a Fast, High-Voltage Drive, 6-Channel Output DecDriver™ Decimating LCD Panel Driver AD8380 FEATURES High-Voltage Drive to Within 1.3 V of Supply Rails 24 V Supply for Fast Output Voltage Drivers High Update Rates: Fast 75 Ms/s 10-Bit Input Word Rate Low Power Dissipation, 550 mW with Power-Down Voltage Controlled Video Reference and Full-Scale (Contrast) Output Levels INV Bit Reverses Polarity of Video Signal Nominal 3.3 V Logic and 15 V Analog Supplies Flexible Logic Addressable or Sequential Channel Loading STSQ/CS Allow Parallel AD8380 Operation for XGA and Greater Resolution Drives Capacitive Loads 26 ns Settling Time to 1% Up to 150 pF Load Slew Rate 270 V/s Available in 44-Lead MQFP APPLICATIONS Poly Si LCD Panel Analog Column Driver FUNCTIONAL BLOCK DIAGRAM AD8380 DB [0:9] 10 10 2-STAGE 10 LATCH DAC VID0 2-STAGE 10 LATCH DAC VID1 2-STAGE 10 LATCH DAC VID2 10 2-STAGE 10 LATCH DAC VID3 10 2-STAGE 10 LATCH DAC VID4 10 2-STAGE 10 LATCH DAC VID5 10 CLK STSQ/CS CHANNEL SELECTOR XFR 10 E/O R/L A[0:2] STBY BYP VREFHI VREFLO 3 BIAS SCALING CONTROL INV PRODUCT DESCRIPTION The AD8380 provides a fast, 10-bit latched decimating digital input that drives 6-channel high voltage drive outputs. The 10bit input word is sequentially muxed into six separate high speed, bipolar DACs. Flexible digital input formats allow several AD8380s to be used in parallel for higher resolution displays. STSQ/CS, in conjunction with 3-bit addressable channel-loading pins, allows loading of the digital words either sequentially or randomly, and R/L control sets loading as either left to right, or vice versa. 6-channel high voltage output drivers drive to within 1.3 V of the rails to rated settling time. The output signal can be adjusted for dc signal reference, signal inversion or contrast for maximum flexibility. VMID The AD8380 is fabricated on ADI’s XFCB26 fast bipolar 26 V process, providing fast input logic, trimmed accuracy bipolar DACs and fast settling, high voltage precision drive amplifiers on the same chip. The AD8380 dissipates nominally 0.55 W of static power. STBY pin reduces power to a minimum, with fast recovery. The AD8380 is offered in a 44-lead 10 × 10 × 2.0 mm MQFP package and operates over the commercial temperature range of 0°C to 85°C. DecDriver is a trademark of Analog Devices, Inc. REV. B Information furnished by Analog Devices is believed to be accurate and reliable. However, no responsibility is assumed by Analog Devices for its use, nor for any infringements of patents or other rights of third parties that may result from its use. No license is granted by implication or otherwise under any patent or patent rights of Analog Devices. One Technology Way, P.O. Box 9106, Norwood, MA 02062-9106, U.S.A. Tel: 781/329-4700 www.analog.com Fax: 781/326-8703 © Analog Devices, Inc., 2001 (@ 25ⴗC, AVCC = 15 V, DVCC = 3.3 V, TMIN = 0ⴗC, TMAX = 85ⴗC, unless AD8380–SPECIFICATIONS otherwise noted) Model Conditions Min Typ Max Unit VIDEO DC PERFORMANCE1 VDE VCME Scale Factor Error Offset Error TMIN to TMAX DAC Code = 450 to 800 DAC Code = 450 to 800 DAC Code = 0 to 1023 DAC Code = 0 to 1023 –7.5 –3.5 –0.25 –7 +1 +0.5 +7.5 +3.5 +0.25 +7 mV mV % mV 6 7 3 5 AVCC – 2.5 VREFHI – 2.5 3.3 0.2 750 7.5 V µA V V V kΩ µA µA REFERENCE INPUTS VMID Range2 VMID Bias Current VFS Range VREFHI VREFLO VREFHI Input Resistance VREFLO Bias Current VREFHI Input Current3 VFS = 5 V RESOLUTION Coding Binary DIGITAL INPUT CHARACTERISTICS Input Data Update Rate Clock to Data Setup Times: t1 Clock to STSQ Setup Times: t3 Clock to XFR Setup Times: t5 Maximum CLK Rise and Fall Time, t7 Clock to A[0:2] Hold Times: t9 Clock to Data Hold Times: t2 Clock to STSQ Hold Times: t4 Clock to XFR Hold Times: t6 Clock to A[0:2] Setup Times: t8 CIN IIN VIH VIL VTH VIDEO OUTPUT CHARACTERISTICS Output Voltage Swing CLK to VID Delay4 Output Current VIDEO OUTPUT