Triple Differential Receiver with
Adjustable Line Equalization
AD8123
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. Specifications subject to change without notice. No
license is granted by implication or otherwise under any patent or patent rights of Analog Devices.
Trademarks and registered trademarks are the property of their respective owners.
One Technology Way, P.O. Box 9106, Norwood, MA 02062-9106, U.S.A.
Tel: 781.329.4700 www.analog.com
Fax: 781.461.3113 ©2007 Analog Devices, Inc. All rights reserved.
FEATURES
Compensates cables to 300 meters for wideband video
Fast rise and fall times
4.9 ns with 2 V step @ 150 meters of UTP cable
8.0 ns with 2 V step @ 300 meters of UTP cable
55 dB peak gain at 100 MHz
Two frequency response gain adjustment pins
High frequency peaking adjustment (VPEAK)
Broadband flat gain adjustment (VGAIN)
Pole location adjustment pin (VPOLE)
Compensates for variations between cables
Can be optimized for either UTP or coaxial cable
DC output offset adjust (VOFFSET)
Low output offset voltage: 24 mV
Compensates both RGB and YPbPr
Two on-chip comparators with hysteresis
Can be used for common-mode sync extraction
Available in 40-lead, 6 mm × 6 mm LFCSP
APPLICATIONS
Keyboard-video-mouse (KVM)
Digital signage
RGB video over UTP cables
Professional video projection and distribution
HD video
Security video
FUNCTIONAL BLOCK DIAGRAM
OUT
R
+IN
R
–IN
R
AD8123
OUT
G
+IN
G
–IN
G
OUT
B
+IN
B
–
IN
B
–IN
CMP1
+IN
CMP1
–IN
CMP2
+IN
CMP2
OUT
CMP2
OUT
CMP1
V
PEAK
V
POLE
V
OFFSET
V
GAIN
0
6814-001
Figure 1.
GENERAL DESCRIPTION
The AD8123 is a triple, high speed, differential receiver and
equalizer that compensates for the transmission losses of UTP
and coaxial cables up to 300 meters in length. Various gain
stages are summed together to best approximate the inverse
frequency response of the cable. Logic circuitry inside the AD8123
controls the gain functions of the individual stages so that the
lowest noise can be achieved at short-to-medium cable lengths.
This technique optimizes its performance for low noise, short-
to-medium range applications, while at the same time provides
the high gain bandwidth required for long cable equalization
(up to 300 meters). Each channel features a high impedance
differential input that is ideal for interfacing directly with the cable.
The AD8123 has three control pins for optimal cable
compensation, as well as an output offset adjust pin. Two
voltage-controlled pins are used to compensate for different
cable lengths; the VPEAK pin controls the amount of high frequency
peaking and the VGAIN pin adjusts the broadband flat gain,
which compensates for the low frequency flat cable loss.
For added flexibility, an optional pole adjustment pin, VPOLE,
allows movement of the pole locations, allowing for the
compensation of different gauges and types of cable as well
as variations between different cables and/or equalizers. The
VOFFSET pin allows the dc voltage at the output to be adjusted,
adding flexibility for dc-coupled systems.
The AD8123 is available in a 6 mm × 6 mm, 40-lead LFCSP
and is rated to operate over the extended temperature range of
−40°C to +85°C.
06814-019
UXGA RESOLUTION IMAGE
AFTER 300 METER CAT-5 CABLE
AFTER AD8123.
UXGA RESOLUTION IMAGE
AFTER 300 METER CAT-5 CABLE
BEFORE AD8123.
Figure 2.
AD8123
Rev. A | Page 2 of 16
TABLE OF CONTENTS
Features .............................................................................................. 1
Applications....................................................................................... 1
Functional Block Diagram .............................................................. 1
General Description ......................................................................... 1
Revision History ............................................................................... 2
Specifications..................................................................................... 3
Absolute Maximum Ratings............................................................ 5
Thermal Resistance ...................................................................... 5
ESD Caution.................................................................................. 5
Pin Configuration and Function Description .............................. 6
Typical Performance Characteristics ............................................. 7
Theory of Operation ...................................................................... 10
Input Common-Mode Voltage Range Considerations ......... 10
Applications Information .............................................................. 11
Basic Operation .......................................................................... 11
Comparators ............................................................................... 11
Sync Pulse Extraction Using Comparators............................. 12
Using the VPEAK, VPOLE, VGAIN, and VOFFSET Inputs ................... 12
Using the AD8123 with Coaxial Cable.................................... 13
Driving 75 Ω Video Cable With the AD8123 ........................ 13
Driving a Capacitive Load......................................................... 13
Filtering the RGB Outputs ........................................................ 13
Power Supply Filtering............................................................... 14
Layout and Power Supply Decoupling Considerations......... 14
Input Common-Mode Range ................................................... 15
Small Signal Frequency Response............................................ 15
Power-Down ............................................................................... 15
Outline Dimensions ....................................................................... 16
Ordering Guide .......................................................................... 16
REVISION HISTORY
11/07—Rev. 0 to Rev. A
Changes to Features.......................................................................... 1
Changes to Ordering Guide .......................................................... 16
8/07—Revision 0: Initial Version
AD8123
Rev. A | Page 3 of 16
SPECIFICATIONS
TA = 25°C, VS = ±5 V, RL = 150 Ω, Belden Cable (BL-7987R), VOFFSET = 0 V, VPEAK, VGAIN, and VPOLE are set to recommended settings shown in
Figure 17, unless otherwise noted.
