ADC083000 8-Bit, 3 GSPS, High Performance, Low Power A/D Converter General Description Features The ADC083000 is a single, low power, high performance CMOS analog-to-digital converter that digitizes signals to 8 bits resolution at sampling rates up to 3.4 GSPS. Consuming a typical 1.9 Watts at 3 GSPS from a single 1.9 Volt supply, this device is guaranteed to have no missing codes over the full operating temperature range. The unique folding and interpolating architecture, the fully differential comparator design, the innovative design of the internal sample-and-hold amplifier and the self-calibration scheme enable a very flat response of all dynamic parameters up to Nyquist, producing a high 7.0 Effective Number Of Bits, (ENOB) with a 748 MHz input signal and a 3 GHz sample rate while providing a 10 -18 Bit Error Rate, (BER). The ADC083000 achieves a 3 GSPS sampling rate by utilizing both the rising and falling edge of a 1.5 GHz input clock. Output formatting is offset binary and the LVDS digital outputs are compatible with IEEE 1596.3-1996, with the exception of an adjustable common mode voltage between 0.8V and 1.15V. The ADC has a 1:4 demultiplexer that feeds four LVDS buses and reduces the output data rate on each bus to a quarter of the sampling rate. The converter typically consumes less than 25 mW in the Power Down Mode and is available in a 128-lead, thermally enhanced exposed pad LQFP and operates over the Industrial (-40°C ≤ TA ≤ +85°C) temperature range. ■ ■ ■ ■ ■ ■ ■ ■ Internal Sample-and-Hold Single +1.9V ±0.1V Operation Choice of SDR or DDR output clocking Guaranteed No Missing Codes Serial Interface for Extended Control Adjustment of Input Full-Scale Range and Offset Duty Cycle Corrected Sample Clock Test pattern Key Specifications ■ ■ ■ ■ ■ ■ ■ Resolution Max Conversion Rate Bit Error Rate (BER) ENOB @ 748 MHz Input SNR @ 748 MHz Full Power Bandwidth Power Consumption — Operating — Power Down Mode 8 Bits 3 GSPS (min) 10-18 (typ) 7.0 Bits (typ) 44.5 dB (typ) 3 GHz (typ) 1.9 W (typ) 25 mW (typ) Applications ■ ■ ■ ■ ■ Direct RF Down Conversion Digital Oscilloscopes Satellite Set-top boxes Communications Systems Test Instrumentation Ordering Information Industrial Temperature Range (-40°C < TA < +85°C) NS Package ADC083000CIYB 128-Pin Exposed Pad LQFP ADC08x3000EB Development Board © 2007 National Semiconductor Corporation 201932 www.national.com ADC083000 8-Bit, 3 GSPS, High Performance, Low Power A/D Converter May 2007 ADC083000 Block Diagram 20193253 Pin Configuration 20193201 Note: The exposed pad on the bottom of the package must be soldered to a ground plane to ensure rated performance. www.national.com 2 ADC083000 Pin Descriptions and Equivalent Circuits Pin Functions Pin No. 3 4 Symbol Equivalent Circuit Description OutV / SCLK Output Voltage Amplitude / Serial Interface Clock (Input):LVCMOS Tie this pin high for normal differential DCLK and data amplitude. Ground this pin for a reduced differential output amplitude and reduced power consumption. See Section 1.1.6. When the extended control mode is enabled, this pin functions as the SCLK input which clocks in the serial data. See Section 1.2 for details on the extended control mode. See Section 1.3 for description of the serial interface. OutEdge / DDR / SDATA Edge Select / Double Data Rate / Serial Data (Input):LVCMOS This input sets the output edge of DCLK+ at which the output data transitions. (See Section 1.1.5.2). When this pin is floating or connected to 1/2 the supply voltage, DDR clocking is enabled. When the extended control mode is enabled, this pin functions as the SDATA input. See Section 1.2 for details on the extended control mode. See Section 1.3 for description of the serial interface. DCLK Reset (Input):LVCMOS A positive pulse on this pin is used to reset and synchronize the DCLK outs of multiple converters. See Section 1.5 for detailed description. When bit 14 in the Configuration Register (address 1h) is set to 0b, this singleended DCLK_RST pin is selected. 15 DCLK_RST 26 PD Power Down (Input):LVCMOS A logic high on the PD pin puts the entire device into the Power Down Mode. CAL Calibration Cycle Initiate (Input):LVCMOS A minimum 80 input clock cycles logic low followed by a minimum of 80 input clock cycles high on this pin initiates the calibration sequence. See Section 2.4.2 for an overview of self-calibration and Section 2.4.2.2 for a description of on-command calibration. FSR/ECE Full Scale Range Select / Extended Control Enable (Input):LVCMOS In non-extended control mode, a logic low on this pin sets the full-scale differential input range to 600 mVP-P. A logic high on this pin sets the full-scale differential input range to 820 mVP-P. See Section 1.1.4. To enable the extended control mode, whereby the serial interface and control registers are employed, allow this pin to float or connect it to a voltage equal to VA/2. See Section 1.2 for information on the extended control mode. 30 14 3 www.national.com ADC083000 Pin Functions Pin No. Symbol Equivalent Circuit Description CalDly / SCS Calibration Delay / Serial Interface Chip Select (Input):LVCMOS With a logic high or low on pin 14, this pin functions as Calibration Delay and sets the number of input clock cycles after power up before calibration begins (See Section 1.1.1). With pin 14 floating, this pin acts as the enable pin for the serial interface input and the CalDly value becomes "0" (short delay with no provision for a long power-up calibration delay). 10 11 CLK+ CLK- Sampling Clock Input (Input):LVDS The differential clock signal must be a.c. coupled to these pins. The input signal is sampled on the rising and falling edge of CLK. See Section 1.1.2 for a description of acquiring the input and Section 2.3 for an overview of the clock inputs. 18 19 VIN+ VIN− Signal Input (Input):Analog The differential full-scale input range is 600 mVP-P when the FSR pin is low, or 820 mVP-P when the FSR pin is high. DCLK_RST+ DCLK_RST- Sample Clock Reset (Input):LVDS A positive pulse on this pin is used to reset and synchronize the DCLK outs of multiple converters. See Section 1.5 for detailed description. When bit 14 in the Configuration Register (address 1h) is set to 1b, these differential DCLK_RST pins are selected. VCMO Common Mode Voltage (Output):Analog - The voltage output at this pin is required to be the common mode input voltage at VIN+ and VIN− when d.c. coupling is used. This pin should be grounded when a.c. coupling is used at the analog input. This pin is capable of sourcing or sinking 100μA and can drive a load up to 80 pF. See Section 2.2. 127 22 23 7 www.national.com 4 Pin No. Symbol Equivalent Circuit Description 31 VBG Bandgap Output Voltage (Output):Analog - Capable of 100 μA source/sink and can drive a load up to 80 pF. 126 CalRun Calibration Running (Output):LVCMOS - This pin is at a logic high when calibration is running. 32 REXT External Bias Resistor Connection Analog - Nominal value is 3.3k-Ohms (±0.1%) to ground. See Section 1.1.1. 34 35 Tdiode_P Tdiode_N Temperature Diode Analog - Positive (Anode) and Negative (Cathode) for die temperature measurements. See Section 2.6.2. 5 www.national.com ADC083000 Pin Functions ADC083000 Pin Functions Pin No. Symbol 36 / 37 38 / 39 43 / 44 45 / 46 47 / 48 49 / 50 54 / 55 56 / 57 58 / 59 60 / 61 65 / 66 67 / 68 69 / 70 71 / 72 75 / 76 77 / 78 Da0+ / Da0Da1+ / Da1Da2+ / Da2− Da3+ / Da3Da4+ / Da4− Da5+ / Da5Da6+ / Da6− Da7+ / Da7Dc0+ / Dc0Dc1+ / Dc1Dc2+ / Dc2− Dc3+ / Dc3Dc4+ / Dc4− Dc5+ / Dc5Dc6+ / Dc6− Dc7+ / Dc7- A and C Data (Output):LVDS Data Outputs from the first internal converter. The data should be extracted in the order ABCD These outputs should always be terminated with a 100Ω differential resistor. 83 / 84 85 / 86 89 / 90 91 / 92 93 / 94 95 / 96 100 / 101 102 / 103 104 / 105 106 / 107 111 / 112 113 / 114 115 / 116 117 / 118 122 / 123 124 / 125 Dd7− / Dd7+ Dd6- / Dd6+ Dd5− / Dd5+ Dd4- / Dd4+ Dd3- / Dd3+ Dd2- / Dd2+ Dd1− / Dd1+ Dd0- / Dd0+ Db7- / Db7+ Db6- / Db6+ Db5− / Db5+ Db4- / Db4+ Db3− / Db3+ Db2- / Db2+ Db1− / Db1+ Db0- / Db0+ B and D Data (Output):LVDS Data Outputs from the second internal converter. The data should be extracted in the order ABCD These outputs should always be terminated with a 100Ω differential resistor. 79 80 82 81 www.national.com Equivalent Circuit Description OR+ OR- Out Of Range (Output):LVDS - A differential high at these pins indicates that the differential input is out of range (outside the range ±325 mV or ±435 mV as defined by the FSR pin). These outputs should always be terminated with a 100Ω differential resistor. DCLK+ DCLK- Differential Clock (Output):LVDS - The Differential Clock outputs are used to latch the output data. Delayed and non-delayed data outputs are supplied synchronous to this signal. DCLK is 1/2 the sample clock rate in SDR mode and 1/4 the sample clock rate in the DDR mode. These outputs should always be terminated with a 100Ω differential resistor. The DCLK outputs are not active during the calibration cycle depending on the setting of Configuration Register (address 1h), bit- 14 (RTD). DCLK is continuously present during the calibration cycle when bit-14 is set high (1b) and is not active during the calibration cycle when set low (0b). 6 ADC083000 Pin Functions Pin No. Symbol 2, 5, 8, 13, 16, 17, 20, 25, 28, 33, 128 Equivalent Circuit Description VA Analog power supply pins (Power) - Bypass these pins to ground. 40, 51 ,62, 73, 88, 99, 110, 121 VDR Output Driver power supply pins (Power) - Bypass these pins to DR GND. 1, 6, 9, 12, 21, 24, 27 GND (Gnd) - Ground return for VA. 42, 53, 64, 74, 87, 97, 108, 119 DR GND (Gnd) - Ground return for VDR. 29,41,52, 63, 98, 109, 120 NC No Connection Make no connection to these pins 7 www.national.com ADC083000 Absolute Maximum Ratings Operating Ratings (Notes 1, 2) −40°C ≤ TA ≤ +85°C (Notes 1, 2) Ambient Temperature Range If Military/Aerospace specified devices are required, please contact the National Semiconductor Sales Office/ Distributors for availability and specifications. Supply Voltage (VA) Driver Supply Voltage (VDR) Analog Input Common Mode Voltage VIN+, VIN- Voltage Range (Maintaining Common Mode) Ground Difference (|GND - DR GND|) CLK Pins Voltage Range Differential CLK Amplitude Supply Voltage (VA, VDR) Supply Difference VA - VDR Voltage on Any Input Pin Ground Difference |GND - DR GND| Input Current at Any Pin (Note 3) Package Input Current (Note 3) 2.2V 0V to -100mV −0.15V to (VA + 0.15V) Power Dissipation at TA ≤ 85°C ESD Susceptibility (Note 4) Human Body Model Machine Model Storage Temperature 0V to 100 mV ±25 mA ±50 mA +1.8V to +2.0V +1.8V to VA VCMO ±50mV 200mV to VA 0V 0V to VA 0.4VP-P to 2.0VP-P Package Thermal Resistance 2.3 W Package 2500V 250V −65°C to +150°C 128-Lead Exposed Pad LQFP θJA 26°C / W θJC (Top of θJ-PAD Package) (Thermal Pad) 10°C / W 2.8°C / W Soldering process must comply with National Semiconductor’s Reflow Temperature Profile specifications. Refer to www.national.com/packaging. (Note 5) Converter Electrical Characteristics The following specifications apply after calibration for VA = VDR = +1.9VDC, OutV = 1.9V, VIN FSR (a.c. coupled) = differential 820mVP-P, CL = 10 pF, Differential, a.c. coupled Sinewave Input Clock, fCLK = 1.5GHz at 0.5VP-P with 50% duty cycle, VBG = Floating, Non-Extended Control Mode, SDR Mode, REXT = 3300Ω ±0.1%, Analog Signal Source Impedance = 100Ω Differential. Boldface limits apply for TA = TMIN to TMAX. All other limits TA = 25°C, unless otherwise noted. (Notes 6, 7) Symbol Parameter Conditions Typical (Note 8) Limits (Note 8) Units (Limits) STATIC CONVERTER CHARACTERISTICS INL Integral Non-Linearity (Best fit) DC Coupled, 1MHz Sine Wave Over Ranged ±0.35 ±0.9 LSB (max) DNL Differential Non-Linearity DC Coupled, 1MHz Sine Wave Over Ranged ±0.20 ±0.6 LSB (max) 8 Bits Resolution with No Missing Codes VOFF Offset Error VOFF_ADJ Input Offset Adjustment Range PFSE Positive Full-Scale Error (Note 9) −1.6 ±25 mV (max) NFSE Negative Full-Scale Error (Note 9) −1.00 ±25 mV (max) FS_ADJ Full-Scale Adjustment Range ±20 ±15 %FS -0.20 Extended Control Mode Extended Control Mode LSB ±45 mV DYNAMIC CONVERTER CHARACTERISTICS FPBW B.E.R. Full Power Bandwidth SINAD GHz 10-18 Error/Sample d.c. to 750 MHz ±0.42 dBFS d.c. to 1500 MHz ±0.83 dBFS Bit Error Rate Gain Flatness ENOB 3 Effective Number of Bits Signal-to-Noise Plus Distortion Ratio www.national.com fIN = 373 MHz, VIN = FSR − 0.5 dB 7.2 6.85 Bits (min) fIN = 748 MHz, VIN = FSR − 0.5 dB 7.0 6.6 Bits (min) fIN = 1498 MHz, VIN = FSR − 0.5 dB 6.5 Bits fIN = 373 MHz, VIN = FSR − 0.5 dB 45 42.2 dB (min) fIN = 