600 MHz, 32 x 16 Buffered Analog Crosspoint Switch AD8104/AD8105 APPLICATIONS Routing of high speed signals including RGB and component video routing KVM Compressed video (MPEG, wavelet) Data communications GENERAL DESCRIPTION The AD8104/AD8105 are high speed, 32 x 16 analog crosspoint switch matrices. They offer 600 MHz bandwidth and slew rate of 1800 V/s for high resolution computer graphics (RGB) signal switching. With less than -70 dB of crosstalk and -90 dB isolation (@ 5 MHz), the AD8104/AD8105 are useful in many high speed applications. The 0.1 dB flatness, which is greater than 50 MHz, makes the AD8104/AD8105 ideal for composite video switching. The AD8104/AD8105 include 16 independent output buffers that can be placed into a high impedance state for paralleling crosspoint outputs so that off-channels present minimal loading to an output bus. The AD8104 has a differential gain of +1, FUNCTIONAL BLOCK DIAGRAM D0 D1 D2 D3 D4 D5 VDD DGND AD8104/ AD8105 A0 A1 A2 A3 SER/PAR 0 192-BIT SHIFT REGISTER WITH 6-BIT PARALLEL LOADING DATA IN 96 96 UPDATE PARALLEL LATCH RESET DATA OUT NO CONNECT 96 DECODE 16 x 6:32 DECODERS INPUT RECEIVER G = +1* G = +2** 512 16 OUTPUT BUFFER G = +1 2 ENABLE/DISABLE 2 SWITCH MATRIX *AD8104 ONLY **AD8105 ONLY VPOS VNEG 16 OUTPUT PAIRS 1 SET INDIVIDUAL, OR RESET ALL OUTPUTS TO OFF WE CLK 32 INPUT PAIRS High channel count, 32 x 16 high speed, nonblocking switch array Differential or single-ended operation Differential G = +1 (AD8104) or G = +2 (AD8105) Pin compatible with AD8117/AD8118, 32 x 32 switch arrays Flexible power supplies Single +5 V supply, or dual 2.5 V supplies Serial or parallel programming of switch array High impedance output disable allows connection of multiple devices with minimal loading on output bus Excellent video performance >50 MHz 0.1 dB gain flatness 0.05% differential gain error (RL = 150 ) 0.05 phase error (RL = 150 ) Excellent ac performance Bandwidth: 600 MHz Slew rate: 1800 V/s Settling time: 2.5 ns to 1% Low power of 1.7 W Low all hostile crosstalk < -70 dB @ 5 MHz < -40 dB @ 600 MHz Reset pin allows disabling of all outputs (connected through a capacitor to ground provides power-on reset capability) 304-ball BGA package (31 mm x 31 mm) VOCM 06612-001 FEATURES Figure 1. while the AD8105 has a differential gain of +2 for ease of use in back-terminated load applications. They operate as fully differential devices or can be configured for single-ended operation. Either a single +5 V supply or dual 2.5 V supplies can be used, while consuming only 340 mA of idle current with all outputs enabled. The channel switching is performed via a double-buffered, serial digital control (which can accommodate daisy-chaining of several devices), or via a parallel control, allowing updating of an individual output without reprogramming the entire array. The AD8104/AD8105 are packaged in a 304-ball BGA package and are available over the extended industrial temperature range of -40C to +85C. Rev. 0 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 (c)2007 Analog Devices, Inc. All rights reserved. AD8104/AD8105 TABLE OF CONTENTS Features .............................................................................................. 1 ESD Caution...................................................................................7 Applications....................................................................................... 1 Pin Configuration and Function Descriptions..............................8 Functional Block Diagram .............................................................. 1 Truth Table and Logic Diagram ............................................... 13 General Description ......................................................................... 1 I/O Schematics................................................................................ 15 Revision History ............................................................................... 2 Typical Performance Characteristics ........................................... 17 Specifications..................................................................................... 3 Theory of Operation ...................................................................... 25 Timing Characteristics (Serial Mode) ....................................... 5 Applications Information .............................................................. 26 Timing Characteristics (Parallel Mode) .................................... 6 Programming.............................................................................. 26 Absolute Maximum Ratings............................................................ 7 Operating Modes........................................................................ 27 Thermal Resistance ...................................................................... 7 Outline Dimensions ....................................................................... 36 Power Dissipation......................................................................... 7 Ordering Guide .......................................................................... 36 REVISION HISTORY 6/07--Revision 0: Initial Version Rev. 0 | Page 2 of 36 AD8104/AD8105 SPECIFICATIONS VS = 2.5 V at TA = 25C, RL, diff = 200 , VOCM = 0 V, differential I/O mode, unless otherwise noted. Table 1. Parameter DYNAMIC PERFORMANCE -3 dB Bandwidth Gain Flatness Propagation Delay Settling Time Slew Rate NOISE/DISTORTION PERFORMANCE Differential Gain Error Differential Phase Error Crosstalk, All Hostile Off Isolation, Input-to-Output Input Voltage Noise DC PERFORMANCE Voltage Gain Gain Error Gain Matching Differential Offset Common-Mode Offset OUTPUT CHARACTERISTICS Output Impedance Output Disable Capacitance Output Leakage Current Output Voltage Range VOCM Input Range Output Swing Limit Output Current INPUT CHARACTERISTICS Input Voltage Range Common-Mode Rejection Ratio Input Capacitance Input Resistance Input Offset Current VOCM Input Bias Current VOCM Input Impedance Conditions Min AD8104/AD8105 Typ Max Unit 200 mV p-p, typical channel 2 V p-p, typical channel 0.1 dB, 200 mV p-p 0.1 dB, 2 V p-p 2 V p-p 1%, 2 V step 2 V step, peak 2 V step, 10% to 90% 600 420/525 100/50 70/50 1.3 2.5 1800 1500 MHz MHz MHz MHz ns ns V/s V/s NTSC or PAL, RL = 150 NTSC or PAL, RL = 150 f = 5 MHz f = 10 MHz f = 100 MHz f = 600 MHz f = 10 MHz, one channel 0.1 MHz to 50 MHz 0.05 0.05 -80/-70 -72/-68 -48/-50 -40/-50 -90 45/53 % Degrees dB dB dB dB dB nV/Hz Differential +1/+2 1 1 1 5 25 V/V % % % mV mV No load Channel-to-channel DC, enabled Disabled, differential Disabled Disabled No load VOUT, diff = 2 V p-p VOUT, diff = 2.8 V p-p Single-ended output Maximum operating signal Common mode, VIN, diff = 2 V p-p Differential f = 10 MHz Any switch configuration Differential Rev. 0 | Page 3 of 36 2.8 -0.5 -0.25 -1.3 3 25 90 0.1 30 4 1 3.8 +0.8 +0.6 +1.3 30 -2 +2 2/1 48 2 5 1 64 4 k pF A V p-p V V V mA V V dB pF k A A k AD8104/AD8105 Parameter SWITCHING CHARACTERISTICS Enable On Time Switching Time, 2 V Step Switching Transient (Glitch) POWER SUPPLIES Supply Current Supply Voltage Range PSRR OPERATING TEMPERATURE RANGE Temperature Range JA JC Conditions Min AD8104/AD8105 Typ Max 50% update to 1% settling 50% update to 1% settling Differential 100 100 40 VPOS, outputs enabled, no load VPOS, outputs disabled VNEG, outputs enabled, no load VNEG, outputs disabled VDD, outputs enabled, no load 340 210 340 210 Unit ns ns mV p-p VNEG, VPOS, f = 1 MHz VOCM, f = 1 MHz 4.5 to 5.5 85 75 mA mA mA mA mA V dB dB Operating (still air) Operating (still air) Operating (still air) -40 to +85 14 1 C C/W C/W Rev. 0 | Page 4 of 36 420 240 420 240 1.2 AD8104/AD8105 TIMING CHARACTERISTICS (SERIAL MODE) Specifications subject to change without notice. Table 2. Parameter Serial Data Setup Time CLK Pulse Width Serial Data Hold Time CLK Pulse Separation CLK to UPDATE Delay UPDATE Pulse Width CLK to DATA OUT Valid Propagation Delay, UPDATE to Switch On or Off RESET Pulse Width RESET Time Symbol t1 t2 t3 t4 t5 t6 t7 Min 40 50 50 150 10 90 120 Limit Typ Max Unit ns ns ns ns ns ns ns ns ns ns 100 60 200 1 WE 0 t2 t4 1 CLK LOAD DATA INTO SERIAL REGISTER ON FALLING EDGE 0 t1 t3 1 DATA IN OUT15 (D5) OUT15 (D4) OUT0 (D0) 0 t5 1 = LATCHED t6 TRANSFER DATA FROM SERIAL REGISTER TO PARALLEL LATCHES DURING LOW LEVEL UPDATE 0 = TRANSPARENT t7 06612-002 1 DATA OUT 0 Figure 2. Timing Diagram, Serial Mode Table 3. Logic Levels VIH RESET, SER/PAR, CLK, DATA IN, UPDATE VIL RESET, SER/PAR, CLK, DATA IN, UPDATE VOH DATA OUT VOL DATA OUT IIH RESET1, SER/PAR, CLK, DATA IN, UPDATE IIL RESET1, SER/PAR, CLK, DATA IN, UPDATE IOH DATA OUT IOL DATA OUT 2.0 V min 0.6 V max VDD - 0.3 V min DGND + 0.5 V max 1 A max -1 A min -1 mA max 1 mA min 1 See Figure 15. Rev. 0 | Page 5 of 36 AD8104/AD8105 TIMING CHARACTERISTICS (PARALLEL MODE) Specifications subject to change without notice. Table 4. Parameter Parallel Data Setup Time WE Pulse Width Parallel Data Hold Time WE Pulse Separation WE to UPDATE Delay UPDATE Pulse Width Propagation Delay, UPDATE to Switch On or Off RESET Pulse Width RESET Time t2 1 Symbol t1 t2 t3 t4 t5 t6 Min 80 110 150 90 10 90 Limit Typ Max Unit ns ns ns ns ns ns ns ns ns 100 60 200 t4 WE 0 t1 t3 1 D0 TO D5 A0 TO A3 0 t5 t6 06612-003 1 = LATCHED UPDATE 0 = TRANSPARENT Figure 3. Timing Diagram, Parallel Mode Table 5. Logic Levels VIH RESET, SER/PAR, WE, D0, D1, D2, D3, D4, D5, A0, A1, A2, A3, UPDATE VIL RESET, SER/PAR, WE, D0, D1, D2, D3, D4, D5, A0, A1, A2, A3, UPDATE VOH DATA OUT VOL DATA OUT IIH RESET 1 , SER/PAR, WE, D0, D1, D2, D3, D4, D5, A0, A1, A2, A3, UPDATE IIL RESET1, SER/PAR, WE, D0, D1, D2, D3, D4, D5, A0, A1, A2, A3, UPDATE IOH DATA OUT IOL DATA OUT 2.0 V min 0.6 V max Disabled Disabled 1 A max -1 A min Disabled Disabled 1 See Figure 15. Rev. 0 | Page 6 of 36 AD8104/AD8105 ABSOLUTE MAXIMUM RATINGS POWER DISSIPATION Table 6. Rating 6V The AD8104/AD8105 are operated with 2.5 V or +5 V supplies and can drive loads down to 100 , resulting in a large range of possible power dissipations. For this reason, extra care must be taken derating the operating conditions based on ambient temperature. 6V +0.5 V to -2.5 V 8V VNEG to VPOS 2 V VDD (VPOS - 1 V) to (VNEG + 1 V) Momentary 80 mA -65C to +125C -40C to +85C 300C Packaged in a 304-ball BGA, the AD8104/AD8105 junction-toambient thermal impedance (JA) is 14C/W. For long-term reliability, the maximum allowed junction temperature of the die should not exceed 150C. Temporarily exceeding this limit may cause a shift in parametric performance due to a change in stresses exerted on the die by the package. Exceeding a junction temperature of 175C for an extended period can result in device failure. The following curve shows the range of allowed internal die power dissipations that meet these conditions over the -40C to +85C ambient temperature range. When using Table 6, do not include external load power in the maximum power calculation, but do include load current dropped on the die output transistors. 8 TJ = 150C 150C MAXIMUM POWER (W) 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. 