KAD5610P Preliminary Dual 10-Bit, 250/210/170/125MSPS A/D Converter General Description The KAD5610P is a family of low-power, highperformance, dual-channel 10-bit, analog-to-digital converters. Designed with Kenet’s proprietary FemtoCharge® technology on a standard CMOS process, the family supports sampling rates of up to 250MSPS. The KAD5610P-25 is the fastest member of this pin-compatible family, which also features sample rates of 210MSPS (KAD5610P-21), 170MSPS (KAD5610P-17) and 125MSPS ( KAD5610P-12). CLKP CLKOUTP CLKN CLKOUTN AINP D[9:0]P AINN D[9:0]N ORP VREF A serial peripheral interface (SPI) port allows for extensive configurability, as well as fine control of gain, skew and offset matching between the two converter cores. VCM ORN OUTFMT BINP OUTMODE BINN Digital output data is presented in selectable LVDS or CMOS formats. The KAD5610P is available in a 72contact QFN package with an exposed paddle. Performance is specified over the full industrial temperature range (-40 to +85°C). VREF 1.25V Features • • • • • • • • • • • Programmable gain, offset and skew control 1.3 GHz analog input bandwidth 52fs Clock Jitter Over-range indicator Selectable Clock Divider: ÷1, ÷2 or ÷4 Clock Phase Selection Nap and Sleep modes Two’s complement, Gray code or Binary data format DDR LVDS-compatible or LVCMOS outputs Programmable Built-in Test Patterns 1.8V Analog and Digital Supplies Applications • • • • • • Power Amplifier Linearization Radar and Satellite Antenna Array Processing Broadband Communications High-Performance Data Acquisition Communications Test Equipment WiMAX and Microwave Receivers Key Specifications • • • SNR = 60.3dBFS for fIN = 124MHz (-1dBFS) SFDR = 80dBc for fIN = 124MHz (-1dBFS) Power consumption • 400mW @ 250MSPS • 312mW @ 125MSPS Pin-Compatible Family Model Resolution Speed (MSPS) KAD5612P-25 12 250 KAD5612P-21 12 210 KAD5612P-17 12 170 KAD5612P-12 12 125 KAD5610P-25 10 250 KAD5610P-21 10 210 KAD5610P-17 10 170 KAD5610P-12 10 125 300 Unicorn Park Dr., Woburn, MA 01801 Sales: 1-781-497-0060 FemtoCharge is a registered trademark of Kenet, Inc. Rev 0.5.1 Preliminary [email protected] Copyright © 2007, Kenet, Inc. Page 1 KAD5610P Table of Contents Section Electrical Specifications Pages 3–7 Section Serial Peripheral Interface Pages 18–24 DC Specifications 3 SPI Physical Interface 19 AC Specifications 4 SPI Configuration 20 Digital Specifications 5 DUT Information 21 Timing Diagrams 5 Indexed DUT Configuration/Control 21 Switching Specifications 6 Global DUT Configuration/Control 22 Absolute Maximum Ratings 6 DUT Test 23 Thermal Impedance 7 SPI Memory Map 24 ESD 7 Pinout/Package Information 8–9 Equivalent Circuits 25 Layout Considerations 25 Pin Descriptions 8 Definitions 26 Pin Configuration 9 Outline Dimensions 27 Typical Performance Characteristics 10–13 Ordering Guide 28 Theory of Operation 14–18 Revision History 28 Functional Description 14 Power-On Calibration 14 User-Initiated Reset 15 Analog Input 15 Clock Input 16 Jitter 16 Voltage Reference 16 Digital Outputs 17 Power Dissipation 17 Nap/Sleep 17 Data Format 18 Rev 0.5.1 Preliminary Page 2 KAD5610P Electrical Specifications All specifications apply under the following conditions unless otherwise noted: AVDD = 1.8V, OVDD = 1.8V, TA = -40°C to +85°C, AIN = -1dBFS, fSAMPLE = Maximum Conversion Rate (per speed grade). DC Specifications KAD5610P-25 Parameter KAD5610P-21 KAD5610P-17 KAD5610P-12 Symbol Conditions Min Typ Max Min Typ Max Min Typ Max Min Typ Max Units VFS Differential 1.38 1.45 1.59 1.38 1.45 1.59 1.38 1.45 1.59 1.38 1.45 1.59 VPP Analog Input Full-Scale Analog Input Range Input Resistance RIN Differential 1000 