DYNAMIC PERFORMANCE VFS = 2 × (VREFHI–VREFLO) 1 VREFLO +0.5 VMID – 0.5 to VREFLO 6 AVCC VREFHI – 0.5 10 Bits 75 1 1 1 4 4 4 4 4 1 3 0.6 2.0 0.8 Threshold Voltage 1.4 AVCC – VOH, VOL – AVEE 50% of VIDx 13.5 30 1.1 15.5 1.3 17.5 Ms/s ns ns ns ns ns ns ns ns ns pF µA V V V V ns mA TMIN to TMAX, VO = 5 V Step, CL = 150 pF, RS = 25 Ω Data Switching Slew Rate Invert Switching Slew Rate Data Switching Settling Time to 1%5 Data Switching Settling Time to 0.25% Invert Switching Settling Time to 1%5 Invert Switching Settling Time to 0.25% CLK Feedthrough6 All-Hostile Crosstalk7 Amplitude Glitch Duration POWER SUPPLY Supply Rejection (VDE) DVCC, Operating Range DVCC, Quiescent Current AVCC, Operating Range Total AVCC Quiescent Current STBY AVCC Current STBY DVCC Current +1 270 625 26 35 30 85 2 32 65 40 100 5 95 40 +VS = 15 V ± 1 V mV p-p ns 1 3 22 9 33 0.5 0.1 STBY = H STBY = H OPERATING TEMPERATURE RANGE 0 V/µs V/µs ns ns ns ns mV p-p 5.5 35 24 44 5 5 mV/V V mA V mA mA mA 85 °C NOTES 1 For definitions of VDE and VCME, see the Transfer Function section. Scale factor error is expressed as percentage of VFS. See Figure 1 for valid ranges of VMID. 3 VREFHI Input Current = (VREFHI – VREFLO)/(VREFHI Input Resistance) = 2.5 V/3.3 kΩ. 4 Delay time from 50% of falling CLK edge to 50% of output change. Measurement is made for both states of INV. 5 For best settling time results, use minimum series output resistance, R S of 25 Ω. 6 An output channel is selected, and glitch is monitored as CLK is driven. STSQ and XFR are set to logic low. 7 Input data is loaded such that any five output channels change by VFS (i.e., 5 V), and the sixth unselected channel is monitored. Measurement is made for both states of INV. 2 Specifications subject to change without notice. –2– REV. B AD8380 PIN FUNCTION DESCRIPTIONS Pin No. Mnemonic Description 1 2–11 12 NC DB[0:9] E/O 13 R/L 14 15, 16 17, 20, 22, 24, 26, 28, 30, 32, 34, 37, 38 18 INV DVEE, DVCC No Connect. Video Data Inputs. DB9 is the MSB. Even/Odd data select, input latches are loaded at the falling edge of CLK if E/O is low or rising edge if E/O is high. Determines starting point of internally generated channel-loading sequence. R/L Low (when address = 111) loads from Channel 0 up to Channel 5. When high, analog video outputs are above the VMID setpoint. See Figure 3. Digital Supplies. Nominally 3.3 V and 0 V, respectively. AVCCxxx, AVEExxx Analog Supplies. Nominally 15 V and 0 V, respectively. STBY Stand By. When high, all digital and analog circuits are “debiased” and the power dissipation drops to a minimum. BYP An external capacitor connected from here to VEE will help to ensure rapid DAC settling time. VMID Externally supplied voltage applied here sets the midpoint reference for the video output. 19 21 23, 25, 27, 29, 31, 33 VID5–VID0 Analog Video Outputs. 36, 35 VREFHI, VREFLO Voltage between these pins sets DAC full-scale range. An external reference must be applied and should be common to all devices to ensure best tracking. 39–41 A[0:2] 3-bit channel address for addressable loading of the digital input latches. 42 STSQ/CS STSQ to start internal sequencing or Chip Select to enable addressable channel addressing. See functional description. Used in conjunction with A[0:2]. 43 XFR If XFR = HIGH at the rising edge of CLK, data is transferred to the DACs on the next falling edge of CLK. See Figures 4, 6, 7, and 8. 