1350 V/s Slew Rate, G = +2
Fast Settling Time: 4 ns
Low Supply Current: <30 mA
Excellent Video Specifications (R
Gain Flatness of 0.1 dB to 50 MHz
0.01% Differential Gain Error
0.01ⴗ Differential Phase Error
“All Hostile“ Crosstalk
–80 dB @ 10 MHz
–50 dB @ 100 MHz
High “OFF” Isolation of 90 dB @ 10 MHz
Low Cost
Fast Output Disable Feature
APPLICATIONS
RGB Buffer in LCD and Plasma Displays
RGB Driver
Video Routers
= 150 ⍀):
L
Video Buffers with Disable
AD8074/AD8075
FUNCTIONAL BLOCK DIAGRAM
AD8074 /AD8075
OE
DGND
IN2
AGND
IN1
AGND
IN0
V
1
2
G =
3
+1/+2
4
G =
5
+1/+2
6
G =
7
+1/+2
8
EE
V
16
CC
V
15
CC
14
OUT2
13
V
EE
12
OUT1
V
11
CC
OUT0
10
V
9
EE
PRODUCT DESCRIPTION
The AD8074/AD8075 are high-speed triple video buffers with
G = +1 and +2 respectively. They have a –3 dB full signal bandwidth in excess of 450 MHz, along with slew rates in excess of
1400 V/µs. With better than –80 dB of all hostile crosstalk and
90 dB isolation, they are useful in many high-speed applications. The differential gain and differential phase error are 0.01%
and 0.01°. Gain flatness of 0.1 dB up to 50 MHz makes the
AD8074/AD8075 ideal for RGB buffering or driving. They
consume less than 30 mA on a ±5 V supply.
Both devices offer a high-speed disable feature that allows the
outputs to be put into a high impedance state. This allows the
building of larger input arrays while minimizing “OFF” channel output loading. The AD8074/AD8075 are offered in a
16-lead TSSOP package.
REV. A
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.
Storage Temperature Range . . . . . . . . . . . . . . –65°C to +150°C
Lead Temperature Range (Soldering 10 sec) . . . . . . . . . . . 300°C
NOTES
1
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.
2
Specification is for device in free air (TA = 25°C).
3
16-lead plastic TSSOP; θJA = 150.4°C/W. Maximum internal power dissipation (P
) should be derated for ambient temperature (TA) such that
D
PD < (150°C – TA)/θJA.
ORDERING GUIDE
TemperaturePackagePackage
ModelRangeDescriptionOption
AD8074ARU–40°C to +85°C16-Lead Plastic TSSOP RU-16
AD8075ARU–40°C to +85°C16-Lead Plastic TSSOP RU-16
AD8074-EVALEvaluation Board
AD8075-EVALEvaluation Board
PIN CONFIGURATION
AD8074 /AD8075
OE
DGND
IN2
AGND
IN1
AGND
IN0
V
1
2
G =
3
+1/+2
4
G =
5
+1/+2
6
G =
7
+1/+2
8
EE
V
16
CC
V
15
CC
14
OUT2
13
V
EE
12
OUT1
V
11
CC
OUT0
10
V
9
EE
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 AD8074/AD8075 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.
MAXIMUM POWER DISSIPATION
The maximum power that can be safely dissipated by the AD8074/
AD8075 is limited by the associated rise in junction temperature.
The maximum safe junction temperature for plastic encapsulated
devices is determined by the glass transition temperature of the
plastic, approximately 150°C. Temporarily exceeding this limit
may cause a shift in parametric performance due to a change in
the stresses exerted on the die by the package. Exceeding a junction temperature of 175°C for an extended period can result in
device failure.
While the AD8074/AD8075 is internally short circuit protected,
this may not be sufficient to guarantee that the maximum junction
temperature (150°C) is not exceeded under all conditions. To
ensure proper operation, it is necessary to observe the maximum
power derating curves shown in Figure 1.