748 MHz, VIN = FSR − 0.5 dB 44.2 40.8 dB (min) fIN = 1498 MHz, VIN = FSR − 0.5 dB 41.1 8 dB SNR THD 2nd Harm 3rd Harm SFDR Typical (Note 8) Limits (Note 8) Units (Limits) fIN = 373 MHz, VIN = FSR − 0.5 dB 45.3 43.2 dB (min) fIN = 748 MHz, VIN = FSR − 0.5 dB 44.5 41.7 dB (min) fIN = 1498 MHz, VIN = FSR − 0.5 dB 41.8 fIN = 373 MHz, VIN = FSR − 0.5 dB -57 -49 dB (max) fIN = 748 MHz, VIN = FSR − 0.5 dB -56 -48 dB (max) fIN = 1498 MHz, VIN = FSR − 0.5 dB -49.5 dB fIN = 373 MHz, VIN = FSR − 0.5 dB −68 dB fIN = 748 MHz, VIN = FSR − 0.5 dB −66 dB fIN = 1498 MHz, VIN = FSR − 0.5 dB −56 dB fIN = 373 MHz, VIN = FSR − 0.5 dB −64 dB fIN = 748 MHz, VIN = FSR − 0.5 dB −58 dB fIN = 1498 MHz, VIN = FSR − 0.5 dB −52 fIN = 373 MHz, VIN = FSR − 0.5 dB 57 49 dB (min) fIN = 748 MHz, VIN = FSR − 0.5 dB 54.5 48 dB (min) fIN = 1498 MHz, VIN = FSR − 0.5 dB 52.0 Parameter Signal-to-Noise Ratio Total Harmonic Distortion Second Harmonic Distortion Third Harmonic Distortion Spurious-Free dynamic Range Conditions dB dB dB ANALOG INPUT AND REFERENCE CHARACTERISTICS FSR pin 14 Low VIN 600 Full Scale Analog Differential Input Range FSR pin 14 High 820 550 mVP-P (min) 650 mVP-P (max) 770 mVP-P (min) 870 mVP-P (max) VCMO − 50 VCMO + 50 mV (min) mV (max) VCMI Analog Input Common Mode Voltage CIN Analog Input Capacitance (Notes Differential 10, 11) Each input pin to ground 1.08 pF 2.2 pF RIN Differential Input Resistance 100 VCMO 95 Ω (min) 105 Ω (max) 0.95 1.45 V (min) V (max) ANALOG OUTPUT CHARACTERISTICS VCMO Common Mode Output Voltage ICMO = ±100 µA 1.26 VCMO_LVL VCMO input threshold to set DC Coupling mode VA = 1.8V 0.60 V VA = 2.0V 0.66 V TC VCMO Common Mode Output Voltage Temperature Coefficient TA = −40°C to +85°C 118 ppm/°C CLOAD VCMO Maximum VCMO load Capacitance VBG Bandgap Reference Output Voltage IBG = ±100 µA TC VBG Bandgap Reference Voltage Temperature Coefficient TA = −40°C to +85°C, IBG = ±100 µA CLOAD VBG Maximum Bandgap Reference load Capacitance 1.26 80 pF 1.20 1.33 V (min) V (max) 28 ppm/°C 80 9 pF www.national.com ADC083000 Symbol ADC083000 Symbol Parameter Conditions Typical (Note 8) Limits (Note 8) Units (Limits) TEMPERATURE DIODE CHARACTERISTICS ΔVBE Temperature Diode Voltage 192 µA vs. 12 µA, TJ = 25°C 71.23 mV 192 µA vs. 12 µA, TJ = 85°C 85.54 mV Sine Wave Clock 0.6 0.4 2.0 VP-P (min) VP-P (max) Square Wave Clock 0.6 0.4 2.0 VP-P (min) VP-P (max) LVDS INPUT CHARACTERISTICS VID II CIN Differential Clock Input Level Input Current Input Capacitance (Notes 10, 11) VIN = 0 or VIN = VA ±1 µA Differential 0.02 pF Each input to ground 1.5 pF Measured differentially, OutV = VA, VBG = Floating (Note 15) 680 LVDS OUTPUT CHARACTERISTICS VOD LVDS Differential Output Voltage Δ VO DIFF Change in LVDS Output Swing Between Logic Levels VOS Output Offset Voltage, see Figure 1 VOS Output Offset Voltage, see Figure 1 Δ VOS Output Offset Voltage Change Between Logic Levels IOS Output Short Circuit Current ZO Differential Output Impedance Measured differentially, OutV = GND, VBG = Floating (Note 15) 520 470 mVP-P (min) 920 mVP-P (max) 380 mVP-P (min) 720 mVP-P (max) ±1 mV VBG = Floating 800 mV VBG = VA (Note 15) 1150 mV ±1 mV ±4 mA 100 Ohms Output+ & Output- connected to 0.8V LVCMOS INPUT CHARACTERISTICS VIH Logic High Input Voltage (Note 12) VIL Logic Low Input Voltage (Note 12) CIN Input Capacitance (Notes 11, 13) Each input to ground 0.85 x VA V (min) 0.15 x VA V (max) 1.2 pF LVCMOS OUTPUT CHARACTERISTICS VOH CMOS H level output IOH = -400uA (Note 12) 1.65 1.5 V VOL CMOS L level output IOH = 400uA (Note 12) 0.15 0.3 V POWER SUPPLY CHARACTERISTICS IA Analog Supply Current PD = Low 734 810 mA (max) IDR Output Driver Supply Current PD = Low 300 410 mA (max) PD = Low 1.9 2.3 W (max) PD = High 25 PD Power Consumption www.national.com 10 mW Parameter Conditions Typical (Note 8) Limits (Note 8) Units (Limits) 1.5 GHz (min) AC ELECTRICAL CHARACTERISTICS - Sampling Clock fCLK1 Maximum Input Clock Frequency Sampling rate is 2x clock input 1.7 fCLK2 Minimum Input Clock Frequency Sampling rate is 2x clock input 500 tCYC Input Clock Duty Cycle 500MHz ≤ Input clock frequency ≤ 1.5 GHz (Note 12) 50 20 80 % (min) % (max) tLC Input Clock Low Time (Note 11) 333 133 ps (min) tHC Input Clock High Time (Note 11) 333 133 ps (min) 45 55 % (min) % (max) MHz DCLK Duty Cycle (Note 11) 50 tAD Sampling (Aperture) Delay Input CLK transition to Acquisition of Data 1.4 ns tAJ Aperture Jitter 0.55 ps rms tOD Input Clock to Data Output Delay 50% of Input Clock transition to 50% of (in addition to Pipeline Delay) Data transition 3.7 ns Dd Outputs Pipeline Delay (Latency) (Notes 11, 14) 13 Db Outputs 14 Dc Outputs 13.5 Da Outputs 14.5 Input Clock Cycles AC ELECTRICAL CHARACTERISTICS - Output Clock and Data (Note 16) tLHT LH Transition Time - Differential 10% to 90% 150 ps tHLT HL Transition Time - Differential 10% to 90% 150 ps tSKEWO DCLK to Data Output Skew 50% of DCLK transition to 50% of Data transition, SDR Mode and DDR Mode, 0° DCLK (Note 11) ±50 ps (max) tOSU Data to DCLK Set-Up Time DDR Mode, 90° DCLK (Note 12) 570 ps tOH DCLK to Data Hold Time DDR Mode, 90° DCLK (Note 12) 555 ps AC ELECTRICAL CHARACTERISTICS - Serial Interface Clock fSCLK Serial Clock Frequency (Note 11) 67 MHz tSS Data to Serial Clock Setup Time (Note 11) 2.5 ns (min) tHS Data to Serial Clock Hold Time 1 ns (min) (Note 11) Serial Clock Low Time 6 ns (min) Serial Clock High Time 6 ns (min) AC ELECTRICAL CHARACTERISTICS - General Signals tSR Setup Time DCLK_RST± tHR Hold Time DCLK_RST± tPWR Pulse Width DCLK_RST± (Note 11) tWU PD low to Rated Accuracy Conversion (Wake-Up Time) (Note 11) tCAL Calibration Cycle Time tCAL_L CAL Pin Low Time See Figure 8 (Note 11) 80 CLK± Cyc. (min) tCAL_H CAL Pin High Time See Figure 8 (Note 11) 80 CLK± Cyc.(min) tCalDly Calibration delay determined by pin 127 See Section 1.1.1, Figure 8, (Note 11) 225 CLK± Cyc.(min) tCalDly Calibration delay determined by pin 127 See Section 1.1.1, Figure 8, (Note 11) 231 CLK± Cyc.(max) (Note 12) 90 ps 30 ps 4 11 CLK± Cyc. (min) 1 µs 1.4 x 105 CLK± Cyc. www.national.com ADC083000 Symbol ADC083000 Note 1: Absolute Maximum Ratings indicate limits beyond which damage to the device may occur. There is no guarantee of operation at the Absolute Maximum Ratings. Operating Ratings indicate conditions for which the device is functional, but do not guarantee specific performance limits. For guaranteed specifications and test conditions, see the Electrical Characteristics. The guaranteed specifications apply only for the test conditions listed. Some performance characteristics may degrade when the device is not operated under the listed test conditions. Note 2: All voltages are measured with respect to GND = DR GND = 0V, unless otherwise specified. Note 3: When the input voltage at any pin exceeds the power supply limits (that is, less than GND or greater than VA), the current at that pin should be limited to 25 mA. The 50 mA maximum package input current rating limits the number of pins that can safely exceed the power supplies with an input current of 25 mA to two. This limit is not placed upon the power, ground and digital output pins. Note 4: Human body model is 100 pF capacitor discharged through a 1.5 kΩ resistor. Machine model is 220 pF discharged through ZERO Ohms. Note 5: Reflow temperature profiles are different for lead-free and non-lead-free packages. Note 6: The analog inputs are protected as shown below. Input voltage magnitudes beyond the Absolute Maximum Ratings may damage this device. 20193204 Note 7: To guarantee accuracy, it is required that VA and VDR be well bypassed. Each supply pin must be decoupled with separate bypass capacitors. Additionally, achieving rated performance requires that the backside exposed pad be well grounded. Note 8: Typical figures are at TA = 25°C, and represent most likely parametric norms. Test limits are guaranteed to National's AOQL (Average Outgoing Quality Level). Note 9: Calculation of Full-Scale Error for this device assumes that the actual reference voltage is exactly its nominal value. Full-Scale Error for this device, therefore, is a combination of Full-Scale Error and Reference Voltage Error. See Figure 2. For relationship between Gain Error and Full-Scale Error, see Specification Definitions for Gain Error. Note 10: The analog and clock input capacitances are die capacitances only. Additional package capacitances of 0.65 pF differential and 0.95 pF each pin to ground are isolated from the die capacitances by lead and bond wire inductances. Note 11: This parameter is guaranteed by design and is not tested in production. Note 12: This parameter is guaranteed by design and/or characterization and is not tested in production. Note 13: The digital control pin capacitances are die capacitances only. Additional package capacitance of 1.6 pF each pin to ground are isolated from the die capacitances by lead and bond wire inductances. Note 14: Each of the two converters of the ADC083000 has two LVDS output buses, which each clock data out at one quarter the sample rate. Bus Db has a pipeline latency that is one Input Clock cycle less than the latency of bus Dd. Likewise, bus Da has a pipeline latency that is one Input Clock cycle less than the latency of bus Dc. Note 15: Tying VBG to the supply rail will increase the output offset voltage (VOS) by 400mv (typical), as shown in the VOS specification above. Tying VBG to the supply rail will also affect the differential LVDS output voltage (VOD), causing it to increase by 40mV (typical). Note 16: All parameters are measured through a transmission line and 100Ω termination using a 0.33pF load oscilloscope probe. FULL POWER BANDWIDTH (FPBW) is a measure of the frequency at which the reconstructed output fundamental drops 3 dB below its low frequency value for a full scale input. GAIN ERROR is the deviation from the ideal slope of the transfer function. It can be calculated from Offset and FullScale Errors: Positive Gain Error = Offset Error − Positive Full-Scale Error Negative Gain Error = −(Offset Error − Negative FullScale Error) Gain Error = Negative Full-Scale Error − Positive FullScale Error = Positive Gain Error + Negative Gain Error INTEGRAL NON-LINEARITY (INL) is a measure of the deviation of each individual code from a straight line through the input to output transfer function. The deviation of any given code from this straight line is measured from the center of that code value. The best fit method is used. INTERMODULATION DISTORTION (IMD) is the creation of additional spectral components as a result of two sinusoidal frequencies being applied to the ADC input at the same time. It is defined as the ratio of the power in the second and third order intermodulation products to the power in one of the original frequencies. IMD is usually expressed in dBFS. Specification Definitions APERTURE (SAMPLING) DELAY is the amount of delay, measured from the sampling edge of the Clock input, after which the signal present at the input pin is sampled inside the device. APERTURE JITTER (tAJ) is the variation in aperture delay from sample to sample. Aperture jitter shows up as input noise. Bit Error Rate (B.E.R.) is the probability of error and is defined as the probable number of errors per unit of time divided by the number of bits seen in that amount of time. A B.E.R. of 10-18 corresponds to a statistical error in one bit about every four (4) years. CLOCK DUTY CYCLE is the ratio of the time that the clock wave form is at a logic high to the total time of one clock period. DIFFERENTIAL NON-LINEARITY (DNL) is the measure of the maximum deviation from the ideal step size of 1 LSB. Measured at 3 GSPS with a ramp input. EFFECTIVE NUMBER OF BITS (ENOB, or EFFECTIVE BITS) is another method of specifying Signal-to-Noise and Distortion Ratio, or SINAD. ENOB is defined as (SINAD − 1.76) / 6.02 and says that the converter is equivalent to a perfect ADC of this (ENOB) number of bits. www.national.com 12 POSITIVE FULL-SCALE ERROR (PFSE) is a measure of how far the last code transition is from the ideal 1-1/2 LSB below a differential +VIN / 2. For the ADC083000 the reference voltage is assumed to be ideal, so this error is a combination of full-scale error and reference voltage error. POWER SUPPLY REJECTION RATIO (PSRR) can be one of two specifications. PSRR1 (DC PSRR) is the ratio of the change in full-scale error that results from a power supply voltage change from 1.8V to 2.0V. PSRR2 (AC PSRR) is a measure of how well an a.c. signal riding upon the power supply is rejected from the output and is measured with a 248 MHz, 50 mVP-P signal riding upon the power supply. It is the ratio of the output amplitude of that signal at the output to its amplitude on the power supply pin. PSRR is expressed in dB. SIGNAL TO NOISE RATIO (SNR) is the ratio, expressed in dB, of the rms value of the input signal at the output to the rms value of the sum of all other spectral components below onehalf the sampling frequency, not including harmonics or d.c. SIGNAL TO NOISE PLUS DISTORTION (S/(N+D) or SINAD) is the ratio, expressed in dB, of the rms value of the input signal at the output to the rms value of all of the other spectral components below half the input clock frequency, including harmonics but excluding d.c. SPURIOUS-FREE DYNAMIC RANGE (SFDR) is the difference, expressed in dB, between the rms values of the input signal at the output and the peak spurious signal, where a spurious signal is any signal present in the output spectrum that is not present at the input, excluding d.c. TOTAL HARMONIC DISTORTION (THD) is the ratio expressed in dB, of the rms total of the first nine harmonic levels at the output to the level of the fundamental at the output. THD is calculated as VFS / 2n where VFS is the differential full-scale amplitude of VIN as set by the FSR input (pin-14) and "n" is the ADC resolution in bits, which is 8 for the ADC083000. LVDS DIFFERENTIAL OUTPUT VOLTAGE (VOD) is the absolute value of the difference between the VD+ & VD- outputs; each measured with respect to Ground. 