4 15 35 45 55 65 75 85 Figure 4. Maximum Die Power Dissipation vs. Ambient Temperature Table 7. Thermal Resistance JC 1 25 AMBIENT TEMPERATURE (C) JA is specified for the worst-case conditions, that is, a device soldered in a circuit board for surface-mount packages. JA 14 6 5 THERMAL RESISTANCE Package Type 304-Ball BGA 7 06612-004 Parameter Analog Supply Voltage (VPOS - VNEG) Digital Supply Voltage (VDD - DGND) Ground Potential Difference (VNEG - DGND) Maximum Potential Difference (VDD - VNEG) Common-Mode Analog Input Voltage Differential Analog Input Voltage Digital Input Voltage Output Voltage (Disabled Analog Output) Output Short-Circuit Duration Output Short-Circuit Current Storage Temperature Range Operating Temperature Range Lead Temperature (Soldering, 10 sec) Junction Temperature ESD CAUTION JB 6.5 JT 0.6 JB 5.7 Unit C/W Rev. 0 | Page 7 of 36 AD8104/AD8105 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 A VPOS VPOS VPOS VPOS NC NC NC NC NC NC NC NC NC NC NC NC NC NC NC NC VPOS VPOS VPOS A B VPOS VPOS VPOS NC NC NC NC NC NC NC NC NC NC NC NC NC NC NC NC VPOS VPOS VPOS VPOS B C VPOS VPOS VPOS VPOS VNEG VNEG VNEG VNEG VNEG VNEG VPOS VPOS VPOS VNEG VNEG VNEG VNEG VNEG VNEG VPOS VPOS VPOS VPOS C D IN16 VPOS VPOS VNEG VOCM VNEG VNEG VNEG VNEG VNEG VPOS VPOS VPOS VNEG VNEG VNEG VNEG VNEG VOCM VNEG VPOS IP0 VPOS D E IP16 IN17 VNEG VOCM VOCM VNEG IN0 IP1 E F IN18 IP17 VNEG VDD VDD VNEG IP2 IN1 F G IP18 IN19 VNEG DGND DGND VNEG IN2 IP3 G H IN20 IP19 VNEG RESET DATA OUT VNEG IP4 IN3 H J IP20 IN21 VNEG CLK VNEG IN4 IP5 J VNEG WE DATA IN VNEG IP6 IN5 K SER/ PAR VPOS IN6 IP7 L DGND VPOS IP8 IN7 M K IN22 IP21 L IP22 IN23 VPOS D5 M IN24 IP23 VPOS D4 N IP24 IN25 VPOS D3 A3 VPOS IN8 IP9 N P IN26 IP25 VNEG D2 A2 VNEG IP10 IN9 P R IP26 IN27 VNEG D1 A1 VNEG IN10 IP11 R T IN28 IP27 VNEG D0 A0 VNEG IP12 IN11 T U IP28 IN29 VNEG VDD VDD VNEG IN12 IP13 U V IN30 IP29 VNEG DGND DGND VNEG IP14 IN13 V W IP30 IN31 VNEG VOCM VOCM VNEG IN14 IP15 W Y VPOS IP31 VPOS VNEG VOCM VNEG VNEG VNEG VNEG VNEG VPOS VPOS VPOS VNEG VNEG VNEG VNEG VNEG VOCM VNEG VPOS VPOS IN15 Y AA VPOS VPOS VPOS VPOS VNEG VNEG VNEG VNEG VNEG VNEG VPOS VPOS VPOS VNEG VNEG VNEG VNEG VNEG VNEG VPOS VPOS VPOS VPOS AA AB VPOS VPOS VPOS VPOS ON14 OP14 ON12 OP12 ON10 OP10 ON8 OP8 ON6 OP6 ON4 OP4 ON2 OP2 ON0 OP0 VPOS VPOS VPOS AB AC VPOS VPOS VPOS ON15 OP15 ON13 OP13 ON11 OP11 ON9 OP9 ON7 OP7 ON5 OP5 ON3 OP3 ON1 OP1 VPOS VPOS VPOS VPOS AC 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 AD8104/AD8105 BOTTOM VIEW (Not to Scale) Figure 5. Package Bottom View Rev. 0 | Page 8 of 36 06612-005 PIN CONFIGURATION AND FUNCTION DESCRIPTIONS 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 A VPOS VPOS VPOS NC NC NC NC NC NC NC NC NC NC NC NC NC NC NC NC VPOS VPOS VPOS VPOS A B VPOS VPOS VPOS VPOS NC NC NC NC NC NC NC NC NC NC NC NC NC NC NC NC VPOS VPOS VPOS B C VPOS VPOS VPOS VPOS VNEG VNEG VNEG VNEG VNEG VNEG VPOS VPOS VPOS VNEG VNEG VNEG VNEG VNEG VNEG VPOS VPOS VPOS VPOS C D VPOS IP0 VPOS VNEG VOCM VNEG VNEG VNEG VNEG VNEG VPOS VPOS VPOS VNEG VNEG VNEG VNEG VNEG VOCM VNEG VPOS VPOS IN16 D E IP1 IN0 VNEG VOCM VOCM VNEG IN17 IP16 E F IN1 IP2 VNEG VDD VDD VNEG IP17 IN18 F G IP3 IN2 VNEG DGND DGND VNEG IN19 IP18 G H IN3 IP4 VNEG DATA OUT RESET VNEG IP19 IN20 H J IP5 IN4 VNEG CLK VNEG IN21 IP20 J VNEG DATA IN WE VNEG IP21 IN22 K D5 VPOS IN23 IP22 L D4 VPOS IP23 IN24 M K IN5 IP6 L IP7 IN6 VPOS SER/ PAR M IN7 IP8 VPOS DGND N IP9 IN8 VPOS A3 D3 VPOS IN25 IP24 N P IN9 IP10 VNEG A2 D2 VNEG IP25 IN26 P R IP11 IN10 VNEG A1 D1 VNEG IN27 IP26 R T IN11 IP12 VNEG A0 D0 VNEG IP27 IN28 T U IP13 IN12 VNEG VDD VDD VNEG IN29 IP28 U V IN13 IP14 VNEG DGND DGND VNEG IP29 IN30 V W IP15 IN14 VNEG VOCM VOCM VNEG IN31 IP30 W Y IN15 VPOS VPOS VNEG VOCM VNEG VNEG VNEG VNEG VNEG VPOS VPOS VPOS VNEG VNEG VNEG VNEG VNEG VOCM VNEG VPOS IP31 VPOS Y AA VPOS VPOS VPOS VPOS VNEG VNEG VNEG VNEG VNEG VNEG VPOS VPOS VPOS VNEG VNEG VNEG VNEG VNEG VNEG VPOS VPOS VPOS VPOS AA AB VPOS VPOS VPOS OP0 ON0 OP2 ON2 OP4 ON4 OP6 ON6 OP8 ON8 OP10 ON10 OP12 ON12 OP14 ON14 VPOS VPOS VPOS VPOS AB AC VPOS VPOS VPOS VPOS OP1 ON1 OP3 ON3 OP5 ON5 OP7 ON7 OP9 ON9 OP11 ON11 OP13 ON13 OP15 ON15 VPOS VPOS VPOS AC 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 AD8104/AD8105 TOP VIEW (Not to Scale) Figure 6. Package Top View Table 8. Ball Grid Description Ball No. A1 A2 A3 A4 A5 A6 A7 A8 A9 A10 A11 A12 A13 A14 Mnemonic VPOS VPOS VPOS NC NC NC NC NC NC NC NC NC NC NC Description Analog Positive Power Supply. Analog Positive Power Supply. Analog Positive Power Supply. No Connect. No Connect. No Connect. No Connect. No Connect. No Connect. No Connect. No Connect. No Connect. No Connect. No Connect. Ball No. A15 A16 A17 A18 A19 A20 A21 A22 A23 B1 B2 B3 B4 B5 Rev. 0 | Page 9 of 36 Mnemonic NC NC NC NC NC VPOS VPOS VPOS VPOS VPOS VPOS VPOS VPOS NC Description No Connect. No Connect. No Connect. No Connect. No Connect. Analog Positive Power Supply. Analog Positive Power Supply. Analog Positive Power Supply. Analog Positive Power Supply. Analog Positive Power Supply. Analog Positive Power Supply. Analog Positive Power Supply. Analog Positive Power Supply. No Connect. 06612-006 AD8104/AD8105 AD8104/AD8105 Ball No. B6 B7 B8 B9 B10 B11 B12 B13 B14 B15 B16 B17 B18 B19 B20 B21 B22 B23 C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 C20 C21 C22 C23 D1 D2 D3 D4 D5 Mnemonic NC NC NC NC NC NC NC NC NC NC NC NC NC NC NC VPOS VPOS VPOS VPOS VPOS VPOS VPOS VNEG VNEG VNEG VNEG VNEG VNEG VPOS VPOS VPOS VNEG VNEG VNEG VNEG VNEG VNEG VPOS VPOS VPOS VPOS VPOS IP0 VPOS VNEG VOCM D6 D7 D8 D9 D10 D11 VNEG VNEG VNEG VNEG VNEG VPOS Description No Connect. No Connect. No Connect. No Connect. No Connect. No Connect. No Connect. No Connect. No Connect. No Connect. No Connect. No Connect. No Connect. No Connect. No Connect. Analog Positive Power Supply. Analog Positive Power Supply. Analog Positive Power Supply. Analog Positive Power Supply. Analog Positive Power Supply. Analog Positive Power Supply. Analog Positive Power Supply. Analog Negative Power Supply. Analog Negative Power Supply. Analog Negative Power Supply. Analog Negative Power Supply. Analog Negative Power Supply. Analog Negative Power Supply. Analog Positive Power Supply. Analog Positive Power Supply. Analog Positive Power Supply. Analog Negative Power Supply. Analog Negative Power Supply. Analog Negative Power Supply. Analog Negative Power Supply. Analog Negative Power Supply. Analog Negative Power Supply. Analog Positive Power Supply. Analog Positive Power Supply. Analog Positive Power Supply. Analog Positive Power Supply. Analog Positive Power Supply. Input Number 0, Positive Phase. Analog Positive Power Supply. Analog Negative Power Supply. Output Common-Mode Reference Supply. Analog Negative Power Supply. Analog Negative Power Supply. Analog Negative Power Supply. Analog Negative Power Supply. Analog Negative Power Supply. Analog Positive Power Supply. Ball No. D12 D13 D14 D15 D16 D17 D18 D19 Mnemonic VPOS VPOS VNEG VNEG VNEG VNEG VNEG VOCM D20 D21 D22 VNEG VPOS VPOS D23 E1 E2 E3 E4 IN16 IP1 IN0 VNEG VOCM Input Number 16, Negative Phase. Input Number 1, Positive Phase. Input Number 0, Negative Phase. Analog Negative Power Supply. Output Common-Mode Reference Supply. E20 VOCM Output Common-Mode Reference Supply. E21 E22 E23 F1 F2 F3 F4 F20 F21 F22 F23 G1 G2 G3 G4 G20 G21 G22 G23 H1 H2 H3 H4 H20 H21 H22 H23 J1 J2 J3 J4 J20 VNEG IN17 IP16 IN1 IP2 VNEG VDD VDD VNEG IP17 IN18 IP3 IN2 VNEG DGND DGND VNEG IN19 IP18 IN3 IP4 VNEG DATA OUT RESET VNEG IP19 IN20 IP5 IN4 VNEG CLK UPDATE Analog Negative Power Supply. Input Number 17, Negative Phase. Input Number 16, Positive Phase. Input Number 1, Negative Phase. Input Number 2, Positive Phase. Analog Negative Power Supply. Logic Positive Power Supply. Logic Positive Power Supply. Analog Negative Power Supply. Input Number 17, Positive Phase. Input Number 18, Negative Phase. Input Number 3, Positive Phase. Input Number 2, Negative Phase. Analog Negative Power Supply. Logic Negative Power Supply. Logic Negative Power Supply. Analog Negative Power Supply. Input Number 19, Negative Phase. Input Number 18, Positive Phase. Input Number 3, Negative Phase. Input Number 4, Positive Phase. Analog Negative Power Supply. Control Pin: Serial Data Out. Control Pin: Second Rank Data Reset. Analog Negative Power Supply. Input Number 19, Positive Phase. Input Number 20, Negative Phase. Input Number 5, Positive Phase. Input Number 4, Negative Phase. Analog Negative Power Supply. Control Pin: Serial Data Clock. Control Pin: Second Rank Write Strobe. Rev. 0 | Page 10 of 36 Description Analog Positive Power Supply. Analog Positive Power Supply. Analog Negative Power Supply. Analog Negative Power Supply. Analog Negative Power Supply. Analog Negative Power Supply. Analog Negative Power Supply. Output Common-Mode Reference Supply. Analog Negative Power Supply. Analog Positive Power Supply. Analog Positive Power Supply. AD8104/AD8105 Ball No. J21 J22 J23 K1 K2 K3 K4 K20 K21 K22 K23 L1 L2 L3 L4 L20 L21 L22 L23 M1 M2 M3 M4 M20 M21 M22 M23 N1 N2 N3 N4 N20 N21 N22 N23 P1 P2 P3 P4 P20 P21 P22 P23 R1 R2 R3 R4 R20 R21 R22 R23 T1 Mnemonic VNEG IN21 IP20 IN5 IP6 VNEG DATA IN WE VNEG IP21 IN22 IP7 IN6 VPOS SER/PAR D5 VPOS IN23 IP22 IN7 IP8 VPOS DGND D4 VPOS IP23 IN24 IP9 IN8 VPOS A3 D3 VPOS IN25 IP24 IN9 IP10 VNEG A2 D2 VNEG IP25 IN26 IP11 IN10 VNEG A1 D1 VNEG IN27 IP26 IN11 Description Analog Negative Power Supply. Input Number 21, Negative Phase. Input Number 20, Positive Phase. Input Number 5, Negative Phase. Input Number 6, Positive Phase. Analog Negative Power Supply. Control Pin: Serial Data In. Control Pin: First Rank Write Strobe. Analog Negative Power Supply. Input Number 21, Positive Phase. Input Number 22, Negative Phase. Input Number 7, Positive Phase. Input Number 6, Negative Phase. Analog Positive Power Supply. Control Pin: Serial/Parallel Mode Select. Control Pin: Input Address Bit 5. Analog Positive Power Supply. Input Number 23, Negative Phase. Input Number 22, Positive Phase. Input Number 7, Negative Phase. Input Number 8, Positive Phase. Analog Positive Power Supply. Logic Negative Power Supply Control Pin: Input Address Bit 4. Analog Positive Power Supply. Input Number 23, Positive Phase. Input Number 24, Negative Phase. Input Number 9, Positive Phase. Input Number 8, Negative Phase. Analog Positive Power Supply. Control Pin: Output Address Bit 3. Control Pin: Input Address Bit 3. Analog Positive Power Supply. Input Number 25, Negative Phase. Input Number 24, Positive Phase. Input Number 9, Negative Phase. Input Number 10, Positive Phase. Analog Negative Power Supply. Control Pin: Output Address Bit 2. Control Pin: Input Address Bit 2. Analog Negative Power Supply. Input Number 25, Positive Phase. Input Number 26, Negative Phase. Input Number 11, Positive Phase. Input Number 10, Negative Phase. Analog Negative Power Supply. Control Pin: Output Address Bit 1. Control Pin: Input Address Bit 1. Analog Negative Power Supply. Input Number 27, Negative Phase. Input Number 26, Positive Phase. Input Number 11, Negative Phase. Ball No. T2 T3 T4 T20 T21 T22 T23 Mnemonic IP12 VNEG A0 D0 VNEG IP27 IN28 Description Input Number 12, Positive Phase. Analog Negative Power Supply. Control Pin: Output Address Bit 0. Control Pin: Input Address Bit 0. Analog Negative Power Supply. Input Number 27, Positive Phase. Input Number 28, Negative Phase. U1 U2 U3 U4 U20 U21 U22 U23 IP13 IN12 VNEG VDD VDD VNEG IN29 IP28 Input Number 13, Positive Phase. Input Number 12, Negative Phase. Analog Negative Power Supply. Logic Positive Power Supply. Logic Positive Power Supply. Analog Negative Power Supply. Input Number 29, Negative Phase. Input Number 28, Positive Phase. V1 V2 V3 V4 V20 V21 V22 V23 W1 W2 W3 W4 IN13 IP14 VNEG DGND DGND VNEG IP29 IN30 IP15 IN14 VNEG VOCM W20 VOCM W21 W22 W23 Y1 Y2 Y3 Y4 Y5 VNEG IN31 IP30 IN15 VPOS VPOS VNEG VOCM Y6 Y7 Y8 Y9 Y10 Y11 Y12 Y13 Y14 Y15 Y16 Y17 Y18 VNEG VNEG VNEG VNEG VNEG VPOS VPOS VPOS VNEG VNEG VNEG VNEG VNEG Input Number 13, Negative Phase. Input Number 14, Positive Phase. Analog Negative Power Supply. Logic Negative Power Supply. Logic Negative Power Supply. Analog Negative Power Supply. Input Number 29, Positive Phase. Input Number 30, Negative