Table 1.
Parameter Conditions Min Typ Max Unit
PEAKING PERFORMANCE (NO CABLE)
Peak Frequency VPEAK = 2 V, VGAIN = 0.6 V, VPOLE = 1 V 100 MHz
V
PEAK = 2 V, VGAIN = 0.6 V, VPOLE = 2 V 105 MHz
Peak Gain VPEAK = 2 V, VGAIN = 0.6 V, VPOLE = 1 V 45 dB
V
PEAK = 2 V, VGAIN = 0.6 V, VPOLE = 2 V 55 dB
DYNAMIC PERFORMANCE
10% to 90% Rise/Fall Time VOUT = 2 V step, 150 meters Cat-5 4.9 ns
V
OUT = 2 V step, 300 meters Cat-5 8.0 ns
Settling Time to 2% VOUT = 2 V step, 150 meters Cat-5 36 ns
V
OUT = 2 V step, 300 meters Cat-5 106 ns
–3 dB Large Signal Bandwidth VOUT = 1 V p-p, <10 meters Cat-5 120 MHz
V
OUT = 2 V p-p, <10 meters Cat-5 110 MHz
V
OUT = 2 V p-p, 150 meters Cat-5 78 MHz
V
OUT = 2 V p-p, 300 meters Cat-5 43 MHz
Integrated Output Voltage Noise 150 meter setting, integrated to 160 MHz 2.5 mV rms
300 meter setting, integrated to 160 MHz 24 mV rms
INPUT DC PERFORMANCE
Input Voltage Range −IN and +IN ±3.0 V
Maximum Differential Voltage Swing 4 V p-p
Voltage Gain ΔVO/ΔVI, VGAIN set for 0 meters of cable 1 V/V
Common-Mode Rejection Ratio (CMRR) At dc, VPEAK = VGAIN = VPOLE = 0 V −86 dB
At dc, VPEAK = VGAIN = VPOLE = 2 V −67 dB
At 1 MHz, VPEAK = VGAIN = VPOLE = 2 V −52 dB
Input Resistance Common mode 4.4 MΩ
Differential 3.7 MΩ
Input Capacitance Common mode 1.0 pF
Differential 0.5 pF
Input Bias Current 2.4 μA
VOFFSET Pin Current 28.9 μA
VGAIN Pin Current 0.5 μA
VPEAK Pin Current 0.4 μA
VPOLE Pin Current 0.4 μA
ADJUSTMENT PINS
VPEAK Input Voltage Range Relative to GND 0 to 2 V
VPOLE Input Voltage Range Relative to GND 0 to 2 V
VGAIN Input Voltage Range Relative to GND 0 to 2 V
VOFFSET to OUT Gain OUT/VOFFSET, range limited by output swing 1 V/V
Maximum Flat Gain VGAIN = 2 V 2 dB
OUTPUT CHARACTERISTICS
Output Voltage Swing 150 Ω load −3.75 to +3.69 V
1 kΩ load −3.66 to +3.69 V
Output Offset Voltage Referred to output, VPEAK = VGAIN = VPOLE = 0 V 24 mV
Referred to output, VPEAK = VGAIN = VPOLE = 2 V 32 mV
Output Offset Voltage Drift Referred to output 33 μV/°C
AD8123
Rev. A | Page 4 of 16
Parameter Conditions Min Typ Max Unit
POWER SUPPLY
Operating Voltage Range ±4.5 ±5.5 V
Positive Quiescent Supply Current 132 mA
Negative Quiescent Supply Current 126 mA
Supply Current Drift, ICC/IEE 80 μA/°C
Positive Power Supply Rejection Ratio DC, referred to output −51 dB
Negative Power Supply Rejection Ratio DC, referred to output −63 dB
Power Down, VIH (Minimum) Minimum Logic 1 voltage 1.1 V
Power Down, VIL (Maximum) Maximum Logic 0 voltage 0.8 V
Positive Supply Current, Powered Down VPEAK = VGAIN = VPOLE = 0 V 1.1 μA
Negative Supply Current, Powered Down VPEAK = VGAIN = VPOLE = 0 V 0.7 μA
COMPARATORS
Output Voltage Levels VOH/VOL 3.33/0.043 V
Hysteresis VHYST 70 mV
Propagation Delay tPD, LH/tPD, HL 17.5/10.0 ns
Rise/Fall Times tRISE/tFALL 9.3/9.3 ns
Output Resistance 0.03 Ω
OPERATING TEMPERATURE RANGE −40 +85 °C
AD8123
Rev. A | Page 5 of 16
ABSOLUTE MAXIMUM RATINGS
Table 2.
Parameter Rating
Supply Voltage 11 V
Power Dissipation See Figure 3
Input Voltage (Any Input) VS− − 0.3 V to VS+ + 0.3 V
Storage Temperature Range −65°C to +125°C
Operating Temperature Range −40°C to +85°C
Lead Temperature (Soldering, 10 sec) 300°C
Junction Temperature 150°C
Stresses above those listed under Absolute Maximum Ratings
may cause permanent damage to the device. This is a stress
rating only; functional operation of the device at these or any
other conditions above those indicated in the operational section of
this specification is not implied. Exposure to absolute maximum
rating conditions for extended periods may affect device reliability.
THERMAL RESISTANCE
θJA is specified for the worst-case conditions, that is, θJA is
specified for the device soldered in a circuit board in still air.