1000 1000 1000 Ω Input Capacitance CIN Differential 4 4 4 4 pF Full Scale Range Temp. Drift AVTC Full Temp 90 90 90 90 ppm/°C Input Offset Voltage VOS ±1.5 ±1.5 ±1.5 ±1.5 mV Gain Error EG ±0.6 ±0.6 ±0.6 ±0.6 % VCM 0.535 0.535 0.535 0.535 V Common-Mode Output Voltage Power Requirements 1.8V Analog Supply Voltage AVDD 1.7 1.8 1.9 1.7 1.8 1.9 1.7 1.8 1.9 1.7 1.8 1.9 V 1.8V Digital Supply Voltage OVDD 1.7 1.8 1.9 1.7 1.8 1.9 1.7 1.8 1.9 1.7 1.8 1.9 V 1.8V Analog Supply Current IAVDD 157 TBD 142 TBD 130 TBD 116 TBD mA 1.8V Digital Supply Current IOVDD 65.0 TBD 63.6 TBD 60.7 TBD 57.2 TBD mA Power Supply Rejection Ratio PSRR -53 Normal Mode PD 400 TBD 371 TBD 345 TBD 312 TBD mW Nap Mode PD 40 TBD 40 TBD 40 TBD 40 TBD mW Sleep Mode PD 10 TBD 10 TBD 10 TBD 10 TBD mW -53 -53 -53 dBFS Power Dissipation Rev 0.5.1 Preliminary Page 3 KAD5610P AC Specifications KAD5610P-25 Parameter Symbol Conditions Min Differential Nonlinearity DNL fIN = 10MHz Integral Nonlinearity INL fIN = 10MHz Minimum Conversion Rate fS MIN Maximum Conversion Rate fS MAX Signal-to-Noise Ratio SNR SINAD ENOB SFDR Two-Tone SFDR IMD 2TSFDR Channel to Channel Isolation Word Error Rate Full Power Bandwidth Rev 0.5.1 Preliminary FPBW Max Min TBD TBD TBD TBD TBD TBD TBD Typ Max Min TBD TBD TBD TBD TBD 210 KAD5610P-12 Typ Max Units TBD TBD LSB TBD TBD LSB TBD MSPS TBD 170 125 MSPS fIN = 10MHz 60.4 dBFS 60.3 dBFS TBD 60.3 TBD TBD TBD dBFS fIN = 230MHz 60.2 fIN = 400MHz 60.0 TBD TBD TBD dBFS fIN = 1000MHz 58.6 TBD TBD TBD dBFS fIN = 10MHz 59.3 dBFS fIN = 70MHz 59.2 dBFS TBD dBFS 59.2 TBD TBD TBD dBFS fIN = 230MHz 59.1 fIN = 400MHz 58.4 TBD TBD TBD dBFS fIN = 1000MHz 52.2 TBD TBD TBD dBFS dBFS fIN = 10MHz 9.6 Bits fIN = 70MHz 9.5 Bits TBD 9.5 TBD TBD TBD Bits fIN = 230MHz 9.5 fIN = 400MHz 9.4 TBD TBD TBD Bits fIN = 1000MHz 8.4 TBD TBD TBD Bits fIN = 10MHz 84 84 85 85 dBc fIN = 70MHz 84 83 82 83 dBc 79 dBc fIN = 140MHz Intermodulation Distortion TBD Typ KAD5610P-17 fIN = 70MHz fIN = 140MHz Spurious-Free Dynamic Range Min 250 fIN = 140MHz Effective Number of Bits Max TBD fIN = 140MHz Signal-to-Noise and Distortion Typ KAD5610P-21 TBD Bits 80 TBD 80 TBD 80 TBD fIN = 230MHz 79 77 77 79 dBc fIN = 400MHz 71 TBD TBD TBD dBc fIN = 1000MHz 55 TBD TBD TBD dBc fIN = 10MHz TBD fIN = 70MHz -83 fIN = 170MHz -84 TBD TBD TBD dBc fIN = 10MHz TBD TBD TBD TBD dBc fIN = 70MHz TBD fIN = 170MHz TBD TBD TBD TBD TBD TBD TBD TBD TBD TBD TBD TBD TBD TBD TBD TBD TBD TBD TBD dBc TBD TBD dBc dBc dBc fIN = 10MHz 90 90 90 90 dB fIN = 124MHz 90 90 90 90 dB fIN = 170MHz TBD TBD TBD TBD dB 10-12 10-12 10-12 10-12 1.3 1.3 1.3 1.3 GHz Page 4 KAD5610P Digital Specifications Parameter Symbol Conditions Min Typ Max Units Inputs Input Current High (RESETN) IIH VIN = 1.8V 0 1 10 µA Input Current Low (RESETN) IIL VIN = 0V 25 50 75 µA Input Current High (OUTMODE, NAP/SLP, CLKDIV, OUTFMT ) IIH TBD 25 TBD µA Input Current Low (OUTMODE, NAP/SLP, CLKDIV, OUTFMT ) IIL TBD 25 TBD µA CDI 3 pF Differential Output Voltage VT 210 mV Output Offset Voltage VOS TBD mV Output Rise Time tR 500 ps Output Fall Time tF 500 ps Voltage Output High VOH OVDD-0.1 V Voltage Output Low VOL 0.1 V Output Rise Time tR TBD ns Output Fall Time tF TBD ns Input Capacitance LVDS Outputs CMOS Outputs Timing Diagrams Figure 1. LVDS Timing Diagram—DDR Rev 0.5.1 Preliminary Figure 2. CMOS Timing Diagram—DDR Page 5 KAD5610P Switching Specifications Parameter Symbol Min Typ Max Units ADC Aperture Delay tA 375 ps RMS Aperture Jitter jA 52 fs Input Clock to Output Clock Propagation Delay tCPD TBD TBD TBD ps Input Clock to Data Propagation Delay tPD TBD TBD TBD ps Output Clock to Data Propagation Delay tDC TBD TBD TBD ps Latency (Pipeline Delay) L 7.5 cycles Over Voltage Recovery tOVR 1 cycles Absolute Maximum Ratings1 Parameter Min Max Units AVDD to AVSS -0.4 2.1 V OVDD to OVSS -0.4 2.1 V AVSS to OVSS -0.3 0.3 V Analog Inputs to AVSS -0.4 AVDD + 0.3 V Clock Inputs to AVSS -0.4 AVDD + 0.3 V Logic Input to AVSS -0.4 OVDD + 0.3 V Logic Inputs to OVSS -0.4 OVDD + 0.3 V Operating Temperature -40 85 °C Storage Temperature -65 150 °C 150 °C Junction Temperature 1. Exposing the device to levels in excess of the maximum ratings may cause permanent damage. Exposure to maximum conditions for extended periods may affect device reliability. Rev 0.5.1 Preliminary Page 6 KAD5610P Thermal Impedance Parameter Symbol Typ Unit Junction to Paddle2 ΦJP 30 °C/W Junction to Case2 ΦJC TBD °C/W Junction to Ambient2 ΦJA TBD °C/W 2. Paddle soldered to ground plane. ESD Electrostatic charge accumulates on humans, tools and equipment and may discharge through any metallic package contacts (pins, balls, exposed paddle, etc.) of an integrated circuit. Industry-standard protection techniques have been utilized in the design of this product. However, reasonable care must be taken in the storage and handling of ESD sensitive products. Contact Kenet for the specific ESD sensitivity rating of this product. Rev 0.5.1 Preliminary Page 7 KAD5610P Pin Descriptions Pin # LVDS [LVCMOS] Name LVDS [LVCMOS] Function 1, 6, 19, 24, 71 AVDD 1.8V Analog Supply 2-5, 17, 18, 28-35 DNC Do Not Connect 7, 10-12, 72 AVSS Analog Ground 8, 9 BINP, BINN B-Channel Analog Input Positive, Negative 13, 14 AINN, AINP A-Channel Analog Input Negative, Positive 15 VCM Common Mode Output 16 CLKDIV Clock Divider Control 20, 21 CLKP, CLKN Clock Input True, Complement 22 OUTMODE Output Mode (LVDS, LVCMOS) 23 NAPSLP Power Control (Nap, Sleep modes) 25 RESETN Power On Reset (Active Low) 26, 45, 55, 65 OVSS Output Ground 27, 36, 56 OVDD 1.8V Output Supply 37, 38 D0N, D0P [NC, D0] LVDS Bit 0 (LSB) Output Complement, True [NC, LVCMOS Bit 0] 39, 40 D1N, D1P [NC, D1] LVDS Bit 1 Output Complement, True [NC, LVCMOS Bit 1] 41, 42 D2N, D2P [NC, D2] LVDS Bit 2 Output Complement, True [NC, LVCMOS Bit 2] 43, 44 D3N, D3P [NC, D3] LVDS Bit 3 Output Complement, True [NC, LVCMOS Bit 3] 46 RLVDS LVDS Bias Resistor (connect to OVSS with a 10kΩ, 1% resistor) 47, 48 CLKOUTN, CLKOUTP [NC, CLKOUT] LVDS Clock Output Complement, True [NC, LVCMOS CLKOUT] 49, 50 D4N, D4P [NC, D4] LVDS Bit 4 Output Complement, True [NC, LVCMOS Bit 4] 51, 52 D5N, D5P [NC, D5] LVDS Bit 5 Output Complement, True [NC, LVCMOS Bit 5] 53, 54 D6N, D6P [NC, D6] LVDS Bit 6 Output Complement, True [NC, LVCMOS Bit 6] 57, 58 D7N, D7P [NC, D7] LVDS Bit 7 Output Complement, True [NC, LVCMOS Bit 7] 59, 60 D8N, D8P [NC, D8] LVDS Bit 8 Output Complement, True [NC, LVCMOS Bit 8] 61, 62 D9N, D9P [NC, D9] LVDS Bit 9 (MSB) Output Complement, True [NC, LVCMOS Bit 9] 63, 64 ORN, ORP [NC, OR] LVDS Over Range Complement, True [NC, LVCMOS Over Range] 66 SDO SPI Serial Data Output (4.7kΩ pull-up to OVDD is required) 67 CSB SPI Chip Select (active low) 68 SCLK SPI Clock 69 SDIO SPI Serial Data Input/Output 70 OUTFMT Output Data Format (Two’s Comp., Gray Code, Offset Binary) Exposed Paddle AVSS Analog Ground LVCMOS Output Mode Functionality is shown in brackets (NC = No Connection) Rev 0.5.1 Preliminary Page 8 KAD5610P 72 71 70 69 68 67 66 65 64 63 62 61 60 59 58 57 56 55 AVSS AVDD OUTFMT SDIO SCLK CSB SDO OVSS ORP ORN D9P D9N D8P D8N D7P D7N OVDD OVSS Pin Configuration 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 KAD5610 72 QFN Top View Not to Scale 54 53 52 51 50 49 48 47 46 45 44 43 42 41 40 39 38 37 D6P D6N D5P D5N D4P D4N CLKOUTP CLKOUTN RLVDS