44 CLK Master Clock Input. CHANNEL SELECTION FUNCTIONALITY MAXIMUM OUTPUT VOLTAGE There are two channel selection modes, addressed channel loading, (in which the user directly controls which DAC is loaded), and internally sequenced loading (in which the user controls the direction and clock phase in which the loading proceeds). The maximum output signal swing is constrained by the output voltage compliance of the DACs and the output dynamic range of the output amplifiers. The minimum voltage allowed at the outputs of the DACs is about 6 V. This constrains the minimum value of VMID to be 6 V. The output amplifiers will swing and settle cleanly, as described on the specification page, for output voltages within 1.5 V from each supply voltage rail. ADDRESSED CHANNEL LOADING: When channel address (A0, A1, A2) = 000 through 101, the video data is loaded into Channels 0 through 5. (STSQ/CS functions as “Chip Selection” this case.) INTERNALLY SEQUENCED LOADING: When channel address = 111 the video data is loaded in a sequence determined internally. The sequencing is initiated by a pulse applied to STSQ/CS input. The count proceeds from 0 to 5 if R/L is LOW or from 5 to 0 if R/L is HIGH. For a given value of VMID, the voltage required to saturate the video output voltages defines the maximum usable full-scale voltage. For example, if VMID is less than AVCC/2, the maximum value of VFS is (VMID – 1.5 V). If VMID is greater than AVCC/2, the maximum useful VFS is (AVCC – 1.5 – VMID). Figure 1 graphically describes these limiting factors. 6 VFS – Volts DAC TRANSFER FUNCTION VOUT = VMID + VFS × (1 – N/1023); if INV is HIGH, VOUT = VMID – VFS × (1 – N/1023); if INV is LOW 4.5 where VFS = 2 × (VREFHI – VREFLO) 6 VMID – Volts 7.5 Figure 1. Valid Range for VMID with Respect to VFS (AVCC = 15 V) REV. B –3– AD8380 ABSOLUTE MAXIMUM RATINGS 1 MAXIMUM POWER DISSIPATION Supply Voltage AVCC–AVEE . . . . . . . . . . . . . . . . . . . . . 26 V Internal Power Dissipation2 Quad Flat Package (S) . . . . . . . . . . . . . . . . . . . . . . . 1.7 W Output Short Circuit Duration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . See Associated Text Storage Temperature Range . . . . . . . . . . . . –65°C to +125°C Operating Temperature Range . . . . . . . . . . . . . . 0°C to 85°C Lead Temperature Range (Soldering 10 sec) . . . . . . . . . 300°C The maximum power that can be safely dissipated by the AD8380 is limited by the associated rise in junction temperature. The maximum safe junction temperature for plastic encapsulated devices is determined by the glass transition temperature of the plastic, approximately 150°C. Exceeding this limit temporarily may cause a shift in parametric performance due to a change in the stresses exerted on the die by the package. Exceeding a junction temperature of 175°C for an extended period can result in device failure. NOTES 1 Stresses above those listed under Absolute Maximum Ratings may cause permanent damage to the device. This is a stress rating only; functional operation of the device at these or any other conditions above those indicated in the operational section of this specification is not implied. Exposure to absolute maximum rating conditions for extended periods may affect device reliability. 2 Specification is for device in free air: 44-Lead MQFP Package: θJA = 73°C/W (Still Air), where P D = (TJ – TA)/θJA. θJC = 22°C/W. 44 43 42 41 40 39 38 37 36 35 34 33 PIN 1 IDENTIFIER VID0 AVCC0,1 31 VID1 32 DB1 3 DB2 4 30 AVEE1,2 DB3 5 AD8380 29 VID2 DB4 6 DB5 7 TOP VIEW (Not to Scale) 28 AVCC2,3 3.0 MAXIMUM POWER DISSIPATION – Watts NC 1 DB0 2 27 VID3 26 AVEE3,4 DB6 8 DB7 9 DB8 10 (MSB) DB9 11 25 VID4 24 AVCC4,5 23 VID5 AVEE5 VMID BYP AVEE BIAS AVCC BIAS STBY DVEE DVCC R/L INV E/O 12 13 14 15 16 17 18 19 20 21 22 NC = NO CONNECT The AD8380’s internal short circuit limitation is not sufficient to protect the device in the event of a direct short circuit between a video output and a power supply voltage rail (VCC or VEE). Temporary short circuits can reduce an output’s ability to source or sink current and, therefore, impact the device’s ability to drive a load. Short circuits of extended duration can cause metal lines to fuse open, rendering the device nonfunctional. To prevent these problems, it is recommended that a series resistor of 25 