20193246 FIGURE 1. LVDS OUTPUT OFFSET VOLTAGE (VOS) is the midpoint between the D+ and D- pins output voltage; ie., [(VD+) + ( VD-)]/2. MISSING CODES are those output codes that are skipped and will never appear at the ADC outputs. These codes cannot be reached with any input value. MSB (MOST SIGNIFICANT BIT) is the bit that has the largest value or weight. Its value is one half of full scale. NEGATIVE FULL-SCALE ERROR (NFSE) is a measure of how far the last code transition is from the ideal 1/2 LSB above a differential -VIN / 2. For the ADC083000 the reference voltage is assumed to be ideal, so this error is a combination of full-scale error and reference voltage error. OFFSET ERROR (VOFF) is a measure of how far the midscale point is from the ideal zero voltage differential input. Offset Error = Actual Input causing average of 8k samples to result in an average code of 127.5. OUTPUT DELAY (tOD) is the time delay from 50% point of the input clock transition, CLK, to the 50% point of the updated data transition at the output pins. OVER-RANGE RECOVERY TIME is the time required after the differential input voltages goes from ±1.2V to 0V for the converter to recover and make a conversion with its rated accuracy. PIPELINE DELAY (LATENCY) is the number of input clock cycles between initiation of conversion and when that data is presented to the output driver stage. New data is available at every clock cycle, but the data lags the conversion by the Pipeline Delay plus the tOD. where Af1 is the RMS power of the fundamental (output) frequency and Af2 through Af10 are the RMS power of the first 9 harmonic frequencies in the output spectrum. – Second Harmonic Distortion (2nd Harm) is the difference, expressed in dB, between the RMS power in the input frequency seen at the output and the power in its 2nd harmonic level at the output. – Third Harmonic Distortion (3rd Harm) is the difference expressed in dB between the RMS power in the input frequency seen at the output and the power in its 3rd harmonic level at the output. 13 www.national.com ADC083000 LSB (LEAST SIGNIFICANT BIT) is the bit that has the smallest value or weight of all bits. This value is ADC083000 Transfer Characteristic 20193222 FIGURE 2. Input / Output Transfer Characteristic www.national.com 14 ADC083000 Timing Diagrams 20193214 FIGURE 3. ADC083000 Timing — SDR Clocking 20193259 FIGURE 4. ADC083000 Timing — DDR Clocking 15 www.national.com ADC083000 20193220 FIGURE 5. Clock Reset Timing in DDR Mode 20193223 FIGURE 6. Clock Reset Timing in SDR Mode with OUTEDGE Low 20193224 FIGURE 7. Clock Reset Timing in SDR Mode with OUTEDGE High www.national.com 16 ADC083000 20193225 FIGURE 8. Calibration and On-Command Calibration Timing 20193219 FIGURE 9. Serial Interface Timing 17 www.national.com ADC083000 Typical Performance Characteristics VA=VDR=1.9V, FCLK=1500MHz, TA=25°C unless otherwise stated. DNL vs. TEMPERATURE POWER DISSIPATION vs. SAMPLE RATE 20193268 20193269 ENOB vs. TEMPERATURE ENOB vs. SUPPLY VOLTAGE 20193270 20193271 ENOB vs. SAMPLE RATE ENOB vs. INPUT FREQUENCY 20193272 www.national.com 20193273 18 ADC083000 SNR vs. TEMPERATURE SNR vs. SUPPLY VOLTAGE 20193274 20193275 SNR vs. SAMPLE RATE SNR vs. INPUT FREQUENCY 20193276 20193277 THD vs. TEMPERATURE THD vs. SUPPLY VOLTAGE 20193278 20193279 19 www.national.com ADC083000 THD vs. SAMPLE RATE THD vs. INPUT FREQUENCY 20193280 20193281 SFDR vs. TEMPERATURE SFDR vs. SUPPLY VOLTAGE 20193282 20193283 SFDR vs. SAMPLE RATE SFDR vs. INPUT FREQUENCY 20193284 www.national.com 20193285 20 Spectral Response at FIN = 1497 MHZ 20193287 20193288 FULL POWER BANDWIDTH 20193290 21 www.national.com ADC083000 Spectral Response at FIN = 748 MHZ ADC083000 In normal operation, calibration is performed just after application of power and whenever a valid calibration command is given, which is holding the CAL pin low for at least 80 input clock cycles, then hold it high for at least another 80 input clock cycles. The time taken by the calibration procedure is specified in the A.C. Characteristics Table. Holding the CAL pin high upon power up will prevent the calibration process from running until the CAL pin experiences the above-mentioned 80 input clock cycles low followed by 80 cycles high. CalDly (pin 127) is used to select one of two delay times after the application of power to the start of calibration. This calibration delay is 225 input clock cycles (about 22 ms at 3 GSPS) with CalDly low, or 231 input clock cycles (about 1.4 seconds at 3 GSPS) with CalDly high. These delay values allow the power supply to come up and stabilize before calibration takes place. If the PD pin is high upon power-up, the calibration delay counter will be disabled until the PD pin is brought low. Therefore, holding the PD pin high during power up will further delay the start of the power-up calibration cycle. The best setting of the CalDly pin depends upon the poweron settling time of the power supply. 1.0 Functional Description The ADC083000 is a versatile A/D Converter with an innovative architecture permitting very high speed operation. The controls available ease the application of the device to circuit solutions. Optimum performance requires adherence to the provisions discussed here and in the Applications Information Section. While it is generally poor practice to allow an active pin to float, pins 4 and 14 of the ADC083000 are designed to be left floating without jeopardy. In all discussions throughout this data sheet, whenever a function is called by allowing a control pin to float, connecting that pin to a potential of one half the VA supply voltage will have the same effect as allowing it to float. 1.1 OVERVIEW The ADC083000 uses a calibrated folding and interpolating architecture that achieves over 7.2 effective bits. The use of folding amplifiers greatly reduces the number of comparators and power consumption. Interpolation reduces the number of front-end amplifiers required, minimizing the load on the input signal and further reducing power requirements. In addition to other things, on-chip calibration reduces the INL bow often seen with folding architectures. The result is an extremely fast, high performance, low power converter. The analog input signal that is within the converter's input voltage range is digitized to eight bits at speeds of 1.0 GSPS to 3.0 GSPS, typical. Differential input voltages below negative full-scale will cause the output word to consist of all zeroes. Differential input voltages above positive full-scale will cause the output word to consist of all ones. Either of these conditions at the analog input will cause the OR (Out of Range) output to be activated. This single OR output indicates when the output code from the converter is below negative full scale or above positive full scale. The ADC083000 demultiplexes the data at 1:4 and is output on all four output busses at a quarter of the ADC sampling rate. The outputs must be interleaved by the user to provide output words at the full conversion rate. The output levels may be selected to be normal or reduced voltage. Using reduced levels saves power but could result in erroneous data capture of some or all of the bits, especially at higher sample rates and in marginally designed systems. Calibration Operation Notes: • During the calibration cycle, the OR output may be active as a result of the calibration algorithm. All data on the output pins and the OR output are invalid during the calibration cycle. • During the power-up calibration and during the oncommand calibration when Resistor Trim Disable (address: 1h, bit 13) is not active (0b) , all clocks are halted on chip, including internal clocks and DCLK, while the input termination resistor is trimmed to a value that is equal to REXT / 33. This is to reduce noise during the input resistor calibration portion of the calibration cycle. See Section 2.4.2.2 for information on maintaining DCLK operation during on-command calibration. This external resistor is located between pin 32 and ground. REXT must be 3300 Ω ±0.1%. With this value, the input termination resistor is trimmed to be 100 Ω. Because REXT is also used to set the proper current for the Track and Hold amplifier, for the preamplifiers and for the comparators, other values of REXT should not be used. • The CalRun output is high whenever the calibration procedure is running. This is true whether the calibration is done at power-up or on-command. 1.1.1 Calibration A calibration is performed upon power-up and can also be invoked by the user upon command. Calibration trims the 100Ω analog input differential termination resistor and minimizes full-scale error, offset error, DNL and INL, resulting in maximizing SNR, THD, SINAD (SNDR) and ENOB. Internal bias currents are also set with the calibration process. All of this is true whether the calibration is performed upon power up or is performed upon command. Running the calibration is an important part of this chip's functionality and is required in order to obtain adequate performance. In addition to the requirement to be run at power-up, calibration must be re-run by the user whenever the state of the FSR pin is changed. For best performance, we recommend an on command calibration be run after initial power up and the device has reached a stable temperature. Also, we recommend that an on command calibration be run whenever the operating temperature changes significantly relative to the specific system performance requirements. See Section 2.4.2.2 for more information. Calibration can not be initiated or run while the device is in the power-down mode. See Section 1.1.7 for information on the interaction between Power Down and Calibration. www.national.com 1.1.2 Acquiring the Input Data is acquired at the rising and falling edge of CLK+ (pin 10) and the digital equivalent of that data is available at the digital outputs 13 input clock cycles later for the Dd output bus, 13.5 input clock cycles later for Dc output bus, 14 input clock cycles later for the Db output bus and 14.5 input clock cycles later for the Da output bus. See Table 1. There is an additional internal delay called tOD before the data is available at the outputs. See Figure 3 and Figure 4. The ADC083000 will convert as long as the input clock signal is present. The fully differential comparator design and the innovative design of the sample-and-hold amplifier, together with calibration, enables a very flat SINAD/ENOB response beyond 1.5 GHz. The ADC083000 output data signaling is LVDS and the output format is offset binary. 1.1.3 Control Modes Much of the user control can be accomplished with several control pins that are provided. Examples include initiation of the calibration cycle, power down mode and full scale range setting. However, the ADC083000 also provides an Extended 22 scribes the use of the clock input pins. A differential LVDS output clock (DCLK) is available for use in latching the ADC output data into whatever device is used to receive the data. The ADC083000 offers options for CLK+/CLK- and DCLK clocking. For DCLK, the clock edge on which output data transitions, and a choice of Single Data Rate (SDR) or Double Data Rate (DDR) outputs are available. The sampling clock CLK has optional duty cycle correction as part of its circuit. This feature is enabled by default and provides improved ADC clocking. This circuitry allows the ADC to be clocked with a signal source having a duty cycle ratio of 80 / 20 % (worst case). 