Figure 1. Maximum Power Dissipation vs. Temperature
TPC 6. AD8075 Frequency Response vs. Capacitive Load
CL = 10pF
CL = 0pF
CL = 5pF
V
C
L
150k⍀
OUT
REV. A
0
FREQUENCY – MHz
0
0.11000
110100
–10
–20
–30
–40
–50
–60
–70
–80
–90
–100
–110
V
OUT
= 2V p-p (ACTIVE CHANNEL(s))
R
L
= 150⍀
R
T
= 37.5⍀
ALL-HOSTILE
ADJACENT
CROSSTALK – dB
V
= 2V p-p (ACTIVE CHANNEL(s))
OUT
–10
= 1k⍀
R
L
= 37.5⍀
R
T
–20
–30
–40
–50
–60
ALL-HOSTILE
ADJACENT
110100
FREQUENCY – MHz
–70
CROSSTALK – dB
–80
–90
–100
–110
0.11000
TPC 7. AD8074 Crosstalk vs. Frequency (All Hostile and
Adjacent R
= 1 kΩ)
L
AD8074/AD8075
TPC 9. AD8075 Crosstalk vs. Frequency (All Hostile and
Adjacent R
= 150 Ω)
L
0
V
= 2V p-p
OUT
–10
= 150⍀
R
L
= 37.5⍀
R
T
–20
–30
–40
–50
–60
DISTORTION – dBc
–70
–80
–90
–100
1100010100
SECOND
HARMONIC
THIRD
HARMONIC
FUNDAMENTAL FREQUENCY – MHz
TPC 8. AD8074 Distortion vs. Frequency
0
V
= 2V p-p
OUT
–10
–20
–30
–40
–50
–60
DISTORTION – dBc
–70
–80
–90
–100
= 150⍀
R
L
= 37.5⍀
R
T
SECOND
HARMONIC
THIRD
HARMONIC
1100010100
FUNDAMENTAL FREQUENCY – MHz
TPC 10. AD8075 Distortion vs. Frequency
REV. A
–5–
AD8074/AD8075
–20
–30
–40
–50
–60
–70
–80
OFF ISOLATION – dB
–90
–100
–110
0.11000
RL = 1k⍀
RL = 150⍀
110100
FREQUENCY – MHz
TPC 11. AD8074 Off Isolation vs. Frequency
10
0
–10
–20
–30
–40
PSRR – dB
–50
–60
–70
–80
0.11000
110100
–PSRR
+PSRR
FREQUENCY – MHz
–20
–30
–40
–50
–60
–70
–80
OFF ISOLATION – dB
–90
–100
–110
0.11000110100
FREQUENCY – MHz
RL = 1k⍀
RL = 150⍀
TPC 14. AD8075 Off Isolation vs. Frequency
20
10
0
–10
–20
–30
PSRR – dB
–40
–50
–60
–70
0.11000
110100
+PSRR
–PSRR
FREQUENCY – MHz
TPC 12. AD8074 PSRR vs. Frequency
350
300
250
200
150
100
VOLTAGE NOISE – nV/ Hz
50
0
101001M1k10k100k
FREQUENCY – Hz
TPC 13. AD8074 Voltage Noise vs. Frequency
TPC 15. AD8075 PSRR vs. Frequency
400
350
300
250
200
150
VOLTAGE NOISE – nV/ Hz
100
50
0
101001M1k10k100k
FREQUENCY – Hz
TPC 16. AD8075 Voltage Noise vs. Frequency
–6–
REV. A
AD8074/AD8075
10000
1000
100
10
1
INPUT IMPEDANCE – k⍀
0.1
0.01
0.11000
110100
FREQUENCY – MHz
TPC 17. AD8074 Input Impedance vs. Frequency
1000
100
10
10000
1000
100
10
1
INPUT IMPEDANCE – k⍀
0.1
0.01
0.11000110100
FREQUENCY – MHz
TPC 20. AD8075 Input Impedance vs. Frequency
1000
100
10
OUTPUT IMPEDANCE – ⍀
1
0.1
0.11000
110100
FREQUENCY – MHz
TPC 18. AD8074 Output Impedance vs. Frequency;
Enabled
1000
100
10
1
0.1
OUTPUT IMPEDANCE – k⍀
0.01
0.001
0.11000
110100
FREQUENCY – MHz
OUTPUT IMPEDANCE – ⍀
1
0.1
0.11000110100
FREQUENCY – MHz
TPC 21. AD8075 Output Impedance vs. Frequency;
Enabled
1000
100
10
1
0.1
OUTPUT IMPEDANCE – k⍀
0.01
0.001
0.11000
110100
FREQUENCY – MHz
TPC 19. AD8074 Output Impedance vs. Frequency;
Disabled
REV. A
TPC 22. AD8075 Output Impedance vs. Frequency;
Disabled
–7–
AD8074/AD8075
0.15
VO = 200mV STEP
0.10
0.05
0
–0.05
–0.10
–0.15
2ns
TPC 23. AD8074 Small Signal Pulse Response (RL = 1 kΩ,
= 5 pF)