Ω or greater be placed as close as possible to the AD8380’s video outputs. This will serve to substantially reduce the magnitude of the fault currents and protect the outputs from damage caused by intermittent short circuits. This may not be enough to guarantee that the maximum junction temperature (150°C) is not exceeded under all conditions. To ensure proper operation, it is necessary to observe the maximum power derating curve in Figure 2 below. VREFLO AVEE0 VREFHI A2 AVCCDAC AVEEDAC A1 STSQ/CS A0 XFR CLK PIN CONFIGURATION Output Short Circuit Limit TJ, MAX = 150ⴗC 2.5 2.0 1.5 1.0 0.5 0 10 20 30 40 50 60 70 AMBIENT TEMPERATURE – ⴗC 80 90 Figure 2. Maximum Power Dissipation vs. Temperature ORDERING GUIDE Model Temperature Range Package Description Package Option AD8380JS 0°C to 85°C 44-Lead MQFP S-44A CAUTION ESD (electrostatic discharge) sensitive device. Electrostatic charges as high as 4000 V readily accumulate on the human body and test equipment and can discharge without detection. Although the AD8380 features proprietary ESD protection circuitry, permanent damage may occur on devices subjected to high-energy electrostatic discharges. Therefore, proper ESD precautions are recommended to avoid performance degradation or loss of functionality. –4– WARNING! ESD SENSITIVE DEVICE REV. B Typical Performance Characteristics–AD8380 VMID + VFS INV = H CODE = 0 VMID = 7V VFS = 5V VIDX VMID = 7V VFS = 5V VMID 25⍀ VIDX CL 150pF 1.25V/DIV 25⍀ CL 150pF 1V/DIV VMID – VFS INV = L VMID – VFS 20ns/DIV 20ns/DIV TPC 1. Invert Switching 10 V Step Response (Rise) at CL TPC 4. Data Switching Full-Scale Step Response (Fall) at CL, INV = L INV = H VMID + VFS VMID = 7V VFS = 5V CODE = 0 VMID = 7V VFS = 5V VIDX VMID + VFS 25⍀ VIDX CL 150pF 25⍀ CL 150pF 1V/DIV 1.25V/DIV VMID INV = L VMID – VFS 20ns/DIV 20ns/DIV TPC 2. Invert Switching 10 V Step Response (Fall) at CL TPC 5. Data Switching Full-Scale Step Response (Rise) at CL, INV = H VMID = 7V VFS = 5V VMID = 7V VFS = 5V VMID + VFS VMID VIDX 25⍀ CL 150pF 1V/DIV 1V/DIV 25⍀ VIDX VMID – VFS CL 150pF VMID 20ns/DIV 20ns/DIV TPC 3. Data Switching Full-Scale Step Response (Rise) at CL, INV = L REV. B TPC 6. Data Switching Full-Scale Step Response (Fall) at CL, INV = H –5– VMID+ VFS VMID = 7V VFS = 5V INV = H VIDX OUTPUT VOLTAGE ERROR – 0.1%/DIV OUTPUT VOLTAGE ERROR – 1%/DIV AD8380 25⍀ CL 150pF VMID + VFS t=0 VMID VMID = 7V VFS = 5V INV = H VIDX 25⍀ CL 150pF VMID + VFS VMID VMID t=0 10ns/DIV 10ns/DIV TPC 10. Output Settling Time Response to 0.25% of Full Scale (Falling Edge) at CL TPC 7. Output Settling Time Response to 1% of Full Scale (Rising Edge) at CL 7.5 3.5 CODE 482 VIDX VDE – mV OUTPUT VOLTAGE ERROR – 1%/DIV 5.5 VMID = 7V VFS = 5V INV = H 25⍀ CL 150pF 1.5 CODE 738 0 –1.5 VMID + VFS VMID –3.5 VMID t=0 –5.5 –7.5 10ns/DIV TPC 8. Output Settling Time Response to 1% of Full Scale (Falling Edge) at CL 0 10 20 30 40 50 60 TEMPERATURE – ⴗC 70 80 90 TPC 11. Differential Error Voltage (VDE) vs. Temperature 3.5 1.5 CODE 738 VMID + VFS VCME – mV OUTPUT VOLTAGE ERROR – 0.1%/DIV 2.5 VMID = 7V VFS = 5V INV = H VIDX t=0 25⍀ 0.5 0 –0.5 CODE 482 –1.5 CL 150pF –2.5 VMID + VFS VMID –3.5 0 10ns/DIV 10 20 60 30 40 50 TEMPERATURE – ⴗC 70 80 90 TPC 12. Common-Mode Error Voltage (VCME) vs. Temperature TPC 9. Output Settling Time Response to 0.25% of Full Scale (Rising Edge) at CL –6– REV. B 0.5 0.5 0.4 0.4 0.3 0.3 0.2 0.2 0.1 0.1 INL – LSB DNL – LSB AD8380 0 –0.1 0 –0.1 –0.2 –0.2 –0.3 –0.3 –0.4 –0.4 –0.5 –0.5 0 128 256 384 512 640 768 896 1024 0 128 256 384 CODE 0.5 0.5 0.4 0.4 0.3 0.3 0.2 0.2 0.1 0.1 0 –0.1 –0.3 –0.3 –0.4 –0.4 256 384 512 CODE 896 1024 –0.1 –0.2 128 768 0 –0.2 0 640 TPC 16. Integral Nonlinearity (INL) vs. Code, INV = H INL – LSB DNL – LSB TPC 13. Differential Nonlinearity (DNL) vs. Code, INV = H –0.5 512 CODE 640 768 896 –0.5 1024 0 128 256 384 512 640 768 896 1024 CODE