1.1.4 The Analog Inputs The ADC083000 must be driven with a differential input signal. Operation with a single-ended signal is not recommended. It is important that the input signals are either a.c. coupled to the inputs with the VCMO pin grounded, or d.c. coupled with the VCMO pin left floating. An input common mode voltage equal to the VCMO output must be provided when d.c. coupling is used. Two full-scale range settings are provided with pin 14 (FSR). A high on pin 14 causes an input full-scale range setting of 820 mVP-P, while grounding pin 14 causes an input full-scale range setting of 600 mVP-P. In the Extended Control mode, the full-scale input range can be set to values between 560 mVP-P and 840 mVP-P through a serial interface. See Section 2.2 1.1.5.1 Output Demultiplexer The ADC083000 utilizes both the rising and falling edge of the input clock resulting in the overall sample rate being twice the input clock frequency, or 3GSPS with a 1.5 GHz input clock. The demultiplexer outputs data on each of the four output busses at 750MHz with a 1.5GHz input clock. All data is available in parallel at the output. The four bytes of parallel data that are output with each clock is in the following sampling order, from the earliest to the latest: Da, Db, Dc, Dd. Table 1 indicates what the outputs represent for the various sampling possibilities. The ADC083000 includes an automatic clock phase background calibration feature which automatically and continuously adjusts the phase of the ADC input clock. This feature removes the need to manually adjust the clock phase and provides optimal ENOB performance. 1.1.5 Clocking The ADC083000 sampling clock (CLK+/CLK-) must be driven with an a.c. coupled, differential clock signal. Section 2.3 de- TABLE 1. Input Channel Samples Produced at Data Outputs Data Outputs* Input/Output Relationship Dd ADC1 sampled with the fall of CLK, 13 cycles earlier Db ADC1 sampled with the fall of CLK, 14 cycles earlier Dc ADC2 sampled with the rise of CLK, 13.5 cycles earlier Da ADC2 sampled with the rise of CLK, 14.5 cycles earlier * Always sourced with respect to fall of DCLK lines and low noise systems, satisfactory performance may be realized with the OutV input low, which results in lower power consumption. If the LVDS lines are long and/or the system in which the ADC083000 is used is noisy, it may be necessary to tie the OutV pin high. The LVDS data outputs have a typical common mode voltage of 800mV when the VBG pin is unconnected and floating. This common mode voltage can be increased to 1.150V by tying the VBG pin to VA if a higher common mode is required. IMPORTANT NOTE: Tying the VBG pin to VA will also increase the differential LVDS output voltage (VOD) by up to 40mV. 1.1.5.2 OutEdge Setting To help ease data capture in the SDR mode, the output data may be caused to transition on either the positive or the negative edge of the output data clock (DCLK). This is chosen with the OutEdge input (pin 4). A high on the OutEdge input pin causes the output data to transition on the rising edge of DCLK, while grounding this input causes the output to transition on the falling edge of DCLK. See Section 2.4.3. 1.1.5.3 Double Data Rate A choice of single data rate (SDR) or double data rate (DDR) output is offered. When the device is in DDR mode, address 1h, bit-8 must be set to 0b. With single data rate the output clock (DCLK) frequency is the same as the data rate of the two output buses. With double data rate the DCLK frequency is half the data rate and data is sent to the outputs on both edges of DCLK. DDR clocking is enabled in non-Extended Control mode by allowing pin 4 to float. 1.1.7 Power Down The ADC083000 is in the active state when the Power Down pin (PD) is low. When the PD pin is high, the device is in the power down mode. In this power down mode the data output pins (positive and negative) including DCLK+/- and OR +/- are put into a high impedance state and the devices power consumption is reduced to a minimal level. If the PD input is brought high while a calibration is running, the device will not go into power down until the calibration sequence is complete. However, if power is applied and PD 1.1.6 The LVDS Outputs The data outputs, the Out Of Range (OR) and DCLK, are LVDS. Output current sources provide 3 mA of output current to a differential 100 Ohm load when the OutV input (pin 3) is high or 2.2 mA when the OutV input is low. For short LVDS 23 www.national.com ADC083000 Control mode whereby a serial interface is used to access register-based control of several advanced features. The Extended Control mode is not intended to be enabled and disabled dynamically. Rather, the user is expected to employ either the normal control mode or the Extended Control mode at all times. When the device is in the Extended Control mode, pin-based control of several features is replaced with registerbased control and those pin-based controls are disabled. These pins are OutV (pin 3), OutEdge/DDR (pin 4), FSR (pin 14). See Section 1.2 for details on the Extended Control mode. ADC083000 is already high, the device will not begin the calibration sequence until the PD input goes low. If a manual calibration is requested while the device is powered down, the calibration will not begin at all. That is, the manual calibration input is completely ignored in the power down state. control of the device through several control pins. The "extended control mode" provides additional configuration and control options through a serial interface and a set of 6 registers. The two control modes are selected with pin 14 (FSR/ ECE: Extended Control Enable). The choice of control modes is required to be a fixed selection and is not intended to be switched dynamically while the device is operational. Table 2 shows how several of the device features are affected by the control mode chosen. 1.2 NORMAL/EXTENDED CONTROL The ADC083000 may be operated in one of two modes. In the Normal Mode, the user affects available configuration and TABLE 2. Features and Modes Feature Normal Control Mode Extended Control Mode SDR or DDR Clocking Selected with pin 4 Selected with nDE in the Configuration Register (1h; bit-10). When the device is in DDR mode, address 1h, bit-8 must be set to 0b. DDR Clock Phase Not Selectable (0° Phase Only) Selected with DCP in the Configuration Register (1h; bit-11). SDR Data transitions with rising or falling Selected with pin 4 DCLK edge Selected with OE in the Configuration Register (1h; bit-8). LVDS output level Selected with pin 3 Selected with the OV in the Configuration Register (1h; bit-9). Power-On Calibration Delay Delay Selected with pin 127 Short delay only. Full-Scale Range Options (600 mVP-P or 820 mVP-P) selected with pin 14. Selected range applies to both channels. Up to 512 step adjustments over a nominal range of 560 mV to 840 mV in the Configuration Register (3h; bits-7 thru 15). Input Offset Adjust Not possible Up to ±45 mV adjustments in 512 steps in the Configuration Register (2h; bits-7 thru 15). Sampling Clock Phase Adjustment The Clock Phase is adjusted automatically The clock phase can be adjusted manually through the Fine & Coarse registers (Dh and Eh). Test Pattern Not Possible A test pattern can be made present at the data outputs by selecting TPO in the Configuration Register (Fh; bit-11). Resistor Trim Disable Not possible The DCLK outputs will continuously be present when RTD is selected in the Configuration Register (1h; bit-13) www.national.com 24 TABLE 3. Extended Control Mode Operation (Pin 14 Floating) Feature Extended Control Mode Default State SDR or DDR Clocking DDR Clocking DDR Clock Phase Data changes with DCLK edge (0° phase) LVDS Output Amplitude Normal amplitude (710 mVP-P) Calibration Delay Short Delay Full-Scale Range 700 mV nominal for both channels No adjustment for either channel TABLE 4. Register Addresses Input Offset Adjust Resistor Trim Disable Trim enabled, DCLK not continuously present at output Loading Sequence: A3 loaded after H0, A0 loaded last A3 A2 A1 A0 Hex Register Addressed Test Pattern Not present at output 0 0 0 0 0h Reserved 0 0 0 1 1h Configuration 0 0 1 0 2h Offset 0 0 1 1 3h Full-Scale Voltage Adjust 0 1 0 0 4h Reserved 0 1 0 1 5h Reserved 0 1 1 0 6h Reserved 0 1 1 1 7h Reserved 1 0 0 0 8h Reserved 1 0 0 1 9h Reserved 1 0 1 0 Ah Reserved 1 0 1 1 Bh Reserved 1 1 0 0 Ch Reserved 1 1 0 1 Dh Extended Clock Phase adjust fine 1 1 1 0 Eh Extended Clock Phase adjust coarse 1 1 1 1 Fh Test Pattern 4-Bit Address 1.3 THE SERIAL INTERFACE The 3-pin serial interface is enabled only when the device is in the Extended Control mode. The pins of this interface are Serial Clock (SCLK), Serial Data (SDATA) and Serial Interface Chip Select (SCS) Eight write only registers are accessible through this serial interface. SCS: This signal should be asserted low while accessing a register through the serial interface. Setup and hold times with respect to the SCLK must be observed. SCLK: Serial data input is accepted with the rising edge of this signal. SDATA: Each register access requires a specific 32-bit pattern at this input. This pattern consists of a header, register address and register value. The data is shifted in MSB first. Setup and hold times with respect to the SCLK must be observed. SeeFigure 9. Each Register access consists of 32 bits, as shown in Figure 9 of the Timing Diagrams. The fixed header pattern is 0000 0000 0001 (eleven zeros followed by a 1). The loading sequence is such that a "0" is loaded first. These 12 bits form the header. The next 4 bits are the address of the register that 25 www.national.com ADC083000 is to be written to and the last 16 bits are the data written to the addressed register. The addresses of the various registers are indicated in Table 4. Refer to the Register Description (Section 1.4) for information on the data to be written to the registers. Subsequent register accesses may be performed immediately, starting with the 33rd SCLK. This means that the SCS input does not have to be de-asserted and asserted again between register addresses. It is possible, although not recommended, to keep the SCS input permanently enabled (at a logic low) when using extended control. IMPORTANT NOTE: The Serial Interface should not be used when calibrating the ADC. Doing so will impair the performance of the device until it is re-calibrated correctly. Programming the serial registers will also reduce dynamic performance of the ADC for the duration of the register access time. The default state of the Extended Control Mode is set upon power-on reset (internally performed by the device) and is shown in Table 3. ADC083000 1.4 REGISTER DESCRIPTION Eight write-only registers provide several control and configuration options in the Extended Control Mode. These registers have no effect when the device is in the Normal Control Mode. Each register description below also shows the PowerOn Reset (POR) state of each control bit. Bit 9 Configuration Register Addr: 1h (0001b) W only (0xB2DF) D15 D14 D13 D12 D11 D10 D9 D8 1 DRE RTD DCS DCP nDE OV OE D7 D6 D5 D4 D3 D2 D1 D0 1 1 1 1 1 1 1 1 Bit 15 Bit 14 Bit 13 Bit 12 Bit 11 Bit 10 www.national.com Bit 8 Must be set to 1b DRE: Differential Reset Enable. When this bit is set to 0b , it enables the single-ended DCLK_RST input. When this bit is set to 1b , it enables the differential DCLK_RST input. POR State: 0b RTD: Resistor Trim Disable. When this bit is set to 1b, the input termination resistor is not trimmed during the calibration cycle and the DCLK output remains enabled. Note that the ADC is calibrated regardless of this setting. POR State: 0b DCS: Duty Cycle Stabilizer. When this bit is set to 1b , a duty cycle stabilization circuit is applied to the clock input. When this bit is set to 0b the stabilization circuit is disabled. POR State: 1b DCP: DDR Clock Phase. This bit only has an effect in the DDR mode. When this bit is set to 0b, the DCLK edges are time-aligned with the data bus edges ("0° Phase"). When this bit is set to 1b, the DCLK edges are placed in the middle of the data bit-cells ("90° Phase"), using the one-half speed DCLK shown in Figure 4 as the phase reference. POR State: 0b nDE: DDR Enable. When this bit is set to 0b, data bus clocking follows the DDR (Dual Data Rate) mode whereby a data word is output with each rising and falling edge of DCLK. When the device is in DDR mode, address 1h, bit-8 must be set to 0b. When this bit is set to a 1b, data bus clocking follows the SDR (single data rate) mode whereby each data word is output with either the rising or falling edge of DCLK , as determined by the OutEdge bit. POR State: 0b Bits 7:0 OV: Output Voltage. This bit determines the LVDS outputs' voltage amplitude and has the same function as the OutV pin that is used in the normal control mode. When this bit is set to 1b, the standard output amplitude of 680 mVP-P is used. When this bit is set to 0b, the reduced output amplitude of 520 mVP-P is used. POR State: 1b OE: Output Edge. This bit has two functions. When the device is in SDR mode, this bit selects the DCLK edge with which the data words transition and has the same effect as the OutEdge pin in the normal control mode. When this bit is set to 1b, the data outputs change with the rising edge of DCLK+. When this bit is set to 0b, the data output changes with the falling edge of DCLK+. When the device is in DDR mode, this bit must be set to 0b. POR State: 0b Must be set to 1b Offset Adjust Addr: 2h (0010b) D15 D14 D13 (MSB) D12 D11 D10 D9 Offset Value D8 (LSB) D7 D6 D5 D4 D3 D2 D1 D0 Sign 1 1 1 1 1 1 1 Bits 15:8 Bit 7 Bit 6:0 26 W only (0x007F) Offset Value. The input offset of the ADC is adjusted linearly and monotonically by the value in this field. 