View AD8104_7813315.PDF datasheet online --- IC-ON-LINE

Datasheet File OCR Text:

600 mhz, 32 16 buffered analog crosspoint switch AD8104/ad8105 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 ?2007 analog devices, inc. all rights reserved. features high channel count, 32 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 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 (r l = 150 ) 0.05 phase error (r l = 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 31 mm) applications routing of high speed signals including rgb and component video routing kvm compressed video (mpeg, wavelet) data communications functional block diagram 2 2 input receiver g = +1* g = +2** output buffer g = +1 16 512 set individual, or reset all outputs to off enable/disable 96 vpos vneg vocm 32 input pairs reset update clk data in we ser/par data out 16 output pairs switch matrix 96 d0 d1 d2 d3 d4 d5 v dd dgnd a0 a1 a2 a3 AD8104/ ad8105 1 0 192-bit shift register with 6-bit parallel loading parallel latch decode 16 6:32 decoders *AD8104 only **ad8105 only 0 6612-001 96 no connect figure 1. general description the AD8104/ad8105 are high speed, 32 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 isola- tion (@ 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, 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 reprogram- ming 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.
AD8104/ad8105 rev. 0 | page 2 of 36 table of contents features .............................................................................................. 1 applications....................................................................................... 1 functional block diagram .............................................................. 1 general description ......................................................................... 1 revision history ............................................................................... 2 specifications..................................................................................... 3 timing characteristics (serial mode) ....................................... 5 timing characteristics (parallel mode) .................................... 6 absolute maximum ratings............................................................ 7 thermal resistance ...................................................................... 7 power dissipation......................................................................... 7 esd caution...................................................................................7 pin configuration and function descriptions..............................8 truth table and logic diagram ............................................... 13 i/o schematics................................................................................ 15 typical performance characteristics ........................................... 17 theory of operation ...................................................................... 25 applications information .............................................................. 26 programming.............................................................................. 26 operating modes........................................................................ 27 outline dimensions ....................................................................... 36 ordering guide .......................................................................... 36 revision history 6/07revision 0: initial version
AD8104/ad8105 rev. 0 | page 3 of 36 specifications v s = 2.5 v at t a = 25c, r l, diff = 200 , v ocm = 0 v, differential i/o mode, unless otherwise noted. table 1. AD8104/ad8105 parameter conditions min typ max unit dynamic performance ?3 db bandwidth 200 mv p-p, typical channel 600 mhz 2 v p-p, typical channel 420/525 mhz gain flatness 0.1 db, 200 mv p-p 100/50 mhz 0.1 db, 2 v p-p 70/50 mhz propagation delay 2 v p-p 1.3 ns settling time 1%, 2 v step 2.5 ns slew rate 2 v step, peak 1800 v/s 2 v step, 10% to 90% 1500 v/s noise/distortion performance differential gain error ntsc or pal, r l = 150 0.05 % differential phase error ntsc or pal, r l = 150 0.05 degrees crosstalk, all hostile f = 5 mhz ?80/?70 db f = 10 mhz ?72/?68 db f = 100 mhz ?48/?50 db f = 600 mhz ?40/?50 db off isolation, input-to-output f = 10 mhz, one channel ?90 db input voltage noise 0.1 mhz to 50 mhz 45/53 nv/hz dc performance voltage gain differential +1/+2 v/v gain error 1 % no load 1 3 % gain matching channel-to-channel 1 % differential offset 5 25 mv common-mode offset 25 90 mv output characteristics output impedance dc, enabled 0.1 disabled, differential 30 k output disable capacitance disabled 4 pf output leakage current disabled 1 a output voltage range no load 2.8 3.8 v p-p v ocm input range v out, diff = 2 v p-p ?0.5 +0.8 v v out, diff = 2.8 v p-p ?0.25 +0.6 v output swing limit single-ended output ?1.3 +1.3 v output current maximum operating signal 30 ma input characteristics input voltage range common mode, v in, diff = 2 v p-p ?2 +2 v differential 2/1 v common-mode rejection ratio f = 10 mhz 48 db input capacitance any switch configuration 2 pf input resistance differential 5 k input offset current 1 a v ocm input bias current 64 a v ocm input impedance 4 k
AD8104/ad8105 rev. 0 | page 4 of 36 AD8104/ad8105 parameter conditions min typ max unit switching characteristics enable on time 50% update to 1% settling 100 ns switching time, 2 v step 50% update to 1% settling 100 ns switching transient (glitch) differential 40 mv p-p power supplies supply current vpos, outputs enabled, no load 340 420 ma vpos, outputs disabled 210 240 ma vneg, outputs enabled, no load 340 420 ma vneg, outputs disabled 210 240 ma vdd, outputs enabled, no load 1.2 ma supply voltage range 4.5 to 5.5 v psrr vneg, vpos, f = 1 mhz 85 db vocm, f = 1 mhz 75 db operating temperature range temperature range operating (still air) ?40 to +85 c ja operating (still air) 14 c/w jc operating (still air) 1 c/w
AD8104/ad8105 rev. 0 | page 5 of 36 timing characteristics (serial mode) specifications subject to change without notice. table 2. limit parameter symbol min typ max unit serial data setup time t 1 40 ns clk pulse width t 2 50 ns serial data hold time t 3 50 ns clk pulse separation t 4 150 ns clk to update delay t 5 10 ns update pulse width t 6 90 ns clk to data out valid t 7 120 ns propagation delay, update to switch on or off 100 ns reset pulse width 60 ns reset time 200 ns load data into serial register on falling edge t 2 t 4 1 0 clk 1 0 1 0 data in out15 (d5) t 1 t 3 out15 (d4) out0 (d0) 1 = latched 0 = transparent update transfer data from serial register to parallel latches during low level t 5 t 7 data out t 6 we 1 0 06612-002 figure 2. timing diagram, serial mode table 3. logic levels v ih v il v oh v ol i ih i il i oh i ol reset , ser /par, clk , data in, update reset , ser /par, clk , data in, update data out data out reset 1 , ser /par, clk , data in, update reset 1 , ser /par, clk , data in, update data out data out 2.0 v min 0.6 v max vdd ? 0.3 v min dgnd + 0.5 v max 1 a max C1 a min ?1 ma max 1 ma min 1 see figure 15.