Phase. Input Number 15, Positive Phase. Input Number 14, Negative Phase. Analog Negative Power Supply. Output Common-Mode Reference Supply. Output Common-Mode Reference Supply. Analog Negative Power Supply. Input Number 31, Negative Phase. Input Number 30, Positive Phase. Input Number 15, Negative Phase. Analog Positive Power Supply. Analog Positive Power Supply. Analog Negative Power Supply. Output Common-Mode Reference Supply. Analog Negative Power Supply. Analog Negative Power Supply. Analog Negative Power Supply. Analog Negative Power Supply. Analog Negative Power Supply. Analog Positive Power Supply. Analog Positive Power Supply. Analog Positive Power Supply. Analog Negative Power Supply. Analog Negative Power Supply. Analog Negative Power Supply. Analog Negative Power Supply. Analog Negative Power Supply. Rev. 0 | Page 11 of 36 AD8104/AD8105 Ball No. Y19 Mnemonic VOCM Y20 Y21 Y22 Y23 AA1 AA2 AA3 AA4 AA5 AA6 AA7 AA8 AA9 AA10 AA11 AA12 AA13 AA14 AA15 AA16 AA17 AA18 AA19 AA20 AA21 AA22 AA23 AB1 AB2 AB3 AB4 AB5 AB6 AB7 AB8 AB9 VNEG VPOS IP31 VPOS VPOS VPOS VPOS VPOS VNEG VNEG VNEG VNEG VNEG VNEG VPOS VPOS VPOS VNEG VNEG VNEG VNEG VNEG VNEG VPOS VPOS VPOS VPOS VPOS VPOS VPOS OP0 ON0 OP2 ON2 OP4 ON4 Description Output Common-Mode Reference Supply. Analog Negative Power Supply. Analog Positive Power Supply. Input Number 31, Positive Phase. Analog Positive Power Supply. Analog Positive Power Supply. Analog Positive Power Supply. Analog Positive Power Supply. Analog Positive Power Supply. Analog Negative Power Supply. Analog Negative Power Supply. Analog Negative Power Supply. Analog Negative Power Supply. Analog Negative Power Supply. Analog Negative Power Supply. Analog Positive Power Supply. Analog Positive Power Supply. Analog Positive Power Supply. Analog Negative Power Supply. Analog Negative Power Supply. Analog Negative Power Supply. Analog Negative Power Supply. Analog Negative Power Supply. Analog Negative Power Supply. Analog Positive Power Supply. Analog Positive Power Supply. Analog Positive Power Supply. Analog Positive Power Supply. Analog Positive Power Supply. Analog Positive Power Supply. Analog Positive Power Supply. Output Number 0, Positive Phase. Output Number 0, Negative Phase. Output Number 2, Positive Phase. Output Number 2, Negative Phase. Output Number 4, Positive Phase. Output Number 4, Negative Phase. Ball No. AB10 AB11 AB12 AB13 AB14 AB15 AB16 AB17 AB18 AB19 AB20 AB21 AB22 AB23 AC1 AC2 AC3 AC4 AC5 AC6 AC7 AC8 AC9 AC10 AC11 AC12 AC13 AC14 AC15 AC16 AC17 AC18 AC19 AC20 AC21 AC22 AC23 Rev. 0 | Page 12 of 36 Mnemonic OP6 ON6 OP8 ON8 OP10 ON10 OP12 ON12 OP14 ON14 VPOS VPOS VPOS VPOS VPOS VPOS VPOS VPOS OP1 ON1 OP3 ON3 OP5 ON5 OP7 ON7 OP9 ON9 OP11 ON11 OP13 ON13 OP15 ON15 VPOS VPOS VPOS Description Output Number 6, Positive Phase. Output Number 6, Negative Phase. Output Number 8, Positive Phase. Output Number 8, Negative Phase. Output Number 10, Positive Phase. Output Number 10, Negative Phase. Output Number 12, Positive Phase. Output Number 12, Negative Phase. Output Number 14, Positive Phase. Output Number 14, Negative Phase. Analog Positive Power Supply. Analog Positive Power Supply. Analog Positive Power Supply. Analog Positive Power Supply. Analog Positive Power Supply. Analog Positive Power Supply. Analog Positive Power Supply. Analog Positive Power Supply. Output Number 1, Positive Phase. Output Number 1, Negative Phase. Output Number 3, Positive Phase. Output Number 3, Negative Phase. Output Number 5, Positive Phase. Output Number 5, Negative Phase. Output Number 7, Positive Phase. Output Number 7, Negative Phase. Output Number 9, Positive Phase. Output Number 9, Negative Phase. Output Number 11, Positive Phase. Output Number 11, Negative Phase. Output Number 13, Positive Phase. Output Number 13, Negative Phase. Output Number 15, Positive Phase. Output Number 15, Negative Phase. Analog Positive Power Supply. Analog Positive Power Supply. Analog Positive Power Supply. AD8104/AD8105 TRUTH TABLE AND LOGIC DIAGRAM Table 9. Operation Truth Table DATA INPUT X DATA OUTPUT X RESET 0 X Datai 1 Datai-192 1 0 X D0...D5 2 A0...A3 3 1 1 0 X X 1 X X X X N/A in parallel mode N/A in parallel mode X 1 1 WE X UPDATE X 1 X 0 X 1 1 CLK X 1 Datai: serial data. D0...D5: data bits. 3 A0...A3: address bits. 2 Rev. 0 | Page 13 of 36 SER/PAR Operation/Comment Asynchronous reset. All outputs are disabled. Remainder of logic in 192-bit shift register is unchanged. Serial mode. The data on the serial DATA IN line is loaded into the serial register. The first bit clocked into the serial register appears at DATA OUT 192 clock cycles later. Parallel mode. The data on parallel lines D0 to D5 are loaded into the shift register location addressed by A0 to A3. Switch matrix update. Data in the 192-bit shift register transfers into the parallel latches that control the switch array. No change in logic. OUTPUT ADDRESS A0 A1 A2 A3 4 TO 16 DECODER CLK WE DATA IN (SERIAL) Figure 7. Logic Diagram Rev. 0 | Page 14 of 36 RESET UPDATE OUT15 EN OUT14 EN OUT13 EN OUT12 EN OUT11 EN OUT10 EN OUT9 EN OUT8 EN OUT7 EN OUT6 EN OUT5 EN OUT4 EN OUT3 EN OUT2 EN OUT1 EN OUT0 EN (OUTPUT ENABLE) SER/PAR D0 D1 D2 D3 D4 D5 OUT0 B0 CLR Q ENA D S D1 Q D Q D0 CLK OUT0 B1 CLR Q ENA D S D1 Q Q D D0 CLK OUT0 B2 CLR Q ENA D S D1 Q D Q D0 CLK 512 OUT0 B4 CLR Q ENA D S D1 Q DQ D0 CLK OUT0 EN CLR Q ENA D S D1 Q D Q D0 CLK SWITCH MATRIX OUT0 B3 CLR Q ENA D S D1 Q Q D D0 CLK OUT1 B0 CLR Q ENA D S D1 Q D Q D0 CLK DECODE OUT14 EN CLR Q ENA D S D1 Q D Q D0 CLK OUT15 B0 CLR Q ENA D S D1 Q DQ D0 CLK OUT15 B2 CLR Q ENA D S D1 Q D Q D0 CLK 16 OUTPUT ENABLE OUT15 B1 CLR Q ENA D S D1 Q D Q D0 CLK OUT15 B3 CLR Q ENA D S D1 Q D Q D0 CLK OUT15 B4 CLR Q ENA D S D1 D Q Q D0 CLK OUT15 EN CLR Q ENA D S D1 Q Q D D0 CLK DATA OUT (SERIAL) 06612-007 PARALLEL DATA AD8104/AD8105 AD8104/AD8105 I/O SCHEMATICS IPn 1.3pF 2500 1.3pF 2500 06612-008 INn 06612-012 0.3pF OPn, ONn Figure 12. AD8104/AD8105 Receiver Simplified Equivalent Circuit When Driving Differentially Figure 8. AD8104/AD8105 Enabled Output (see also ESD Protection Map, Figure 18) OPn IPn 3.4pF 3.33k AD8104 G = +1 3.76k AD8105 G = +2 1.6pF 30k 06612-013 0.4pF INn 3.4pF 06612-009 ONn Figure 13. AD8104/AD8105 Receiver Simplified Equivalent Circuit When Driving Single-Ended Figure 9. AD8104/AD8105 Disabled Output (see also ESD Protection Map, Figure 18) 2500 IPn 2538 1.3pF VOCM 2538 VNEG Figure 10. AD8104 Receiver (see also ESD Protection Map, Figure 18) 2500 IPn 5075 VDD 1.3pF 25k 0.3pF 2500 5075 06612-011 RESET 1.3pF INn Figure 14. VOCM Input (see also ESD Protection Map, Figure 18) 1k DGND Figure 11. AD8105 Receiver (see also ESD Protection Map, Figure 18) Rev. 0 | Page 15 of 36 06612-015 2500 06612-010 1.3pF INn 06612-014 0.3pF Figure 15. Reset Input (see also ESD Protection Map, Figure 18) AD8104/AD8105 VPOS 1k DGND VNEG VDD 06612-017 DATA OUT DGND DGND Figure 18. ESD Protection Map Figure 16. Logic Input (see also ESD Protection Map, Figure 18) Figure 17. Logic Output (see also ESD Protection Map, Figure 18) Rev. 0 | Page 16 of 36 06612-018 CLK, RESET, SER/PAR, WE, UPDATE, DATA IN, DATA OUT, A[3:0], D[5:0] IPn, INn, OPn, ONn, VOCM 06612-016 CLK, SER/PAR, WE, UPDATE, DATA IN, A[3:0], D[5:0] VDD AD8104/AD8105 TYPICAL PERFORMANCE CHARACTERISTICS VS = 2.5 V at TA = 25C, RL, diff = 200 , VOCM = 0 V, differential I/O mode, unless otherwise noted. 200 10 180 AD8105 6 160 4 140 120 AD8104 0 100 -2 80 -4 60 -6 40 -8 -10 1 100 10 06612-022 COUNT 2 06612-019 GAIN (dB) 8 20 0 1000 540 560 580 FREQUENCY (MHz) 640 660 700 800 AD8105 6 4 2 AD8104 0 -2 -4 06612-020 -6 -8 1 10 -1.0 -1.5 -2.0 1000 100 -0.5 06612-023 NORMALIZED BANDWIDTH ERROR (%) 0 8 0 2 FREQUENCY (MHz) 4 6 8 10 12 14 16 NUMBER OF ENABLED CHANNELS Figure 20. AD8104, AD8105 Large Signal Frequency Response, 2 V p-p Figure 23. AD8104 Bandwidth Error vs. Enabled Channels 10 0 DIFFERENTIAL OUT 8 -10 6 10pF 4 5pF -20 2pF 2 CMR (dB) NORMALIZED GAIN (dB) 680 Figure 22. AD8104 -3 dB Bandwidth Histogram, One Device, All 512 Channels 10 GAIN (dB) 620 FREQUENCY (MHz) Figure 19. AD8104, AD8105 Small Signal Frequency Response, 200 mV p-p -10 600 0 -2 0pF -4 -30 -40 -50 -6 -10 0 10 100 -70 300k 1000 FREQUENCY (MHz) 06612-024 -60 06612-021 -8 1M 10M 100M 1G FREQUENCY (Hz) Figure 21. AD8104 Small Signal Frequency Response with Capacitive Loads, 200 mV p-p Rev. 0 | Page 17 of 36 Figure 24. AD8104, AD8105 Common-Mode Rejection 2G AD8104/AD8105 0 -15 DIFFERENTIAL OUT DIFFERENTIAL IN/OUT VNEG AGGRESSOR -25 -20 -35 CROSSTALK (dB) VPOS AGGRESSOR PSR (dB) -45 -55 VOCM AGGRESSOR -65 -40 -60 -75 06612-025 -95 0.1 1 10 100 -100 300k 1000 06612-028 -80 -85 1M Figure 25. AD8104 Power Supply Rejection 10 10M 0 SINGLE-ENDED OUT 0 DIFFERENTIAL IN/OUT -20 -5 VOCM AGGRESSOR CROSSTALK (dB) -15 -20 VNEG AGGRESSOR -30 -35 -40 -60 VPOS AGGRESSOR -80 06612-026 -40 -45 1 10 100 -100 300k 1000 06612-029 PSR (dB) -10 -50 0.1 1M FREQUENCY (MHz) 10M 100M 1G FREQUENCY (Hz) Figure 29. AD8105 Crosstalk, One Adjacent Channel Figure 26. AD8104 Power Supply Rejection, Single-Ended 0 180 SINGLE-ENDED IN/OUT DIFFERENTIAL OUT 160 -20 140 AD8105 80 60 40 AD8104 10k -60 -80 20 0 1k -40 100k -100 300k 1M 06612-030 100 CROSSTALK (dB) 120 06612-027 NOISE SPECTRAL DENSITY (nV/ Hz) 1G Figure 28. AD8104 Crosstalk, One Adjacent Channel 5 -25 100M FREQUENCY (Hz) FREQUENCY (MHz) 1M 10M 100M 1G FREQUENCY (Hz) FREQUENCY (Hz) Figure 30. AD8104 Crosstalk, One Adjacent Channel, Single-Ended Figure 27. AD8104, AD8105 Noise Spectral Density, RTO Rev. 0 | Page 18 of 36 AD8104/AD8105 0 0 SINGLE-ENDED IN/OUT -20 CROSSTALK (dB) -20 -40 -60 -80 -40 -60 -80 -100 300k 06612-031 -100 1M 10M 100M -120 300k 1G 06612-034 CROSSTALK (dB) SINGLE-ENDED IN/OUT 1M Figure 31. AD8105 Crosstalk, One Adjacent Channel, Single-Ended 0 SINGLE-ENDED IN/OUT -20 -40 CROSSTALK (dB) CROSSTALK (dB) -20 -60 -80 -40 -60 -80 06612-032 -100 1M 10M 100M -100 300k 1G 1M FREQUENCY (Hz) 0 100M 1G Figure 35. AD8105 Crosstalk, All Hostile, Single-Ended 0 DIFFERENTIAL IN/OUT DIFFERENTIAL IN/OUT -20 FEEDTHROUGH (dB) -20 -40 -60 -40 -60 1M 10M 100M -100 300k 1G 06612-036 -80 -80 06612-033 CROSSTALK (dB) 10M FREQUENCY (Hz) Figure 32. AD8104 Crosstalk, All Hostile -100 300k 1G Figure 34. AD8104 Crosstalk, All Hostile, Single-Ended DIFFERENTIAL IN/OUT -120 300k 100M 06612-035 0 10M FREQUENCY (Hz) FREQUENCY (Hz) 1M 10M 100M FREQUENCY (Hz) FREQUENCY (Hz) Figure 33. AD8105 Crosstalk, All Hostile Figure 36. AD8104 Crosstalk, Off Isolation Rev. 0 | Page 19 of 36 1G 2G AD8104/AD8105 0 30000 SINGLE-ENDED IN/OUT 25000 OUTPUT IMPEDANCE () -40 -60 -80 20000 15000 10000 06612-037 5000 1M 10M 100M 1G 06612-040 FEEDTHROUGH (dB) -20 -100 300k DIFFERENTIAL OUT 0 100k 2G 1M 10M FREQUENCY (Hz) Figure 37. AD8104 Crosstalk, Off Isolation, Single-Ended 6000 100M 1G FREQUENCY (Hz) Figure 40. AD8104, AD8105 Output Impedance, Disabled 1000 DIFFERENTIAL IN AD8105 5000 OUTPUT IMPEDANCE () INPUT IMPEDANCE () AD8104 4000 3000 2000 100 10 1 1M 10M 100M 06612-041 0 300k 06612-038 1000 0.1 100k 1G 1M 10M FREQUENCY (Hz) 100M 1G FREQUENCY (Hz) Figure 38. AD8104, AD8105 Input Impedance Figure 41. AD8104, AD8105 Output Impedance, Enabled 0.4 4500 SINGLE-ENDED IN 4000 0.3 AD8104 0.2 VOUT (V, DIFF) 3000 2500 2000 1500 0 -0.1 -0.3 1M 10M 100M -0.4 1G 06612-042 500 0 300k 0.1 -0.2 1000 06612-039 INPUT IMPEDANCE () 3500 AD8105 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 TIME (ns) FREQUENCY (Hz) Figure 39. AD8104, AD8105 Input Impedance, Single-Ended Figure 42. AD8104 Small Signal Pulse Response, 200 mV p-p Rev. 0 | Page 20 of 36 AD8104/AD8105 3 1.5 0.20 UPDATE 0.15 1.0 2 0.5 1 0 0 0 -0.05 P-CHANNEL -1.0 -0.20 06612-043 -0.15 0 1 2 3 4 5 6 7 8 9 -1 -0.5 -0.10 