C
L
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
–0.1
–0.2
VO = 700mV STEP
2ns
TPC 24. AD8074 Video Amplitude Pulse Response
(R
= 1 kΩ, CL = 5 pF)
L
0.15
VO = 200mV STEP
0.10
0.05
0
–0.05
–0.10
–0.15
2ns
TPC 26. AD8075 Small Signal Pulse Response (RL = 150 kΩ)
0.8
0.7
V
0.6
0.5
0.4
0.3
0.2
0.1
0
–0.1
= 700mV STEP
O
2ns
TPC 27. AD8075 Video Amplitude Pulse Response
(R
= 150 Ω)
L
1.5
VO= 2V STEP
1.0
0.5
0
–0.5
–1.0
–1.5
2ns
TPC 25. AD8074 Large Signal Pulse Response
(R
= 1 kΩ, CL = 5 pF)
L
1.5
VO = 2V STEP
1.0
0.5
0
–0.5
–1.0
2ns
–1.5
TPC 28. AD8075 Large Signal Pulse Response (RL = 150 Ω)
–8–
REV. A
AD8074/AD8075
THEORY OF OPERATION
The AD8074 (G = +1) and AD8075 (G = +2) are triple-channel,
high-speed buffers with TTL-compatible output enable control.
Optimized for buffering RGB (red, green, blue) video sources,
the devices have high peak slew rates, maintaining their bandwidth for large signals. Additionally, the buffers are compensated
for high phase margin, minimizing overshoot for good pixel
resolution. The buffers also have video specifications that are
suitable for buffering NTSC or PAL composite signals.
The buffers are organized as three independent channels, each
with an input transconductance stage and an output transimpedance stage. Each channel is characterized by low input
capacitance and high input impedance. The transconductance
stages, NPN differential pairs, source signal current into the folded
cascode output stages. Each output stage contains a compensating network and emitter follower output buffer. Internal voltage
feedback sets the gain, the AD8074 being configured as a unity
gain follower, and the AD8075 as a gain-of-two amplifier with a
feedback network. The architecture provides drive for a reverseterminated video load (150 Ω) with low differential gain and
phase error for relatively low power consumption. Careful chip
design and layout allow excellent crosstalk isolation between
channels.
One logic pin, OE, controls whether the three outputs are
enabled, or disabled to a high-impedance state. The high impedance disable allows larger matrices to be built when busing the
outputs together. When disabled, the AD8074 and AD8075 consume a fifth the power as when enabled. In the case of the
AD8075 (G = +2), a feedback isolation scheme is used so that
the impedance of the gain-of-two feedback network does not
load the output.
Full power bandwidth for an undistorted sinusoid is often calculated using peak slew rate from the equation:
Full Power Bandwidth
=
Peak Slew Rate
Sinusoidal Amplitude
××2 π
Peak slew rate is not the same as average slew rate (25% to
75%) which is typically specified. For a natural response, peak
slew rate may be 2.7 times larger than average slew rate. Therefore, calculating a full power bandwidth with a specified average
slew rate will give a pessimistic result.