AD8104/ad8105 rev. 0 | page 6 of 36 timing characteristics (parallel mode) specifications subject to change without notice. table 4. limit parameter symbol min typ max unit parallel data setup time t 1 80 ns we pulse width t 2 110 ns parallel data hold time t 3 150 ns we pulse separation t 4 90 ns we to update delay t 5 10 ns update pulse width t 6 90 ns propagation delay, update to switch on or off 100 ns reset pulse width 60 ns reset time 200 ns t 2 t 4 1 0 we 1 0 t 1 t 3 1 = latched 0 = transparent update t 6 d0 to d5 a0 to a3 t 5 06612-003 figure 3. timing diagram, parallel mode table 5. logic levels v ih v il v oh v ol i ih i il i oh i ol reset , ser /par, we , d0, d1, d2, d3, d4, d5, a0, a1, a2, a3, update reset , ser /par, we , d0, d1, d2, d3, d4, d5, a0, a1, a2, a3, update data out data out reset 1 , ser /par, we, d0, d1, d2, d3, d4, d5, a0, a1, a2, a3, update reset 1 , ser /par, we , d0, d1, d2, d3, d4, d5, a0, a1, a2, a3, update data out data out 2.0 v min 0.6 v max disabled disabled 1 a max C1 a min disabled disabled 1 see figure 15.
AD8104/ad8105 rev. 0 | page 7 of 36 absolute maximum ratings table 6. parameter rating analog supply voltage (vpos C vneg) 6 v digital supply voltage (vdd C dgnd) 6 v ground potential difference (vneg C dgnd) +0.5 v to ?2.5 v maximum potential difference (vdd C vneg) 8 v common-mode analog input voltage vneg to vpos differential analog input voltage 2 v digital input voltage vdd output voltage (disabled analog output) (vpos ? 1 v) to (vneg + 1 v) output short-circuit duration momentary output short-circuit current 80 ma storage temperature range ?65c to +125c operating temperature range ?40c to +85c lead temperature (soldering, 10 sec) 300c junction temperature 150c stresses above those listed under absolute maximum ratings may cause permanent damage to the device. this is a stress rating only; functional operation of the device at these or any other conditions above those indicated in the operational section of this specification is not implied. exposure to absolute maximum rating conditions for extended periods may affect device reliability. thermal resistance ja is specified for the worst-case conditions, that is, a device soldered in a circuit board for surface-mount packages. table 7. thermal resistance package type ja jc jb jt jb unit 304-ball bga 14 1 6.5 0.6 5.7 c/w power dissipation 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. packaged in a 304-ball bga, the AD8104/ad8105 junction-to- ambient thermal impedance ( ja ) is 14c/w. for long-term reliability, the maximum allowed junction temperature of the die should not exceed 150c. te mporarily 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 4 15 85 maximum power (w) ambient temperature (c) t j = 150c 7 6 5 25 35 45 55 65 75 06612-004 figure 4. maximum die power dissipation vs. ambient temperature esd caution
AD8104/ad8105 rev. 0 | page 8 of 36 pin configuration and fu nction descriptions 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 a nc nc nc nc nc nc nc nc nc a b nc nc nc nc nc nc nc nc nc b c vneg vpos vpos vpos vneg vneg vneg vneg vneg c d vneg vpos vpos vpos vneg vneg vneg vneg vneg d e e f f g g h h j j k k l l m m n n p p r r t t u u v v w w y vneg vpos vpos vpos vneg vneg vneg vneg vneg y aa vneg vpos vpos vpos vneg vneg vneg vneg vneg aa a b op10 on8 op8 on6 op6 on4 op4 on2 op2 ab a c vpos vpos vpos in16 ip16 in18 ip18 in20 ip20 in22 ip22 in24 ip24 in26 ip26 in28 ip28 in30 ip30 vpos vpos vpos vpos vpos vpos vpos vpos in17 ip17 in19 ip19 in21 ip21 in23 ip23 in25 ip25 in27 ip27 in29 ip29 in31 ip31 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 nc vpos vneg vocm vdd dgnd reset we d5 d4 d3 d2 d1 d0 vdd dgnd vocm vneg vpos vpos on15 nc nc vneg vocm vocm vneg on14 op15 nc nc vneg vneg vneg vneg op14 on13 nc nc vneg vneg vneg vneg on12 op13 nc nc vneg vneg vneg vneg op12 on11 nc nc vneg vneg vneg vneg on10 op11 on9 op9 on7 op7 on5 op5 on3 op3 on1 nc nc vneg vocm vocm vneg on0 op1 nc vpos vpos vneg vocm vdd dgnd data out clk data in ser/ par dgnd a3 a2 a1 a0 vdd dgnd vocm vneg vpos op0 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 ip0 in0 ip2 in2 ip4 in4 ip6 in6 ip8 in8 ip10 in10 ip12 in12 ip14 in14 vpos vpos vpos vpos vpos vpos vpos vpos ip1 in1 ip3 in3 ip5 in5 ip7 in7 ip9 in9 ip11 in11 ip13 in13 ip15 in15 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) 0 6612-005 figure 5. package bottom view
AD8104/ad8105 rev. 0 | page 9 of 36 21 22 23 20 1918 17 1615 14 13121110 987654321 a nc nc nc nc nc nc nc nc nc a b nc nc nc nc nc nc nc nc nc b c vneg vpos vpos vpos vneg vneg vneg vneg vneg c d vneg vpos vpos vpos vneg vneg vneg vneg vneg d e e f f g g h h j j k k l l m m n n p p r r t t u u v v w w y vneg vpos vpos vpos vneg vneg vneg vneg vneg y aa vneg vpos vpos vpos vneg vneg vneg vneg vneg aa ab op10 on8 op8 on6 op6 on4 op4 on2 op2 ab ac vpos vpos vpos in16 ip16 in18 ip18 in20 ip20 in22 ip22 in24 ip24 in26 ip26 in28 ip28 in30 ip30 vpos vpos vpos vpos vpos vpos vpos vpos in17 ip17 in19 ip19 in21 ip21 in23 ip23 in25 ip25 in27 ip27 in29 ip29 in31 ip31 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 nc vpos vneg vocm vdd dgnd reset we d5 d4 d3 d2 d1 d0 vdd dgnd vocm vneg vpos vpos on15 nc nc vneg vocm vocm vneg on14 op15 nc nc vneg vneg vneg vneg op14 on13 nc nc vneg vneg vneg vneg on12 op13 nc nc vneg vneg vneg vneg op12 on11 nc nc vneg vneg vneg vneg on10 op11 on9 op9 on7 op7 on5 op5 on3 op3 on1 nc nc vneg vocm vocm vneg on0 op1 nc vpos vpos vneg vocm vdd dgnd data out clk data in ser/ par dgnd a3 a2 a1 a0 vdd dgnd vocm vneg vpos op0 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 ip0 in0 ip2 in2 ip4 in4 ip6 in6 ip8 in8 ip10 in10 ip12 in12 ip14 in14 vpos vpos vpos vpos vpos vpos vpos vpos ip1 in1 ip3 in3 ip5 in5 ip7 in7 ip9 in9 ip11 in11 ip13 in13 ip15 in15 vpos vpos vpos ac 2322212019181716151413121110 987654321 AD8104/ad8105 top view (not to scale) 0 6612-006 figure 6. package top view table 8. ball grid description ball no. mnemonic description a1 vpos analog positive power supply. a2 vpos analog positive power supply. a3 vpos analog positive power supply. a4 nc no connect. a5 nc no connect. a6 nc no connect. a7 nc no connect. a8 nc no connect. a9 nc no connect. a10 nc no connect. a11 nc no connect. a12 nc no connect. a13 nc no connect. a14 nc no connect. ball no. mnemonic description a15 nc no connect. a16 nc no connect. a17 nc no connect. a18 nc no connect. a19 nc no connect. a20 vpos analog positive power supply. a21 vpos analog positive power supply. a22 vpos analog positive power supply. a23 vpos analog positive power supply. b1 vpos analog positive power supply. b2 vpos analog positive power supply. b3 vpos analog positive power supply. b4 vpos analog positive power supply. b5 nc no connect.