TPC 14. Differential Nonlinearity (DNL) vs. Code, INV = L TPC 17. Integral Nonlinearity (INL) vs. Code, INV = L 7.5 3.5 5.0 2.5 VCME – mV VDE – mV 2.5 0 0 –2.5 –2.5 –5.0 –7.5 0 128 256 384 512 640 768 896 –3.5 1024 CODE TPC 15. Differential Error Voltage (VDE) vs. Code REV. B 0 128 256 384 512 CODE 640 768 896 1024 TPC 18. Common-Mode Error Voltage (VCME) vs. Code –7– AD8380 1mV/DIV VFS = 5V VMID = 7V INV = L RS = 25⍀ CL = 150pF OUTPUT @ CODE 0 2.5V/DIV 20mV/DIV CLK 2V/DIV VFS = 5V VMID = 7V INV = L RS = 25⍀ CL = 150pF VID0,1,2,3,4 VID5 (QUIET) 20ns/DIV 10ns/DIV TPC 19. Clock Switching Transient (Feedthrough) at CL TPC 21. All-Hostile Crosstalk at CL 60 40 20 PSRR – dB 5mV/DIV VFS = 5V VMID = 7V INV = L RS = 25⍀ CL = 150pF OUTPUT @ CODE 1023 2V/DIV DATA 0 –20 VOUTP (INV = L) –40 VOUTN (INV = H) –60 20ns/DIV –80 10k TPC 20. Data Switching Transient (Feedthrough) at CL 100k FREQUENCY – Hz 1M 5M TPC 22. AVCC Power Supply Rejection vs. Frequency –8– REV. B AD8380 THEORY OF OPERATION The AD8380 is a system building block designed to directly drive the columns of poly-silicon LCD panels of the type popularized for use in data projectors. It comprises six channels of precision 10-bit digital-to-analog converters loaded from a single, high speed, 10-bit parallel input. Precision current feedback amplifiers providing well-damped pulse responses and rapid voltage settling into large capacitive loads buffer the six outputs. Excellent linearity performance and laser trimming of scale factors and output offsets at the wafer level ensure low absolute output errors over all input codes. Tight channel-tochannel matching in high channel count systems is guaranteed by reliance on an externally-applied voltage reference. AD8380 DB [0:9] 10 10 2-STAGE 10 LATCH DAC VID0 2-STAGE 10 LATCH DAC VID1 2-STAGE 10 LATCH DAC VID2 10 2-STAGE 10 LATCH DAC VID3 10 2-STAGE 10 LATCH DAC VID4 10 CLK STSQ/CS CHANNEL SELECTOR XFR 10 E/O R/L A[0:2] STBY BYP 3 BIAS VREFLO For each value of input code there are then two possible values for the output voltage, depending on the status of INV. When INV is low the output is defined as VOUTP(N) where N refers to the input code, and the P refers to the positive slope of the voltage variation with code. When INV is high, the output is defined as VOUTN(N). To best correlate transfer function errors to image artifacts, the overall accuracy of the AD8380 is defined by comparing the output voltages, VOUTP(N) and VOUTN(N), to each other and to their ideal values. Two parameters are defined, one dependent on the difference between the signal amplitudes at a particular code, and one dependent on their average value. These are VDE and VCME. Their defining expressions are: VDE = [VOUTN(N) – VOUTP(N)]/2 – [(1 – N/1023) × VFS] where N = input code, and VFS = 2 × (VREFHI – VREFLO) VCME = [[VOUTN(N) +VOUTP(N)]/2 – VMID] × (1/2) where 10 2-STAGE 10 LATCH VREFHI The region over which the output voltage varies with input code is defined by the status of the INV input. When INV is low, the video output voltages rise from (VMID – VFS), (where VFS = the full-scale output voltage), to VMID as the input code increases from 0 to 1023. When INV is high, the output voltages drop from (VMID + VFS) to VMID with increasing code (see Figure 4). DAC VID5 VMID = midpoint reference voltage for the video outputs. Setting the Full-Scale Output SCALING CONTROL INV The full-scale output voltage (VFS), which defines the maximum output voltage excursion for a full code input transition, is defined as twice the voltage difference between the VREFHI and VREFLO inputs. VMID Figure 3. Top Level Block Diagram Transfer Function Operating Modes, Control Logic and DAC Latches The transfer function of the AD8380 is made up of two regions of operation, in which the video output voltages are either above or below an output reference