00h provides a nominal zero offset, while FFh provides a nominal 45 mV of offset. Thus, each code step provides 0.176 mV of offset. POR State: 0000 0000 b (no adjustment) Sign bit. 0b gives positive offset, 1b gives negative offset. POR State: 0b Must be set to 1b D15 D14 D13 Addr: Eh (1110b) W only (0x807F) D12 (MSB) D11 D10 D9 D8 D15 Adjust Value D14 D13 ENA W only (0x07FF) D12 D11 CAM D10 D9 D8 LFS 1 1 D7 D6 D5 D4 D3 D2 D1 D0 D7 D6 D5 D4 D3 D2 D1 D0 (LSB) 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Bit 15:7 Bits 6:0 Full Scale Voltage Adjust Value. The input fullscale voltage or gain of the ADC is adjusted linearly and monotonically with a 9 bit data value. The adjustment range is ±20% of the nominal 700 mVP-P differential value. Bit 15 Bit 14:11 0000 0000 0 560mVP-P Bit 10 1000 0000 0 Default Value 1111 1111 1 700mVP-P 840mVP-P For best performance, it is recommended that the value in this field be limited to the range of 0110 0000 0b to 1110 0000 0b. i.e., limit the amount of adjustment to ±15%. The remaining ±5% headroom allows for the ADC's own full scale variation. A gain adjustment does not require ADC re-calibration. POR State: 1000 0000 0b Must be set to 1b Bits 9:0 Test Pattern Register Addr: Fh (1111b) Extended Clock Phase Adjust Fine Addr: Dh (1101b) D15 D14 D13 D12 D11 D10 D9 D8 FAM D7 D6 D5 D4 D3 D2 D1 D0 (LSB) 1 1 1 1 1 1 1 Bit 15:7 Bit 6:0 W only (0xF7FF) D15 D14 D13 D12 D11 D10 D9 D8 1 1 1 1 TPO 1 1 1 D7 D6 D5 D4 D3 D2 D1 D0 1 1 1 1 1 1 1 1 Bits 15:12 Must be set to 1b Bit 11 TPO: Test Pattern Output enable. When this bit is set 1b, the ADC is disengaged and a test pattern generator is connected to the outputs including OR. This test pattern will work with the device in the SDR and DDR modes. POR State: 0b Bit 10:0 Must be set to 1b W only (0x3FFF) (MSB) Enable, default is 0b Coarse Adjust Magnitude. Each LSB results in approximately 70ps of clock adjust. Refer to Section 2.3.1. POR State: 0000b Low Frequency Sample clock. When this bit is set 1b, the dynamic performance of the device is improved when the sample clock is less than 900MHz. POR State: 0b Must be set to 1b Fine Adjust Magnitude. With all bits set, total adjust = 110ps of non-linear clock adjust. Refer to Section 2.3.1. POR State: 000 0000 0b Must be set to 1b 27 www.national.com ADC083000 Extended Clock Phase Adjust Coarse Full-Scale Voltage Adjust Addr: 3h (0011b) ADC083000 ferential DCLK_RST is not activated, the inputs should be grounded. Differential DCLK_RST has an internal 100 ohm termination resistor and should not be AC coupled. The DCLK_RST signal must observe some timing requirements that are shown in Figure 5, Figure 6 and Figure 7 of the Timing Diagrams. The DCLK_RST pulse must be of a minimum width and its deassertion edge must observe setup and hold times with respect to the CLK input rising edge. These times are specified in the AC Electrical Characteristics Table. The DCLK_RST signal can be asserted asynchronous to the input clock. If DCLK_RST is asserted, the DCLK output is held in a designated state. The state in which DCLK is held during the reset period is determined by the mode of operation (SDR/ DDR) and the setting of the Output Edge configuration pin or bit. (Refer to Figure 5, Figure 6 and Figure 7 for the DCLK reset state conditions). Therefore, depending upon when the DCLK_RST signal is asserted, there may be a narrow pulse on the DCLK line during this reset event. When the DCLK_RST signal is de-asserted in synchronization with the CLK rising edge, the next CLK falling edge synchronizes the DCLK output with those of other ADC083000s in the system. The DCLK output is enabled again after a constant delay (relative to the input clock frequency) which is equal to the CLK input to DCLK output delay (tSD). The device always exhibits this delay characteristic in normal operation. If the device is not programmed to allow DCLK to run continuously, DCLK will become inactive during a calibration cycle. Therefore, it is strongly recommended that DCLK only be used as a data capture clock and not as a system clock. The DCLK_RST pin should NOT be brought high while the calibration process is running (while CalRun is high). Doing so could cause a digital glitch in the digital circuitry, resulting in corruption and invalidation of the calibration. 1.4.1 Note Regarding Extended Mode Offset Correction When using the Offset Adjust register, the following information should be noted. For offset values of +0000 0000 and -0000 0000, the actual offset is not the same. By changing only the sign bit in this case, an offset step in the digital output code of about 1/10th of an LSB is experienced. This is shown more clearly in the Figure below. 20193230 FIGURE 10. Extended Mode Offset Behavior 1.5 MULTIPLE ADC SYNCHRONIZATION The ADC083000 has the capability to precisely reset its sampling clock (CLK) to synchronize its output clock (DCLK) and data with multiple ADCs in a system. This allows multiple ADCs in a system to have their DCLK (and data) outputs transition at the same time with respect to the shared CLK input that they all use for sampling. The ADC083000 has been designed to accommodate systems which require a single-ended (LVCMOS) DCLK_RST or a differential (LVDS) DCLK_RST. Single-Ended (LVCMOS) DCLK_RST: The Power on Reset state of DCLK_RST is to have single-ended DCLK_RST activated. Bit 14, (DRE) in the Configuration Register is asserted low, 0b. When not using singled-ended DCLK_RST, the input should be grounded. Differential (LVDS) DCLK_RST: Activated by asserting bit 14, (DRE) in the configuration register high, 1b. When the dif- 1.6 ADC TEST PATTERN To aid in system debug, the ADC083000 has the capability of providing a test pattern at the four output ports completely independent of the input signal. The test pattern is selected by setting bit-11 (TPO) in the Test Pattern Register (address Fh). The test pattern will appear at the digital output about 10 DCLK cycles after the last write to the Test Pattern Register. The ADC is disengaged and a test pattern generator is connected to the outputs including OR. Each port is given a unique 8-bit word, alternating between 1's and 0's as described in theTable 5. TABLE 5. Test Pattern by Output Port Time Da Db Dc Dd OR T0 01h 02h 03h 04h 0 T1 FEh FDh FCh FBh 1 T2 01h 02h 03h 04h 0 T3 FEh FDh FCh FBh 1 T4 01h 02h 03h 04h 0 T5 01h 02h 03h 04h 0 T6 FEh FDh FCh FBh 1 T7 01h 02h 03h 04h 0 T8 FEh FDh FCh FBh 1 T9 01h 02h 03h 04h 0 T10 01h 02h 03h 04h 0 T11 ... ... ... ... ... Note: The same bit pattern repeats when the test pattern sequence is concatenated. www.national.com 28 Comments Pattern Sequence n Pattern Sequence n+1 Pattern Sequence n+2 2.1 THE REFERENCE VOLTAGE The voltage reference for the ADC083000 is derived from a 1.254V bandgap reference, a buffered version of which is made available at pin 31, VBG for user convenience and has an output current capability of ±100 μA. This reference voltage should be buffered if more current is required. The internal bandgap-derived reference voltage has a nominal value of 600 mV or 820 mV, as determined by the FSR pin and described in Section 1.1.4. There is no provision for the use of an external reference voltage, but the full-scale input voltage can be adjusted through a Configuration Register in the Extended Control mode, as explained in Section 1.2. Differential input signals up to the chosen full-scale level will be digitized to 8 bits. Signal excursions beyond the full-scale range will be clipped at the output. These large signal excursions will also activate the OR output for the time that the signal is out of range. See Section 2.2.2. One extra feature of the VBG pin is that it can be used to raise the common mode voltage level of the LVDS outputs. The output offset voltage (VOS) is typically 800mV when the VBG pin is used as an output or left unconnected. To raise the LVDS offset voltage to a typical value of 1150mV the VBG pin can be connected directly to the supply rails. 20193244 FIGURE 11. Differential Data Input Connection When the d.c. coupled mode is used, a precise common mode voltage must be provided at the differential inputs. This common mode voltage should track the VCMO output pin. Note that the VCMO output potential will change with temperature. The common mode output of the driving device should track this change. Full-scale distortion performance falls off rapidly as the input common mode voltage deviates from VCMO. This is a direct result of using a very low supply voltage to minimize power. Keep the input common voltage within 50 mV of VCMO. Performance is as good in the d.c. coupled mode as it is in the a.c. coupled mode, provided the input common mode voltage at both analog inputs remain within 50 mV of VCMO. 2.2 THE ANALOG INPUT The analog input is a differential one to which the signal source may be a.c. coupled or d.c. coupled. The full-scale input range is selected with the FSR pin to be 600 mVP-P or 820 mVP-P, or can be adjusted to values between 560 mVP-P and 840 mVP-P in the Extended Control mode through the Serial Interface. For best performance, it is recommended that the full-scale range be kept between 595 mVP-P and 805 mVP-P in the Extended Control mode. Table 6 gives the input to output relationship with the FSR pin high and the normal (non-extended) mode is used. With the FSR pin grounded, the millivolt values in Table 6 are reduced to 75% of the values indicated. In the Enhanced Control Mode, these values will be determined by the full scale range and offset settings in the Control Registers. 2.2.1 Handling Single-Ended Input Signals There is no provision for the ADC083000 to adequately process single-ended input signals. The best way to handle single-ended signals is to convert them to differential signals before presenting them to the ADC. 2.2.1.1 A.C. Coupled Input The easiest way to accomplish single-ended a.c. input to differential a.c. signal is with an appropriate balun-connected transformer, as shown in Figure 12. TABLE 6. DIFFERENTIAL INPUT TO OUTPUT RELATIONSHIP (Non-Extended Control Mode, FSR High) VIN+ VIN− Output Code VCM − 205mV VCM + 205mV 0000 0000 VCM − 102.5 mV VCM + 102.5 mV 0100 0000 VCM VCM 0111 1111 / 1000 0000 VCM + 102.5 mV VCM −102.5 mV 1100 0000 VCM + 205mV VCM − 205mV 1111 1111 20193243 FIGURE 12. Single-Ended to Differential signal conversion with a balun-connected transformer The buffered analog inputs simplify the task of driving these inputs and the RC pole that is generally used at sampling ADC inputs is not required. If it is desired to use an amplifier circuit before the ADC, use care in choosing an amplifier with adequate noise and distortion performance and adequate gain at the frequencies used for the application. The Input impedance of VIN+ / VIN- in the d.c. coupled mode (VCMO pin not grounded) consists of a precision 100Ω resistor across the inputs and a capacitance from each of these inputs to ground. In the a.c. coupled mode, the input appears the Figure 12 is a generic depiction of a single-ended to differential signal conversion using a balun. The circuitry specific to the balun will depend on the type of balun selected and the overall board layout. It is recommended that the system designer contact the manufacturer of the balun they have selected to aid in designing the best performing single-ended to differential conversion circuit using that particular balun. 