AD8104/ad8105 rev. 0 | page 10 of 36 ball no. mnemonic description b6 nc no connect. b7 nc no connect. b8 nc no connect. b9 nc no connect. b10 nc no connect. b11 nc no connect. b12 nc no connect. b13 nc no connect. b14 nc no connect. b15 nc no connect. b16 nc no connect. b17 nc no connect. b18 nc no connect. b19 nc no connect. b20 nc no connect. b21 vpos analog positive power supply. b22 vpos analog positive power supply. b23 vpos analog positive power supply. c1 vpos analog positive power supply. c2 vpos analog positive power supply. c3 vpos analog positive power supply. c4 vpos analog positive power supply. c5 vneg analog negative power supply. c6 vneg analog negative power supply. c7 vneg analog negative power supply. c8 vneg analog negative power supply. c9 vneg analog negative power supply. c10 vneg analog negative power supply. c11 vpos analog positive power supply. c12 vpos analog positive power supply. c13 vpos analog positive power supply. c14 vneg analog negative power supply. c15 vneg analog negative power supply. c16 vneg analog negative power supply. c17 vneg analog negative power supply. c18 vneg analog negative power supply. c19 vneg analog negative power supply. c20 vpos analog positive power supply. c21 vpos analog positive power supply. c22 vpos analog positive power supply. c23 vpos analog positive power supply. d1 vpos analog positive power supply. d2 ip0 input number 0, positive phase. d3 vpos analog positive power supply. d4 vneg analog negative power supply. d5 vocm output common-mode reference supply. d6 vneg analog negative power supply. d7 vneg analog negative power supply. d8 vneg analog negative power supply. d9 vneg analog negative power supply. d10 vneg analog negative power supply. d11 vpos analog positive power supply. ball no. mnemonic description d12 vpos analog positive power supply. d13 vpos analog positive power supply. d14 vneg analog negative power supply. d15 vneg analog negative power supply. d16 vneg analog negative power supply. d17 vneg analog negative power supply. d18 vneg analog negative power supply. d19 vocm output common-mode reference supply. d20 vneg analog negative power supply. d21 vpos analog positive power supply. d22 vpos analog positive power supply. d23 in16 input number 16, negative phase. e1 ip1 input number 1, positive phase. e2 in0 input number 0, negative phase. e3 vneg analog negative power supply. e4 vocm output common-mode reference supply. e20 vocm output common-mode reference supply. e21 vneg analog negative power supply. e22 in17 input number 17, negative phase. e23 ip16 input number 16, positive phase. f1 in1 input number 1, negative phase. f2 ip2 input number 2, positive phase. f3 vneg analog negative power supply. f4 vdd logic positive power supply. f20 vdd logic positive power supply. f21 vneg analog negative power supply. f22 ip17 input number 17, positive phase. f23 in18 input number 18, negative phase. g1 ip3 input number 3, positive phase. g2 in2 input number 2, negative phase. g3 vneg analog negative power supply. g4 dgnd logic negative power supply. g20 dgnd logic negative power supply. g21 vneg analog negative power supply. g22 in19 input number 19, negative phase. g23 ip18 input number 18, positive phase. h1 in3 input number 3, negative phase. h2 ip4 input number 4, positive phase. h3 vneg analog negative power supply. h4 data out control pin: serial data out. h20 reset control pin: second rank data reset. h21 vneg analog negative power supply. h22 ip19 input number 19, positive phase. h23 in20 input number 20, negative phase. j1 ip5 input number 5, positive phase. j2 in4 input number 4, negative phase. j3 vneg analog negative power supply. j4 clk control pin: serial data clock. j20 update control pin: second rank write strobe.