VOUT -1.5 -40 10 11 12 13 14 15 -2 -20 0 20 40 60 80 100 -3 120 06612-046 VOUT (V, DIFF) VOUT (V, SE) N-CHANNEL 0.05 UPDATE (V) 0.10 TIME (ns) TIME (ns) Figure 46. AD8104 Switching Time Figure 43. AD8104 Small Signal Pulse Response, Single-Ended, 200 mV p-p 2.0 1.5 2 5000 1 4000 0 3000 0.5 0 -0.5 VOUT -1 2000 -2 1000 SLEW RATE (V/s) VOUT (V, DIFF) VOUT (V, DIFF) 1.0 -1.0 -3 0 1 2 3 4 5 6 7 8 9 -1000 -4 10 11 12 13 14 15 0 1 2 3 4 5 TIME (ns) TIME (ns) 06612-047 -2.0 0 SLEW RATE 06612-044 -1.5 Figure 47. AD8104 Large Signal Rising Edge and Slew Rate Figure 44. AD8104 Large Signal Pulse Response, 2 V p-p 3500 2 1.0 0.8 VOUT 1 2500 0.2 0 -0.2 P-CHANNEL -0.4 06612-045 -0.6 -0.8 -1.0 0 1 2 3 4 5 6 7 8 9 0 1500 -1 500 SLEW RATE -2 -500 -3 -1500 -2500 -4 0 10 11 12 13 14 15 1 2 3 4 5 1ns/DIV TIME (ns) Figure 45. AD8104 Large Signal Pulse Response, Single-Ended, 2 V p-p Rev. 0 | Page 21 of 36 Figure 48. AD8104 Large Signal Falling Edge and Slew Rate 06612-048 N-CHANNEL VOUT (V, DIFF) VOUT (V, SE) 0.4 SLEW RATE (V/s) 0.6 AD8104/AD8105 0.014 5 DIFFERENTIAL PHASE ERROR (%) 0.012 3 2 1 0 10 20 30 40 50 60 70 80 0.008 0.006 0.004 0.002 0 -0.002 06612-049 0 -40 -30 -20 -10 0.010 -0.004 -700 90 100 06612-052 OFFSET (mV) 4 -500 -300 -100 100 300 500 700 VIN, DIFF (mV) TEMPERATURE (C) Figure 49. AD8104 VOS vs. Temperature with All Outputs Enabled Figure 52. AD8104 Phase vs. DC Voltage, Carrier Frequency = 3.58 MHz, Subcarrier Amplitude = 600 mV p-p, Differential 50 2.0 40 1.5 10pF 5pF 1.0 2pF VOUT (V, DIFF) 20 10 0 0.5 0pF 0 -0.5 -1.0 -20 -0.10 -0.08 -0.06 -0.04 -0.02 -1.5 06612-050 -10 0 0.02 0.04 0.06 0.08 06612-053 VOUT (mV, DIFF) 30 -2.0 0.10 0 2 4 6 TIME (s) 8 10 12 14 16 18 TIME (ns) Figure 50. AD8104 Switching Transient (Glitch) Figure 53. AD8104 Large Signal Pulse Response with Capacitive Loads 0.020 0.4 5pF 0.2 2pF VOUT (V, DIFF) 0.010 0.005 0 0.1 0pF 0 -0.1 -0.2 -0.005 -500 -300 -100 100 300 500 06612-054 -0.010 -700 -0.3 06612-051 DIFFERENTIAL GAIN ERROR (%) 10pF 0.3 0.015 -0.4 700 0 VIN, DIFF (mV) 2 4 6 8 10 12 14 16 18 TIME (ns) Figure 51. AD8104 Gain vs. DC Voltage, Carrier Frequency = 3.58 MHz, Subcarrier Amplitude = 600 mV p-p, Differential Figure 54. AD8104 Small Signal Pulse Response with Capacitive Loads Rev. 0 | Page 22 of 36 AD8104/AD8105 2.8 IDD (SERIAL MODE) 0.8 1.6 0.6 1.2 0.4 0.8 0.4 VOUT IPOS AND INEG CURRENT (mA) 2.0 UPDATE (V) 2.4 1.0 0.2 500 750 IDD (PARALLEL MODE) 400 -0.2 -50 -30 -10 10 30 50 70 90 110 130 -0.4 150 TIME (ns) 650 IPOS AND INEG (ALL OUTPUTS ENABLED) 300 200 550 450 IPOS AND INEG (ALL OUTPUTS DISABLED) 0 0 100 -35 -25 -15 -5 5 15 25 35 45 55 65 75 85 350 95 TEMPERATURE (C) Figure 55. AD8104 Enable Time Figure 58. AD8104, AD8105 Quiescent Supply Currents vs. Temperature 1.4 1.2 850 IDD CURRENT (A) UPDATE 06612-055 VOUT (V, DIFF) 1.2 600 06612-058 1.4 UPDATE 2.8 360 2.4 340 1040 960 IPOS, INEG 0.4 0.8 0.2 0 -10 10 30 50 70 90 110 130 800 IDD SERIAL 280 720 260 0.4 240 560 0 220 480 -0.4 150 200 TIME (ns) Figure 56. AD8104 Disable Time 1 2 3 5 6 7 8 400 10 11 12 13 14 15 16 9 CHANNELS 65 3.25 60 -0.01 2.75 OUTPUT ERROR (%) 50 -0.02 -0.03 06612-057 -0.04 -25 0 25 50 75 3.00 (VOUT - VIN)/VOUT 55 GAIN (dB) 4 Figure 59. AD8104, AD8105 Quiescent Supply Currents vs. Enabled Outputs 0 -0.05 -50 640 IDD PARALLEL 2.50 VIN 45 2.25 40 2.00 35 1.75 30 1.50 VOUT 25 1.25 20 1.00 15 0.75 10 0.50 5 0.25 0 0 -5 100 -0.25 0 TEMPERATURE (C) 1 2 3 4 5 TIME (ns) Figure 57. AD8104 DC Gain vs. Temperature Figure 60. AD8104 Settling Time Rev. 0 | Page 23 of 36 6 7 (V, DIFF) -30 300 06612-060 -0.2 -50 880 IDD (A) 1.2 IPOS AND INEG (mA) 0.6 UPDATE (V) 1.6 320 06612-056 VOUT (V, DIFF) 0.8 VOUT 06612-059 2.0 1.0 AD8104/AD8105 5 2.0 4 1.5 VOUTP 1.0 2 1 VOUT (V, SE) 0 -1 0.5 VINP 0 -0.5 -2 -1.0 -3 -1.5 06612-061 -4 -5 0 1 2 3 4 5 6 VOUTN VINN 06612-063 OUTPUT/INPUT (%) 3 -2.0 7 100 0 200 300 TIME (ns) Figure 61. AD8104 Settling Time (Zoom) -30 VINP 700 VOUT = 2V p-p, DIFF -50 DISTORTION (dBc) 1.0 0.5 0 -0.5 -1.0 THIRD HARMONIC -60 -70 SECOND HARMONIC -80 VOUTN -2.0 VINN -2.5 0 100 200 300 400 500 -90 600 -100 0.1 700 TIME (ns) 06612-064 -1.5 06612-062 VOUT (V, SE) 600 -40 VOUTP 1.5 500 Figure 63. AD8105 Overdrive Recovery, Single-Ended 2.5 2.0 400 TIME (ns) 1 10 100 FREQUENCY (MHz) Figure 62. AD8104 Overdrive Recovery, Single-Ended Figure 64. AD8104 Harmonic Distortion Rev. 0 | Page 24 of 36 1000 AD8104/AD8105 THEORY OF OPERATION The AD8104/AD8105 are fully differential crosspoint arrays with 16 outputs, each of which can be connected to any one of 32 inputs. Organized by output row, 32 switchable input transconductance stages are connected to each output buffer to form 32-to-1 multiplexers. There are 16 of these multiplexers, each with its inputs wired in parallel, for a total array of 512 transconductance stages forming a multicast-capable crosspoint switch. Decoding logic for each output selects one (or none) of the transconductance stages to drive the output stage. The enabled transconductance stage drives the output stage, and feedback forms a closed-loop amplifier with a differential gain of +1 (the difference between the output voltages is equal to the difference between the input voltages). A second feedback loop controls the common-mode output level, forcing the average of the differential output voltages to match the voltage on the VOCM reference pin. Although each output has an independent common-mode control loop, the VOCM reference is common for the entire chip, and as such needs to be driven with a low impedance to avoid crosstalk. Each differential input to the AD8104/AD8105 is buffered by a receiver. The purpose of this receiver is to provide an extended input common-mode range, and to remove this common mode from the signal chain. Like the output multiplexers, the input receiver has both a differential loop and a common-mode control loop. A mask-programmable feedback network sets the closed-loop differential gain. For the AD8104, this differential gain is +1, and for the AD8105, this differential gain is +2. The receiver has an input stage that does not respond to the common mode of the signal. This architecture, along with the attenuating feedback network, allows the user to apply input voltages that extend from rail to rail. Excess differential loop gain bandwidth product reduces the effect of the closed-loop gain on the bandwidth of the device. The output stage of the AD8104/AD8105 is designed for low differential gain and phase error when driving composite video signals. It also provides slew current for fast pulse response when driving component video signals. Unlike many multiplexer designs, these requirements are balanced such that large signal bandwidth is very similar to small signal bandwidth. The design load is 150 , but provisions are made to drive loads as low as 75 as long as on-chip power dissipation limits are not exceeded. The outputs of the AD8104/AD8105 can be disabled to minimize on-chip power dissipation. When disabled, there is a feedback network of 25 k between the differential outputs. This high impedance allows multiple ICs to be bussed together without additional buffering. Care must be taken to reduce output capacitance, which results in more overshoot and frequency domain peaking. A series of internal amplifiers drive internal nodes such that a wideband high impedance is presented at the disabled output, even while the output bus is under large signal swings. When the outputs are disabled and driven externally, the voltage applied to them should not exceed the valid output swing range for the AD8104/AD8105 in order to keep these internal amplifiers in their linear range of operation. Applying excess differential voltages to the disabled outputs can cause damage to the AD8104/AD8105 and should be avoided (see the Absolute Maximum Ratings section for guidelines). The connection of the AD8104/AD8105 is controlled by a flexible TTL-compatible logic interface. Either parallel or serial loading into a first rank of latches preprograms each output. A global update signal moves the programming data into the second rank of latches, simultaneously updating all outputs. In serial mode, a serial-out pin allows devices to be daisy-chained together for single-pin programming of multiple ICs. A poweron reset pin is available to avoid bus conflicts by disabling all outputs. This power-on reset clears the second rank of latches, but does not clear the first rank of latches. In parallel mode, to quickly clear the first rank, a broadcast parallel programming feature is available. In serial mode, preprogramming individual inputs is not possible and the entire shift register needs to be flushed. The AD8104/AD8105 can operate on a single +5 V supply, powering both the signal path (with the VPOS/VNEG supply pins), and the control logic interface (with the VDD/DGND supply pins). However, to easily interface to ground-referenced video signals, split supply operation is possible with 2.5 V supplies. In this case, a flexible logic interface allows the control logic supplies (VDD/DGND) to be run off +2 V/0 V to +5 V/0 V while the core remains on split supplies. Additional flexibility in the analog output common-mode level facilitates unequal split supplies. If +3 V/-2 V supplies to +2 V/-3 V supplies are desired, the VOCM pin can still be set to 0 V for ground-referenced video signals. Rev. 0 | Page 25 of 36 AD8104/AD8105 APPLICATIONS INFORMATION PROGRAMMING The AD8104/AD8105 have two options for changing the programming of the crosspoint matrix. In the first option, a serial word of 192 bits can be provided to update the entire matrix each time. The second option allows for changing the programming of a single output via a parallel interface. The serial option requires fewer signals, but more time (clock cycles) for changing the programming, while the parallel programming technique requires more signals, but can change a single output at a time and requires fewer clock cycles to complete programming. Therefore, the data for the last device in the chain should come at the beginning of the programming sequence. The length of the programming sequence is 192 bits times the number of devices in the chain. Parallel Programming Description Serial Programming Description When using the parallel programming mode, it is not necessary to reprogram the entire device when making changes to the matrix. In fact, parallel programming allows the modification of a single output at a time. Because this takes only one WE/UPDATE cycle, significant time savings can be realized by using parallel programming. The serial programming mode uses the CLK, DATA IN, UPDATE, and SER/PAR device pins. The first step is to assert a low on SER/PAR in order to enable the serial programming mode. The parallel clock WE should be held high during the entire serial programming operation. One important consideration in using parallel programming is that the RESET signal does not reset all registers in the AD8104/ AD8105. When taken low, the RESET signal only sets each output to the disabled state. This is helpful during power-up to ensure that two parallel outputs are not active at the same time. The UPDATE signal should be high during the time that data is shifted into the serial port of the device. Although the data still shifts in when UPDATE is low, the transparent, asynchronous latches allow the shifting data to reach the matrix. This causes the matrix to try to update to every intermediate state as defined by the shifting data. After initial power-up, the internal registers in the device generally have random data, even though the RESET signal has been asserted. If parallel programming