Table 3. Thermal Resistance with the Underside Pad
Connected to the Plane
Package Type/PCB Type θJA Unit
40-Lead LFCSP/4-Layer 29 °C/W
Maximum Power Dissipation
The maximum safe power dissipation in the AD8123 package
is limited by the associated rise in junction temperature (TJ) on
the die. At approximately 150°C, which is the glass transition
temperature, the plastic changes its properties. Even temporarily
exceeding this temperature limit can change the stresses that the
package exerts on the die, permanently shifting the parametric
performance of the AD8123. Exceeding a junction temperature
of 175°C for an extended time can result in changes in the
silicon devices, potentially causing failure.
The power dissipated in the package (PD) is the sum of the
quiescent power dissipation and the power dissipated in the
package due to the load drive for all outputs. The quiescent
power is the voltage between the supply pins (VS) times the
quiescent current (IS). The power dissipation due to each load
current is calculated by multiplying the load current by the
voltage difference between the associated power supply and the
output voltage. The total power dissipation due to load currents
is then obtained by taking the sum of the individual power
dissipations. RMS output voltages must be used when dealing
with ac signals.
Airflow reduces θJA. In addition, more metal directly in contact
with the package leads from metal traces, through holes, ground,
and power planes reduces the θJA. The exposed paddle on the
underside of the package must be soldered to a pad on the PCB
surface that is thermally connected to a solid plane (usually the
ground plane) to achieve the specified θJA.
Figure 3 shows the maximum safe power dissipation in the
package vs. the ambient temperature for the 40-lead LFCSP
(29°C/W) on a JEDEC standard 4-layer board with the underside
paddle soldered to a pad that is thermally connected to a PCB
plane. θJA values are approximations.
06814-025
0
–40 –20 0 20 40 60 80
AMBIENT TEMPERATURE (°C)
MAXIMUM POWER DISSIPATION (W)
1
2
3
4
5
6
7
Figure 3. Maximum Power Dissipation vs. Temperature for a 4-Layer Board
ESD CAUTION
AD8123
Rev. A | Page 6 of 16
PIN CONFIGURATION AND FUNCTION DESCRIPTION
1
NC
GND
+IN
CMP1
OUT
CMP1
V
S+
_CMP
V
S–
V
S+
V
S–
V
OFFSET
GND
V
POLE
V
PEAK
V
GAIN
V
S+
–IN
G
+IN
B
–IN
B
–IN
CMP1
OUT
CMP2
–IN
CMP2
+IN
CMP2
NC
OUT
B
OUT
G
OUT
R
V
S–
V
S+
V
S–
V
S+
V
S–
V
S+
NC
V
S–
_CMP
NC
PD
NC
+IN
R
+IN
G
NC
1
2
3
4
5
6
7
8
9
10
23
24
25
26
27
28
29
30
22
21
11
12
13
15
17
16
18
19
20
14
33
34
35
36
37
38
39
40
32
31
TOP VIEW
(Not to Scale)
AD8123
–IN
R
06814-002
2
NC = NO CONNECT
NOTES
1. EXPOSED PADDLE ON THE BOTTOM OF THE PACKAGE
MUST BE CONNECTED TO A PCB PLANE TO ACHIEVE
SPECIFIED THERMAL RESISTANCE.
Figure 4. Pin Configuration
Table 4. Pin Function Descriptions
Pin No. Mnemonic Description
1, 10, 20, 21, 30, 40 NC No Internal Connection.
2 +INCMP1 Positive Input, Comparator 1.
3 −INCMP1 Negative Input, Comparator 1.
4 OUTCMP1 Output, Comparator 1.
5 VS+_CMP Positive Power Supply, Comparator. Must be connected to VS+.
6 OUTCMP2 Output, Comparator 2.
7 −INCMP2 Negative Input, Comparator 2.
8 +INCMP2 Positive Input, Comparator 2.
9 VS−_CMP Negative Power Supply, Comparator. Must be connected to VS−.
11, 14, 17, 22, 33 VS− Negative Power Supply, Equalizer Sections.
12 OUTBOutput, Blue Channel.
13, 16, 19, 29, 36 VS+ Positive Power Supply, Equalizer Sections.
15 OUTGOutput, Green Channel.
18 OUTROutput, Red Channel.
23 VOFFSET Output Offset Control Voltage.
24, 39 GND Signal Ground Reference.
25 VGAIN Broadband Flat Gain Control Voltage.
26 VPEAK Equalizer High Frequency Boost Control Voltage.
27 VPOLE Equalizer Pole Location Adjustment Control Voltage.
28 PD Power Down.
31 +INRPositive Input, Red Channel.
32 −INRNegative Input, Red Channel.
34 +INGPositive Input, Green Channel.
35 −INGNegative Input, Green Channel.
37 +INBPositive Input, Blue Channel.
38 −INBNegative Input, Blue Channel.
Exposed Underside Pad Thermal Plane Connection. Connect to any PCB plane with voltage between VS+ and VS−.
AD8123
Rev. A | Page 7 of 16
TYPICAL PERFORMANCE CHARACTERISTICS
TA = 25°C, VS = ±5 V, RL = 150 Ω, Belden Cable (BL-7987R), VOFFSET = 0 V, VPEAK, VGAIN, and VPOLE are set to recommended settings shown in
Figure 17, unless otherwise noted.