29 www.national.com ADC083000 same except there is also a resistor of 50KΩ between each analog input pin and the on-chip VCMO potential. When the inputs are a.c. coupled, the VCMO output must be grounded, as shown in Figure 11. This causes the on-chip VCMO voltage to be connected to the inputs through on-chip 50KΩ resistors. 2.0 Applications Information ADC083000 When selecting a balun, it is important to understand the input of the ADC. The ADC083000 has an on-chip 100 Ohm differential input termination resistor. The range of this input termination resistor is described in the electrical table as the specification RIN. Also, as a result of the ADC architecture, the phase and amplitude balance are critical. The lowest possible phase and amplitude imbalance is desired when selecting a balun. The ADC can tolerate no more than ±2.5° of phase imbalance and the amplitude imbalance should be limited to less than 1dB at the desired input frequency. Since the ADC architecture is based on interleaving, if the phase and amplitude imbalance are greater than what the ADC can tolerate, interleaving spurs will appear in the FFT spectrum and degrade the dynamic performance of the device. Finally, when selecting a balun, the VSWR and insertion loss of the balun should also be considered. The VSWR aids in determining the overall transmission line termination capability of the balun when interfacing to the ADC input. The insertion loss should be considered so that the signal at the balun output is within the specified input range of the ADC as described in the electrical table as the specification VIN. TABLE 7. D.C. Coupled Offset Adjustment Unadjusted Offset Reading Resistor Value 0mV to 10mV no resistor needed 11mV to 30mV 20.0kΩ 31mV to 50mV 10.0kΩ 51mV to 70mV 6.81kΩ 71mV to 90mV 4.75kΩ 91mV to 110mV 3.92kΩ 2.2.2 Out Of Range (OR) Indication When the conversion result is clipped the Out of Range output is activated such that OR+ goes high and OR- goes low. This output is active as long as accurate data on either or both of the buses would be outside the range of 00h to FFh. During a calibration cycle, the OR output is invalid. Refer to 1.1 OVERVIEW for more details. 2.2.3 Full-Scale Input Range As with all A/D Converters, the input range is determined by the value of the ADC's reference voltage. The reference voltage of the ADC083000 is derived from an internal band-gap reference. In the normal mode, the FSR pin controls the effective reference voltage of the ADC083000 such that the differential full-scale input range at the analog inputs is 820 mVP-P with the FSR pin high, or is 600 mVP-P with FSR pin low. In the Extended Control Mode, the Full Scale Range can be set anywhere from 560mV to 840mV. Best SNR is obtained with higher Full Scale Ranges, but better distortion and SFDR are obtained with lower Full Scale Ranges. The LMH6555 of Figure 13 is suitable for any Full Scale Range. 2.2.1.2 D.C. Coupled Input When d.c. coupling to the ADC083000 analog inputs is required, single-ended to differential conversion may be easily accomplished with the LMH6555. An example of this type of circuit is shown in Figure 13. In such applications, the LMH6555 performs the task of single-ended to differential conversion while delivering low distortion and noise, as well as output balance, that supports the operation of the ADC083000. Connecting the ADC083000 VCMO pin to the VCM_REF pin of the LMH6555, through the appropriate buffer, will ensure that the common mode input voltage is as needed for optimum performance of the ADC083000. The LMV321 was chosen to buffer VCMO for its low voltage operation and reasonable offset voltage. Be sure that the current drawn from the VCMO output does not exceed 100 μA. 2.3 THE SAMPLE CLOCK INPUT The ADC083000 has a differential LVDS clock input, CLK+ / CLK-, which must be driven with an a.c. coupled, differential clock signal. Although the ADC083000 is tested and its performance is guaranteed with a differential 1.5 GHz clock, it typically will function well with input clock frequencies indicated in the Electrical Characteristics Table. The clock inputs are internally terminated and biased. The input clock signal must be capacitively coupled to the clock pins as indicated in Figure 14. Operation up to the sample rates indicated in the Electrical Characteristics Table is typically possible if the maximum ambient temperatures indicated are not exceeded. Operating at higher sample rates than indicated for the given ambient temperature may result in reduced device reliability and product lifetime. This is because of the higher power consumption and die temperatures at high sample rates. Important also for reliability is proper thermal management . See Section 2.6.2. 20193255 FIGURE 13. Example of Servoing the Analog Input with VCMO In , RADJ- and RADJ+ are used to adjust the differential offset that can be measured at the ADC inputs VIN+ / VIN-. An unadjusted positive offset with reference to VIN- greater than |15mV| should be reduced with a resistor in the RADJ- position. Likewise, an unadjusted negative offset with reference to VIN- greater than |15mV| should be reduced with a resistor in the RADJ+ position. Table 7 gives suggested RADJ- and RADJ+ values for various unadjusted differential offsets to bring the VIN+ / VIN- offset back to within |15mV|. 20193247 FIGURE 14. Differential Sample Clock Connection www.national.com 30 and fIN is the maximum input frequency, in Hertz, to the ADC analog input. Note that the maximum jitter described above is the arithmetic sum of the jitter from all sources, including that in the ADC input clock, that added by the system to the ADC input clock and input signals and that added by the ADC itself. Since the effective jitter added by the ADC is beyond user control, the best the user can do is to keep the sum of the externally added input clock jitter and the jitter added by the analog circuitry to the analog signal to a minimum. Input clock amplitudes above those specified in the Electrical Characteristics Table may result in increased input offset voltage. This would cause the converter to produce an output code other than the expected 127/128 when both input pins are at the same potential. 2.3.1 Manual Sample Clock Phase Adjust The sample clock phase can be manually adjusted in the Extended Control Mode to accommodate subtle layout differences when synchronizing multiple ADCs. Register addresses Dh and Eh provide extended mode access to fine and coarse adjustments. Use of Low Frequency Sample Clock control, (register Eh; bit-10) is not supported while using manual sample clock phase adjustments. It should be noted that by just enabling the phase adjust capability (register Eh; bit-15), degradation of dynamic performance is expected, specifically SFDR. It is intended that very small adjustments are used. Larger increases in phase adjustments will begin to affect SNR and ultimately ENOB. Therefore, the use of coarse phase adjustment should be minimized in favor of better system design. tJ(MAX) = (VIN(P-P)/VINFSR) x (1/(2(N+1) x π x fIN)) where tJ(MAX) is the rms total of all jitter sources in seconds, VIN(P-P) is the peak-to-peak analog input signal, VINFSR is the full-scale range of the ADC, "N" is the ADC resolution in bits Fine Clock Phase Adjust Range Coarse Clock Phase Adjust Range 20193291 20193292 anywhere from 560 mVP-P to 840 mVP-P. See Section 2.2 for more information. 2.4 CONTROL PINS Six control pins (without the use of the serial interface) provide a wide range of possibilities in the operation of the ADC083000 and facilitate its use. These control pins provide Full-Scale Input Range setting, Calibration, Calibration Delay, Output Edge Synchronization choice, LVDS Output Level choice and a Power Down feature. 2.4.2 Calibration The ADC083000 calibration must be run to achieve specified performance. The calibration procedure is run upon power-up and can be run any time on command. The calibration procedure is exactly the same whether there is an input clock present upon power up or if the clock begins some time after application of power. The CalRun output indicator is high while a calibration is in progress. Note that the DCLK outputs are not active during a calibration cycle, therefore it is not recommended as a system clock. 2.4.1 Full-Scale Input Range Setting The input full-scale range can be selected to be either 600 mVP-P or 820 mVP-P, as selected with the FSR control input (pin 14) in the Normal Mode of operation. In the Extended Control Mode, the input full-scale range may be set to be 31 www.national.com ADC083000 The differential sample clock line pair should have a characteristic impedance of 100Ω and (when using a balun), be terminated at the clock source in that (100Ω) characteristic impedance. The input clock line should be as short and as direct as possible. The ADC083000 clock input is internally terminated with an untrimmed 100Ω resistor. Insufficient input clock levels will result in poor dynamic performance. Excessively high input clock levels could cause a change in the analog input offset voltage. To avoid these problems, keep the input clock level within the range specified in the Electrical Characteristics Table. The low and high times of the input clock signal can affect the performance of any A/D Converter. The ADC083000 features a duty cycle clock correction circuit which can maintain performance over temperature. The ADC will meet its performance specification if the input clock high and low times are maintained within the range (20/80% ratio) as specified in the Electrical Characteristics Table. High speed, high performance ADCs such as the ADC083000 require a very stable input clock signal with minimum phase noise or jitter. ADC jitter requirements are defined by the ADC resolution (number of bits), maximum ADC input frequency and the input signal amplitude relative to the ADC input full scale range. The maximum jitter (the sum of the jitter from all sources) allowed to prevent a jitter-induced reduction in SNR is found to be ADC083000 Extended Control Mode registers (addresses: Dh and Eh). Refer to section 1.4 for register programming information. 2.4.2.1 Power-On Calibration Power-on calibration begins after a time delay following the application of power. This time delay is determined by the setting of CalDly, as described in the Calibration Delay Section, below. The calibration process will be not be performed if the CAL pin is high at power up. In this case, the calibration cycle will not begin until the on-command calibration conditions are met. The ADC083000 will function with the CAL pin held high at power up, but no calibration will be done and performance will be impaired. A manual calibration, however, may be performed after powering up with the CAL pin high. See OnCommand Calibration Section 2.4.2.2. The internal power-on calibration circuitry comes up in an unknown logic state. If the input clock is not running at power up and the power on calibration circuitry is active, it will hold the analog circuitry in power down and the power consumption will typically be less than 25 mW. The power consumption will be normal after the clock starts. 2.4.2.3 Calibration Delay The CalDly input (pin 127) is used to select one of two delay times after the application of power to the start of calibration, as described in Section 1.1.1. The calibration delay values allow the power supply to come up and stabilize before calibration takes place. With no delay or insufficient delay, calibration would begin before the power supply is stabilized at its operating value and result in non-optimal calibration coefficients. If the PD pin is high upon power-up, the calibration delay counter will be disabled until the PD pin is brought low. Therefore, holding the PD pin high during power up will further delay the start of the power-up calibration cycle. The best setting of the CalDly pin depends upon the power-on settling time of the power supply. Note that the calibration delay selection is not possible in the Extended Control mode and the short delay time is used. 2.4.3 Output Edge Synchronization DCLK signals are available to help latch the converter output data into external circuitry. The output data can be synchronized with either edge of these DCLK signals. That is, the output data transition can be set to occur with either the rising edge or the falling edge of the DCLK signal, so that either edge of that DCLK signal can be used to latch the output data into the receiving circuit. When OutEdge (pin 4) is high, the output data is synchronized with (changes with) the rising edge of the DCLK+ (pin 82). When OutEdge is low, the output data is synchronized with the falling edge of DCLK+. At the very high speeds of which the ADC083000 is capable, slight differences in the lengths of the DCLK and data lines can mean the difference between successful and erroneous data capture. The OutEdge pin is used to capture data on the DCLK edge that best suits the application circuit and layout. 