OVSS D3P D3N D2P D2N D1P D1N D0P D0N AVDD CLKP CLKN OUTMODE NAPSLP AVDD RESETN OVSS OVDD DNC DNC DNC DNC DNC DNC DNC DNC OVDD 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 AVDD DNC DNC DNC DNC AVDD AVSS BINP BINN AVSS AVSS AVSS AINN AINP VCM CLKDIV DNC DNC Figure 3. Pin Configuration Rev 0.5.1 Preliminary Page 9 KAD5610P Typical Performance Curves SNRFS (dBFS) & SFDR (dBc) 90 85 SFDR 80 75 70 TBD 65 60 55 SNRFS 50 0 200 400 600 800 1000 INPUT FREQUENCY (MHz) Figure 4. SNR & SFDR vs. fIN Figure 5. HD2 & HD3 vs. fIN TBD TBD Figure 6. SNR & SFDR vs. AIN Figure 7. HD2 & HD3 vs. AIN TBD TBD Figure 8. SNR & SFDR vs. fSAMPLE Figure 9. HD2 & HD3 vs. fSAMPLE Rev 0.5.1 Preliminary Page 10 KAD5610P Typical Performance Curves TBD TBD Figure 10. Power vs. fSAMPLE Figure 11. Differential Nonlinearity TBD TBD Figure 12. Integral Nonlinearity Figure 13. SNR & SFDR vs. VCM TBD TBD Figure 14. Noise Histogram Figure 15. Single Tone Spectrum @ 10 MHz Rev 0.5.1 Preliminary Page 11 KAD5610P Typical Performance Curves TBD TBD Figure 16. Single Tone Spectrum @ 70 MHz Figure 17. Single Tone Spectrum @ 140 MHz TBD TBD Figure 18. Single Tone Spectrum @ 240 MHz Figure 19. Single Tone Spectrum @ 500 MHz TBD TBD Figure 20. Two-Tone Spectrum @ 10 MHz Figure 21. Two-Tone Spectrum @ 70 MHz Rev 0.5.1 Preliminary Page 12 KAD5610P Typical Performance Curves TBD TBD Figure 22. Two-Tone Spectrum @ 140 MHz Figure 23. Two-Tone Spectrum @ 240 MHz TBD TBD Figure 24. Two-Tone Spectrum @ 500 MHz Figure 25. SNR & SFDR vs. Temperature TBD Figure 26. SNR & SFDR vs. Power Supply Voltage Rev 0.5.1 Preliminary Page 13 KAD5610P Power-On Calibration Functional Description The KAD5610P is based upon a 10-bit, 250MSPS A/D converter core that utilizes a pipelined successive approximation architecture (Figure 27). The input voltage is captured by a Sample-Hold Amplifier (SHA) and converted to a unit of charge. Proprietary charge domain techniques are used to successively compare the input to a series of reference charges. Decisions made during the successive approximation operations determine the digital code for each input value. The converter pipeline requires six samples to produce a result. Digital error correction is also applied, resulting in a total latency of seven and a half clock cycles. This is evident to the user as a latency between the start of a conversion and the data being available on the digital outputs. The device contains two A/D converter cores with carefully matched transfer characteristics. At startup, each core performs a self-calibration to minimize gain and offset errors. The reset pin (RESETN) is initially set high at power-up and will remain in that state until the calibration is complete. The clock frequency should remain fixed during this time, and no SPI communications should be attempted. Recalibration can be initiated via the SPI port at any time after the initial self-calibration. At start-up, the core performs a self-calibration to minimize gain and offset errors. An internal power-onreset (POR) circuit detects the supply voltage ramps and initiates the calibration when the analog and digital supply voltages are above a threshold. The following conditions must be adhered to for the power-on calibration to execute successfully: • A frequency-stable conversion clock must be applied to the CLKP/CLKN pins • DNC pins (especially 3, 4 and 18) must not be pulled up or down • SDO (pin 66) must be high • RESETN (pin 25) must begin low • SPI communications must not be attempted A user-initiated reset can subsequently be invoked in the event that the above conditions cannot be met at power-up. The SDO pin