–6
–5
–4
–3
–2
–1
0
1
2
3
4
100k 1M 10M 100M
V
GAIN
= 0V
V
GAIN
= 1V
V
GAIN
= 2V
V
GAIN
= 0V
V
POLE
= 0V
V
O
= 1V p-p
06814-003
FREQUENCY (Hz)
GAIN (dB)
Figure 5. Frequency Response for Various VGAIN Without Cable
–60
–40
–20
0
20
40
60
V
PEAK
= 0V
V
PEAK
= 1V
V
PEAK
= 2V
100k 1M 10M 100M
06814-004
FREQUENCY (Hz)
GAIN (dB)
V
GAIN
= 0.6V
V
POLE
= 2V
V
O
= 1V p-p
Figure 6. Frequency Response for Various VPEAK Without Cable
–60
–50
–40
–30
–20
–10
0
10
20
30
40
VPOLE = 0V
VPOLE = 1V
VPOLE = 2V
100k 1M 10M 100M
06814-005
FREQUENCY (Hz)
GAIN (dB)
VGAIN = 0.6V
VPEAK = 1V
VO = 1V p-p
Figure 7. Frequency Response for Various VPOLE Without Cable
–12
–9
–10
–11
–6
–3
–4
–5
–7
–8
–2
–1
0
1
2
3
50m
100m
150m
200m
300m
V
O
= 2V p-p
100k 1M 10M 100M
06814-006
FREQUENCY (Hz)
GAIN (dB)
Figure 8. Equalized Frequency Response for Various Cable Lengths
0
20
40
60
80
100
120
0 50 100 150 200 250 300
06814-007
CABLE LENGTH (meters)
BANDWIDTH (MHz)
V
O
= 2V p-p
Figure 9. Equalized −3 dB Bandwidth vs. Cable Length
–6
–4
–2
0
2
4
6
0 50 100 150 200 250 300 350 400 450 500
06814-008
TIME (ns)
VOLTAGE (V)
V
GAIN
= 0.6V
V
PEAK
= 0V
V
POLE
= 0V
INPUT
OUTPUT
Figure 10. Overdrive Recovery
AD8123
Rev. A | Page 8 of 16
–1.5
–1.0
–0.5
0
0.5
1.0
1.5
0 50 100 150 200 250 300 350 400 450 500
06814-009
TIME (ns)
OUTPUT VOLTAGE (V)
150m
300m
50m
Figure 11. Pulse Response for Various Cable Lengths (2 MHz)
10
100
1000
10000
0m
150m
300m
100k 1M 10M 100M
06814-010
FREQUENCY (Hz)
OUTPUT VOLTAGE NOISE (nV/ Hz)
Figure 12. Output Voltage Noise vs. Frequency for Various Cable Length
–90
–80
–70
–60
–50
–40
–30
–20
–10
0
10
20
V
GAIN
= 0V, V
PEAK
= 0V, V
POLE
= 0V
V
GAIN
= 1.85V, V
PEAK
= 1.65V, V
POLE
= 1.75V
100k 1M 10M 100M
06814-011
FREQUENCY (Hz)
CMRR (dB)
Figure 13. CMRR vs. Frequency
–1.5
–1.0
–0.5
0
0.5
1.0
1.5
012345678910
50m
150m
300m
06814-012
TIME (µs)
OUTPUT VOLTAGE (V)
Figure 14. Pulse Response for Various Cable Lengths (100 kHz)
5
10
15
20
25
0
30
25 50 75 100 125 150 175 200 225 250 275 300
06814-013
CABLE LENGTH (meters)
INTEGRATED OUTPUT VOLTAGE NOISE FROM
100kHz TO 160MHz (mV rms)
Figure 15. Integrated Output Voltage Noise vs. Cable Length
100k 1M 10M 100M
06814-014
FREQUENCY (Hz)
V
GAIN
= 0V, V
PEAK
= 0V, V
POLE
= 0V
V
GAIN
= 1.85V, V
PEAK
= 1.65V, V
POLE
= 1.75V
–80
–70
–60
–50
–40
–30
–20
–10
0
10
20
CROSSTALK (dB)
Figure 16. Crosstalk vs. Frequency
AD8123
Rev. A | Page 9 of 16
0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
25 50 75 100 125 150 175 200 225 250 275 300
06814-016
CABLE LENGTH (meters)
CONTROL VOLTAGE (V)
V
PEAK
V
POLE
V
GAIN
0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
25 50 75 100 125 150 175 200 225 250 275 300
06814-015
CABLE LENGTH (meters)
CONTROL VOLTAGE (V)
V
PEAK
V
POLE
V
GAIN
Figure 17. Recommended Settings for UTP Cable
Figure 18. Recommended Settings for Coaxial Cable
AD8123
Rev. A | Page 10 of 16
THEORY OF OPERATION
The AD8123 is a unity-gain, triple, wideband, low noise analog
line equalizer that compensates for losses in UTP and coaxial
cables up to 300 meters in length. The 3-channel architecture is
targeted at high resolution RGB applications but can be used in
HD YPbPr applications as well.
Three continuously adjustable control voltages, common
to the RGB channels, are available to the designer to provide
compensation for various cable lengths as well as for variations
in the cable itself. The VPEAK input is used to control the amount
of high frequency peaking. VPEAK is the primary control that is
used to compensate for frequency and cable-length dependent,
high frequency losses that are present due to the skin effect of
the cable. A second control pin, VGAIN, is used to adjust broadband
gain to compensate for low frequency flat losses present in the
cable. A third control, VPOLE, is used to move the positions of the
equalizer poles and can be linearly derived from VPEAK, as illustrated
in the Typical Performance Characteristics and Applications
Information sections, for UTP and coaxial cables. Finally, an
output offset adjust control, VOFFSET, allows the designer to shift
the output dc level.
