June 2006 ADC083000 High Performance, Low Power, 8-Bit, 3 GSPS A/D Converter General Description Features Note: This product is currently in development. - ALL specifications are design targets and are subject to change. n n n n n n n n n n 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 1.8 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-andhold amplifier and the self-calibration scheme enable a very flat response of all dynamic parameters up to Nyquist, producing a high 7.0 ENOB with a 748 MHz input signal and a 3 GHz sample rate while providing a 10-18 B.E.R. The ADC083000 achieves a 3GSPS sampling rate by utilizing both the rising and falling edge of a 1.5 GSPS input clock. Output formatting is offset binary and the LVDS digital outputs are compliant with IEEE 1596.3-1996, with the exception of an adjustable common mode voltage between 0.8V and 1.2V. 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 ADC can be programmed into the 1:2 Output Mode where the data is output on the Dc and Dd channels at the rate of the input clock. The converter typically consumes less than 20 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. © 2006 National Semiconductor Corporation DS201932 Internal Sample-and-Hold Single +1.9V ± 0.1V Operation Choice of SDR or DDR output clocking 1:2 or 1:4 Selectable Output Demux Clock Phase Adjust for Multiple ADC Synchronization Guaranteed No Missing Codes Serial Interface for Extended Control Fine Adjustment of Input Full-Scale Range and Offset Duty Cycle Corrected Sample Clock Test pattern Key Specifications n n n n n n n Resolution Max Conversion Rate Bit Error Rate ENOB @ 748 MHz Input SNR @ 748MHz Full Power Bandwidth Power Consumption — Operating — Power Down Mode 8 Bits 3 GSPS (min) 10-18 (typ) 7.0 Bits (typ) 44 dB (typ) 3 GHz (typ) 1.8 W (typ) 20 mW (typ) Applications n n n n n Direct RF Down Conversion Digital Oscilloscopes Satellite Set-top boxes Communications Systems Test Instrumentation www.national.com ADC083000 High Performance, Low Power, 8-Bit, 3 GSPS A/D Converter ADVANCE INFORMATION ADC083000 Block Diagram 20193253 www.national.com 2 ADC083000 Pin Configuration 20193201 * Exposed pad on back of package must be soldered to ground plane to ensure rated performance. 3 www.national.com ADC083000 Pin Descriptions and Equivalent Circuits Pin Functions Pin No. Symbol Equivalent Circuit Description OutV / SCLK Output Voltage Amplitude and Serial Interface Clock. 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. 4 OutEdge / DDR / SDATA DCLK Edge Select, Double Data Rate Enable and Serial Data Input. 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. 15 DCLK_RST DCLK Reset. 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. 26 PD Power Down. A logic high on the PD pin puts the entire device into the Power Down Mode. CAL Calibration Cycle Initiate. A minimum 80 input clock cycles logic low followed by a minimum of 80 input clock cycles high on this pin initiates the self 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 and Extended Control Enable. In non-extended control mode, a logic low on this pin sets the full-scale differential input range to 650 mVP-P. A logic high on this pin sets the full-scale differential input range to 870 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. CalDly / SCS Calibration Delay and Serial Interface Chip Select. 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). 3 30 14 127 www.national.com 4 ADC083000 Pin Descriptions and Equivalent Circuits (Continued) Pin Functions Pin No. Symbol Equivalent Circuit Description 10 11 CLK+ CLK- LVDS Clock input pins for the ADC. The differential clock signal must be a.c. coupled to these pins. The input signal is sampled on the 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− Analog signal inputs to the ADC. The differential full-scale input range is 650 mVP-P when the FSR pin is low, or 870 mVP-P when the FSR pin is high. 7 VCMO Common Mode Voltage. 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 inputs. This pin is capable of sourcing or sinking 100µA. See Section 2.2. 31 VBG Bandgap output voltage capable of 100 µA source/sink. 126 CalRun Calibration Running indication. This pin is at a logic high when calibration is running. 32 REXT External bias resistor connection. Nominal value is 3.3k-Ohms ( ± 0.1%) to ground. See Section 1.1.1. 34 35 Tdiode_P Tdiode_N Temperature Diode Positive (Anode) and Negative (Cathode) for die temperature measurements. See Section 2.6.2. 5 www.national.com ADC083000 Pin Descriptions and Equivalent Circuits (Continued) Pin Functions Pin No. Symbol Equivalent Circuit Description 36 38 43 45 47 49 54 56 58 60 65 67 69 71 75 77 / / / / / / / / / / / / / / / / 37 39 44 46 48 50 55 57 59 61 66 68 70 72 76 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 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 85 89 91 93 95 100 102 104 106 111 113 115 117 122 124 / / / / / / / / / / / / / / / / 84 86 90 92 94 96 101 103 105 107 112 114 116 118 123 125 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 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 OR+ OR- Out Of Range output. 