The primary cause of overshoot in these amplifiers is the presence of large reactive loads at the output and insufficient series
isolation of the load. However, it is possible to overdrive these
amplifiers with 1 V, subnanosecond input-pulse edges. The
ensuing dynamics may give rise to subnanosecond overshoot. To
reduce these effects, an edge-rate limiting network at the input
should be considered for input transition times less than 0.5 ns.
APPLICATIONS
Response Tuning
It has been mentioned in passing that the primary cause of overshoot for the AD8074 and AD8075 is the presence of large
reactive loads at the output. If the system exhibits excessive
ringing while settling, a 10 Ω–50 Ω series resistor may be used
at the output to isolate the emitter-follower output buffer from
the reactive load. If the output exhibits an overdamped response,
the system designer may add a few pF shunt capacitance at the
output to tune for a faster edge transition. A system with a small
degree of overshoot will settle faster than an overdamped system.
2.0
1.5
1.0
0.5
–0.5
–1.0
–1.5
–2.0
0
RS = 0⍀
= 5pF
C
L
V
IN
75⍀
RS = 10⍀
= 10pF
C
L
RS = 20⍀
= 15pF
C
L
R
S
C
L
1k⍀
V
OUT
2ns
Figure 2. Driving Capacitive Loads
Single Supply Operation
The AD8074 and AD8075 may be operated from a single 10 V
supply. In this configuration, the AD8075’s AGND pins must
be tied near midsupply, as AGND provides the reference for the
ground buffer, to which the internal gain network is terminated.
Logic is referenced to DGND. The buffers are disabled in single
supply operation for V
V
OE
< V
+ 0.8 V. TTL logic levels are expected. The fol-
DGND
OE
> V
+ ~2.0 V and enabled for
DGND
lowing restrictions are placed upon the digital ground potential:
3512.–VVVV
≤≤
AVCCDGND
V
≥ V
DGND
AVEE
The architecture of the output buffer is such that the output
voltage can swing to within ~2.3 V of either rail. For example, if
the output need swing only 2 V, then the buffers could be operated on dual 3.5 V or single 7 V supplies. It is cautioned that
saturation effects may become noticeable when the output swings
within 2.6 V of either rail. The system designer may opt to
use this characteristic to his or her advantage by using the
soft-saturation regime, (2.2 V–2.6 V from the supply rails), to
tame excessive overshoot. The designer is cautioned that a
charge storage associated time delay of several nanoseconds is
incurred when recovering from soft-saturation. This effect
results in longer settling tails.
REV. A
–9–
AD8074/AD8075
RGB Buffer for Second Monitor
The RGB signals for PC monitors are driven through coax
cables whose characteristic impedance is 75 Ω. The graphics
chip will generally have current-source output drivers that should
be double terminated with a 75 Ω shunt termination at each end.
On the transmit end, the shunt terminations are provided to
ground close to the graphics IC, while the monitor terminates
its end via internal termination resistors. While this scheme works
well and is virtually foolproof for a single monitor, it leaves no
means for passively connecting a second monitor to the same source.
A second monitor that is connected simply in parallel will provide an extra set of terminations that will upset the signal levels.
To keep costs low, most computer monitors do not have the ability
to open-circuit the terminations in order that an additional monitor
PC GRAPHICS IC
R
75⍀
CURRENT SOURCE
OUTPUT DRIVERS
G
75⍀
B
75⍀
can be connected to the same signal, as is done in some studiotype TV monitors.
A way around this problem is to connect the first monitor to the
RGB channels in the standard fashion, and then to provide a
triple gain-of-two buffer to drive the second monitor. The AD8075
is designed to provide this function and also provide excellent
high-frequency performance for high-resolution graphics signals.
Figure 3 shows a schematic of this circuit.