The AD8123 has a high impedance differential input that makes
termination simple and allows dc-coupled signals to be received
directly from the cable. The AD8123 input can also be used in a
single-ended fashion in coaxial cable applications. For differential
systems that require very high CMRR, a triple differential
receiver, such as the AD8143 or AD8145, can be placed in
front of the AD8123.
The AD8123 has a low impedance output that is capable of
driving a 150 Ω load. For systems where the AD8123 has to
drive a high impedance capacitive load, it is recommended that
a small series resistor be placed between the output and load to
buffer the capacitance. The resistor should not be so large as to
reduce the overall bandwidth to an unacceptable level.
The AD8123 is designed such that systems that use short-to-
medium-length cables do not pay a noise penalty for excess gain
that they do not require. The high gain is only available for
longer length systems where it is required. This feature is built
into the VPEAK control and is transparent to the user.
Two comparators are provided on-chip that can be used for
sync pulse extraction in systems that use sync-on-common
mode encoding. Each comparator has very low output impedance
and can therefore be used in a source-only cable termination
scheme by placing a series resistor equal to the cable characteristic
impedance directly on the comparator output. Additional
details are provided in the Applications Information section.
INPUT COMMON-MODE VOLTAGE RANGE
CONSIDERATIONS
When using the AD8123 as a receiver, it is important to ensure
that its input common-mode voltage stays within the specified
range. The received common-mode level is calculated by adding
the common-mode level of the driver, the single-ended peak
amplitude of the received signal, the amplitude of any sync
pulses, and the other induced common-mode signals, such as
ground shifts between the driver and the AD8123 and pickup
from external sources, such as power lines and fluorescent
lights. See the Applications Information section for more details.
AD8123
Rev. A | Page 11 of 16
APPLICATIONS INFORMATION
BASIC OPERATION
The AD8123 is easy to apply because it contains everything
on-chip needed for cable loss compensation. Figure 20 shows a
basic application circuit (power supplies not shown) with
common-mode sync pulse extraction that is compatible with
the common-mode sync pulse encoding technique used in the
AD8134, AD8147, and AD8148 triple differential drivers. If
sync extraction is not required, the terminations can be single
100 Ω resistors, and the comparator inputs can be left floating.
In Figure 20, the AD8123 is feeding a high impedance input,
such as a delay line or crosspoint switch, and the additional gain
of two that makes up for double termination loss is not required.
COMPARATORS
In addition to general-purpose applications, the two on-chip
comparators can be used to extract video sync pulses from the
received common-mode voltages or to receive differential digital
information. Built-in hysteresis helps to eliminate false triggers
from noise. The Sync Pulse Extraction Using Comparators
section describes the sync extraction details.
The comparator outputs have nearly 0 Ω output impedance and
are designed to drive source-terminated transmission lines. The
source termination technique uses a resistor in series with each
comparator output such that the sum of the comparator source
resistance (≈0 Ω) and the series resistor equals the transmission
line characteristic impedance. The load end of the transmission
line is high impedance. When the signal is launched into the source
termination, its initial value is one-half of its source value because
its amplitude is divided by two in the voltage divider formed by
the source termination and the transmission line. At the load,
the signal experiences nearly 100% positive reflection due to the
high impedance load and is restored to nearly its full value. This
technique is commonly used in PCB layouts that involve high
speed digital logic.
Figure 19 shows how to apply the comparators with source
termination when driving a 50 Ω transmission line that is high
impedance at its receive end.
06814-021
49.9
Ω
HIGH-Z
Z
0
= 50Ω
Figure 19. Using Comparator with Source Termination
18
RED
BLUE
GREEN
CMV
RED CMV
BLUE CMV
AD8123
15
12
6
4
V
PEAK
26
25
27
23
28
V
POLE
V
OFFSET
V
GAIN
06814-020
HSYNC OUT
VSYNC OUT
32
31
GREEN
35
34
38
37
3
2
7
8
RECEIVED
RED VIDEO
RECEIVED
GREEN VIDEO
RECEIVED
BLUE VIDEO
RED VIDEO OUT
GREEN VIDEO OUT
BLUE VIDEO OUT
ANALOG
CONTROL
INPUTS
POWER DOWN
CONTROL
GND REFERENCE
24, 39
PD
1
2
1kΩ
1kΩ
49.9Ω
49.9Ω
49.9Ω
475Ω
49.9Ω
49.9Ω
49.9Ω
47pF47pF
Figure 20. Basic Application Circuit with Common-Mode Sync Extraction
AD8123
Rev. A | Page 12 of 16
SYNC PULSE EXTRACTION USING COMPARATORS
The AD8123 is useful in many systems that transport computer
video signals, which are typically comprised of red, green, and
blue (RGB) video signals and separate horizontal and vertical
sync signals. Because the sync signals are separate and not
embedded in the color signals, it is advantageous to transmit
them using a simple scheme that encodes them among the three
common-mode voltages of the RGB signals. The AD8134,
AD8147, and AD8148 triple differential drivers are natural
complements to the AD8123 because they perform the sync
pulse encoding with the necessary circuitry on-chip.
The sync encoding equations follow:
[
HV
K
VRed CM −= 2
]
(1)
[]
(2)
V2
2−= K
VGreen CM
[
HV
K
VBlue CM += 2
]
(3)
where:
Red VCM, Green VCM, and Blue VCM are the transmitted common-
mode voltages of the respective color signals.
K is an adjustable gain constant that is set by the driver.