is used to program one output, then that output will be properly programmed, but the rest of the device will have a random program state depending on the internal register content at power-up. Therefore, when using parallel programming, it is essential that all outputs be programmed to a desired state after power-up. This ensures that the programming matrix is always in a known state. From then on, parallel programming can be used to modify a single output or more at a time. The data at DATA IN is clocked in at every falling edge of CLK. A total of 192 bits must be shifted in to complete the programming. For each of the 16 outputs, there are five bits (D0 to D4) that determine the source of its input followed by one bit (D5) that determines the enabled state of the output. If D5 is low (output disabled), the five associated bits (D0 to D4) do not matter, because no input is switched to that output. These comprise the first 96 bits of DATA IN. The remaining 96 bits of DATA IN should be set to zero. If a string of 96 zeros is not suffixed to the first 96 bits of DATA IN, a certain test mode is employed that can cause the device to draw up to 40% more supply current. The most significant output address data, the enable bit (D5), is shifted in first, followed by the input address (D4 to D0) entered sequentially with D4 first and D0 last. Each remaining output is programmed sequentially, until the least significant output address data is shifted in. At this point, UPDATE can be taken low, which programs the device according to the data that was just shifted in. The UPDATE latches are asynchronous and when UPDATE is low, they are transparent. If more than one AD8104/AD8105 device is to be serially programmed in a system, the DATA OUT signal from one device can be connected to the DATA IN of the next device to form a serial chain. All of the CLK, UPDATE, and SER/PAR pins should be connected in parallel and operated as described previously. The serial data is input to the DATA IN pin of the first device of the chain, and it ripples through to the last. In similar fashion, if UPDATE is taken low after initial powerup, the random power-up data in the shift register will be programmed into the matrix. Therefore, in order to prevent the crosspoint from being programmed into an unknown state, do not apply a low logic level to UPDATE after power is initially applied. Programming the full shift register one time to a desired state, by either serial or parallel programming after initial power-up, eliminates the possibility of programming the matrix to an unknown state. To change the programming of an output via parallel programming, SER/PAR and UPDATE should be taken high. The serial programming clock, CLK, should be left high during parallel programming. The parallel clock, WE, should start in the high state. The 4-bit address of the output to be programmed should be put on A0 to A3. The first five data bits (D0 to D4) should contain the information that identifies the input that is programmed to the output that is addressed. The sixth data bit (D5) determines the enabled state of the output. If D5 is low (output disabled), then the data on D0 to D4 does not matter. After the desired address and data signals have been established, they can be latched into the shift register by a high to low transition of the WE signal. The matrix is not programmed, Rev. 0 | Page 26 of 36 AD8104/AD8105 mode rejection ratio (CMRR). Differential operation offers a great noise benefit for signals that are propagated over distance in a noisy environment. RF RG IN+ OUT- RCVR VOCM IN- Reset OUT+ RG When powering up the AD8104/AD8105, it is usually desirable to have the outputs come up in the disabled state. The RESET pin, when taken low, causes all outputs to be in the disabled state. However, the UPDATE signal does not reset all registers in the AD8104/AD8105. This is important when operating in the parallel programming mode. Refer to the Parallel Programming Description section for information about programming internal registers after power-up. Serial programming programs the entire matrix each time; therefore, no special considerations apply. Since the data in the shift register is random after power-up, it should not be used to program the matrix, or the matrix can enter unknown states. To prevent this, do not apply a logic low signal to UPDATE initially after power-up. The shift register should first be loaded with the desired data, and then UPDATE can be taken low to program the device. The RESET pin has a 20 k pull-up resistor to VDD that can be used to create a simple power-up reset circuit. A capacitor from RESET to ground holds RESET low for some time while the rest of the device stabilizes. The low condition causes all the outputs to be disabled. The capacitor then charges through the pull-up resistor to the high state, thus allowing full programming capability of the device. Because the AD8104/AD8105 have random data in the internal registers at power-up, the device may power up in a test state where the supply current is larger than typical. Therefore, the RESET pin should be used to disable all outputs and bring the device out of any test mode. OPERATING MODES The AD8104/AD8105 has fully differential inputs and outputs. The inputs and outputs can also be operated in a single-ended fashion. This presents several options for circuit configurations that require different gains and treatment of terminations, if they are used. Differential Input TO SWITCH MATRIX RF 06612-065 however, until the UPDATE signal is taken low. It is thus possible to latch in new data for several or all of the outputs first via successive negative transitions of WE while UPDATE is held high, and then have all the new data take effect when UPDATE goes low. This technique should be used when programming the device for the first time after power-up when using parallel programming. Figure 65. Input Receiver Equivalent Circuit The circuit configuration used by the differential input receivers is similar to that of several Analog Devices, Inc. general-purpose differential amplifiers, such as the AD8131. It is a voltage feedback amplifier with internal gain setting resistors. The arrangement of feedback makes the differential input impedance appear to be 5 k across the inputs. R IN , dm = 2 x RG = 5 k This impedance creates a small differential termination error if the user does not account for the 5 k parallel element, although this error is less than 1% in most cases. Additionally, the source impedance driving the AD8104/AD8105 appears in parallel with the internal gain-setting resistors, such that there may be a gain error for some values of source resistance. The AD8104/ AD8105 are adjusted such that its gains are correct when driven by a back-terminated 75 source impedance at each input phase (37.5 effective impedance to ground at each input pin, or 75 differential source impedance across pairs of input pins). If a different source impedance is presented, the differential gain of the AD8104/AD8105 can be calculated by Gdm = VOUT,dm VIN ,dm = RF RG + RS where: RG = 2.5 k. RS is the user single-ended source resistance (such as 37.5 for a back-terminated 75 source). RF = 2.538 k for the AD8104 and 5.075 k for the AD8105. In the case of the AD8104, Gdm = 2.538 k 2.5 k + RS In the case of the AD8105, Each differential input to the AD8104/AD8105 is applied to a differential receiver. These receivers allow the user to drive the inputs with a differential signal with an uncertain commonmode voltage, such as from a remote source over twisted pair. The receivers respond only to the difference in input voltages, and will restore a common-mode voltage suitable for the internal signal path. Noise or crosstalk that is present in both inputs is rejected by the input stage, as specified by its commonRev. 0 | Page 27 of 36 Gdm = 5.075 k 2.5 k + RS AD8104/AD8105 When operating with a differential input, care must be taken to keep the common mode, or average, of the input voltages within the linear operating range of the AD8104/AD8105 receiver. This common-mode range can extend rail-to-rail, provided the differential signal swing is small enough to avoid forward biasing the ESD diodes (it is safest to keep the common mode plus differential signal excursions within the supply voltages of the part). See the Specifications section for guaranteed input range. The differential output of the AD8104/AD8105 receiver is linear for a peak of 1.4 V of output voltage difference (1.4 V peak input difference for the AD8104, and 0.7 V peak input difference for the AD8105). Taking the output differentially, using the two output phases, this allows 2.8 V p-p of linear output signal swing. Beyond this level, the signal path can saturate and limits the signal swing. This is not a desired operation, as the supply current increases and the signal path is slow to recover from clipping. The absolute maximum allowed differential input signal is limited by the long-term reliability of the input stage. The limits in the Absolute Maximum Ratings section should be observed in order to avoid degrading device performance permanently. R IN = INn In most cases, a single-ended input signal is referred to midsupply, typically ground. In this case, the undriven differential input can be connected to ground. For best dynamic performance and lowest offset voltage, this unused input should be terminated with an impedance matching the driven input, instead of being directly shorted to ground. Due to the differential feedback of the receiver, there is high frequency signal current in the undriven input and it should be treated as a signal line in the board design. AD8104 IPn INn 75 OPn RCVR ONn 06612-067 75 (OR 37.5) OPn RCVR ONn Figure 67. Example of Input Driven Single-Ended 50 AC Coupling of Inputs 06612-066 50 2 x (RG + RS + RF ) where: RG = 2.5 k. RS is the user single-ended source resistance (such as 37.5 for a back-terminated 75 source). RF = 2.538 k for the AD8104 and 5.075 k for the AD8105. AD8104 IPn 1- RG + RS RF Figure 66. Example of Input Driven Differentially Single-Ended Input The AD8104/AD8105 input receivers can be driven singleendedly (unbalanced). From the standpoint of the receiver, there is very little difference between signals applied positive and negative in two phases to the input pair vs. a signal applied to one input only with the other input held at a constant potential. One small difference is that the common mode between the input pins is changing if only one input is moving, and there is a very small common-mode to differential conversion gain in the receiver that adds an additional gain error to the output (see the common-mode rejection ratio for the input stage in the Specifications section). For low frequencies, this gain error is negligible. The common-mode rejection ratio degrades with increasing frequency. When operating the AD8104/AD8105 receivers single-endedly, the observed input resistance at each input pin is lower than in the differential input case, due to a fraction of the receiver internal output voltage appearing as a common-mode signal on its input terminals, bootstrapping the voltage on the input resistance. This single-ended input resistance can be calculated by the equation It is possible to ac couple the inputs of the AD8104/AD8105 receiver. This is simplified because the bias current does not need to be supplied externally. A capacitor in series with the inputs to the AD8104/AD8105 creates a high-pass filter with the input impedance of the device. This capacitor needs to be sized such that the corner frequency is low