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 outputs used to latch the output data. Delayed and non-delayed data outputs are supplied synchronous to this signal. This signal is at 1/2 the input clock rate in SDR mode and at 1/4 the input 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 bit 14 of the RTRIM Disable register (address 9h). DCLK is continuously present during the calibration cycle when bit 14 is set high and is not active during the calibration cycle when set low. When the device is in the 1:1 Output Mode, the signal is at the input clock rate. 6 ADC083000 Pin Descriptions and Equivalent Circuits (Continued) Pin Functions Pin No. Symbol 2, 5, 8, 13, 16, 17, 20, 25, 28, 33, 128 Equivalent Circuit Description VA 40, 51 ,62, 73, 88, 99, 110, 121 VDR Output Driver power supply pins. Bypass these pins to DR GND. 1, 6, 9, 12, 21, 24, 27, 41 GND Ground return for VA. 42, 53, 64, 74, 87, 97, 108, 119 DR GND 22,23,29,52, 63, 98, 109, 120 NC Analog power supply pins. Bypass these pins to ground. Ground return for VDR. No Connection. Make no connection to these pins. 7 www.national.com ADC083000 Absolute Maximum Ratings Operating Ratings (Notes 1, 2) (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) −40˚C ≤ TA ≤ +85˚C +1.8V to +2.0V Driver Supply Voltage (VDR) +1.8V to VA Supply Voltage (VA, VDR) 2.2V Analog Input Common Mode Voltage VCMO ± 50mV Voltage on Any Input Pin −0.15V to (VA +0.15V) VIN+, VIN- Voltage Range (Maintaining Common Mode) 200mV to VA 0V to 100 mV Ground Difference (|GND - DR GND|) Ground Difference |GND - DR GND| Package Input Current (Note 3) ± 25 mA ± 50 mA Power Dissipation at TA ≤ 85˚C TBD W Input Current at Any Pin (Note 3) ESD Susceptibility (Note 4) Human Body Model Machine Model Storage Temperature 0V to VA Differential CLK Amplitude 0.4VP-P to 2.0VP-P Package Thermal Resistance 2500V 250V Soldering Temperature, Infrared, 10 seconds, (Note 5), (Applies to standard plated package only) 0V CLK Pins Voltage Range Package θJA θJC (Top of θJ-PAD Package) (Thermal 10˚C / W 2.8˚C / W Pad) 128-Lead Exposed Pad LQFP 235˚C −65˚C to +150˚C 26˚C / W Converter Electrical Characteristics NOTE: This product is currently in development and the parameters specified in this section are DESIGN TARGETS. The specifications in this section cannot be guaranteed until device characterization has taken place. The following specifications apply after calibration for VA = VDR = +1.9VDC, OutV = 1.9V, VIN FSR (a.c. coupled) = differential 870mVP-P, CL = 10 pF, Differential, a.c. coupled Sinewave Input Clock, fCLK = 1.5GHz at 0.5VP-P with 50% duty cycle, VBG = Floating, NonExtended 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 Overanged ± 0.35 ± TBD LSB (max) DNL Differential Non-Linearity DC Coupled, 1MHz Sine Wave Overanged ± 0.25 ± TBD LSB (max) 8 Bits −TBD TBD LSB (min) LSB (max) Resolution with No Missing Codes VOFF Offset Error -0.45 ± 45 VOFF_ADJ Input Offset Adjustment Range PFSE Positive Full-Scale Error (Note 9) −0.6 ± TBD mV (max) NFSE Negative Full-Scale Error (Note 9) −1.31 ± TBD mV (max) FS_ADJ Full-Scale Adjustment Range ± 20 ± 15 %FS Extended Control Mode Extended Control Mode mV DYNAMIC CONVERTER CHARACTERISTICS FPBW Full Power Bandwidth TBD GHz B.E.R. Bit Error Rate 10-18 Error/Sample ± TBD ± TBD dBFS Gain Flatness ENOB www.national.com Effective Number of Bits d.c. to 1500 MHz d.c. to 3 GHz dBFS fIN = 248 MHz, VIN = FSR − 0.5 dB 7.4 TBD Bits (min) fIN = 748 MHz, VIN = FSR − 0.5 dB 7.0 TBD Bits (min) fIN = 1498 MHz, VIN = FSR − 0.5 dB 6.5 TBD Bits (min) 8 (Continued) NOTE: This product is currently in development and the parameters specified in this section are DESIGN TARGETS. The specifications in this section cannot be guaranteed until device characterization has taken place. The following specifications apply after calibration for VA = VDR = +1.9VDC, OutV = 1.9V, VIN FSR (a.c. coupled) = differential 870mVP-P, CL = 10 pF, Differential, a.c. coupled Sinewave Input Clock, fCLK = 1.5GHz at 0.5VP-P with 50% duty cycle, VBG = Floating, NonExtended 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 Typical (Note 8) Limits (Note 8) Units (Limits) fIN = 248 MHz, VIN = FSR − 0.5 dB 46.3 TBD dB (min) fIN = 748 MHz, VIN = FSR − 0.5 dB 44.5 TBD dB (min) fIN = 1498 MHz, VIN = FSR − 0.5 dB 40.9 TBD dB (min) fIN = 248 MHz, VIN = FSR − 0.5 dB 46.9 TBD dB (min) fIN = 748 MHz, VIN = FSR − 0.5 dB 44.0 TBD dB (min) Parameter Conditions DYNAMIC CONVERTER CHARACTERISTICS SINAD SNR THD 2nd Harm 3rd Harm SFDR IMD Signal-to-Noise Plus Distortion Ratio Signal-to-Noise Ratio Total Harmonic Distortion Second Harmonic Distortion Third Harmonic Distortion Spurious-Free dynamic Range Intermodulation Distortion Out of Range Output Code (In addition to OR Output high) fIN = 1498 MHz, VIN = FSR − 0.5 dB 42.0 TBD dB (min) fIN = 248 MHz, VIN = FSR − 0.5 dB -55 -TBD dB (max) fIN = 748 MHz, VIN = FSR − 0.5 dB -54 -TBD dB (max) fIN = 1498 MHz, VIN = FSR − 0.5 dB -47.4 -TBD dB (max) fIN = 248 MHz, VIN = FSR − 0.5 dB −60 dB fIN = 748 MHz, VIN = FSR − 0.5 dB −60 dB fIN = 1498 MHz, VIN = FSR − 0.5 dB −60 dB fIN = 248 MHz, VIN = FSR − 0.5 dB −65 dB fIN = 748 MHz, VIN = FSR − 0.5 dB −65 dB fIN = 1498 MHz, VIN = FSR − 0.5 dB −65 dB fIN = 248 MHz, VIN = FSR − 0.5 dB 55 TBD dB (min) fIN = 748 MHz, VIN = FSR − 0.5 dB 49 TBD dB (min) fIN = 1498 MHz, VIN = FSR − 0.5 dB 43 TBD dB (min) fIN1 = 321 MHz, VIN = FSR − 7 dB fIN2 = 326 MHz, VIN = FSR − 7 dB -50 dB (VIN+) − (VIN−) > + Full Scale 255 (VIN+) − (VIN−) < − Full Scale 0 ANALOG INPUT AND REFERENCE CHARACTERISTICS VIN VCMI CIN RIN Full Scale Analog Differential Input Range FSR pin 14 Low 600 FSR pin 14 High 800 Analog Input Common Mode Voltage VCMO 520 mVP-P (min) 680 mVP-P (max) 720 mVP-P (min) 880 mVP-P (max) VCMO − 50 VCMO + 50 mV (min) mV (max) Analog Input Capacitance, Normal operation (Notes 10, 11) Differential 0.02 pF Each input pin to ground 1.6 pF Analog Input Capacitance, DES Mode (Notes 10, 11) Differential 0.08 pF Each input pin to ground 2.2 Differential Input Resistance 100 pF 94 Ω (min) 106 Ω (max) 0.95 1.45 V (min) V (max) ANALOG OUTPUT CHARACTERISTICS VCMO Common Mode Output Voltage VCMO_LVL VCMO input threshold to set DC Coupling mode TC VCMO Common Mode Output Voltage Temperature Coefficient 1.26 VA = 1.8V 0.60 V VA = 2.0V 0.66 V TA = −40˚C to +85˚C 118 ppm/˚C 9 www.national.com ADC083000 Converter Electrical