V and H are the vertical and horizontal sync pulses, defined
with a weight of −1 when the pulses are in their low states, and a
weight of +1 when they are in their high states.
The AD8134 and AD8146/AD8147/AD8148 data sheets contain
further details regarding the encoding scheme. Figure 20 illustrates
how the AD8123 comparators can be used to extract the horizontal
and vertical sync pulses that are encoded on the RGB common-
mode voltages by the aforementioned drivers.
USING THE VPEAK, VPOLE, VGAIN, AND VOFFSET INPUTS
The VPEAK input is the main peaking control and is used to
compensate for the low-pass roll-off in the cable response. The
VPOLE input is a secondary frequency response shaping control
that shifts the positions of the equalizer poles. The VGAIN input
controls the wideband flat gain and is used to compensate for
the low frequency cable loss that is nominally flat. The VOFFSET
input is used to produce an offset at the AD8123 output. The
output offset is equal to the voltage applied to the VOFFSET input,
limited by the output swing limits.
The VPEAK and VPOLE controls can be used independently or they
can be coupled to form a single peaking control. While Figure 17
and Figure 18 show recommended settings vs. cable length,
designers may find other combinations that they prefer. These
two controls give designers extra freedom, as well as the ability
to compensate for different cable types (such as UTP and coaxial
cable), as opposed to having only a single frequency shaping
control.
In some cases, as would likely be with automatic control, the
VPEAK control is derived from a low impedance source, such as
an op amp. Figure 21 shows how to derive VPOLE from VPEAK in a
UTP application according to the recommended curves shown
in Figure 17, when VPEAK originates from a low impedance
source. Clearly, the 5 V supply must be clean to provide a clean
VPOLE voltage.
06814-026
VPEAK
2+ 0.9V
20Ω
5.11kΩ
VPEAK
VPOLE ≈
5V
14kΩ
8.25kΩ
VPEAK
Figure 21. Deriving VPOLE from VPEAK with Low-Z Source for UTP Cable
The 20 Ω series resistor in the VPEAK path provides capacitive
load buffering for the op amp. This value can be modified,
depending on the actual capacitive load.
In automatic equalization circuits that place the control voltages
inside feedback loops, attention must be paid to the poles
produced by the summing resistors and load capacitances.
The peaking can also be adjusted by a mechanical or digitally
controlled potentiometer. In these cases, if the resistance of the
potentiometer is a couple of orders of magnitude lower than the
values of the resistors used to develop VPOLE, its resistance can be
ignored. Figure 22 shows how to use a 500 Ω potentiometer with
the resistor values shown in Figure 21 scaled up by a factor of 10.
06814-027
V
PEAK
2+ 0.9V
51.1kΩ
V
PEAK
V
POLE
≈
5V
5V
140kΩ
82.5kΩ
750Ω
500Ω
Figure 22. Deriving VPOLE from VPEAK with Potentiometer for UTP Cable
Many potentiometers have wide tolerances. If a wide tolerance
potentiometer is used, it may be necessary to change the value
of the 750 Ω resistor to obtain a full swing for VPEAK.
The VGAIN input is essentially a contrast control and can be set
by adjusting it to produce the correct amplitude of a known test
signal (such as a white screen) at the AD8123 output.
VGAIN can also be derived from VPEAK according to the linear
relationships shown in Figure 17 and Figure 18. Figure 23 shows
how to derive VPOLE and VGAIN from VPEAK in a UTP application
that originates from a low-Z source.
0
6814-028
V
PEAK
2+ 0.9V
20Ω
5.11kΩ
V
PEAK
V
POLE
≈
5V
14kΩ
8.25kΩ
5.11kΩ
V
GAIN
≈ 0.89 × V
PEAK
+ 0.38V
5V
60.4kΩ
133kΩ
V
PEAK
Figure 23. Deriving VPOLE and VGAIN from VPEAK with Low-Z Source for UTP Cable
AD8123
Rev. A | Page 13 of 16
USING THE AD8123 WITH COAXIAL CABLE
The VPOLE control allows the AD8123 to be used with other
types of cable, including coaxial cable. Figure 18 presents the
recommended settings for VPEAK, VPOLE, and VGAIN when the
AD8123 is used with good quality 75 Ω video cable. Figure 24
shows how to derive VPOLE and VGAIN from VPEAK in a coaxial
cable application where VPEAK originates from a low-Z source.
20Ω
5.11kΩ
20kΩ
V
PEAK
V
PEAK
–5V
+5V
24.3kΩ
47.5kΩ
1.16kΩ
V
GAIN
≈ 1.06 × V
PEAK
– 0.62V
V
POLE
≈ 0.76 × V
PEAK
– 0.41V
10kΩ
1.24kΩ
0
6814-029
Figure 24. Deriving VPOLE and VGAIN from VPEAK with Low-Z Source for Coaxial Cable
The op amp in the circuit that develops VGAIN is required to
insert the offset of −0.62 V with a gain from VPEAK to VGAIN that
is close to unity. A passive offset circuit would require an offset
injection voltage that is much larger in magnitude than the
available −5 V supply. Clearly, the VGAIN control voltage can
also be developed independently.
The AD8123 differential input can accept signals carried over
unbalanced cable, as shown in Figure 25, for an unbalanced
75 Ω coaxial cable termination.