requires an external 4.7kΩ pull-up to OVDD. If the SDO pin is pulled low externally during power-up, calibration will not be executed properly. After the power supply has stabilized the internal POR releases RESETN and an internal pull-up pulls it high, which starts the calibration sequence. The RESETN pin should be connected to an open-drain driver with a drive strength of less than 0.5mA. Figure 27. ADC Core Block Diagram Rev 0.5.1 Preliminary Page 14 KAD5610P The calibration sequence is initiated on the rising edge of RESETN, as shown in Figure 28. The overrange output (OR) is set high once RESETN is pulled low, and remains in that state until calibration is complete. The OR output returns to normal operation at that time, so it’s important that the analog input be within the converter’s full-scale range in order to observe the transition. If the input is in an over-range condition the OR pin will stay high and it will not be possible to detect the end of the calibration cycle. While RESETN is low, the output clock (CLKOUTP/CLKOUTN) stops toggling and is set low. Normal operation of the output clock resumes at the next input clock edge (CLKP/CLKN) after RESETN is deasserted. At 250MSPS the nominal calibration time is 300ms. Figure 29. Analog Input Range An RF transformer will give the best noise and distortion performance for wideband and/or high intermediate frequency (IF) inputs. Two different transformer input schemes are shown in Figures 30 and 31. Figure 30. Transformer Input for General Purpose Applications Figure 28. Calibration Timing User-Initiated Reset Recalibration of the ADC can be initiated at any time by driving the RESETN pin low for a minimum of one clock cycle. An open-drain driver with a drive strength of less than 0.5mA is recommended. As is the case during power-on reset, the SDO, RESETN and DNC pins must be in the proper state for the calibration to successfully execute. Analog Input Each ADC core contains a fully differential input (AINP/AINN, BINP/BINN) to the sample and hold amplifier (SHA). The ideal full-scale input voltage is 1.45V, centered at the VCM voltage of 0.535V as shown in Figure 29. Best performance is obtained when the analog inputs are driven differentially. The common mode output voltage, VCM, should be used to properly bias the inputs as shown in Figures 30 through 32. Rev 0.5.1 Preliminary Figure 31. Transmission-line Transformer Input for High IF Applications A back-to-back transformer scheme is used to improve common mode rejection, which keeps the common mode level of the input matched to VCM. The value of the shunt resistor should be determined based on the desired load impedance. The differential input resistance of the KAD5610P is 1000Ω. The SHA design uses a switched capacitor input stage, which creates charge kick-back when the sampling capacitance is reconnected to the input voltage. This kick-back creates a disturbance at the input which must settle before the next sampling point. Lower source impedance will result in faster settling and improved performance. Therefore a 1:1 transformer and low shunt resistance are recommended for optimal performance. Page 15 KAD5610P tails on this are contained in the Serial Peripheral Interface section. A delay-locked loop (DLL) generates internal clock signals for various stages within the charge pipeline. If the frequency of the input clock changes, the DLL may take up to 52µs to regain lock at 250MSPS. The lock time is inversely proportional to the sample rate. Jitter Figure 32. Differential Amplifier Input A differential amplifier, as shown in Figure 32, can be used in applications that require