The outputs of the AD8075 are low impedance voltage sources
and are therefore series-terminated with 75 Ω resistors. The
internal resistors in Monitor #2 provide the terminations at its
end. The overall effect of this type of termination scheme is to
divide the signal amplitude by two. This is compensated by the
gain of two provided by the AD8075.
MONITOR #1
75⍀
INTERNAL
75⍀
TERMINATIONS
75⍀
25F
0.1F
+5V
0.1F
MONITOR #2
75⍀
75⍀
+5V
+
+5V
0.1F
AD8075
75⍀
INTERNAL
75⍀
TERMINATIONS
75⍀
75⍀
25F
+
0.1F
–5V
0.1F
–5V
0.1F
–5V
Figure 3. Buffer
–10–
REV. A
AD8074/AD8075
Triple Video Multiplexer
The AD8074 and AD8075 each have an output-enable function
that can be used to disable the outputs and put them in a highimpedance state. Usually, for a unity-gain device, it is relatively
easy to provide high disabled impedance, because the feedback
path is from the output to a high-impedance input. However, for a
non-unity-gain part, the feedback provides a resistive path to
ground. This will usually dominate the disabled output impedance, and make it a much lower value than the unity-gain device.
The AD8075 has an internal buffer that provides a low-impedance,
ground level output that terminates the feedback path during
enabled operation. In the disabled state, both this buffer output
0.1F
+5V
0.1F
+5V
+
25F
R
75⍀
AD8075
SOURCE 1
G
B
75⍀
75⍀
and the amplifier output are disabled to a high impedance to
provide a high-impedance disabled state.
To construct a multiplexer, the outputs from one or more devices
are connected in parallel and only one device is enabled at a
time while all of the others are disabled. The two sets of inputs
are applied individually to each of the separate device inputs.
Figure 4 shows the circuit details for this function. The first RGB
Source 1 is input to the first AD8075. Each of the individual
signals is terminated to ground with 75 Ω to provide proper
termination for the input cables. In a similar fashion, the Source
2 signals are input to the second AD8075.
+5V
0.1F
75⍀
75⍀
75⍀
OE
R
G
B
OUTPUT
SEL1/SEL2
SOURCE 2
25F
+
+5V
+
25F
R
75⍀
0.1F
0.1F
–5V
+5V
0.1F
–5V
+5V
0.1F
0.1F
–5V
0.1F
OE
75⍀
AD8075
G
B
75⍀
75⍀
25F
+
0.1F
0.1F
75⍀
75⍀
0.1F
REV. A
–5V
–5V
Figure 4. Mux
–11–
–5V
AD8074/AD8075
Each of the six outputs has a 75 Ω series resistor that is used to
reverse-terminate the output transmission line. The corresponding outputs are then wired in parallel and delivered to the output
cable. The termination resistors in this position help to isolate
the off capacitance of the disabled device’s outputs from loading
the enabled device’s outputs. The gain-of-two of the AD8075
compensates for the signal halving that occurs as a result of the
output terminations.
A select signal is provided directly to the OE of the second
AD8075 and an inverted version is used to drive the other device’sOE. This will ensure that only one device is active at a time. Since
there is a total of 150 Ω in series between any two outputs, it is
not essential to be overly concerned about the exact timing of
the making and breaking of the enable signals.
Additional inputs can easily be added to the circuit shown to
make wider multiplexers. The outputs of all of the devices will
be wired in parallel, and the logic must allow that only one output
be enabled at a time.
If it is desired to make a triple 3:1 multiplexer, a triple 2:1 multiplexer, like the AD8185 can be used along with the AD8075.
The same general guidelines for input and output treatment
should be followed and the logic must perform the proper function.
If it is desired to design such a multiplexer at unity gain, the
AD8074 should be used. For a triple 3:1 multiplexer, an
AD8183 (triple 2:1 mux) can be combined with an AD8074 to
provide this function.