75Ω
06814-030
INPUT FROM
75Ω CABLE
AD8123
INPUT STAGE
Figure 25. Terminating a 75 Ω Cable
DRIVING 75 Ω VIDEO CABLE WITH THE AD8123
When the RGB outputs must drive a 75 Ω line rather than a
high impedance load, an additional gain of two is required to
make up for the double termination loss (75 Ω source and load
terminations). There are two options available for this.
One option is to place the additional gain of 2 at the drive end
by using the AD8148 triple differential driver to drive the cable.
The AD8148 has a fixed gain of 4 instead of the usual gain of 2
and thereby provides the required additional gain of 2 without
having to add additional amplifiers to the signal chain. The
AD8148 also contains sync-on-common-mode encoding. If
sync-on-common-mode is not required, it can be deactivated
on the AD8148 by connecting its SYNC LEVEL input to ground.
The other option is to include a triple gain-of-2 buffer, such as the
ADA4862-3, on the AD8123 RGB outputs, as shown in Figure 26
for one channel (power supplies not shown). The ADA4862-3
provides the gain of 2 that compensates for the double-
termination loss.
06814-022
ONE VIDEO
OUTPUT
FROM AD8123
ONE CHANNEL OF ADA4862-3
75Ω
75
Ω
500
Ω
500
Ω
Z
0
= 75Ω
Figure 26. Using ADA4862-3 on AD8123 Outputs
DRIVING A CAPACITIVE LOAD
When driving a high impedance capacitive input, it is necessary
to place a small series resistor between each of the three AD8123
video outputs and the load to buffer the input capacitance of the
device being driven. Clearly, the resistor value must be small
enough to preserve the required bandwidth.
FILTERING THE RGB OUTPUTS
In some cases, it is desirable to place low-pass filters on the
AD8123 video outputs to reduce high frequency noise. A 3-pole
Butterworth filter with cutoff frequency in the neighborhood of
140 MHz is sufficient in most applications. Figure 27 and Figure 28
present filters for the high impedance load case (driving a delay
line, crosspoint switch, ADA4862-3) and the double-termination
case (75 Ω source and load resistances), respectively. In the high
impedance load case, the load capacitance must be absorbed in
the capacitor that is placed across the load. For example, in
Figure 27, if the high-Z load were the input to an ADA4862-3,
which has an input capacitance of 2 pF, the filter capacitor value
in parallel with the input would be 15 pF to obtain 17 pF.
06814-023
*INPUT CAPACITANCE OF LOAD MUST BE
ABSORBED INTO THIS VALUE.
HIGH-Z
150nH
100
Ω
5.6pF 17pF*
AD8123
OUTPUT
Figure 27. 140 MHz Low-Pass Filter on AD8123 Output Feeding High-Z Load
06814-024
180nH
75Ω
75
ΩZ0 = 75Ω
15pF 15pF
AD8123
OUTPUT
Figure 28. 135 MHz Low-Pass Filter on AD8123 Output Feeding
Doubly Terminated Load
These filters are by no means the only choices but are presented
here as examples. In the high-Z load case, it is important to
keep the filter source resistance large enough to buffer the
capacitive loading presented by the first capacitor in the filter.
AD8123
Rev. A | Page 14 of 16
POWER SUPPLY FILTERING
External power supply filtering between the system power
supplies and the AD8123 is required in most applications to
prevent supply noise from contaminating the received signal as
well as to prevent unwanted feedback through the supplies that
could cause instability. Figure 29 shows that the AD8123 power
supply rejection decreases with increasing frequency. These
plots are for the lowest control settings and shift upward as the
peaking is increased.
–60
–50
–40
–30
–20
–10
0
10
+PSRR
–PSRR
100k 1M 10M 100M
06814-017
FREQUENCY (Hz)
PSRR (dB)
V
GAIN
= 0V
V
PEAK
= 0V
V
POLE
= 0V
Figure 29. AD8123 PSRR vs. Frequency
A suitable filter that uses a surface-mount ferrite bead is shown
in Figure 30, and its frequency response is shown in Figure 31.
Because the frequency response was taken using a 50 Ω network
analyzer and with only one 0.1 μF capacitor on the AD8123
side, the actual amount of rejection provided by the filter in a
real-world application will be different from that shown in
Figure 31. The general shape of the rejection curve, however,
matches Figure 31, providing substantially increased overall
PSRR from approximately 5 MHz to 500 MHz, where it is most
needed. One filter is required on each of the two supplies (not one
filter per supply pin).
FAIR-RITE
2743021447
*ALL AD8123 SUPPLY PINS ARE INDIVIDUALLY
DECOUPLED WITH A 0.1µF CAPACITOR.
06814-031
0.1µF 4700pF 4700pF
SYSTEM
SUPPLY TO AD8123*
Figure 30. Power Supply Filter
–120
–100
–80
–60
–40
–20
0
10k 100k 1M 10M 100M
06814-018
FREQUENCY (Hz)
OUTPUT RESPONSE (dB)
Figure 31. Power Supply Filter Frequency Response in a 50 Ω System
LAYOUT AND POWER SUPPLY DECOUPLING
CONSIDERATIONS
Standard high speed PCB layout practices should be adhered
to when designing with the AD8123. A solid ground plane is
required and controlled impedance traces should be used when
interconnecting the high speed signals. Source termination
resistors on all of the outputs must be placed as close as possible
to the output pins.
The exposed paddle on the underside of the AD8123 must be
connected to a pad that connects to at least one PCB plane.
Several thermal vias should be used to make the connection
between the pad and the plane(s).
High quality 0.1 μF power supply decoupling capacitors should
be placed as close as possible to all of the supply pins. Small
surface-mount ceramic capacitors should be used for these, and
tantalum capacitors are recommended for bulk supply decoupling.