dc-coupling. In this configuration the amplifier will typically dominate the achievable SNR and distortion performance. In a sampled data system, clock jitter directly impacts the achievable SNR performance. The theoretical relationship between clock jitter (tJ) and SNR is shown in Equation 1 and is illustrated in Figure 34. ⎛ 1 SNR = 20 log 10 ⎜⎜ 2 π f IN t J ⎝ Clock Input Equation 1. The clock input circuit is a differential pair (see Figure 47). Driving these inputs with a high level (up to 1.8VPP on each input) sine or square wave will provide the lowest jitter performance. A transformer with 4:1 impedance ratio will provide increased drive levels. 100 95 tj=0.1ps 90 14 Bits 85 SNR - dB The recommended drive circuit is shown in Figure 33. The clock can be driven single-ended, but this will reduce the edge rate and may impact SNR performance. The clock inputs are internally self-biased to AVDD/2 to facilitate ac coupling. ⎞ ⎟⎟ ⎠ 80 tj=1ps 12 Bits 75 70 tj=10ps 65 60 10 Bits tj=100ps 55 50 1 10 100 1000 Input Frequency - MHz Figure 34. SNR vs. Clock Jitter Figure 33. Recommended Clock drive A selectable 2X/4X divider is provided in series with the clock input. The divider can be used in the 2X mode with a sample clock equal to twice the desired sample rate. This will result in a clock input with 50% duty cycle and will maximize the converter’s performance. CLKDIV Pin Divide Ratio AVSS 2 Float 1 AVDD 4 Table 1. CLKDIV Pin Settings The clock divider can also be controlled through the SPI port, which overrides the CLKDIV pin setting. DeRev 0.5.1 Preliminary This relationship shows the SNR that would be achieved if clock jitter were the only non-ideal factor. In reality, achievable SNR is limited by internal factors such as linearity, aperture jitter and thermal noise. Internal aperture jitter is the uncertainty in the sampling instant shown in Figure 1. The internal aperture jitter combines with the input clock jitter in a rootsum-square fashion, since they are not statistically correlated, and this determines the total jitter in the system. The total jitter, combined with other noise sources, then determines the achievable SNR. Voltage Reference A temperature compensated voltage reference provides the reference charges used in the successive approximation operations. The full-scale range of each A/D is proportional to the reference voltage. The nominal value of the voltage reference is 1.25V. Page 16 KAD5610P Digital Outputs Output data is available as a parallel bus in LVDScompatible or CMOS modes. In either case, the data is presented in double data rate (DDR) format with the A and B channel data available on alternating clock edges. When CLKOUT is low channel A data is output, while on the high phase channel B data is presented. Figures 1 and 2 show the timing relationships for LVDS and CMOS modes, respectively. Additionally, the drive current for LVDS mode can be set to a nominal 3 mA or a power-saving 2 mA. The lower current setting can be used in designs where the receiver is in close physical proximity to the ADC. The applicability of this setting is dependent upon the PCB layout, therefore the user should experiment to determine if performance degradation is observed. TBD Figure 35. Power vs. Sample Rate, LVDS Mode The output mode and LVDS drive current are selected via the OUTMODE pin as shown in Table 2. OUTMODE Pin Mode AVSS LVCMOS Float LVDS, 3 mA AVDD LVDS, 2 mA Table 2. OUTMODE Pin Settings The output mode can also be controlled through the SPI port, which overrides the OUTMODE pin setting. Details on this are contained in the Serial Peripheral Interface section. An external resistor