Layout and Grounding
The AD8074 and AD8075 are extreme bandwidth, high-slew-rate
devices that are designed to drive up to the highest resolution
monitors and provide excellent resolution. To realize their full
performance potential, it is essential to adhere to the best practices of high-speed PCB layout.
A major area of focus should be the power distribution system.
There should be a full ground plane that provides the reference
and return paths for both the inputs and outputs. The ground
also provides isolation between the input signals to minimize the
crosstalk. This ground plane should cover as wide an area as
possible and be minimally interrupted in order to keep its
impedance to a minimum.
The power planes should also be as broad as possible to provide
minimal inductance, which is required for high-slew-rate signals. These power planes layers should be spaced closely to the
ground plane to increase the interplane capacitance between the
supplies and ground.
Each supply pin should be bypassed with a low inductance
0.1 µF ceramic capacitance with minimal excess circuit length
to minimize the series impedance. A 25 µF tantalum electro-
lytic capacitor will supply a charge reservoir for lower frequency,
high-amplitude transitions.
The input and output signals should be run as directly as possible in order to minimize the effects of parasitics. If they must
run over a longer distance of more than a few centimeters, controlled impedance PCB traces should be used to minimize the
effect of reflections due to mismatches in impedance and the
proper termination should be provided.
To avoid excess crosstalk, the above recommendations should
be followed carefully. The power system and signal routing are
the most important aspects of preventing excess crosstalk.
Beyond these techniques, shielding can be provided by ground
traces between adjacent signals, especially those that travel
parallel over long distances.
–12–
REV. A
AD8074/AD8075
IN2
IN1
IN0
DISOUT
IN2
AGND
IN1
AGND
IN0
AGND
75⍀
R2
75⍀
R3
75⍀
TP5
DISOUT
R1
V
CC
R16
20k⍀
W2
AGND
50⍀ IMPEDANCE LINE
DO NOT INSTALL
75⍀ IMPEDANCE LINE
75⍀ IMPEDANCE LINE
75⍀ IMPEDANCE LINE
TP2
V
1P1
EE
AGND
V
CC
R11
50⍀
AGND
1
OE
AGND
DGND
2
3
IN2
AGND
4
AGND
5
IN1
AGND
6
AGND
IN0
7
V
EE
AGND
C6
0.1F
AGND
C8
0.01F
V
8
EE
2P1
TP1
3P1
DUT
AD8074
C2
10F
+
AGND
AGND
+
C1
10F
AGND
DO NOT INSTALL
AGND
V
16
CC
V
15
CC
OUT2
14
V
13
EE
OUT1
12
V
11
CC
OUT0
10
V
EE
9
DO NOT INSTALL
V
EE
50⍀ IMPEDANCE LINE
AGND
DO NOT INSTALL
V
CC
50⍀ IMPEDANCE LINE
C3
0.1F
C7
0.1F
C11
0.01F
AGND
V
EE
V
CC
AGNDAGND
V
EE
C15
0.01F
AGND
AGND
C13
0.01F
C12
0.01F
AGND AGND
V
CC
C14
0.01F
V
EE
DO NOT INSTALL
V
CC
DO NOT INSTALL
TP4TP3
75⍀ IMPEDANCE LINE
R7
75⍀
R6
150⍀
75⍀ IMPEDANCE LINE
R8
75⍀
R10
150⍀
R9
75⍀ IMPEDANCE LINE
75⍀
R12
DO NOT INSTALL
150⍀
AGND
AGND
AGND
OUT2
OUT1
OUT0
OUT2
AGND
OUT1
AGND
OUT0
AGND
Figure 5. Evaluation Board Schematic
REV. A
–13–
AD8074/AD8075
Figure 6. Component Side
Figure 7. Circuit Side
Figure 8. Silkscreen Top
Figure 9. Silkscreen Bottom
Figure 10. Internal 2
–14–
REV. A
OUTLINE DIMENSIONS
Dimensions shown in inches and (mm).
Controlling Dimension: Metric, shown in parentheses.