AD8123
Rev. A | Page 15 of 16
INPUT COMMON-MODE RANGE When used, common-mode sync signals are generally applied
with a peak deviation of 500 mV and thereby increase the
common-mode level from 2.675 V to 3.175 V. This common-
mode level exceeds the specified input voltage swing limits of
±3.0 V; therefore, the AD8123 cannot be used with a system
that uses common-mode sync encoding with 500 mV sync peak
deviation and 2.5 V common-mode line level. While it is possible
to operate a driver powered from a single 5 V supply at a common-
mode voltage of <2.5 V to obtain a received voltage swing that is
within the specified limits, there is not much margin for other
shifts in the common-mode level due to interference pickup and
differing ground potentials. There are two ways to increase the
common-mode range of the overall system. One is to power the
driver from ±5 V supplies, and the other is to place an AD8143
in front of the AD8123, as shown in Figure 32. These techniques
may be combined or applied separately.
Most applications that use the AD8123 as a receiver use a driver
(such as one from the AD8146/AD8147/AD8148 family, the
AD8133, or the AD8134) powered from ±5 V supplies. This
places the common-mode voltage on the line nominally at 0 V
relative to the ground potential at the driver and provides optimum
immunity from any common-mode anomalies picked up along
the cable (including ground shifts between the driver and receiver
ends). In many of these applications, the AD8123 input voltage
range of typically ±3.0 V is sufficient. If wider input range is
required, the AD8143 triple receiver (input common-mode
range equals ±10.5 V on ±12 V supplies) may be placed in front of
the AD8123. Figure 32 illustrates how this is done for one channel.
06814-033
100Ω
49.9Ω
1
2
3
RECEIVED
SIGNAL
+5V
ONE AD8123
INPUT
ONE AD8143 CHANNE L
POWER SUPPLIES = ±12V
–5V
HBAT-540C
SMALL SIGNAL FREQUENCY RESPONSE
Though the AD8123 large signal frequency response
(VO = 1 V p-p) is of most concern, occasionally designers are
interested in the small signal frequency response. The AD8123
frequency response for VO = 300 m V p-p is shown in Figure 33
for 200 meter and 300 meter cable lengths.
Figure 32. Optional Use of AD8143 in Front of AD8123 for
Wide Input Common-Mode Range
06814-032
FREQUENCY (MHz)
GAIN (dB)
–12
–11
–10
–9
–8
–7
–6
–5
–4
–3
–2
–1
0
1
2
3
0.1 1 10 1000.01
V
O
= 300mV p-p
200 METERS
300 METERS
The Schottky diodes are required to protect the AD8123 from
any AD8143 outputs that may exceed the AD8123 input limits.
The 49.9 Ω resistor limits the fault current and produces a pole
at approximately 800 MHz with the effective diode capacitance of
3 pF and the AD8123 input capacitance of 1 pF. The pole drops
the response by only 0.07 dB at 100 MHz and therefore has a
negligible effect on the signal.
When using a single 5 V supply on the driver side, the common-
mode voltage at the driver is typically midsupply, or VCM = 2.5 V.
The largest received differential video signal is approximately
700 mV p-p, and this therefore adds 175 mVPEAK to the common-
mode voltage, resulting in a worst-case peak voltage of 2.675 V
on an AD8123 input (presuming there is no ground shift between
driver and receiver). This is within the AD8123 input voltage
swing limits, and such a system works well as long as the difference
in ground potential between driver and receiver does not cause
the input voltage swing to exceed its specified limits.
Figure 33. Small Signal Frequency Response for Various Cable Lengths
POWER-DOWN
The power-down feature is intended to be used to reduce power
consumption when a particular device is not in use and does
not place the output in a high-Z state when asserted. The input
logic levels and supply current in power-down mode are presented
in the Power Supply section of Table 1 .
AD8123
Rev. A | Page 16 of 16
OUTLINE DIMENSIONS
COMPLIANT TO JEDEC STANDARDS MO-220-VJJD-2
080107-A
1
40
10
11
29
28
20
19
4.45
4.30 SQ
4.15
TOP
VIEW
6.00
BSC SQ
5.75
BCS SQ
COPLANARITY
0.08
4.50
REF
0.50
0.40
0.30
0.50
BSC
PIN 1
INDICATOR
0.60 MAX
0.60 MAX
0.25 MIN
EXPOSED
PAD
(BOT TOM VIEW)
PIN 1
INDICATOR
0.30
0.23
0.18
0.20 REF
12° MAX 0.80 MAX
0.65 TYP
1.00
0.85
0.80 0.05 MAX
0.02 NOM
S
EATING
PLANE
Figure 34. 40-Lead Lead Frame Chip Scale Package [LFCSP_VQ]
6 mm × 6 mm, Very Thin Quad
(CP-40-4)
Dimensions shown in millimeters
ORDERING GUIDE
Model Temperature Range Package Description Package Option
AD8123ACPZ1−40°C to +85°C 40-Lead LFCSP_VQ CP-40-4
AD8123ACPZ-R71−40°C to +85°C 40-Lead LFCSP_VQ CP-40-4
AD8123ACPZ-RL1−40°C to +85°C 40-Lead LFCSP_VQ CP-40-4
1 Z = RoHS Compliant Part.
©2007 Analog Devices, Inc. All rights reserved. Trademarks and
registered trademarks are the property of their respective owners.
D06814-0-11/07(A)