AD8104/ad8105 rev. 0 | page 11 of 36 ball no. mnemonic description j21 vneg analog negative power supply. j22 in21 input number 21, negative phase. j23 ip20 input number 20, positive phase. k1 in5 input number 5, negative phase. k2 ip6 input number 6, positive phase. k3 vneg analog negative power supply. k4 data in control pin: serial data in. k20 we control pin: first rank write strobe. k21 vneg analog negative power supply. k22 ip21 input number 21, positive phase. k23 in22 input number 22, negative phase. l1 ip7 input number 7, positive phase. l2 in6 input number 6, negative phase. l3 vpos analog positive power supply. l4 ser /par control pin: serial/parallel mode select. l20 d5 control pin: input address bit 5. l21 vpos analog positive power supply. l22 in23 input number 23, negative phase. l23 ip22 input number 22, positive phase. m1 in7 input number 7, negative phase. m2 ip8 input number 8, positive phase. m3 vpos analog positive power supply. m4 dgnd logic negative power supply m20 d4 control pin: input address bit 4. m21 vpos analog positive power supply. m22 ip23 input number 23, positive phase. m23 in24 input number 24, negative phase. n1 ip9 input number 9, positive phase. n2 in8 input number 8, negative phase. n3 vpos analog positive power supply. n4 a3 control pin: output address bit 3. n20 d3 control pin: input address bit 3. n21 vpos analog positive power supply. n22 in25 input number 25, negative phase. n23 ip24 input number 24, positive phase. p1 in9 input number 9, negative phase. p2 ip10 input number 10, positive phase. p3 vneg analog negative power supply. p4 a2 control pin: output address bit 2. p20 d2 control pin: input address bit 2. p21 vneg analog negative power supply. p22 ip25 input number 25, positive phase. p23 in26 input number 26, negative phase. r1 ip11 input number 11, positive phase. r2 in10 input number 10, negative phase. r3 vneg analog negative power supply. r4 a1 control pin: output address bit 1. r20 d1 control pin: input address bit 1. r21 vneg analog negative power supply. r22 in27 input number 27, negative phase. r23 ip26 input number 26, positive phase. t1 in11 input number 11, negative phase. ball no. mnemonic description t2 ip12 input number 12, positive phase. t3 vneg analog negative power supply. t4 a0 control pin: output address bit 0. t20 d0 control pin: input address bit 0. t21 vneg analog negative power supply. t22 ip27 input number 27, positive phase. t23 in28 input number 28, negative phase. u1 ip13 input number 13, positive phase. u2 in12 input number 12, negative phase. u3 vneg analog negative power supply. u4 vdd logic positive power supply. u20 vdd logic positive power supply. u21 vneg analog negative power supply. u22 in29 input number 29, negative phase. u23 ip28 input number 28, positive phase. v1 in13 input number 13, negative phase. v2 ip14 input number 14, positive phase. v3 vneg analog negative power supply. v4 dgnd logic negative power supply. v20 dgnd logic negative power supply. v21 vneg analog negative power supply. v22 ip29 input number 29, positive phase. v23 in30 input number 30, negative phase. w1 ip15 input number 15, positive phase. w2 in14 input number 14, negative phase. w3 vneg analog negative power supply. w4 vocm output common-mode reference supply. w20 vocm output common-mode reference supply. w21 vneg analog negative power supply. w22 in31 input number 31, negative phase. w23 ip30 input number 30, positive phase. y1 in15 input number 15, negative phase. y2 vpos analog positive power supply. y3 vpos analog positive power supply. y4 vneg analog negative power supply. y5 vocm output common-mode reference supply. y6 vneg analog negative power supply. y7 vneg analog negative power supply. y8 vneg analog negative power supply. y9 vneg analog negative power supply. y10 vneg analog negative power supply. y11 vpos analog positive power supply. y12 vpos analog positive power supply. y13 vpos analog positive power supply. y14 vneg analog negative power supply. y15 vneg analog negative power supply. y16 vneg analog negative power supply. y17 vneg analog negative power supply. y18 vneg analog negative power supply.
AD8104/ad8105 rev. 0 | page 12 of 36 ball no. mnemonic description y19 vocm output common-mode reference supply. y20 vneg analog negative power supply. y21 vpos analog positive power supply. y22 ip31 input number 31, positive phase. y23 vpos analog positive power supply. aa1 vpos analog positive power supply. aa2 vpos analog positive power supply. aa3 vpos analog positive power supply. aa4 vpos analog positive power supply. aa5 vneg analog negative power supply. aa6 vneg analog negative power supply. aa7 vneg analog negative power supply. aa8 vneg analog negative power supply. aa9 vneg analog negative power supply. aa10 vneg analog negative power supply. aa11 vpos analog positive power supply. aa12 vpos analog positive power supply. aa13 vpos analog positive power supply. aa14 vneg analog negative power supply. aa15 vneg analog negative power supply. aa16 vneg analog negative power supply. aa17 vneg analog negative power supply. aa18 vneg analog negative power supply. aa19 vneg analog negative power supply. aa20 vpos analog positive power supply. aa21 vpos analog positive power supply. aa22 vpos analog positive power supply. aa23 vpos analog positive power supply. ab1 vpos analog positive power supply. ab2 vpos analog positive power supply. ab3 vpos analog positive power supply. ab4 op0 output number 0, positive phase. ab5 on0 output number 0, negative phase. ab6 op2 output number 2, positive phase. ab7 on2 output number 2, negative phase. ab8 op4 output number 4, positive phase. ab9 on4 output number 4, negative phase. ball no. mnemonic description ab10 op6 output number 6, positive phase. ab11 on6 output number 6, negative phase. ab12 op8 output number 8, positive phase. ab13 on8 output number 8, negative phase. ab14 op10 output number 10, positive phase. ab15 on10 output number 10, negative phase. ab16 op12 output number 12, positive phase. ab17 on12 output number 12, negative phase. ab18 op14 output number 14, positive phase. ab19 on14 output number 14, negative phase. ab20 vpos analog positive power supply. ab21 vpos analog positive power supply. ab22 vpos analog positive power supply. ab23 vpos analog positive power supply. ac1 vpos analog positive power supply. ac2 vpos analog positive power supply. ac3 vpos analog positive power supply. ac4 vpos analog positive power supply. ac5 op1 output number 1, positive phase. ac6 on1 output number 1, negative phase. ac7 op3 output number 3, positive phase. ac8 on3 output number 3, negative phase. ac9 op5 output number 5, positive phase. ac10 on5 output number 5, negative phase. ac11 op7 output number 7, positive phase. ac12 on7 output number 7, negative phase. ac13 op9 output number 9, positive phase. ac14 on9 output number 9, negative phase. ac15 op11 output number 11, positive phase. ac16 on11 output number 11, negative phase. ac17 op13 output number 13, positive phase. ac18 on13 output number 13, negative phase. ac19 op15 output number 15, positive phase. ac20 on15 output number 15, negative phase. ac21 vpos analog positive power supply. ac22 vpos analog positive power supply. ac23 vpos analog positive power supply.