voltage externally applied at the VMID input. Control logic included on the AD8380 chip facilitates channel loading in ascending or descending order (for image mirroring), data loading on rising or falling clock edges (for even/odd word loading), and addressing and loading individual channels (for system testing or debugging). The on-chip logic makes it easy to build systems requiring more than six drive channels per color. (VMID + VFS) VIDEO OUTPUT VOLTAGE INV = H DAC latches are of a two-stage master-slave design that guarantees all channel outputs are updated simultaneously. VOUTN VMID VOUTP INV = L (VMID – VFS) 0 INPUT CODE 1023 Figure 4. Definition of Output Transfer Function REV. B –9– AD8380 SVGA System Operation 1 COLOR OF ‘EVEN/ODD’ XGA An SVGA system is characterized by the requirement of six channels of panel drive for each displayed color. Such a system would use a single AD8380 per color. STSQ/CS STSQ_A STSQ_B XFR E/O_A PANEL R/L CONTROLLER INV E/O_B CLKIN With E/O and all address bits A[0:2] set high, channel loading commences on the first rising edge of CLK following a valid assertion of the Start Sequence (STSQ) input. The second stage latches, and therefore the video outputs, are updated on the first falling edge of the clock following a valid Transfer (XFR) signal. (See Figure 5 for signal timing details.) XFR E/O R/L INV 5 0 5 t1 t7 2.0V CLK STSQ/CS t3 VIDEO DB[0:9] t7 t5 t6 Figure 5. Sequenced SVGA Timing (A[0:2] = HIGH, E/ O = HIGH, See Table I ) Data Byte Number E/O = HIGH R/L = LOW 0 1 2 3 4 5 VID0 VID1 VID2 VID3 VID4 VID5 DB[0:9] CLK DCLK/2 DB[0:9] 10 0 11 0 9 10 t1 Table I. Sequenced SVGA Data Byte to Channel Assignment Channel Number 6 3 6 Figure 6. Even/Odd: Outputs of Devices A and B are Configured as Even and Odd Data Channels and Loading Sequence Is Defined by Status of E / O and R / L Inputs t4 XFR VIDEO OUT 10 2.0V 0.8V 0.8V AD8380 DEVICE “A” STSQ/CS A[0:2] XFR AD8380 E/O DEVICE “B” R/L VIDEO OUT INV DB[0:9] CLK 0 t2 3 A[0:2] DVCC IMAGE PROCESSOR DB[0:9] DVCC CLK (EVEN CHIP) 11 t2 t4 t3 STSQ/CS (EVEN CHIP) t1 CLK (ODD CHIP) t3 t4 STSQ/CS (ODD CHIP) Load Sequence Switching (Right/Left Control) To facilitate image mirroring, the order in which channels are loaded can be easily switched. When the voltage on the right/left control input (R/L) is low, the internal sequencer will load data starting with Channel 0 and counting up to Channel 5. When this voltage is high, channel loading will be in reverse order, from Channel 5 down to Channel 0. t5 t6 XFR A0:A2 = HIGH Figure 7. Sequenced Even/Odd XGA Timing, A[0:2] = HIGH (See Table II) XGA System Operation In an XGA system, twelve column drivers (two AD8380s) are required for each color (refer to Figure 6). An “even/odd” system, in which one AD8380 drives even numbered columns and another drives odd numbered columns, can be easily implemented as detailed in Figures 7 and 8. A clock at one-half the pixel rate is applied to the CLK input. Even bytes are loaded on the rising edge of the clock, while odd bytes are loaded on the falling edge. Identifying whether a chip is to load on rising or falling edges is done by setting the proper level on the E/O input. t2 Table II. Sequenced Even/Odd XGA Data Byte to Channel Assignment Channel Number Data Byte Number R/L = LOW R/L = HIGH E/O = HIGH VID0 VID1 VID2 VID3 VID4 VID5 0 2 4 6 8 10 10 8 6 4 2 0 E/O = LOW VID0 VID1 VID2 VID3 VID4 VID5 1 3 5 7 9 11 11 9 7 5 3 1 –10– REV. B AD8380 DCLK AD8380 INPUT DATA 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 CLK (AD8380) STSQ/CS (EVEN) STSQ/CS (ODD) XFR 12 0 24 14 2 16 4 18 6 20 8 AD8380 #1 22 10 EVEN CHANNEL (E/O = HIGH) (R/L = LOW) 0 12 2 14 4 16 6 18 8 20 10 22 13 1 15 3 17 5 19 7 21 9 AD8380 #2 ODD CHANNEL (E/O = LOW) (R/L = LOW) 11 23 1 13 3 15 5 17 7 19 9 21 11 23 Figure 8. Operation of Even/Odd XGA System REV. B –11– AD8380 