2.4.2.2 On-Command Calibration To initiate an on-command calibration, bring the CAL pin high for a minimum of 80 input clock cycles after it has been low for a minimum of 80 input clock cycles. Holding the CAL pin high upon power up will prevent execution of power-on calibration until the CAL pin is low for a minimum of 80 input clock cycles, then brought high for a minimum of another 80 input clock cycles. The calibration cycle will begin 80 input clock cycles after the CAL pin is thus brought high. The CalRun signal should be monitored to determine when the calibration cycle has completed. The minimum 80 input clock cycle sequences are required to ensure that random noise does not cause a calibration to begin when it is not desired. As mentioned in section 1.1 for best performance, a calibration should be performed 20 seconds or more after power up and repeated when the operating temperature changes significantly relative to the specific system design performance requirements. ENOB changes slightly with increasing junction temperature and can be easily corrected by performing an on-command calibration. 2.4.4 LVDS Output Level Control The output level can be set to one of two levels with OutV (pin3). The strength of the output drivers is greater with OutV high. With OutV low there is less power consumption in the output drivers, but the lower output level means decreased noise immunity. For short LVDS lines and low noise systems, satisfactory performance may be realized with the OutV input low. If the LVDS lines are long and/or the system in which the ADC083000 is used is noisy, it may be necessary to tie the OutV pin high. Considerations for a continuous DCLK and proper CalRun operation: • During a Power-On calibration cycle, both the ADC and the input termination resistor are calibrated. As ENOB changes slightly with junction temperature, an OnCommand calibration can be executed to bring the performance of the ADC inline. By default, On-Command calibration includes calibrating the input termination resistance and the ADC. However, since the input termination resistance changes marginally with temperature, the user has the option to disable the input termination resistor trim (address: 1h, bit: 13, set to 1b), which will guarantee that the DCLK is continuously present at the output during calibration. The Resistor Trim Disable can be programmed in register (address: 1h, bit 13) when in the Extended Control mode. Refer to section 1.4 for register programming information. • When an on-command calibration is requested while using the Aperture Adjust Circuitry through the Extended Control Mode registers, we recommend that the Resistor Trim Disable bit be disabled (address: 1h, bit: 13, set to 1b). This allows continuous operation of all clocks in the ADC including DCLK and proper operation of CalRun output. The Aperture Adjust Circuitry control is resident in the www.national.com 2.4.5 Power Down Feature The Power Down pin (PD) allows the ADC083000 to be entirely powered down. See Section 1.1.7 for details on the power down feature. The digital data (+/-) output pins are put into a high impedance state when the PD pin for the respective channel is high. Upon return to normal operation, the pipeline will contain meaningless information and must be flushed. If the PD input is brought high while a calibration is running, the device will not go into power down until the calibration sequence is complete. However, if power is applied and PD is already high, the device will not begin the calibration sequence until the PD input goes low. If a manual calibration is requested while the device is powered down, the calibration will not begin at all. That is, the manual calibration input is completely ignored in the power down state. 32 20193254 FIGURE 15. Non-Spiking Power Supply The output drivers should have a supply voltage, VDR, that is within the range specified in the Operating Ratings table. This voltage should not exceed the VA supply voltage and should never spike to a voltage greater than ( VA + 100mV). If the power is applied to the device without an input clock signal present, the current drawn by the device might be below 200 mA. This is because the ADC083000 gets reset through clocked logic and its initial state is unknown. If the reset logic comes up in the "on" state, it will cause most of the analog circuitry to be powered down, resulting in less than 100 mA of current draw. This current is greater than the power down current because not all of the ADC is powered down. The device current will be normal after the input clock is established. 2.6 POWER CONSIDERATIONS A/D converters draw sufficient transient current to corrupt their own power supplies if not adequately bypassed. A 33 µF capacitor should be placed within an inch (2.5 cm) of the A/D converter power pins. A 0.1 µF capacitor should be placed as close as possible to each VA pin, preferably within one-half centimeter. Leadless chip capacitors are preferred because they have low lead inductance. The VA and VDR supply pins should be isolated from each other to prevent any digital noise from being coupled into the analog portions of the ADC. A ferrite choke, such as the JW Miller FB20009-3B, is recommended between these supply lines when a common source is used for them. As is the case with all high speed converters, the ADC083000 should be assumed to have little power supply noise rejection. Any power supply used for digital circuitry in a system where a lot of digital power is being consumed should not be used to supply power to the ADC083000. The ADC supplies should be the same supply used for other analog circuitry, if not a dedicated supply. 2.6.2 Thermal Management The ADC083000 is capable of impressive speeds and performance at very low power levels for its speed. However, the power consumption is still high enough to require attention to thermal management. For reliability reasons, the die temperature should be kept to a maximum of 130°C. That is, TA (ambient temperature) plus ADC power consumption times θJA (junction to ambient thermal resistance) should not exceed 130°C. This is not a problem if the ambient temperature is kept to a maximum of +85°C as specified in the Operating Ratings section. Please note that the following are general recommendations for mounting exposed pad devices onto a PCB. This should be considered the starting point in PCB and assembly process development. It is recommended that the process be developed based upon past experience in package mounting. The package of the ADC083000 has an exposed pad on its back that provides the primary heat removal path as well as 2.6.1 Supply Voltage The ADC083000 is specified to operate with a supply voltage of 1.9V ±0.1V. It is very important to note that, while this device will function with slightly higher supply voltages, these higher supply voltages may reduce product lifetime. No pin should ever have a voltage on it that is in excess of the supply voltage or below ground by more than 150 mV, not even on a transient basis. This can be a problem upon application of power and power shut-down. Be sure that the supplies to circuits driving any of the input pins, analog or digital, 33 www.national.com ADC083000 do not come up any faster than does the voltage at the ADC083000 power pins. The Absolute Maximum Ratings should be strictly observed, even during power up and power down. A power supply that produces a voltage spike at turn-on and/or turn-off of power can destroy the ADC083000. The circuit of Figure 15 will provide supply overshoot protection. Many linear regulators will produce output spiking at poweron unless there is a minimum load provided. Active devices draw very little current until their supply voltages reach a few hundred millivolts. The result can be a turn-on spike that can destroy the ADC083000, unless a minimum load is provided for the supply. The 100Ω resistor at the regulator output provides a minimum output current during power-up to ensure there is no turn-on spiking. Whether a linear or switching regulator is used, it is advisable to provide a slow start circuit to prevent overshoot of the supply. In the circuit of Figure 15, an LM317 linear regulator is satisfactory if its input supply voltage is 4V to 5V . If a 3.3V supply is used, an LM1086 linear regulator is recommended. 2.5 THE DIGITAL OUTPUTS The ADC083000 demultiplexes the output data of each of the two ADCs on the die onto two LVDS output buses (total of four buses, two for each ADC). For each of the two converters, the results of successive conversions start on the falling edges of the CLK+ pin and are available on one of the two LVDS buses. The results of conversions that start on the rising edges of the CLK+ pin are available on the other LVDS bus. This means that, the word rate at each LVDS bus is 1/2 the ADC083000 input clock rate and the two buses must be multiplexed to obtain the entire 3 GSPS conversion result. Since the minimum recommended input clock rate for this device is 500 MHz, the sampling rate can be reduced to as low as 1 GSPS by using the results available on all four LVDS busses. The effective sampling rate can be reduced to as low as 250 MSPS by decimating the data by the use of one bus. There is one LVDS output clock pair (DCLK+/-) available for use to latch the LVDS outputs on all buses. Whether the data is sent at the rising or falling edge of DCLK is determined by the sense of the OutEdge pin, as described in Section 2.4.3. DDR (Double Data Rate) clocking can also be used. In this mode a word of data is presented with each edge of DCLK, reducing the DCLK frequency to 1/4 the input clock frequency. When the device is in DDR mode, address 1h, bit-8 must be set to 0b. See the Timing Diagram section for details. The OutV pin is used to set the LVDS differential output levels. See Section 2.4.4. The output format is Offset Binary. Accordingly, a full-scale input level with VIN+ positive with respect to VIN− will produce an output code of all ones, a full-scale input level with VIN− positive with respect to VIN+ will produce an output code of all zeros and when VIN+ and VIN− are equal, the output code will vary between codes 127 and 128. ADC083000 excellent electrical grounding to the printed circuit board. The land pattern design for lead attachment to the PCB should be the same as for a conventional LQFP, but the exposed pad must be attached to the board to remove the maximum amount of heat from the package, as well as to ensure best product parametric performance. To maximize the removal of heat from the package, a thermal land pattern must be incorporated on the PC board within the footprint of the package. The exposed pad of the device must be soldered down to ensure adequate heat conduction out of the package. The land pattern for this exposed pad should be at least as large as the 5 x 5 mm of the exposed pad of the package and be located such that the exposed pad of the device is entirely over that thermal land pattern. This thermal land pattern should be electrically connected to ground. A clearance of at least 0.5 mm should separate this land pattern from the mounting pads for the package pins. taining a maximum pad temperature reading of 123.7°C will ensure that the die temperature does not exceed 130°C, assuming that the exposed pad of the ADC083000 is properly soldered down and the thermal vias are adequate. (The inaccuracy of the temperature sensor is additional to the above calculation). 