AD8104/ad8105 rev. 0 | page 13 of 36 truth table and logic diagram table 9. operation truth table we update clk data input data output reset ser /par operation/comment x x x x x 0 x asynchronous reset. all outputs are disabled. remainder of logic in 192-bit shift register is unchanged. 1 x data i 1 data i-192 1 0 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. 0 x x d0d5 2 a0a3 3 n/a in parallel mode 1 1 parallel mode. the data on parallel lines d0 to d5 are loaded into the shift register location addressed by a0 to a3. 1 0 x x n/a in parallel mode 1 x switch matrix update. data in the 192-bit shift register transfers into the parallel latches that control the switch array. 1 x x x x 1 1 no change in logic. 1 data i : serial data. 2 d0d5: data bits. 3 a0a3: address bits.
AD8104/ad8105 rev. 0 | page 14 of 36 d1 d0 q s dq clk d0 d1 d2 d3 d4 d5 d1 d0 q s d q clk d1 d0 q s dq clk d1 d0 q s d q d1 d0 q s dq clk d1 d0 q s dq clk update reset a3 a2 a1 a0 d1 d0 q s dq out1 en out0 en out15 en out2 en out3 en out4 en out5 en out6 en out7 en out8 en out9 en out10 en out11 en out12 en out13 en out14 en d1 d0 q s dq clk d1 d0 q s dq clk d1 d0 q s dq clk d1 d0 q s dq clk d1 d0 q s dq clk d1 d 0 q s d q clk decode d1 d0 q s dq clk switch matrix output enable 512 16 d clr q out0 b0 ena d clr q out0 b1 ena d clr q out0 b2 ena d clr q out0 b3 ena d clr q out0 b4 ena d clr q out0 en ena d clr q out1 b0 ena d clr q out14 en ena d clr q out15 b0 ena d clr q out15 b1 ena d clr q out15 b2 ena d clr q out15 b3 ena d clr q out15 b4 ena d clr q out15 en ena data out (serial) parallel data (output enable) ser/par we data in (serial) clk output address 4 to 16 decoder clk clk 0 6612-007 figure 7. logic diagram
AD8104/ad8105 rev. 0 | page 15 of 36 i/o schematics opn, onn 06612-008 figure 8. AD8104/ad8105 enabled output (see also esd protection map, figure 18 ) 0.4pf 30k ? 3.4pf 3.4pf opn onn 06612-009 figure 9. AD8104/ad8105 disabled output (see also esd protection map, figure 18 ) 1.3pf 1.3pf 0.3pf ipn inn 2500 ? 2500 ? 2538 ? 2538 ? 06612-010 figure 10. AD8104 receiver (see also esd protection map, figure 18 ) 2500 ? 5075 ? 0.3pf 1.3pf 1.3pf 2500 ? 5075 ? ipn inn 0 6612-011 figure 11. ad8105 receiver (see also esd protection map, figure 18 ) 1.3pf 1.3pf 0.3pf ip n inn 2500? 2500? 0 6612-012 figure 12. AD8104/ad8105 receiver simplified equivalent circuit when driving differentially 1.6pf ip n inn 3.33k ? AD8104 g = +1 3.76k ? ad8105 g = +2 0 6612-013 figure 13. AD8104/ad8105 receiver simplified equivalent circuit when driving single-ended v ocm vneg 0 6612-014 figure 14. vocm input (see also esd protection map, figure 18 ) reset dgnd v dd 1k? 25k ? 0 6612-015 figure 15. reset input (see also esd protection map, figure 18 )
AD8104/ad8105 rev. 0 | page 16 of 36 v pos vneg ipn, inn, opn, onn, vocm v dd dgnd clk, reset, ser/par, we, update, data in, data out, a[3:0], d[5:0] 06612-018 clk, ser/par, we, update, data in, a[3:0], d[5:0] dgnd 1k? 0 6612-016 figure 16. logic input (see also esd protection map, figure 18 ) figure 18. esd protection map v dd dgnd data out 0 6612-017 figure 17. logic output (see also esd protection map, figure 18 )
AD8104/ad8105 rev. 0 | page 17 of 36 typical performance characteristics v s = 2.5 v at t a = 25c, r l, diff = 200 , v ocm = 0 v, differential i/o mode, unless otherwise noted. 10 ?10 1000 gain (db) frequency (mhz) ?2 0 2 ?8 ?6 ?4 4 6 8 1 10 100 ad8105 AD8104 06612-019 figure 19. AD8104, ad8105 small signal frequency response, 200 mv p-p 10 ?10 1000 gain (db) frequency (mhz) ?2 0 2 ?8 ?6 ?4 4 6 8 1 10 100 ad8105 AD8104 06612-020 figure 20. AD8104, ad8105 large signal frequency response, 2 v p-p 10 ?10 0 1000 normalized gain (db) frequency (mhz) 8 6 4 2 0 ?2 ?4 ?6 ?8 100 10 10pf 5pf 2pf 0pf 06612-021 figure 21. AD8104 small signal frequency response with capacitive loads, 200 mv p-p 200 0 540 06612-022 frequency (mhz) count 180 160 140 120 100 80 60 40 20 560 580 600 620 640 660 680 700 800 figure 22. AD8104 ?3 db bandwidth histogram, one device, all 512 channels 0 ?2.0 normalized bandwidth error (%) number of enabled channels ?1.0 ?1.5 ?0.5 12 864 02 1 14 10 6 06612-023 figure 23. AD8104 bandwidt h error vs. enabled channels 0 ?70 ?60 ?50 ?40 ?30 ?20 ?10 300k 1m 10m 100m 1g 2g cmr (db) frequency (hz) differential out 06612-024 figure 24. AD8104, ad8105 common-mode rejection
AD8104/ad8105 rev. 0 | page 18 of 36 ? 15 ?95 0.1 1000 psr (db) frequency (mhz) ?25 ?35 ?45 ?55 ?65 ?75 ?85 100 10 1 vneg aggressor differential out vpos aggressor vocm aggressor 06612-025 figure 25. AD8104 power supply rejection 10 ?50 0.1 1000 psr (db) frequency (mhz) 5 0 ?5 ?30 ?35 ?40 ?45 100 10 1 ?10 ?15 ?20 ?25 vneg aggressor vpos aggressor vocm aggressor single-ended out 06612-026 figure 26. AD8104 power suppl y rejection, single-ended 140 160 180 0 20 40 60 80 100 120 1k 10k 100k 1m noise spectral density (nv/ hz) frequency (hz) differential out ad8105 AD8104 06612-027 figure 27. AD8104, ad8105 noise spectral density, rto 0 ?100 ?60 ?80 ?40 ?20 300k 1m 10m 100m 1g crosstalk (db) frequency (hz) differential in/out 06612-028 figure 28. AD8104 crosstalk, one adjacent channel 300k 1m 10m 100m 1g crosstalk (db) frequency (hz) ?100 ?80 ?60 ?40 ?20 0 differential in/out 06612-029 figure 29. ad8105 crosstalk, one adjacent channel 0 ?100 ?60 ?80 ?40 ?20 300k 1m 10m 100m 1g crosstalk (db) frequency (hz) single-ended in/out 06612-030 figure 30. AD8104 crosstalk, on e adjacent channel, single-ended