enough for frequencies of interest. Differential Output Benefits of Differential Operation The AD8104/AD8105 have a fully differential switch core, with differential outputs. The two output voltages move in opposite polarity, with a differential feedback loop maintaining a fixed output stage differential gain of +1 (the different overall signal path gains between the AD8104 and AD8105 are set in the input stage for best signal-to-noise ratio). This differential output stage provides a benefit of crosstalk-canceling due to parasitic coupling from one output to another being equal and out of phase. Additionally, if the output of the device is utilized in a differential design, noise, crosstalk, and offset voltages generated on-chip that are coupled equally into both outputs are cancelled by the common-mode rejection ratio of the next device in the signal chain. By utilizing the AD8104/AD8105 outputs in a differential application, the best possible noise and offset specifications can be realized. Rev. 0 | Page 28 of 36 AD8104/AD8105 Differential Gain The specified signal path gain of the AD8104/AD8105 refers to its differential gain. For the AD8104, the gain of +1 means that the difference in voltage between the two output terminals is equal to the difference applied between the two input terminals. For the AD8105, the ratio of output difference voltage to applied input difference voltage is +2. back-termination, and helps shorten settling time by terminating reflected signals when driving a load that is not accurately terminated at the load end. A side effect of back-termination is an attenuation of the output signal by a factor of two. In this case, a gain of two is usually necessary somewhere in the signal path to restore the signal. AD8104/ AD8105 The common mode, or average voltage of the pair of output signals is set by the voltage on the VOCM pin. This voltage is typically set to midsupply (often ground), but can be moved approximately 0.5 V to accommodate cases where the desired output common-mode voltage may not be midsupply (as in the case of unequal split supplies). Adjusting VOCM can limit differential swing internally below the specifications listed in Table 1. Regardless of the differential gain of the device, the commonmode gain for the AD8104 and AD8105 is +1 to the output. This means that the common mode of the output voltages directly follows the reference voltage applied to the VOCM input. The VOCM reference is a high speed signal input, common to all output stages on the device. It requires only small amounts of bias current, but noise appearing on this pin is buffered to the outputs of all the output stages. As such, the VOCM node should be connected to a low noise, low impedance voltage to avoid being a source of noise, offset, and crosstalk in the signal path. Termination The AD8104/AD8105 are designed to drive 150 on each output (or an effective 300 differential), but the output stage is capable of supplying the current to drive 100 loads (200 differential) over the specified operating temperature range. If care is taken to observe the maximum power derating curves, the output stage can drive 75 loads with slightly reduced slew rate and bandwidth (an effective 150 differential load). Termination at the load end is recommended for best signal integrity. This load termination is often a resistor to a ground reference on each individual output. By terminating to the same voltage level that drives the VOCM reference, the power dissipation due to dc termination current is reduced. In differential signal paths, it is often desirable to terminate differentially, with a single resistor across the differential outputs at the load end. This is acceptable for the AD8104/ AD8105, but when the device outputs are placed in a disabled state, a small amount of dc bias current is required if the output is to present as a high impedance over an excursion of output bus voltages. If the AD8104/AD8105 disabled outputs are floated (or simply tied together by a resistor), internal nodes saturate and an increase in disabled output current may be observed. For best pulse response, it is often desirable to place a series resistor in each output to match the characteristic impedance and termination of the output trace or cable. This is known as OPn 50 ONn 06612-068 50 + 100 - Figure 68. Example of Back-Terminated Differential Load Single-Ended Output Usage The AD8104/AD8105 output pairs can be used single-endedly, taking only one output and not using the second. This is often desired to reduce the routing complexity in the design, or because a single-ended load is being driven directly. This mode of operation produces good results, but has some shortcomings when compared to taking the output differentially. When observing the single-ended output, noise that is common to both outputs appears in the output signal. This includes thermal noise in the chip biasing, as well as crosstalk that is coupled into the signal path. This component noise and crosstalk is equal in both outputs, and as such can be ignored by a differential receiver with a high common-mode rejection ratio. However, when taking the output single-ended, this noise is present with respect to the ground (or VOCM) reference and is not rejected. When observing the output single-ended, the distribution of offset voltages appears greater. In the differential case, the difference between the outputs when the difference between the inputs is zero is a small differential offset. This offset is created from mismatches in components of the signal path, which must be corrected by the finite differential loop gain of the device. In the single-ended case, this differential offset is still observed, but an additional offset component is also relevant. This additional component is the common-mode offset, which is a difference between the average of the outputs and the VOCM reference. This offset is created by mismatches that affect the signal path in a common-mode manner, and is corrected by the finite common-mode loop gain of the device. A differential receiver would reject this common-mode offset voltage, but in the single-ended case, this offset is observed with respect to the signal ground. The single-ended output sums half the differential offset voltage and all of the common-mode offset voltage for a net increase in observed offset. Rev. 0 | Page 29 of 36 AD8104/AD8105 Single-Ended Gain The AD8104/AD8105 operate as a closed-loop differential amplifier. The primary control loop forces the difference between the output terminals to be a ratio of the difference between the input terminals. One output increases in voltage, while the other decreases an equal amount to make the total difference correct. The average of these output voltages is forced to be equal to the voltage on the VOCM terminal by a second control loop. If only one output terminal is observed with respect to the VOCM terminal, only half of the difference voltage is observed. This implies that when using only one output of the device, half of the differential gain is observed. An AD8104 taken with single-ended output appears to have a gain of +0.5. An AD8105 has a single-ended gain of +1. This factor of one half in the gain increases the noise of the device when referred to the input, contributing to higher noise specifications for single-ended output designs. output draws current from the positive supply, the other output draws current from the negative supply. When the phase alternates, the first output draws current from the negative supply and the second from the positive supply. The effect is that a more constant current is drawn from each supply, such that the crosstalk-inducing supply fluctuation is minimized. A third benefit of driving balanced loads can be seen if one considers that the output pulse response changes as load changes. The differential signal control loop in the AD8104/ AD8105 forces the difference of the outputs to be a fixed ratio to the difference of the inputs. If the two output responses are different due to loading, this creates a difference that the control loop sees as signal response error, and it attempts to correct this error. This distorts the output signal from the ideal response if the two outputs were balanced. AD8104/ AD8105 Termination One component of crosstalk is magnetic, coupling by mutual inductance between output package traces and bond wires that carry load current. In a differential design, there is coupling from one pair of outputs to other adjacent pairs of outputs. The differential nature of the output signal simultaneously drives the coupling field in one direction for one phase of the output, and in an opposite direction for the other phase of the output. These magnetic fields do not couple exactly equal into adjacent output pairs due to different proximities, but they do destructively cancel the crosstalk to some extent. If the load current in each output is equal, this cancellation is greater, and less adjacent crosstalk is observed (regardless if the second output is actually being used). A second benefit of balancing the output loads in a differential pair is to reduce fluctuations in current requirements from the power supply. In single-ended loads, the load currents alternate from the positive supply to the negative supply. This creates a parasitic signal voltage in the supply pins due to the finite resistance and inductance of the supplies. This supply fluctuation appears as crosstalk in all outputs, attenuated by the power supply rejection ratio (PSRR) of the device. At low frequencies, this is a negligible component of crosstalk, but PSRR falls off as frequency increases. With differential, balanced loads, as one 75 75 ONn When operating the AD8104/AD8105 with a single-ended output, the preferred output termination scheme is a resistor at the load end to the VOCM voltage. A back-termination can be used, at an additional cost of one half the signal gain. In single-ended output operation, the complementary phase of the output is not used, and may or may not be terminated locally. Although the unused output can be floated to reduce power dissipation, there are several reasons for terminating the unused output with a load resistance matched to the load on the signal output. OPn 06612-069 150 Figure 69. Example of Back-Terminated Single-Ended Load Decoupling The signal path of the AD8104/AD8105 is based on high openloop gain amplifiers with negative feedback. Dominant-pole compensation is used on-chip to stabilize these amplifiers over the range of expected applied swing and load conditions. To guarantee this designed stability, proper supply decoupling is necessary with respect to both the differential control loops and the common-mode control loops of the signal path. Signalgenerated currents must return to their sources through low impedance paths at all frequencies in which there is still loop gain (up to 700 MHz at a minimum). A wideband parallel capacitor arrangement is necessary to properly decouple the AD8104/AD8105. The signal path compensation capacitors in the AD8104/ AD8105 are connected to the VNEG supply. At high frequencies, this limits the power supply rejection ratio (PSRR) from the VNEG supply to a lower value than that from the VPOS supply. If given a choice, an application board should be designed such that the VNEG power is supplied from a low inductance plane, subject to a least amount of noise. The VOCM should be considered a reference pin and not a power supply. It is an input to the high speed, high gain common-mode control loop of all receivers and output drivers. In the single-ended output sense, there is no rejection from noise on the VOCM net to the output. For this reason, care must be taken to produce a low noise VOCM source over the entire range of frequencies of interest. This is not only important to single-ended operation, but to differential Rev. 0 | Page 30 of 36 AD8104/AD8105 VPOS operation as well, as there is a common-mode-to-differential gain conversion that becomes greater at higher frequencies. IOUTPUT, QUIESCENT During operation of the AD8104/AD8105, transient currents flow into the VOCM net from the amplifier control loops. Although the magnitude of these currents are small (10 A to 20 A per output), they can contribute to crosstalk if they flow through significant impedances. Driving VOCM with a low impedance, low noise source is desirable. QNPN QPNP IOUTPUT VNEG Figure 71. Simplified Output Stage Example 8 MAXIMUM POWER (W) TJ = 150C For the AD8104/AD8105, in an ambient temperature of 85C, with all 16 outputs driving 1 V rms into 100 loads and power supplies at 2.5 V, follow these steps: 7 1. 