SXGA and Beyond APPLICATION Very high resolution display systems can be built using the E/O XGA system as a model. By using four AD8380s, twenty-four columns can be driven together for an SXGA display. Two would be designated for even columns and two for odd. Four separate STSQ signals would be used to coordinate data loading with a single XFR to synchronize updating of output voltages. The AD8380 is a mixed-signal, high speed, very accurate device with multiple channels. In order to realize its specifications, it is essential to use a properly designed circuit board. Layout and Grounding Using a single external voltage source to drive the VREF inputs on all drivers for a particular color and a single voltage source for all their VMID inputs, will guarantee matching for all channels. The exceptional accuracy of the AD8380’s transfer function will ensure that high channel count systems can be built without fear of image artifacts resulting from channel-to-channel matching errors. Direct Channel Loading For debug or characterization purposes, it may be desirable to load data directly into a single channel without requiring exercise of the STSQ and XFR inputs. This can be done by applying dc logic high levels to the STSQ and XFR inputs, and addressing the desired channel through the A[0:2] inputs. Data will then be loaded into the selected channel on each falling edge of the CLK signal. The analog and digital sections of the AD8380 are pinned out on approximately opposite sides of the package. When laying out a circuit board, please keep these sections separate from each other to minimize crosstalk and noise coupling of the digital input signals into the analog outputs. All signal trace lengths should be made as short and direct as possible to prevent signal degradation due to parasitic effects. Please note that digital signals should not cross or be routed near analog signals. The maximum rate at which a channel can be updated will be limited by the settling time of the output amplifiers. It is imperative to provide a solid ground plane under and around the part. All of the ground pins of the part should be directly connected to the ground plane with no extra signal path length. For conventional operation, this includes the pins DVEE, AVEEDAC, AVEEBIAS, AVEE0, AVEE1,2; AVEE3,4; and AVEE5. The return currents for any of the signals for the part should be routed close to the ground pin for that section to prevent stray signals from appearing on other ground pins. Addressed Channel Loading Power Supply Bypassing The direct channel loading method can be extended. Channels may be loaded in an arbitrary sequence through the use of an active XFR signal with STSQ set to a high level. Use the A[0:2] inputs to define the desired channel sequence. Data will be loaded on the falling edge of CLK into the channel whose address was valid on the preceding rising edge of CLK. All channel outputs are then updated together by qualification of a valid XFR signal. See Figure 9 for timing details. The AD8380 has several power supply and reference voltages that must be properly bypassed to the ground plane for optimum performance. The bypass capacitor for each supply pin, as well as VREFHI, VREFLO, and VMID, should be connected as close as possible to the IC pins and directly to the ground plane. A 0.1 µF capacitor, preferably a ceramic chip, should be used to minimize lead length. DB[0:9] 5 0 1 t1 5 0 t2 CLK t5 t6 XFR t8 t9 To provide low frequency, high current bypassing, larger value tantalum capacitors should also be used. These should be connected from the supply to ground, but it is not necessary to place these close to the IC pins. Stray inductance will not greatly affect their performance. The high current outputs should be bypassed with these capacitors. It is recommended that two 22 µF tantalum capacitors be