AD8104/ad8105 rev. 0 | page 19 of 36 300k 1m 10m 100m 1g crosstalk (db) frequency (hz) ?100 ?80 ?60 ?40 ?20 0 single-ended in/out 06612-031 figure 31. ad8105 crosstalk, on e adjacent channel, single-ended 0 ?120 ?60 ?80 ?100 ?40 ?20 300k 1m 10m 100m 1g crosstalk (db) frequency (hz) 06612-032 differential in/out figure 32. AD8104 crosstalk, all hostile 300k 1m 10m 100m 1g crosstalk (db) frequency (hz) ?100 ?80 ?60 ?40 ?20 0 differential in/out 06612-033 figure 33. ad8105 crosstalk, all hostile 0 ?120 ?60 ?80 ?100 ?40 ?20 300k 1m 10m 100m 1g crosstalk (db) frequency (hz) 06612-034 single-ended in/out figure 34. AD8104 crosstalk, all hostile, single-ended 300k 1m 10m 100m 1g crosstalk (db) frequency (hz) ?100 ?80 ?60 ?40 ?20 0 single-ended in/out 06612-035 figure 35. ad8105 crosstalk, all hostile, single-ended 0 ?100 ?60 ?80 ?40 ?20 300k 1m 10m 100m 1g 2g feedthrough (db) frequency (hz) differential in/out 06612-036 figure 36. AD8104 crosstalk, off isolation
AD8104/ad8105 rev. 0 | page 20 of 36 0 ?100 ?60 ?80 ?40 ?20 300k 1m 10m 100m 1g 2g feedthrough (db) frequency (hz) single-ended in/out 06612-037 figure 37. AD8104 crosstalk, off isolation, single-ended 6000 5000 0 2000 1000 3000 4000 300k 1m 10m 100m 1g input impedance ( ? ) frequency (hz) differential in ad8105 AD8104 06612-038 figure 38. AD8104, ad8105 input impedance 300k 1m 10m 100m 1g input impedance ( ? ) frequency (hz) 0 500 1000 1500 2000 2500 3000 3500 4000 4500 ad8105 AD8104 single-ended in 06612-039 figure 39. AD8104, ad8105 in put impedance, single-ended 30000 25000 20000 15000 10000 5000 0 100k 1m 10m 100m 1g output impedance ( ? ) frequency (hz) differential out 06612-040 figure 40. AD8104, ad8105 output impedance, disabled 1000 0.1 100k 1g output impedance ( ? ) frequency (hz) 10 1 100 10m 100m 1m 06612-041 figure 41. AD8104, ad8105 output impedance, enabled 0.4 ?0.4 01 v out (v, diff) time (ns) ?0.2 0 0.2 35791113 5 0.3 ?0.3 ?0.1 0.1 4 6 8 10 12 14 12 06612-042 figure 42. AD8104 small signal pulse response, 200 mv p-p
AD8104/ad8105 rev. 0 | page 21 of 36 0.20 ?0.20 01 5 v out (v, se) time (ns) ?0.10 0 0.10 35791113 0.15 ?0.15 ?0.05 0.05 4 6 8 10 12 14 12 n-channel p-channel 06612-043 figure 43. AD8104 small signal pulse response, single-ended, 200 mv p-p 2.0 ?2.0 01 5 v out (v, diff) time (ns) ?1.0 0 1.0 35791113 1.5 ?1.5 ?0.5 0.5 4 6 8 10 12 14 12 06612-044 figure 44. AD8104 large si gnal pulse response, 2 v p-p 1.0 ?1.0 01 5 v out (v, se) time (ns) ?0.6 0.2 35791113 0.6 ?0.2 0 4 6 8 101214 12 n-channel p-channel ?0.4 0.4 0.8 ?0.8 06612-045 1.5 ?1.5 ?40 120 v out (v, diff) time (ns) 1.0 0.5 0 ?0.5 ?1.0 3 ?3 2 1 0 ?1 ?2 100 806040 20 0 ?20 update update (v) v out 06612-046 2 ?4 0 v out (v, diff) time (ns) figure 45. AD8104 larg e signal pulse response, single-ended, 2 v p-p figure 46. AD8104 switching time 1 0 ?1 ?2 ?3 5 4 3 2 1 5000 ?1000 slew rate (v/s) 4000 3000 2000 1000 0 slew rate v out 06612-047 2 ?4 0 v out (v, diff) 1ns/div figure 47. AD8104 large signal rising edge and slew rate 1 0 ?1 ?2 ?3 5 4 3 2 1 3500 ?2500 slew rate (v/s) 2500 1500 500 ?500 ?1500 slew rate v out 06612-048 figure 48. AD8104 large signal falling edge and slew rate
AD8104/ad8105 rev. 0 | page 22 of 36 5 0 ?40 100 offset (mv) temperature (oc) 4 3 1 2 908070 60 5040 30 2010 0 ?10?20?30 06612-049 figure 49. AD8104 v os vs. temperature with all outputs enabled 50 ?20 ?0.10 0.10 v out (mv, diff) time (s) 40 30 20 10 0 ?10 0.080.06 0.04 0.02 0 ?0.02 ?0.04 ?0.06 ?0.08 06612-050 figure 50. AD8104 switching transient (glitch) 0.020 ?0.010 ?700 700 differential gain error (%) v in , diff (mv) 0.015 0.010 0.005 0 ?0.005 ?500 500 300 100 ?100 ?300 06612-051 figure 51. AD8104 gain vs. dc voltage, carrier frequency = 3.58 mhz, subcarrier amplitude = 600 mv p-p, differential 0.014 ?0.004 differential phase error (%) v in , diff (mv) 0.012 0.010 0.008 0.006 0.004 0.002 0 ?0.002 ?700 700 ?500 500 300 100 ?100 ?300 06612-052 figure 52. AD8104 phase vs. dc voltage, carrier frequency = 3.58 mhz, subcarrier amplitude = 600 mv p-p, differential 2.0 ?2.0 01 v out (v, diff) time (ns) 8 16 1412 10 8642 1.5 1.0 0.5 0 ?0.5 ?1.0 ?1.5 10pf 5pf 2pf 0pf 06612-053 figure 53. AD8104 large signal pulse response with capacitive loads 0.4 ?0.4 01 v out (v, diff) time (ns) 8 0.3 0.2 0.1 0 ?0.1 ?0.2 ?0.3 1614 12 10 8642 10pf 5pf 2pf 0pf 06612-054 figure 54. AD8104 small signal pulse response with capacitive loads
