A)
0.3937 (10.00)
0.3859 (9.80)
169
0.2440 (6.20)
0.2284 (5.80)
81
0.325 (8.25)
0.300 (7.62)
0.015 (0.381)
0.008 (0.204)
0.195 (4.95)
0.115 (2.93)
C00541–0–5/01(B)
PIN 1
0.0098 (0.25)
0.0040 (0.10)
0.050 (1.27)
BSC
0.0192 (0.49)
0.0138 (0.35)
0.0688 (1.75)
0.0532 (1.35)
SEATING
PLANE
0.0099 (0.25)
0.0075 (0.19)
0.0196 (0.50)
0.0099 (0.25)
8ⴗ
0ⴗ
0.0500 (1.27)
0.0160 (0.41)
ⴛ 45ⴗ
PRINTED IN U.S.A.
–12–
REV. B
Loading...
+ hidden pages
You need points to download manuals.
1 point = 1 manual.
You can buy points or you can get point for every manual you upload.