placed from the AVCC supply to ground at either end of the output side of the IC. AVCCBIAS and AVCCDAC should each have a 10 µF tantalum bypass capacitor to ground. See Figure 10. VREFHI Reference Distribution A[0:2] IN THIS CASE, INPUT LATCHES ARE LOADED IN THE ORDER SELECTED BY A[0:2], THEN DACs ARE UPDATED TOGETHER BY XFR. Figure 9. Addressed Channel Timing (E/O = HIGH, STSQ/CS = HIGH) Standby Mode A high level applied to the standby (STBY) input will turn off most of the internal circuitry, dropping the quiescent power dissipation to a few milliwatts. Since both digital and analog circuits are debiased, all stored data will be lost. Upon returning STBY to a low level, normal operation is restored. In a system that uses more than one AD8380 per color, it is important that all of the AD8380 devices operate from equal reference voltages to ensure that the video outputs are well matched. VREFLO is not a concern due to its high input resistance and very low bias current. Therefore, it is not likely that there will be significant dc voltage drops in the circuit traces to that supply. It is recommended to have good local supply bypassing at each AD8380 from their respective VREFLOs to ground. The higher input current that flows in the VREFHI circuit requires that this be laid out more carefully. VREFHI connects internally to a 20 kΩ resistor for each of the six channels to provide an input resistance of about 3.3 kΩ. Thus with a (VREFHI – VREFLO) voltage of 2.5 V (to yield a VFS of 5.0 V ), about 750 µA will flow into each VREFHI circuit. –12– REV. B AD8380 In order to obtain the best matching, the traces to each of the VREFHI pins of the AD8380s should be connected by an approximately same length and same width circuit trace in a “star” configuration. The source of the VREFHI voltage should be at the center of the “star.” Therefore, the VREFHI currents for two devices will not share a significant length of circuit trace, and each trace will provide an approximately equal voltage drop. For example, if a trace length is 5 in. long (13 cm.), then the trace width for a 1 oz. copper foil should be wider than 0.025 in. (0.7 mm) in order to keep the trace impedance below 100 mΩ. Driving a Capacitive Load A purely capacitive load can react with output impedance of the AD8380 resulting in overshoot and ringing in its step response. To minimize this effect, and optimize settling time, it is recommended that a 25 Ω resistor be placed in series with each of the driver’s outputs as shown in Figure 10. In addition, if the VREFHI traces must be long, then the traces should be widened to minimize differences in the voltage drops due to differences in the VREFHI input currents of different AD8380s. The dc resistance of these traces should be less than 100 mΩ. If the VREFHI input current is about 1 mA, then the voltage drop will be about 100 µV. 10V 10F CLK 10F 2.500V 3.3V DVCC 0.1F 0.1F VREFHI VREFLO 10F 22F 22F 15V 0.1F 7V 0.1F AVCCDAC 0.1F AVCCBIAS 0.1F AVCC0,1 0.1F AVCC2,3 AVCC4,5 0.1F 0.1F 10F VMID 10 DB[0:9] XFR E/O R/L LOGIC BIAS DAC 6 DRIVERS VID0 – VID5 STSQ/CS 3 A[0:2] INV STBY DVEE AVEEDAC BYP AVEEBIAS AVEE0 0.1F Figure 10. Interface Drawing REV. B RS 24.9⍀ (TYP FOR 6) –13– AVEE1,2 AVEE3,4 AVEE5 AD8380 OUTLINE DIMENSIONS Dimensions shown in inches and (mm). 44-Lead MQFP (S-44A) 0.530 (13.45) SQ 0.510 (12.95) 0.398 (10.10) SQ 0.390 (9.90) 0.096 (2.45) MAX 0.041 (1.03) 0.029 (0.73) 44 34 1 33 SEATING PLANE 0.315 (8.00) REF TOP VIEW (PINS DOWN) 0.010 (0.25) MAX 0.009 (0.23) 0.005 (0.13) 0.083 (2.10) 0.077 (1.95) 11 23 22 12 0.031 (0.80) BSC 0.018 (0.45) 0.012 (0.30) AD8380–Revision History Location Page Data sheet changed from REV. A to REV B. Single-Channel Block Diagram section and graphic removed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 –14– REV. B –15– –16– PRINTED IN U.S.A. C01091–0–4/01(B)