AD8104/ad8105 rev. 0 | page 23 of 36 1.4 ?0.2 ?50 150 v out (v, diff) time (ns) 1.2 1.0 0.8 0.6 0.4 0.2 0 2.8 ?0.4 2.4 2.0 1.6 1.2 0.8 0.4 0 130110 9070 503010 ?10 ?30 update v out update (v) 06612-055 figure 55. AD8104 enable time 1.4 ?0.2 ?50 150 v out (v, diff) time (ns) 1.2 1.0 0.8 0.6 0.4 0.2 0 2.8 ?0.4 2.4 2.0 1.6 1.2 0.8 0.4 0 1301109070503010?10?30 v out update update (v) 06612-056 figure 56. AD8104 disable time 0 ?0.05 ?50 100 gain (db) temperature (c) ?250 255075 ?0.01 ?0.02 ?0.03 ?0.04 06612-057 figure 57. AD8104 dc gain vs. temperature 06 612-058 600 100 ?35 95 i pos and i neg current (ma) temperature (c) i dd current (a) 500 400 300 200 850 350 750 650 550 450 857565554535 25 15 5 ?5?15?25 i dd (serial mode) i dd (parallel mode) i pos and i neg (all outputs enabled) i pos and i neg (all outputs disabled) figure 58. AD8104, ad8105 quiescent supply currents vs. temperature 06612-059 360 200 11 channels i pos and i neg (ma) i dd (a) 6 340 320 300 280 260 240 220 23456789101112131415 400 480 560 640 720 800 880 960 1040 i pos , i neg i dd serial i dd parallel figure 59. AD8104, ad8105 quiescent supply currents vs. enabled outputs 65 ?5 07 output error (%) time (ns) 45 36 21 55 45 35 25 15 5 0 60 50 40 30 20 10 3.25 ?0.25 2.75 2.25 1.75 1.25 0.75 0.25 0 3.00 2.50 2.00 1.50 1.00 0.50 (v, diff) v out v in (v out ? v in )/v out 06612-060 figure 60. AD8104 settling time
AD8104/ad8105 rev. 0 | page 24 of 36 5 ?5 07 output/input (%) time (ns) 654321 ?4 4 3 2 1 0 ?1 ?2 ?3 06612-061 figure 61. AD8104 settling time (zoom) 2.5 ?2.5 07 0 0 v out (v, se) time (ns) 600 500 400 300 200 100 ?1.5 ?1.0 ?2.0 0 0.5 ?0.5 1.5 2.0 1.0 v inn v outn v outp v inp 06612-062 figure 62. AD8104 overdrive recovery, single-ended 07 v out (v, se) time (ns) 600 500 400 300 200 100 ?1.5 ?2.0 ?1.0 0 0.5 ?0.5 1.5 2.0 1.0 0 0 v outp v inp v inn v outn 06612-063 figure 63. ad8105 overdrive recovery, single-ended ? 30 ?100 1000 distortion (dbc) frequency (mhz) ?90 ?80 ?70 ?60 ?50 ?40 0.1 10 100 v out = 2v p-p, diff third harmonic second harmonic 06612-064 1 figure 64. AD8104 harmonic distortion
AD8104/ad8105 rev. 0 | page 25 of 36 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 multi- plexer 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 power- on 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/C2 v supplies to +2 v/C3 v supplies are desired, the vocm pin can still be set to 0 v for ground-referenced video signals.
AD8104/ad8105 rev. 0 | page 26 of 36 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, wh ile the parallel programming technique requires more signals, but can change a single output at a time and requires fewer clock cycles to complete programming. serial programming description 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. 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. 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 program- ming. 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. 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 programm ing 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. 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. 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. in similar fashion, if update is taken low after initial power- up, 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 program- ming, 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 pro- grammed 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,
AD8104/ad8105 rev. 0 | page 27 of 36 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. reset 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 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 common- mode 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 common- mode rejection ratio (cmrr). differential operation offers a great noise benefit for signals that are propagated over distance in a noisy environment. in+ vocm in? r g r g rcvr r f r f out? out+ to switch matrix 06612-065 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 imped- ance appear to be 5 k across the inputs. k 52 , = = g dmin r r 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 mo st 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 sg f dmin out,dm dm rr r v v g + == , where: r g = 2.5 k. r s is the user single-ended source resistance (such as 37.5 for a back-terminated 75 source). r f = 2.538 k for the AD8104 and 5.075 k for the ad8105. in the case of the AD8104, s dm r g + = k 5.2 k 538.2 in the case of the ad8105, s dm r g + = k 5.2 k 075.5
AD8104/ad8105 rev. 0 | page 28 of 36 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. rcvr AD8104 opn onn ipn inn 50? 50? 06612-066 figure 66. example of inpu t driven differentially single-ended input the AD8104/ad8105 input receivers can be driven single- endedly (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 ) (2 1 f sg f sg in rrr r rr r where: r g = 2.5 k. r s is the user single-ended source resistance (such as 37.5 for a back-terminated 75 source). r f = 2.538 k for the AD8104 and 5.075 k for the ad8105. in most cases, a single-ended input signal is referred to midsup- ply, 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. rcvr opn onn ipn inn 75 ? 75? (or 37.5 ? ) AD8104 06612-067 figure 67. example of input driven single-ended ac coupling of inputs 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.
AD8104/ad8105 rev. 0 | page 29 of 36 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. 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 common- mode 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. term i n ati on 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 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. opn onn 50 ? 50 ? 100 ? + ? AD8104/ ad8105 06612-068 figure 68. example of back-t erminated 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 differen- tial offset voltage and all of the common-mode offset voltage for a net increase in observed offset.