6 Calculate power dissipation of AD8104/AD8105 using data sheet quiescent currents. Disregard VDD current, as it is insignificant. PD , QUIESCENT = (VPOS x I VPOS ) + (VNEG x I VNEG ) PD ,QUIESCENT = (2.5 Vx 340 mA ) + (2.5 Vx 340 mA ) = 1.7 W 06612-070 5 25 35 45 55 65 75 2. 85 AMBIENT TEMPERATURE (C) JA PD ,OUTPUT = (2.5 V - 1 V ) x (1 V / 100 ) = 15 mW (1) As an example, if the AD8104/AD8105 is enclosed in an environment at 45C (TA), the total on-chip dissipation under all load and supply conditions must not be allowed to exceed 7.0 W. There are 16 output pairs, or 32 output currents. nPD ,OUTPUT = 32 x 15 mW = 0.48 W 3. When calculating on-chip power dissipation, it is necessary to include the rms current being delivered to the load, multiplied by the rms voltage drop on the AD8104/AD8105 output devices. For a sinusoidal output, the on-chip power dissipation due to the load can be approximated by ( ) PD , OUTPUT = VPOS - VOUTPUT , RMS x I OUTPUT , RMS The curve in Figure 70 was calculated from TJUNCTION , MAX - TAMBIENT Calculate power dissipation from loads. For a differential output and ground-referenced load, the output power is symmetrical in each output phase. ( Figure 70. Maximum Die Power Dissipation vs. Ambient Temperature PD , MAX = 06612-071 IOUTPUT, QUIESCENT Power Dissipation Calculation of Power Dissipation 4 15 VOUTPUT Subtract the quiescent output stage current for number of loads (32 in this example). The output stage is either standing, or driving a load, but the current only needs to be counted once (valid for output voltages > 0.5 V). PDQ , OUTPUT = (VPOS - VNEG ) x I OUTPUT , QUIESCENT PDQ , OUTPUT = (2.5 V - (-2.5 V)) x 1.65 mA = 8.25 mW ) There are 16 output pairs, or 32 output currents. PD , OUTPUT = VPOS - VO UTPUT , RMS x I OUTPUT , RMS For nonsinusoidal output, the power dissipation should be calculated by integrating the on-chip voltage drop multiplied by the load current over one period. The user can subtract the quiescent current for the Class AB output stage when calculating the loaded power dissipation. For each output stage driving a load, subtract a quiescent power according to PDQ , OUTPUT = (VPOS - VNEG ) x I OUTPUT , QUIESCENT where IOUTPUT, QUIESCENT = 1.65 mA for each single-ended output pin. nPDQ ,OUTPUT = 32 x 8.25 mW = 0.26 W 4. Verify that the power dissipation does not exceed the maximum allowed value. PD , ON - CHIP = PD , QUIESCENT + nPD , OUTPUT - nPDQ , OUTPUT PD ,ON -CHIP = 1.7 W + 0.48 W - 0.26 W = 1.9 W From Figure 70 or Equation 1, this power dissipation is below the maximum allowed dissipation for all ambient temperatures up to and including 85C. For each disabled output, the quiescent power supply current in VPOS and VNEG drops by approximately 9 mA. Rev. 0 | Page 31 of 36 AD8104/AD8105 Short-Circuit Output Conditions Although there is short-circuit current protection on the AD8104/AD8105 outputs, the output current can reach values of 80 mA into a grounded output. Any sustained operation with too many shorted outputs can exceed the maximum die temperature and can result in device failure (see the Absolute Maximum Ratings section). Crosstalk Many systems, such as broadcast video and KVM switches, that handle numerous analog signal channels, have strict requirements for keeping the various signals from influencing any of the others in the system. Crosstalk is the term used to describe the coupling of the signals of other nearby channels to a given channel. When there are many signals in close proximity in a system, as is undoubtedly the case in a system that uses the AD8104/AD8105, the crosstalk issues can be quite complex. A good understanding of the nature of crosstalk and some definition of terms is required in order to specify a system that uses one or more crosspoint devices. Types of Crosstalk Crosstalk can be propagated by means of any of three methods. These fall into the categories of electric field, magnetic field, and sharing of common impedances. This section explains these effects. Every conductor can be both a radiator of electric fields and a receiver of electric fields. The electric field crosstalk mechanism occurs when the electric field created by the transmitter propagates across a stray capacitance (for example, free space), couples with the receiver, and induces a voltage. This voltage is an unwanted crosstalk signal in any channel that receives it. Currents flowing in conductors create magnetic fields that circulate around the currents. These magnetic fields then generate voltages in any other conductors whose paths they link. The undesired induced voltages in these other channels are crosstalk signals. The channels that crosstalk can be said to have a mutual inductance that couples signals from one channel to another. are common mode to the signal and can be rejected by a differential receiver. Areas of Crosstalk A practical AD8104/AD8105 circuit must be mounted to some sort of circuit board in order to connect it to power supplies and measurement equipment. Great care has been taken to create an evaluation board that adds minimum crosstalk to the intrinsic device. This, however, raises the issue that a system's crosstalk is a combination of the intrinsic crosstalk of the devices in addition to the circuit board to which they are mounted. It is important to try to separate these two areas when attempting to minimize the effect of crosstalk. In addition, crosstalk can occur among the inputs to a crosspoint and among the outputs. It can also occur from input to output. Techniques are discussed in the following sections for diagnosing which part of a system is contributing to crosstalk. Measuring Crosstalk Crosstalk is measured by applying a signal to one or more channels and measuring the relative strength of that signal on a desired selected channel. The measurement is usually expressed as dB down from the magnitude of the test signal. The crosstalk is expressed by A (s ) XT = 20 log 10 SEL A ( ) s TEST where: s = j, the Laplace transform variable. ASEL(s) is the amplitude of the crosstalk induced signal in the selected channel. ATEST(s) is the amplitude of the test signal. It can be seen that crosstalk is a function of frequency, but not a function of the magnitude of the test signal (to first order). In addition, the crosstalk signal will have a phase relative to the test signal associated with it. A network analyzer is most commonly used to measure crosstalk over a frequency range of interest. It can provide both magnitude and phase information about the crosstalk signal. The power supplies, grounds, and other signal return paths of a multichannel system are generally shared by the various channels. When a current from one channel flows in one of these paths, a voltage that is developed across the impedance becomes an input crosstalk signal for other channels that share the common impedance. As a crosspoint system or device grows larger, the number of theoretical crosstalk combinations and permutations can become extremely large. For example, in the case of the 32 x 16 matrix of the AD8104/AD8105, look at the number of crosstalk terms that can be considered for a single channel, for example, the input IN00. IN00 is programmed to connect to one of the AD8104/AD8105 outputs where the measurement can be made. All these sources of crosstalk are vector quantities; therefore, the magnitudes cannot simply be added together to obtain the total crosstalk. In fact, there are conditions where driving additional circuits in parallel in a given configuration can actually reduce the crosstalk. Because the AD8104/AD8105 are fully differential designs, many sources of crosstalk either destructively cancel, or First, the crosstalk terms associated with driving a test signal into each of the other 31 inputs can be measured one at a time, while applying no signal to IN00. Then the crosstalk terms associated with driving a parallel test signal into all 31 other inputs can be measured two at a time in all possible combinations, then three at a time, and so on, until, finally, Rev. 0 | Page 32 of 36 AD8104/AD8105 there is only one way to drive a test signal into all 31 other inputs in parallel. Each of these cases is legitimately different from the others and may yield a unique value, depending on the resolution of the measurement system, but it is hardly practical to measure all these terms and then specify them. In addition, this describes the crosstalk matrix for just one input channel. A similar crosstalk matrix can be proposed for every other input. In addition, if the possible combinations and permutations for connecting inputs to the other outputs (not used for measurement) are taken into consideration, the numbers rather quickly grow to astronomical proportions. If a larger crosspoint array of multiple AD8104/AD8105s is constructed, the numbers grow larger still. Obviously, some subset of all these cases must be selected to be used as a guide for a practical measure of crosstalk. One common method is to measure all-hostile crosstalk; this means that the crosstalk to the selected channel is measured while all other system channels are driven in parallel. In general, this yields the worst crosstalk number, but this is not always the case, due to the vector nature of the crosstalk signal. Other useful crosstalk measurements are those created by one nearest neighbor or by the two nearest neighbors on either side. These crosstalk measurements are generally higher than those of more distant channels, so they can serve as a worst-case measure for any other one-channel or two-channel crosstalk measurements. Input and Output Crosstalk Capacitive coupling is voltage-driven (dV/dt), but is generally a constant ratio. Capacitive crosstalk is proportional to input or output voltage, but this ratio is not reduced by simply reducing signal swings. Attenuation factors must be changed by changing impedances (lowering mutual capacitance), or destructive canceling must be utilized by summing equal and out of phase components. For high input impedance devices such as the AD8104/AD8105, capacitances generally dominate inputgenerated crosstalk. Inductive coupling is proportional to current (dI/dt), and often scales as a constant ratio with signal voltage, but also shows a dependence on impedances (load current). Inductive coupling can also be reduced by constructive canceling of equal and out of phase fields. In the case of driving low impedance video loads, output inductances contribute highly to output crosstalk. The flexible programming capability of the AD8104/AD8105 can be