2.7 LAYOUT AND GROUNDING Proper grounding and proper routing of all signals are essential to ensure accurate conversion. A single ground plane should be used, instead of splitting the ground plane into analog and digital areas. Since digital switching transients are composed largely of high frequency components, the skin effect tells us that total ground plane copper weight will have little effect upon the logic-generated noise. Total surface area is more important than is total ground plane volume. Coupling between the typically noisy digital circuitry and the sensitive analog circuitry can lead to poor performance that may seem impossible to isolate and remedy. The solution is to keep the analog circuitry well separated from the digital circuitry. High power digital components should not be located on or near any linear component or power supply trace or plane that services analog or mixed signal components as the resulting common return current path could cause fluctuation in the analog input “ground” return of the ADC, causing excessive noise in the conversion result. Generally, we assume that analog and digital lines should cross each other at 90° to avoid getting digital noise into the analog path. In high frequency systems, however, avoid crossing analog and digital lines altogether. The input clock lines should be isolated from ALL other lines, analog AND digital. The generally accepted 90° crossing should be avoided as even a little coupling can cause problems at high frequencies. Best performance at high frequencies is obtained with a straight signal path. The analog input should be isolated from noisy signal traces to avoid coupling of spurious signals into the input. This is especially important with the low level drive required of the ADC083000. Any external component (e.g., a filter capacitor) connected between the converter's input and ground should be connected to a very clean point in the analog ground plane. All analog circuitry (input amplifiers, filters, etc.) should be separated from any digital components. 20193221 FIGURE 16. Recommended Package Land Pattern Since a large aperture opening may result in poor release, the aperture opening should be subdivided into an array of smaller openings, similar to the land pattern of Figure 16. To minimize junction temperature, it is recommended that a simple heat sink be built into the PCB. This is done by including a copper area of about 2 square inches (6.5 square cm) on the opposite side of the PCB. This copper area may be plated or solder coated to prevent corrosion, but should not have a conformal coating, which could provide some thermal insulation. Thermal vias should be used to connect these top and bottom copper areas. These thermal vias act as "heat pipes" to carry the thermal energy from the device side of the board to the opposite side of the board where it can be more effectively dissipated. The use of 9 to 16 thermal vias is recommended. The thermal vias should be placed on a 1.2 mm grid spacing and have a diameter of 0.30 to 0.33 mm. These vias should be barrel plated to avoid solder wicking into the vias during the soldering process as this wicking could cause voids in the solder between the package exposed pad and the thermal land on the PCB. Such voids could increase the thermal resistance between the device and the thermal land on the board, which would cause the device to run hotter. If it is desired to monitor die temperature, a temperature sensor may be mounted on the heat sink area of the board near the thermal vias. Allow for a thermal gradient between the temperature sensor and the ADC083000 die of θJ-PAD times typical power consumption = 2.8 x 1.9 = 5.3°C. Allowing for 6.3°C, including some margin for temperature drop from the pad to the temperature sensor, then, would mean that main- www.national.com 2.8 DYNAMIC PERFORMANCE The ADC083000 is a.c. tested and its dynamic performance is guaranteed. To meet the published specifications and avoid jitter-induced noise, the clock source driving the CLK input must exhibit low rms jitter. The allowable jitter is a function of the input frequency and the input signal level, as described in Section 2.3. It is good practice to keep the ADC input clock line as short as possible, to keep it well away from any other signals and to treat it as a transmission line. Other signals can introduce jitter into the input clock signal. The clock signal can also introduce noise into the analog path if not isolated from that path. Best dynamic performance is obtained when the exposed pad at the back of the package has a good connection to ground. This is because this path from the die to ground is a lower impedance than offered by the package pins. 2.9 USING THE SERIAL INTERFACE The ADC083000 may be operated in the non-extended control (non-Serial Interface) mode or in the extended control 34 than 150 mV below the ground pins or 150 mV above the supply pins. Exceeding these limits on even a transient basis may not only cause faulty or erratic operation, but may impair device reliability. It is not uncommon for high speed digital circuits to exhibit undershoot that goes more than a volt below ground. Controlling the impedance of high speed lines and terminating these lines in their characteristic impedance should control overshoot. Care should be taken not to overdrive the inputs of the ADC083000. Such practice may lead to conversion inaccuracies and even to device damage. Incorrect analog input common mode voltage in the d.c. coupled mode. As discussed in section 1.1.4 and 2.2, the Input common mode voltage must remain within 50 mV of the VCMO output , which has a variability with temperature that must also be tracked. Distortion performance will be degraded if the input common mode voltage is more than 50 mV from VCMO . Using an inadequate amplifier to drive the analog input. Use care when choosing a high frequency amplifier to drive the ADC083000 as many high speed amplifiers will have higher distortion than will the ADC083000, resulting in overall system performance degradation. Driving the VBG pin to change the reference voltage. As mentioned in Section 2.1, the reference voltage is intended to be fixed to provide one of two different full-scale values (600 mVP-P and 820 mVP-P). Over driving this pin will not change the full scale value, but can be used to change the LVDS common mode voltage from 0.8V to 1.2V by tying the VBG pin to VA. Driving the clock input with an excessively high level signal. The ADC input clock level should not exceed the level described in the Operating Ratings Table or the input offset could change. Inadequate input clock levels. As described in Section 2.3, insufficient input clock levels can result in poor performance. Excessive input clock levels could result in the introduction of an input offset. Using a clock source with excessive jitter, using an excessively long input clock signal trace, or having other signals coupled to the input clock signal trace. This will cause the sampling interval to vary, causing excessive output noise and a reduction in SNR performance. Failure to provide adequate heat removal. As described in Section 2.6.2, it is important to provide adequate heat removal to ensure device reliability. This can either be done with adequate air flow or the use of a simple heat sink built into the board. The backside pad should be grounded for best performance. 2.9.1 Non-Extended Control Mode Operation Non-extended control mode operation means that the Serial Interface is not active and all controllable functions are controlled with various pin settings. That is, the output voltage, full-scale range and output edge selections are all controlled with pin settings. The non-extended control mode is used by setting pin 14 high or low, as opposed to letting it float. Table 8 indicates the pin functions of the ADC083000 in the nonextended control mode. TABLE 8. Non-Extended Control Mode Operation (Pin 14 High or Low) Pin Low High Floating 3 0.52 VP-P Output 0.68 VP-P Output n/a 4 OutEdge = Neg OutEdge = Pos DDR 14 600 mVP-P input range 820 mVP-P input range Extended Control Mode 127 CalDly Low CalDly High Serial Interface Enable Pin 3 can be either high or low in the non-extended control mode. Pin 14 must not be left floating to select this mode. See Section 1.2 for more information. Pin 4 can be high or low or can be left floating in the nonextended control mode. In the non-extended control mode, pin 4 high or low defines the edge at which the output data transitions. See Section 2.4.3 for more information. If this pin is floating, the output clock (DCLK) is a DDR (Double Data Rate) clock (see Section 1.1.5.3) and the output edge synchronization is irrelevant since data is clocked out on both DCLK edges. Pin 127, if it is high or low in the non-extended control mode, sets the calibration delay. If pin 127 is floating, the calibration delay is the same as it would be with this pin low and this pin acts as the enable pin for the serial interface input. TABLE 9. Extended Control Mode Operation (Pin 14 Floating) Pin Function 3 SCLK (Serial Clock) 4 SDATA (Serial Data) 127 SCS (Serial Interface Chip Select) 2.10 COMMON APPLICATION PITFALLS Driving the inputs (analog or digital) beyond the power supply rails. For device reliability, no input should go more 35 www.national.com ADC083000 mode. Table 8 and Table 9 describe the functions of pins 3, 4, 14 and 127 in the non-extended control mode and the extended control mode, respectively. ADC083000 Physical Dimensions inches (millimeters) unless otherwise noted NOTES: UNLESS OTHERWISE SPECIFIED REFERENCE JEDEC REGISTRATION MS-026, VARIATION BFB. 128-Lead Exposed Pad LQFP NS Package Number VNX128A www.national.com 36 ADC083000 Notes 37 www.national.com ADC083000 8-Bit, 3 GSPS, High Performance, Low Power A/D Converter Notes THE CONTENTS OF THIS DOCUMENT ARE PROVIDED IN CONNECTION WITH NATIONAL SEMICONDUCTOR CORPORATION (“NATIONAL”) PRODUCTS. NATIONAL MAKES NO REPRESENTATIONS OR WARRANTIES WITH RESPECT TO THE ACCURACY OR COMPLETENESS OF THE CONTENTS OF THIS PUBLICATION AND RESERVES THE RIGHT TO MAKE CHANGES TO SPECIFICATIONS AND PRODUCT DESCRIPTIONS AT ANY TIME WITHOUT NOTICE. NO LICENSE, WHETHER EXPRESS, IMPLIED, ARISING BY ESTOPPEL OR OTHERWISE, TO ANY INTELLECTUAL PROPERTY RIGHTS IS GRANTED BY THIS DOCUMENT. TESTING AND OTHER QUALITY CONTROLS ARE USED TO THE EXTENT NATIONAL DEEMS NECESSARY TO SUPPORT NATIONAL’S PRODUCT WARRANTY. EXCEPT WHERE MANDATED BY GOVERNMENT REQUIREMENTS, TESTING OF ALL PARAMETERS OF EACH PRODUCT IS NOT NECESSARILY PERFORMED. NATIONAL ASSUMES NO LIABILITY FOR APPLICATIONS ASSISTANCE OR BUYER PRODUCT DESIGN. BUYERS ARE RESPONSIBLE FOR THEIR PRODUCTS AND APPLICATIONS USING NATIONAL COMPONENTS. PRIOR TO USING OR DISTRIBUTING ANY PRODUCTS THAT INCLUDE NATIONAL COMPONENTS, BUYERS SHOULD PROVIDE ADEQUATE DESIGN, TESTING AND OPERATING SAFEGUARDS. EXCEPT AS PROVIDED IN NATIONAL’S TERMS AND CONDITIONS OF SALE FOR SUCH PRODUCTS, NATIONAL ASSUMES NO LIABILITY WHATSOEVER, AND NATIONAL DISCLAIMS ANY EXPRESS OR IMPLIED WARRANTY RELATING TO THE SALE AND/OR USE OF NATIONAL PRODUCTS INCLUDING LIABILITY OR WARRANTIES RELATING TO FITNESS FOR A PARTICULAR PURPOSE, MERCHANTABILITY, OR INFRINGEMENT OF ANY PATENT, COPYRIGHT OR OTHER INTELLECTUAL PROPERTY RIGHT. LIFE SUPPORT POLICY NATIONAL’S PRODUCTS ARE NOT AUTHORIZED FOR USE AS CRITICAL COMPONENTS IN LIFE SUPPORT DEVICES OR SYSTEMS WITHOUT THE EXPRESS PRIOR WRITTEN APPROVAL OF THE CHIEF EXECUTIVE OFFICER AND GENERAL COUNSEL OF NATIONAL SEMICONDUCTOR CORPORATION. As used herein: Life support devices or systems are devices which (a) are intended for surgical implant into the body, or (b) support or sustain life and whose failure to perform when properly used in accordance with instructions for use provided in the labeling can be reasonably expected to result in a significant injury to the user. A critical component is any component in a life support device or system whose failure to perform can be reasonably expected to cause the failure of the life support device or system or to affect its safety or effectiveness. National Semiconductor and the National Semiconductor logo are registered trademarks of National Semiconductor Corporation. All other brand or product names may be trademarks or registered trademarks of their respective holders. Copyright© 2007 National Semiconductor Corporation For the most current product information visit us at www.national.com National Semiconductor Americas Customer Support Center Email: [email protected] Tel: 1-800-272-9959 www.national.com National Semiconductor Europe Customer Support Center Fax: +49 (0) 180-530-85-86 Email: [email protected] Deutsch Tel: +49 (0) 69 9508 6208 English Tel: +49 (0) 870 24 0 2171 Français Tel: +33 (0) 1 41 91 8790 National Semiconductor Asia Pacific Customer Support Center Email: [email protected] National Semiconductor Japan Customer Support Center Fax: 81-3-5639-7507 Email: [email protected] Tel: 81-3-5639-7560