creates the bias for the LVDS drivers. A 10kΩ, 1% resistor must be connected from the RLVDS pin to OVSS. Power Dissipation The power dissipated by the KAD5610P is primarily dependent on the sample rate, but is also related to the input signal in CMOS output mode. There is a static bias in the analog supply, while the remaining power dissipation is linearly related to the sample rate. The output supply dissipation is approximately constant in LVDS mode, but linearly related to the clock frequency in CMOS mode. Figures 35 and 36 illustrate these relationships. TBD Figure 36. Power vs. Sample Rate, CMOS Mode Nap/Sleep Portions of the device may be shut down to save power during times when operation of the ADC is not required. Two power saving modes are available: nap, and sleep. Nap mode reduces power dissipation to 40mW and recovers to normal operation in approximately 1µs. Sleep mode reduces power dissipation to 10mW but requires 1ms to recover. The clock should remain running and at a fixed frequency during Nap or Sleep. Recovery time from Nap mode will increase if the clock is stopped, since the internal DLL can take up to 52µs to regain lock at 250MSPS. By default after the device is powered on, the nap and sleep state is controlled by the NAPSLP pin as shown in Table 3. NAPSLP Pin Mode AVSS Normal Float Sleep AVDD Nap Table 3. NAPSLP Pin Settings Rev 0.5.1 Preliminary Page 17 KAD5610P The power down mode can also be controlled through the SPI port, which overrides the NAPSLP pin setting. Details on this are contained in the Serial Peripheral Interface section. This is an indexed function when controlled from the SPI, but a global function when driven from the pin. Data Format Output data can be presented in three formats: two’s complement, Gray code and offset binary. The data format is selected via the OUTFMT pin as shown in Table 4. OUTFMT Pin Mode AVSS Offset Binary Float Two’s Complement AVDD Gray Code Table 4. OUTFMT Pin Settings Figure 38. Gray Code to Binary Conversion The data format can also be controlled through the SPI port, which overrides the OUTFMT pin setting. Details on this are contained in the Serial Peripheral Interface section. Mapping of the input voltage to the various data formats is shown in Table 5. Offset binary coding maps the most negative input voltage to code 0x000 (all zeros) and the most positive input to 0xFFF (all ones). Two’s complement coding simply complements the MSB of the offset binary representation. When calculating Gray code the MSB is unchanged. The remaining bits are computed as the XOR of the current bit position and the next most significant bit. Figure 37 shows this operation. Input Voltage Offset Binary Two’s Complement Gray Code –Full Scale 0000000000 1000000000 0000000000 –Full Scale + 1LSB 0000000001 1000000001 0000000001 Mid–Scale 1000000000 0000000000 1100000000 +Full Scale – 1LSB 1111111110 0111111110 1000000001 +Full Scale 1111111111 0111111111 1000000000 Table 5. Input Voltage to Output Code Mapping Serial Peripheral Interface Figure 37. Binary to Gray Code Conversion Converting back to offset binary from gray code must be done recursively, using the result of each bit for the next lower bit as shown in Figure 38. A serial peripheral interface (SPI) bus is used to facilitate configuration of the device and to optimize performance. The SPI bus consists of chip select (CSB), serial clock (SCLK) and serial data input/output (SDIO). The maximum SCLK rate is equal to the ADC sample rate (fSAMPLE) divided by 16 for write operations and fSAMPLE divided by 66 for reads. At fSAMPLE = 250MHz, maximum SCLK is 15.63MHz for writing and 3.79MHz