Characteristics ADC083000 Converter Electrical Characteristics (Continued) NOTE: This product is currently in development and the parameters specified in this section are DESIGN TARGETS. The specifications in this section cannot be guaranteed until device characterization has taken place. The following specifications apply after calibration for VA = VDR = +1.9VDC, OutV = 1.9V, VIN FSR (a.c. coupled) = differential 870mVP-P, CL = 10 pF, Differential, a.c. coupled Sinewave Input Clock, fCLK = 1.5GHz at 0.5VP-P with 50% duty cycle, VBG = Floating, NonExtended 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) 80 pF 1.20 1.33 V (min) V (max) ANALOG OUTPUT CHARACTERISTICS 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 28 ppm/˚C 80 pF TEMPERATURE DIODE CHARACTERISTICS ∆VBE 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) CLOCK INPUT CHARACTERISTICS VID II CIN Differential Clock Input Level Input Current VIN = 0 or VIN = VA ±1 µA Input Capacitance (Notes 10, 11) Differential 0.02 pF Each input to ground 1.5 pF DIGITAL CONTROL PIN CHARACTERISTICS VIH Logic High Input Voltage (Note 12) 0.85 x VA V (min) VIL Logic Low Input Voltage (Note 12) 0.15 x VA V (max) CIN Input Capacitance (Notes 11, 13) Each input to ground 1.2 Measured differentially, OutV = VA, VBG = Floating (Note 15) 710 pF DIGITAL OUTPUT CHARACTERISTICS VOD LVDS Differential Output Voltage ∆ VO DIFF Change in LVDS Output Swing Between Logic Levels VOS Output Offset Voltage, see Figure 1 VOS Output Offset Voltage, see Figure 1 ∆ VOS Output Offset Voltage Change Between Logic Levels IOS Output Short Circuit Current ZO Differential Output Impedance www.national.com Measured differentially, OutV = GND, VBG = Floating (Note 15) 510 400 mVP-P (min) 920 mVP-P (max) 280 mVP-P (min) 720 mVP-P (max) ±1 mV VBG = Floating 800 mV VBG = VA (Note 15) 1200 mV ±1 mV ±4 mA 100 Ohms Output+ & Output- connected to 0.8V 10 (Continued) NOTE: This product is currently in development and the parameters specified in this section are DESIGN TARGETS. The specifications in this section cannot be guaranteed until device characterization has taken place. The following specifications apply after calibration for VA = VDR = +1.9VDC, OutV = 1.9V, VIN FSR (a.c. coupled) = differential 870mVP-P, CL = 10 pF, Differential, a.c. coupled Sinewave Input Clock, fCLK = 1.5GHz at 0.5VP-P with 50% duty cycle, VBG = Floating, NonExtended 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) POWER SUPPLY CHARACTERISTICS IA Analog Supply Current PD = Low PD = High 745 10.2 TBD TBD mA (max) mA (max) IDR Output Driver Supply Current PD = Low PD = High 200 0.012 TBD TBD mA (max) mA (max) PD Power Consumption PD = Low PD = High 1.8 20 TBD TBD W (max) mW (max) PSRR1 D.C. Power Supply Rejection Ratio Change in Full Scale Error with change in VA from 1.8V to 2.0V 30 dB PSRR2 A.C. Power Supply Rejection Ratio 248 MHz, 50mVP-P riding on VA 51 dB AC ELECTRICAL CHARACTERISTICS 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 Input Clock Duty Cycle 500MHz ≤ Input clock frequency ≤ 1.5 GHz (Note 12) 50 20 80 % (min) % (max) tCL Input Clock Low Time (Note 11) 333 133 ps (min) tCH Input Clock High Time (Note 11) 333 133 ps (min) 45 55 % (min) % (max) 1.5 GHz (min) MHz DCLK Duty Cycle (Note 11) 50 tRS Reset Setup Time (Note 11) 150 ps tRH Reset Hold Time (Note 11) 250 ps tSD Syncronizing Edge to DCLK Output Delay fCLKIN = 1.5 GHz fCLKIN = 500 MHz TBD TBD ns tRPW Reset Pulse Width (Note 11) tLHT Differential Low to High Transition Time 10% to 90%, CL = 2.5 pF 250 ps tHLT Differential High to Low Transition Time 10% to 90%, CL = 2.5 pF 250 ps tOSK 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) tSU Data to DCLK Set-Up Time DDR Mode, 90˚ DCLK (Note 11) 667 ps tH DCLK to Data Hold Time DDR Mode, 90˚ DCLK (Note 11) 667 ps tAD Sampling (Aperture) Delay Input CLK+ Fall to Acquisition of Data 1.3 ns tAJ Aperture Jitter 0.4 ps rms tOD Input Clock to Data Output Delay (in addition to Pipeline Delay) 3.1 ns 4 50% of Input Clock transition to 50% of Data transition 11 Clock Cycles (min) www.national.com ADC083000 Converter Electrical Characteristics ADC083000 Converter Electrical Characteristics (Continued) NOTE: This product is currently in development and the parameters specified in this section are DESIGN TARGETS. The specifications in this section cannot be guaranteed until device characterization has taken place. The following specifications apply after calibration for VA = VDR = +1.9VDC, OutV = 1.9V, VIN FSR (a.c. coupled) = differential 870mVP-P, CL = 10 pF, Differential, a.c. coupled Sinewave Input Clock, fCLK = 1.5GHz at 0.5VP-P with 50% duty cycle, VBG = Floating, NonExtended 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 Typical (Note 8) Conditions Limits (Note 8) Units (Limits) AC ELECTRICAL CHARACTERISTICS Dd Outputs Pipeline Delay (Latency) (Notes 11, 14) Over Range Recovery Time 13 Db Outputs 14 Dc Outputs 13.5 Da Outputs 14.5 Differential VIN step from ± 1.2V to 0V to get accurate conversion Input Clock Cycles 1 Input Clock Cycle 500 ns tWU PD low to Rated Accuracy Conversion (Wake-Up Time) fSCLK Serial Clock Frequency (Note 11) 100 MHz tSSU Data to Serial Clock Setup Time (Note 11) 2.5 ns (min) tSH Data to Serial Clock Hold Time (Note 11) 1 ns (min) Serial Clock Low Time 4 Serial Clock High Time 4 1.4 x 105 ns (min) ns (min) tCAL Calibration Cycle Time tCAL_L CAL Pin Low Time See Figure 9 (Note 11) 80 Clock Cycles (min) tCAL_H CAL Pin High Time See Figure 9 (Note 11) 80 Clock Cycles (min) tCalDly Calibration delay determined by pin 127 See Section 1.1.1, Figure 9, (Note 11) 225 Clock Cycles (min) tCalDly Calibration delay determined by pin 127 See Section 1.1.1, Figure 9, (Note 11) 231 Clock Cycles (max) Clock Cycles 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: See AN-450, “Surface Mounting Methods and Their Effect on Product Reliability”. 