AD8104/ad8105 rev. 0 | page 30 of 36 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. term i n ati on 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. 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 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. opn onn 75? 75 ? 150 ? AD8104/ ad8105 06612-069 figure 69. example of back-ter minated single-ended load decoupling the signal path of the AD8104/ad8105 is based on high open- loop 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. signal- generated 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
AD8104/ad8105 rev. 0 | page 31 of 36 operation as well, as there is a common-mode-to-differential gain conversion that becomes gr eater at higher frequencies. 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. power dissipation calculation of power dissipation 8 4 15 85 maximum power (w) ambient temperature (c) t j = 150c 7 6 5 25 35 45 55 65 75 06612-070 figure 70. maximum die power dissipation vs. ambient temperature the curve in figure 70 was calculated from ja ambient max junction maxd t t p ? = , , (1) as an example, if the AD8104/a d8105 is enclosed in an envi- ronment at 45c (t a ), the total on-chip dissipation under all load and supply conditions must not be allowed to exceed 7.0 w. 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 ( ) rms output rms utput o pos outputd i vv p , , , ?= 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 ( ) quiescent output neg pos output dq ivv p , , ?= where i output, quiescent = 1.65 ma for each single-ended output pin. for each disabled output, the quiescent power supply current in vpos and vneg drops by approximately 9 ma. qnpn qpnp v neg v pos v output i output i output, quiescent i output, quiescent 06612-071 figure 71. simplified output stage example 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: 1. calculate power dissipation of AD8104/ad8105 using data sheet quiescent currents. disregard vdd current, as it is insignificant. ( ) ( ) vneg neg vpos pos quiescent d iviv p + = , ( ) ( ) w7.1ma340v5.2ma340v5.2 , = + = quiescent d p 2. calculate power dissipation from loads. for a differential output and ground-referenced load, the output power is symmetrical in each output phase. ( ) rms output rms output pos outputd i vv p , , , ? = ( ) () mw15100/v1v1v5.2 , = ? = outputd p there are 16 output pairs, or 32 output currents. w48.0mw1532 , = = outputd np 3. 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). ( ) quiescent output neg pos output dq ivv p , , ? = ( ) mw25.8ma65.1v)5.2(v5.2 , = ? ? = output dq p there are 16 output pairs, or 32 output currents. w26.0mw25.832 , = = output dq np 4. verify that the power dissipation does not exceed the maximum allowed value. output dq outputd quiescent d chipond np np p p , , , , ? + = ? w9.1w26.0w48.0w7.1 , =? + = ? chipond p from figure 70 or equation 1, this power dissipation is below the maximum allowed dissipation for all ambient temperatures up to and including 85c.
AD8104/ad8105 rev. 0 | page 32 of 36 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 require- ments 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. 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. 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 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 systems 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 cross- point 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 ? ? ? ? ? ? ? ? = )( )( log20 10 sa sa xt test sel where: s = j , the laplace transform variable. a sel ( s ) is the amplitude of the crosstalk induced signal in the selected channel. a test ( 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. 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 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. 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,
AD8104/ad8105 rev. 0 | page 33 of 36 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 measure- ment) 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 input- generated 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 termi- nation 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 [ ] scr xt m s = )(log20 10 where: r s is the source resistance. c m 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.
AD8104/ad8105 rev. 0 | page 34 of 36 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. 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 ? ? ? ? ? ? ? ? = l xy r s m xt 10 log20 where: m xy is the mutual inductance of output x to output y. r l is the load resistance on the measured output. this crosstalk mechanism can be minimized by keeping the mutual inductance low and increasing r l . 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. 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 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 termi- nation 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 (v p ) 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 (v p = 144 106 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 10 ?2 m in the calculations. ( ) d vn f p peak 4 12 + = 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. opn onn ipn inn 7 5 ? AD8104/ ad8105 06612-072 figure 72. fly-by input termination, grounds for the two transmission lines shown must be tied together close to the inn pin
AD8104/ad8105 rev. 0 | page 35 of 36 if multiple AD8104/ad8105s are to be driven in parallel, a fly- by input termination scheme is very useful, but the distance from each AD8104/ad8105 input to the driven input transmis- sion line is a stub that should be minimized in length and parasitics using the discussed guidelines. 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 system). this is not often practical as trace widths become large. in most cases, the best practical solution is to place the half- characteristic 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. although the examples discussed so far are for input termina- tion, the theory is similar for output back-termination. taking the AD8104/ad8105 as an ideal voltage source, any distance of routing between the AD8104/ad 8105 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. gnd vocm vdd AD8104/ ad8105 on[15:0], op[15:0] in[31:0], ip[31:0] clk reset we update data in data out j3 pld_vdd pc_vdd pc_gnd sma sma v pos vneg dgnd vdd vpos vneg cpld 50 ? 50? logic pc parallel port in[31:0], ip[31:0] logic isolators j8, w3 to w7 d0 to d5 a0 to a3 on[15:0], op[15:0] analog 06612-073 figure 73. evaluation board simplified schematic
AD8104/ad8105 rev. 0 | page 36 of 36 outline dimensions * compliant to jedec standards mo-192-ban-2 with the exception to package height. detail a a b c d e f g h j k l m n p r t u v w y aa ab ac 1 35 7 9 11 15 17 19 21 23 13 4 6 8 1012 2 16 182022 14 27.94 bsc sq bottom view a 1 corner index area 1.27 bsc top view 31.00 bsc sq ball a1 indicator 0.10 min 0.70 0.63 0.56 1.07 0.99 0.92 coplanarity 0.20 0.90 0.75 0.60 seating plane ball diameter detail a * 1.765 max 022206-a 0.25 min (4 ) figure 74. 304-ball ball grid array, thermally enhanced [bga_ed] (bp-304) dimensions shown in millimeters ordering guide model temperature range package description package option AD8104abpz 1 ?40c to +85c 304-ball ball grid array package, thermally enhanced [bga_ed] bp-304 ad8105abpz 1 ?40c to +85c 304-ball ball grid array package, thermally enhanced [bga_ed] bp-304 AD8104-eval evaluation board ad8105-eval evaluation board 1 z = rohs compliant part. ?2007 analog devices, inc. all rights reserved. trademarks and registered trademarks are the property of their respective owners. d06612-0-6/07(0)

▲Up To Search▲

Price & Availability of AD8104

	To Download AD8104 Datasheet File
If you can't view the Datasheet, Please click here to try to view without PDF Reader .