used to diagnose whether crosstalk is occurring more on the input side or the output side. Some examples are illustrative. A given input pair (IN07 in the middle for this example) can be programmed to drive OUT07 (also in the middle). The inputs to IN07 are just terminated to ground (via 50 or 75 ) and no signal is applied. All the other inputs are driven in parallel with the same test signal (practically provided by a distribution amplifier), with all other outputs except OUT07 disabled. Since grounded IN07 is programmed to drive OUT07, no signal should be present. Any signal that is present can be attributed to the other 31 hostile input signals, because no other outputs are driven (they are all disabled). Thus, this method measures the all hostile input contribution to crosstalk into IN07. Of course, the method can be used for other input channels and combinations of hostile inputs. For output crosstalk measurement, a single input channel is driven (IN00, for example) and all outputs other than a given output (IN07 in the middle) are programmed to connect to IN00. OUT07 is programmed to connect to IN15 (far away from IN00), which is terminated to ground. Thus OUT07 should not have a signal present since it is listening to a quiet input. Any signal measured at the OUT07 can be attributed to the output crosstalk of the other 16 hostile outputs. Again, this method can be modified to measure other channels and other crosspoint matrix combinations. Effect of Impedances on Crosstalk The input side crosstalk can be influenced by the output impedance of the sources that drive the inputs. The lower the impedance of the drive source, the lower the magnitude of the crosstalk. The dominant crosstalk mechanism on the input side is capacitive coupling. The high impedance inputs do not have significant current flow to create magnetically induced crosstalk. However, significant current can flow through the input termination resistors and the loops that drive them. Thus, the PC board on the input side can contribute to magnetically coupled crosstalk. From a circuit standpoint, the input crosstalk mechanism looks like a capacitor coupling to a resistive load. For low frequencies, the magnitude of the crosstalk is given by XT = 20 log 10 [(RS C M ) x s ] where: RS is the source resistance. CM is the mutual capacitance between the test signal circuit and the selected circuit. s is the Laplace transform variable. From the preceding equation, it can be observed that this crosstalk mechanism has a high-pass nature; it can also be minimized by reducing the coupling capacitance of the input circuits and lowering the output impedance of the drivers. If the input is driven from a 75 terminated cable, the input crosstalk can be reduced by buffering this signal with a low output impedance buffer. Rev. 0 | Page 33 of 36 AD8104/AD8105 From a circuit standpoint, this output crosstalk mechanism looks like a transformer with a mutual inductance between the windings that drive a load resistor. For low frequencies, the magnitude of the crosstalk is given by s XT = 20 log 10 M XY x RL where: MXY is the mutual inductance of Output X to Output Y. RL is the load resistance on the measured output. This crosstalk mechanism can be minimized by keeping the mutual inductance low and increasing RL. The mutual inductance can be kept low by increasing the spacing of the conductors and minimizing their parallel length. PCB Layout Extreme care must be exercised to minimize additional crosstalk generated by the system circuit board(s). The areas that must be carefully detailed are grounding, shielding, signal routing, and supply bypassing. The packaging of the AD8104/AD8105 is designed to help keep the crosstalk to a minimum. On the BGA substrate, each pair is carefully routed to predominately couple to each other, with shielding traces separating adjacent signal pairs. The ball grid array is arranged such that similar board routing can be achieved. Only the outer two rows are used for signals, such that vias can be used to take the input rows to a lower signal plane if desired. The input and output signals have minimum crosstalk if they are located between ground planes on layers above and below, and separated by ground in between. Vias should be located as close to the IC as possible to carry the inputs and outputs to the inner layer. The input and output signals surface at the input termination resistors and the output series back-termination resistors. To the extent possible, these signals should also be separated as soon as they emerge from the IC package. applications are generally 50 single-ended (and board manufacturers have the most experience with this application). CAT-5 cabling is usually driven as differential pairs of 100 differential impedance. For flexibility, the AD8104/AD8105 do not contain on-chip termination resistors. This flexibility in application comes with some board layout challenges. The distance between the termination of the input transmission line and the AD8104/AD8105 die is a high impedance stub, and causes reflections of the input signal. With some simplification, it can be shown that these reflections cause peaking of the input at regular intervals in frequency, dependent on the propagation speed (VP) of the signal in the chosen board material and the distance (d) between the termination resistor and the AD8104/AD8105. If the distance is great enough, these peaks can occur in-band. In fact, practical experience shows that these peaks are not high-Q, and should be pushed out to three or four times the desired bandwidth in order to not have an effect on the signal. For a board designer using FR4 (VP = 144 x 106 m/s), this means the AD8104/AD8105 input should be placed no farther than 1.5 cm after the termination resistors, and preferably should be placed even closer. The BGA substrate routing inside the AD8104/ AD8105 is approximately 1 cm in length and adds to the stub length, so 1.5 cm PCB routing equates to d = 2.5 x 10-2 m in the calculations. f PEAK = (2n + 1) x VP 4d where n = {0, 1, 2, 3, ...}. In some cases, it is difficult to place the termination close to the AD8104/AD8105 due to space constraints, differential routing, and large resistor footprints. A preferable solution in this case is to maintain a controlled transmission line past the AD8104/ AD8105 inputs and terminate the end of the line. This is known as fly-by termination. The input impedance of the AD8104/ AD8105 is large enough and stub length inside the package is small enough that this works well in practice. Implementation of fly-by input termination often includes bringing the signal in on one routing layer, then passing through a filled via under the AD8104/AD8105 input ball, then back out to termination on another signal layer. In this case, care must be taken to tie the reference ground planes together near the signal via if the signal layers are referenced to different ground planes. PCB Termination Layout As frequencies of operation increase, the importance of proper transmission line signal routing becomes more important. The bandwidth of the AD8104/AD8105 is large enough that using high impedance routing does not provide a flat in-band frequency response for practical signal trace lengths. It is necessary for the user to choose a characteristic impedance suitable for the application and properly terminate the input and output signals of the AD8104/AD8105. Traditionally, video applications have used 75 single-ended environments. RF IPn INn 75 AD8104/ AD8105 OPn ONn 06612-072 On the output side, the crosstalk can be reduced by driving a lighter load. Although the AD8104/AD8105 are specified with excellent differential gain and phase when driving a standard 150 video load, the crosstalk is higher than the minimum obtainable due to the high output currents. These currents induce crosstalk via the mutual inductance of the output pins and bond wires of the AD8104/AD8105. Figure 72. Fly-By Input Termination, Grounds for the Two Transmission Lines Shown Must be Tied Together Close to the INn Pin Rev. 0 | Page 34 of 36 AD8104/AD8105 If multiple AD8104/AD8105s are to be driven in parallel, a flyby input termination scheme is very useful, but the distance from each AD8104/AD8105 input to the driven input transmission line is a stub that should be minimized in length and parasitics using the discussed guidelines. system). This is not often practical as trace widths become large. In most cases, the best practical solution is to place the halfcharacteristic impedance resistor as close as possible (preferably less than 1.5 cm away) and to reduce the parasitics of the stub (by removing the ground plane under the stub, for example). In either case, the designer must decide if the layout complexity created by a balanced, terminated solution is preferable to simply grounding the undriven input at the ball with no trace. When driving the AD8104/AD8105 single-endedly, the undriven input is often terminated with a resistance to balance the input stage. It can be seen that by terminating the undriven input with a resistor of one half the characteristic impedance, the input stage is perfectly balanced (37.5 , for example, to balance the two parallel 75 terminations on the driven input). However, due to the feedback in the input receiver, there is high speed signal current leaving the undriven input. To terminate this high speed signal, proper transmission line techniques should be used. One solution is to adjust the trace width to create a transmission line of half the characteristic impedance and terminate the far end with this resistance (37.5 in a 75 Although the examples discussed so far are for input termination, the theory is similar for output back-termination. Taking the AD8104/AD8105 as an ideal voltage source, any distance of routing between the AD8104/AD8105 and a back-termination resistor will be an impedance mismatch that potentially creates reflections. For this reason, back-termination resistors should also be placed close to the AD8104/AD8105. In practice, because back-termination resistors are series elements, they can be placed close to the AD8104/AD8105 outputs. VPOS VDD J3 PC_VDD PLD_VDD VPOS VDD ANALOG SMA 50 IN[31:0], IP[31:0] PC PARALLEL PORT IN[31:0], IP[31:0] 50 LOGIC ISOLATORS CLK RESET WE UPDATE DATA IN DATA OUT D0 TO D5 A0 TO A3 CPLD ON[15:0], OP[15:0] AD8104/ AD8105 SMA ON[15:0], OP[15:0] VOCM LOGIC DGND VNEG J8, W3 TO W7 VNEG GND Figure 73. Evaluation Board Simplified Schematic Rev. 0 | Page 35 of 36 06612-073 PC_GND AD8104/AD8105 OUTLINE DIMENSIONS A1 CORNER INDEX AREA 31.00 BSC SQ 23 6 4 2 22 20 18 16 14 12 10 8 21 19 17 15 13 11 9 5 3 7 1 A B C D E F G H J K L M N P R T U V W Y AA AB AC BALL A1 INDICATOR 27.94 BSC SQ TOP VIEW BOTTOM VIEW 1.27 BSC DETAIL A DETAIL A *1.765 MAX 1.07 0.99 0.92 0.70 0.63 0.56 0.10 MIN COPLANARITY 0.20 SEATING PLANE *COMPLIANT TO JEDEC STANDARDS MO-192-BAN-2 WITH THE EXCEPTION TO PACKAGE HEIGHT. 022206-A 0.25 MIN (4 ) 0.90 0.75 0.60 BALL DIAMETER Figure 74. 304-Ball Ball Grid Array, Thermally Enhanced [BGA_ED] (BP-304) Dimensions shown in millimeters ORDERING GUIDE Model AD8104ABPZ 1 AD8105ABPZ1 AD8104-EVAL AD8105-EVAL 1 Temperature Range -40C to +85C -40C to +85C Package Description 304-Ball Ball Grid Array Package, Thermally Enhanced [BGA_ED] 304-Ball Ball Grid Array Package, Thermally Enhanced [BGA_ED] Evaluation Board Evaluation Board Z = RoHS Compliant Part. (c)2007 Analog Devices, Inc. All rights reserved. Trademarks and registered trademarks are the property of their respective owners. D06612-0-6/07(0) Rev. 0 | Page 36 of 36 Package Option BP-304 BP-304