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). www.national.com 12 (Continued) 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 half the sample rate. The data at each bus is clocked out at one half the sample rate. The second bus (D0 through D7) has a pipeline latency that is one Input Clock cycle less than the latency of the first bus (Dd0 through Dd7). 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). 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 that time required after the fall of the clock input for the sampling switch to open. The Sample/Hold circuit effectively stops capturing the input signal and goes into the “hold” mode the aperture delay time (tAD) after the input clock goes low. 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. LSB (LEAST SIGNIFICANT BIT) is the bit that has the smallest value or weight of all bits. This value is VFS / 2n where VFS is the differential full-scale amplitude of 650 mV or 870 mV as set by the FSR input 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+ & VDoutputs; each measured with respect to Ground. 20193246 FIGURE 1. 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. 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 −435 mV with the FSR pin high, or 1/2 LSB above a differential −325 mV with the FSR pin low. 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 after the falling edge of DCLK before the data update is present at the output pins. 13 www.national.com ADC083000 Converter Electrical Characteristics ADC083000 Specification Definitions 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. (Continued) 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. 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 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 +435 mV with the FSR pin high, or 1-1/2 LSB below a differential +325 mV with the FSR pin low. 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 one-half the sampling frequency, not including harmonics or d.c. www.national.com 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. 14 ADC083000 Transfer Characteristic 20193222 FIGURE 2. Input / Output Transfer Characteristic 15 www.national.com ADC083000 Timing Diagrams 20193214 FIGURE 3. ADC083000 Timing — SDR Clocking 20193259 FIGURE 4. ADC083000 Timing — DDR Clocking www.national.com 16 ADC083000 Timing Diagrams (Continued) 20193219 FIGURE 5. Serial Interface Timing 20193220 FIGURE 6. Clock Reset Timing in DDR Mode 20193223 FIGURE 7. Clock Reset Timing in SDR Mode with OUTEDGE Low 17 www.national.com ADC083000 Timing Diagrams (Continued) 20193224 FIGURE 8. Clock Reset Timing in SDR Mode with OUTEDGE High 20193225 FIGURE 9. Self Calibration and On-Command Calibration Timing www.national.com 18 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, 14 and 127 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. 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 abovementioned 80 input clock cycles low followed by 80 cycles high. 1.1 OVERVIEW The ADC083000 uses a calibrated folding and interpolating architecture that achieves over 7.4 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 500 MSPS 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 has a selectable 1:2 or 1:4 demultiplexer. When the 1:4 demultiplexer is selected, data is output on all four output busses at a quarter of the ADC sampling rate. When the 1:2 demultiplexer is selected, data is output on only the Dc and Dd output busses at half of the ADC sampling rate. In both cases, 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. 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 power-on settling time of the power supply. 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. 1.1.2 Acquiring the Input Data is acquired at the falling edge of CLK+ (pin 18) and the digital equivalent of that data is available at the digital outputs 13 input clock cycles later for the Dd and Dc output buses and 14 input clock cycles later for the Db and Da output buses. There is an additional internal delay called tOD before the data is available at the outputs. See the Timing Diagram. 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 self 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. 1.1.1 Self-Calibration A self-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 self 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, self calibration must be re-run whenever the sense of the FSR pin is changed. For best performance, we recommend that self calibration be run 20 seconds or more after application of power and whenever the operating temperature changes significantly relative to the specific system performance re- 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 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 register-based control and those pin-based 19 www.national.com ADC083000 quirements. 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. During the calibration process, the input termination resistor is trimmed to a value that is equal to REXT / 33. 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. 1.0 Functional Description ADC083000 1.0 Functional Description 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). (Continued) 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. 1.1.5.1 Selectable Output Demultiplexer 1.1.4 The Analog Inputs 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. To ease the timing requirements of a system interfacing to the ADC output ports, a selectable output demultiplexer is provided. The demultiplexer selection is programmed in the extended control mode using the Configuration Register. Refer to section 1.4. The default setting selects the 1:4 demultiplexer with data on each of the four output busses at 750MHz with a 1.5GHz input clock. When Bit 0 in the Configuration Register is programmed to 1b, the data is present on Dc and Dd output busses at a rate of 1.5GHz with a 1.5GHz input clock. NOTE: Bit 0 can only be programmed 1b when the ADC083000 is in DDR mode. If the device is not in DDR mode and Bit 0 is set to 1b, the data outputs will not change. 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 870 mVP-P, while grounding pin 14 causes an input full-scale range setting of 650 mVP-P. The full-scale range setting operates equally on both ADCs. 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 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 must be driven with an a.c. coupled, differential clock signal. Section 2.3 describes the use of the clock input pins. A differential LVDS output clock is available for use in latching the ADC output data into whatever device is used to receive the data. The ADC083000 offers options for input and output clocking. These options include a choice of on which DCLK (DCLK) edge the output data transitions, and a choice of Single Data Rate (SDR) or Double Data Rate (DDR) outputs. The ADC083000 also has the option to use a duty cycle corrected clock receiver as part of the input clock circuit. This TABLE 1. Input Channel Samples Produced at Data Outputs Data Outputs (Always sourced with respect to fall of DCLK) 1:4 Demultiplex Dd ADC1 sampled with fall of CLK 13 cycles earlier Db ADC1 sampled with fall of CLK 14 cycles earlier Dc ADC2 sampled with rise of CLK 13.5 cycles earlier Da ADC2 sampled with rise of CLK 14.5 cycles earlier 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. ADC1 sampled with fall of CLK 13 cycles earlier ADC2 sampled with rise of CLK 13.5 cycles earlier 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. 1.1.6 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 14) is high or 2.2 mA when the OutV input is low. For short LVDS lines and low noise systems, satisfactory performance may be realized with the OutV input low, which results in lower 1.1.5.3 Double Data Rate A choice of single data rate (SDR) or double data rate (DDR) output is offered. With single data rate the output clock (DCLK) frequency is the same as the data rate of the two www.national.com 1:2 Demultiplex 20 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. (Continued) 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 output have a typical common mode voltage of 800mV when the VBG pin is unconnected and floating. This common mode voltage can be increased to 1.2V by tying the VBG pin to VA if a higher common mode is required. 1.2 NORMAL/EXTENDED CONTROL The ADC083000 may be operated in one of two modes. In the simpler standard control mode, the user affects available configuration and 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. 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. Table 2 shows how several of the device features are affected by the control mode chosen. 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 TABLE 2. Features and modes Feature Normal Control Mode Extended Control Mode SDR or DDR Clocking Selected with pin 4 Selected with DE bit in the Configuration Register DDR Clock Phase Not Selectable (0˚ Phase Only) Selected with DCP bit in the Configuration Register. See Section 1.4 REGISTER DESCRIPTION SDR Data transitions with rising or falling DCLK edge Selected with pin 4 Selected with the OE bit in the Configuration Register LVDS output level Selected with pin 3 Selected with the OV bit (9)in the Configuration Register Power-On Calibration Delay Delay Selected with pin 127 Short delay only. Full-Scale Range Options (650 mVP-P or 870 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. Selected using registers 3H. Input Offset Adjust Not possible Separate ± 45 mV adjustments in 512 steps using registers 2h. Sampling Clock Phase Adjustment The Clock Phase is adjusted automatically The clock phase can be adjusted manually through the Coarse & Fine registers (Eh and Dh). Test Pattern Not Possible A test pattern can be made present at the data outputs by programming register Fh. Resistor Trim Disable Not possible The DCLK outputs will continuously be present when bit 5 is set 1b in register 1h. Not possible If the device is set in DDR, the output can be programmed to be 1:2 demultiplexed. Bit 0 in register 1h is set 1b, this selects 1:2 demultiplex. If bit 0 is 0b, this selects 1:4 demultiplex. Selectable Output Demultiplexer 21 www.national.com ADC083000 1.0 Functional Description ADC083000 1.0 Functional Description 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. (Continued) 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. TABLE 3. Extended Control Mode Operation (Pin 14 Floating) TABLE 4. Register Addresses Feature Extended Control Mode Default State SDR or DDR Clocking DDR Clocking DDR Clock Phase Data changes with DCLK edge (0˚ phase) A3 A2 A1 A0 Hex Register Addressed 0 0 0 0 0h Reserved LVDS Output Amplitude Normal amplitude (710 mVP-P) 0 0 0 1 1h Configuration 0 0 1 0 2h Offset Calibration Delay Short Delay 0 0 1 1 3h Full-Scale Range 700 mV nominal for both channels Full-Scale Voltage Adjust 0 1 0 0 4h Reserved Input Offset Adjust No adjustment for either channel 0 1 0 1 5h Reserved 0 1 1 0 6h Reserved Selectable Output Demultiplexer 1:4 demultiplex 0 1 1 1 7h Reserved 1 0 0 0 8h Reserved Resistor Trim Disable Trim enabled, DCLK not continuously present at output 1 0 0 1 9h Reserved 1 0 1 0 Ah Reserved Test Pattern Not present at output 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 Loading Sequence: A3 loaded after H0, A0 loaded last 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. 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 Power-On Reset (POR) state of each control bit. 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. See the Timing Diagram. Each Register access consists of 32 bits, as shown in Figure 5 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 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. Configuration Register Addr: 1h (0001b) W only (0xB2DF) D15 D14 D13 1 0 1 D11 D10 D9 D8 DCS DCP D12 nDE OV OE D7 D6 D5 D4 D3 D2 D1 D0 1 1 RTD 1 1 1 1 nOT Bit 15 Must be set to 1b Bit 14 Must be set to 0b Bit 13 Must be set to 1b Bit 12 DCS:Duty Cycle Stabilizer. When this bit is set to 1b , a duty cycle stabilzation circuit is applied to the clock input. When this bit is set to 0b the stabilzation circuit is disabled. POR State: 1b www.national.com 22 Bit 11 Bit 10 (Continued) Bits 0 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. nTO: 1:2 Outpu Enablt. The setting of this bit is contingent on the device being in the DDR mode which is set by bit 10. If bit 10 is 0b, bit 0 can be set 1b. By doing this, the output data rate on the Dc and Dd channels will be that of the input clock rate. DCLK will be in DDR mode. The Da and Db channels will be held in Hi-Z. If the device is not in DDR mode and bit 0 is set to 1b, the data outputs will not change. POR State: 0b 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 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. Offset Adjust Addr: 2h (0010b) D15 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 710 mVP-P is used. When this bit is set to 0b, the reduced output amplitude of 510 mVP-P is used. Bits 5 RTD: Resistor Trim Disable. When this bit is set 0b, the DCLK outputs will not be active during a calibration cycle. When this bit is set 1b, the DCLK outputs will be continuously present during a calibration cycle. Note that in this setting, only the ADC is calibrated. D10 D9 D8 (LSB) D5 D4 D3 D2 D1 D0 Sign 1 1 1 1 1 1 1 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 Sign bit. 0b gives positive offset, 1b gives negative offset. POR State: 0b Bit 6:0 Must be set to 1b Full-Scale Voltage Adjust Addr: 3h (0011b) D15 D14 D13 (MSB) POR State: 0b Must be set to 1b. D11 D6 Bit 7 OE: Output Edge. This bit selects the DCLK edge with which the data words transition in the SDR mode and has the same effect as the OutEdge pin in the normal control mode. When this bit is 1, the data outputs change with the rising edge of DCLK+. When this bit is 0, the data output change with the falling edge of DCLK+. Bits 7:6 D12 D7 POR State: 1b Bit 8 D13 Offset Value Bits 15:8 POR State: 0b Bit 9 D14 (MSB) W only (0x007F) W only (0x807F) D12 D11 D10 D9 D8 Adjust Value D7 D6 D5 D4 D3 D2 D1 D0 (LSB) 1 1 1 1 1 1 1 POR State: 0b Bits 4:1 Must be set to 1b. 23 www.national.com ADC083000 1.0 Functional Description ADC083000 1.0 Functional Description Bit 15:7 (Continued) Bit 6:0 Full Scale Voltage Adjust Value. The input full-scale 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. 0000 0000 0 560mVP-P 1000 0000 0 Default Value 700mVP-P 1111 1111 1 840mVP-P Extended Clock Phase Adjust Coarse Addr: Eh (1110b) D15 D13 (MSB) D11 D8 1 1 1 D6 D5 D4 D3 D2 D1 D0 1 1 1 1 1 1 1 1 Enable, default is 0b Addr: Fh (1111b) W only (0xFFFF) D15 D14 D13 D12 D11 D10 D9 D8 1 1 1 1 nTP 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 D10 D9 D8 D7 D6 D5 D4 D3 D2 D1 D0 (LSB) 1 1 1 1 1 1 1 www.national.com D9 D7 Bit 11 FAM Bit 15:7 D10 Test Pattern Register W only (0x3FFF) D12 D11 Bits 10:0 Must be set to 1b Extended Clock Phase Adjust Fine D14 D12 Bit 14:11 Coarse Adjust Magnitude. Each LSB results in approximately 70ps of clock adjust. The default setting is 0000b Must be set to 1b Addr: Dh (1101b) D13 W only (0x07FF) CAM Bit 15 POR State: 1000 0000 0b (no adjustment) D15 D14 ENA 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. Bits 6:0 Must be set to 1b nTP: Test Pattern Output Enable. When this bit is set 0b, 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, DDR and the 1:2 output modes. POR State: 1b Fine Adjust Magnitude. Each LSB results in approximately 110ps of non-linear clock adjust. The default setting is 000000000b Bit 10:0 24 Must be set to 1b 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 6, Figure 7 and Figure 8 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. (Continued) 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. 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. 20193230 FIGURE 10. Extended Mode Offset Behaviour 1.5 MULTIPLE ADC SYNCHRONIZATION The ADC083000 has the capability to precisely reset its sampling clock input to DCLK output relationship as determined by the user-supplied DCLK_RST pulse. 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 DCLK_RST signal must observe some timing requirements that are shown in Figure 6, Figure 7 and Figure 8 of 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 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 in 1:4 Demultiplex Mode Time Da Db Dc Dd T0 01h 02h 03h 04h T1 FEh FDh FCh FBh T2 01h 02h 03h 04h T3 FEh FDh FCh FBh T4 01h 02h 03h 04h T5 01h 02h 03h 04h T6 FEh FDh FCh FBh T7 01h 02h 03h 04h T8 FEh FDh FCh FBh T9 01h 02h 03h 04h T10 01h 02h 03h 04h T11 ... ... ... ... Comments Pattern Sequence n Pattern Sequence n+1 Pattern Sequence n+2 With the part programmed into the 1:2 demutliplex mode, the test pattern’s order will be as described in Table 6 25 www.national.com ADC083000 1.0 Functional Description ADC083000 1.0 Functional Description (Continued) TABLE 6. Test Pattern by Output Port in 1:2 Demultiplex Mode Time Dc Dd T0a 01h 02h Comments T0b 03h 04h T1a FEh FDh T1b FCh FBh T2a 01h 02h T2b 03h 04h T3a FEh FDh T3b FCh FBh T4a 01h 02h T4b 03h 04h T5a 01h 02h T5b 03h 04h T6a FEh FDh T6b FCh FBh T7a 01h 02h Pattern Sequence n Pattern Sequence n+1 Note: The same bit pattern repeats when the test pattern sequence is concatenated. 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 7 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 7 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. The test pattern is selected by setting bit 15 (TPO) in the Extended Configuration Register. The start of the test pattern’s appearance at the digital output will not be all at once, but will stabilize to the following fixed-pattern within several DCLK cycles. 2.0 Applications Information 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. TABLE 7. DIFFERENTIAL INPUT TO OUTPUT RELATIONSHIP (Non-Extended Control Mode, FSR High) The internal bandgap-derived reference voltage has a nominal value of 650 mV or 870 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 1200mV the VBG pin can be connected directly to the supply rails. VIN− Output Code VCM + 217.5mV 0000 0000 VCM − 109 mV VCM + 109 mV 0100 0000 VCM VCM 0111 1111 / 1000 0000 VCM + 109 mV VCM −109 mV 1100 0000 VCM + 217.5mV VCM − 217.5mV 1111 1111 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. Note that a precise d.c. common mode voltage must be present at the ADC inputs. This common mode voltage, VCMO, is provided on-chip when a.c. input coupling is used and the input signal is a.c. coupled to the ADC. 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 50k-Ohm resistors. IMPORTANT NOTE: An Analog input channel that is not used should be left floating when the input is a.c. coupled. Do not connect an unused analog input to ground. 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 650 mVP-P or 870 mVP-P, or can be adjusted to values between 560 mVP-P and 840 mVP-P in the Extended Control mode through the www.national.com VIN+ VCM − 217.5mV 26 Be sure that the current drawn from the VCMO output does not exceed 100 µA. The Input impedance in the d.c. coupled mode (VCMO pin not grounded) consists of a precision 100Ω resistor between VIN+ and VIN− and a capacitance from each of these inputs to ground. In the a.c. coupled mode the input appears the same except there is also a resistor of 50K between each analog input pin and the VCMO potential. (Continued) Driving the inputs beyond full scale will result in a saturation or clipping of the reconstructed output. 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. The easiest way to accomplish single-ended to differential signal conversion is with an appropriate balun-connected transformer, as shown in Figure 13. 20193244 FIGURE 11. Differential Input Drive When the d.c. coupled mode is used, a 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. IMPORTANT NOTE: An analog input channel that is not used should be tied to the VCMO voltage when the input is d.c coupled. Do not connect unused analog inputs to ground. 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. If d.c. coupling is used, it is best to servo the input common mode voltage, using the VCMO pin, to maintain optimum performance. An example of this type of circuit is shown in Figure 12. 20193243 FIGURE 13. Single-Ended to Differential signal conversion with a balun-connected transformer The 100 Ohm external resistor placed accross the output terminals of the balun in parallel with the ADC083000’s on-chip 100 Ohm resistor makes a 50 Ohms differential impedance at the balun output. Or, 25 Ohms to virtual ground at each of the balun output terminals. Looking into the balun, the source sees the impedance of the first coil in series with the impedance at the output of that coil. Since the transformer has a 1:1 turns ratio, the impedance across the first coil is exactly the same as that at the output of the second coil, namely 25 Ohms to virtual ground. So, the 25 Ohms across the first coil in series with the 25 Ohms at its output gives 50 Ohms total impedance to match the source. 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. 20193255 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 bandgap reference. The FSR pin controls the effective reference voltage of the ADC083000 such that the differential full-scale input range at the analog inputs is 870 mVP-P with the FSR pin high, or is 650 mVP-P with FSR pin low. Best SNR is obtained with FSR high, but better distortion and SFDR are obtained with the FSR pin low. FIGURE 12. Example of Servoing the Analog Input with VCMO One such circuit should be used in front of the VIN+ input and another in front of the VIN− input. In that figure, RD1, RD2 and RD3 are used to divide the VCMO potential so that, after being gained up by the amplifier, the input common mode voltage is equal to VCMO from the ADC. RD1 and RD2 are split to allow the bypass capacitor to isolate the input signal from VCMO. RIN, RD2 and RD3 will divide the input signal, if necessary. Capacitor "C" in Figure 12 should be chosen to keep any component of the input signal from affecting VCMO. 27 www.national.com ADC083000 2.0 Applications Information ADC083000 2.0 Applications Information 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 and fIN is the maximum input frequency, in Hertz, to the ADC analog input. (Continued) 2.3 THE CLOCK INPUTS The ADC083000 has differential LVDS clock inputs, CLK+ and 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. 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. 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. 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, Self Calibration, Calibration Delay, Output Edge Synchronization choice, LVDS Output Level choice and a Power Down feature. 2.4.1 Full-Scale Input Range Setting The input full-scale range can be selected to be either 650 mVP-P or 870 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 anywhere from 560 mVP-P to 840 mVP-P. See Section 2.2 for more information. 20193247 2.4.2 Self Calibration The ADC083000 self-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. FIGURE 14. Differential (LVDS) Input Clock Connection The differential input 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 tJ(MAX) = (VIN(P-P)/VINFSR) x (1/(2(N+1) x π x fIN)) www.national.com 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 On-Command 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 200 mW. The power consumption will be normal after the clock starts. 28 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. (Continued) 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 self 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. 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. During a Power-On calibration cycle, both the ADC and the input termination resistor are calibrated. As ENOB changes slightly with junction temperature, an On-Command calibration can be executed to bring the performance of the ADC inline. By default, this 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, which will guarantee that the DCLK is continuously present at the output. The Resistor Trim Disable can be programmed in register 1h when in the Extended Control mode. Refer to section 1.4 for register programming information. 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. 2.5 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 started on the odd falling edges of the CLK+ pin are available on one of the two LVDS buses, while the results of conversions started on the even falling 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 200 MSPS, the effective rate can be reduced to as low as 100 MSPS by using the results available on just one of the the two LVDS buses and a 200 MHz input clock, decimating the 200 MSPS data by two. 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. See the Timing Diagram section for details. 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. 29 www.national.com ADC083000 2.0 Applications Information ADC083000 2.0 Applications Information (Continued) 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. 20193254 FIGURE 15. Non-Spiking Power Supply 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. 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. 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. 2.6.2 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. 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, 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 power-on 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. 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. www.national.com 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 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. 30 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. (Continued) 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 maintaining 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 addtional to the above calculation). 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. 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 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. 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 full-scale range, single-ended or differential input and input coupling (a.c. or d.c.) 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 non-extended control mode. 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 31 www.national.com ADC083000 2.0 Applications Information ADC083000 2.0 Applications Information volt below ground. Controlling the impedance of high speed lines and terminating these lines in their characteristic impedance should control overshoot. (Continued) 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 . TABLE 8. Non-Extended Control Mode Operation (Pin 14 High or Low) Pin Low High Floating 3 0.50 VP-P Output 0.70 VP-P Output n/a 4 OutEdge = Neg OutEdge = Pos DDR 14 650 mVP-P input range 870 mVP-P input range Extended Control Mode 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 (650 mVP-P and 870 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. 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 non-extended 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. 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 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 www.national.com 32 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 National does not assume any responsibility for use of any circuitry described, no circuit patent licenses are implied and National reserves the right at any time without notice to change said circuitry and specifications. For the most current product information visit us at www.national.com. LIFE SUPPORT POLICY NATIONAL’S PRODUCTS ARE NOT AUTHORIZED FOR USE AS CRITICAL COMPONENTS IN LIFE SUPPORT DEVICES OR SYSTEMS WITHOUT THE EXPRESS WRITTEN APPROVAL OF THE PRESIDENT AND GENERAL COUNSEL OF NATIONAL SEMICONDUCTOR CORPORATION. As used herein: 1. Life support devices or systems are devices or systems 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. 2. A critical component is any component of 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. BANNED SUBSTANCE COMPLIANCE National Semiconductor follows the provisions of the Product Stewardship Guide for Customers (CSP-9-111C2) and Banned Substances and Materials of Interest Specification (CSP-9-111S2) for regulatory environmental compliance. Details may be found at: www.national.com/quality/green. Lead free products are RoHS compliant. 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: +44 (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 ADC083000 High Performance, Low Power, 8-Bit, 3 GSPS A/D Converter Physical Dimensions