for write operations. There is no minimum SCLK rate. The following sections describe various registers that are used to configure the SPI or adjust performance Rev 0.5.1 Preliminary Page 18 KAD5610P or functional parameters. Many registers in the available address space (0x00 to 0xFF) are not defined in this document. Additionally, within a defined register there may be certain bits or bit combinations that are reserved. Undefined registers and undefined values within defined registers are reserved and should not be selected. Setting any reserved register or value may produce indeterminate results. SPI Physical Interface The SPI port operates in a half or full duplex master/slave configuration, with the KAD5610P functioning as a slave. Multiple slave devices can interface to a single master. The chip-select bar (CSB) pin determines when a slave device is being addressed. Multiple slave devices can be written to concurrently, but only one slave device can be read from at a given time. If multiple slave devices are selected for reading at the same time, the results will be indeterminate. The serial clock pin (SCLK) provides synchronization for the data transfer. By default, all data is presented on the serial data input/output (SDIO) pin. The state of the SDIO pin is set automatically in the communication protocol (described below). A dedicated serial data output pin (SDO) can be activated by setting 0x00[7] high to allow operation in full duplex mode. mand. Data can be presented in MSB-first order or LSB-first order. The default is MSB-first, but this can be changed by setting 0x00[6] high. Figures 39 and 40 show the appropriate bit ordering for the MSB-first and LSB-first modes, respectively. In MSB-first mode the address is incremented for multi-byte transfers, while in LSB-first mode it’s decremented. In the default mode the MSB is R/W, which determines if the data is to be read (active high) or written. The next two bits, W1 and W0, determine the number of data bytes to be read or written (see Table 6). The lower 13 bits contain the first address for the data transfer. This relationship is illustrated in Figure 41, and timing values are given in the Switching Specifications section. After the instruction/address bytes have been read, the appropriate number of data bytes are written to or read from the ADC (based on the R/W bit status). The data transfer will continue as long as CSB remains low and SCLK is active. Stalling of the CSB pin is allowed at any byte boundary (instruction/address or data) if the number of bytes being transferred is three or less. For transfers of four bytes or more, CSB is allowed stall in the middle of the instruction/address bytes or before the first data byte. If CSB transitions to a high state after that point the state machine will reset and terminate the data transfer. The communication protocol begins with an instruction/address phase. The first rising SCLK edge following a high to low transition on CSB determines the beginning of the two-byte instruction/address com- Figure 39. MSB-First Addressing Figure 40. LSB-First Addressing Rev 0.5.1 Preliminary Page 19 KAD5610P [W1:W0] Bytes Transferred 00 1 01 2 10 3 11 or LSB to MSB (LSB first) to accommodate various microcontrollers. Bit 7 Bit 6 SDO Active LSB First Setting this bit high configures the SPI to interpret serial data as arriving in LSB to MSB order. 4 or more Table 6. Byte Transfer Selection Bit 5 Figures 42 and 43 illustrate the timing relationships for 2-byte and N-byte transfers, respectively. The operation for a 3-byte transfer can be inferred from these diagrams. SPI Configuration Soft Reset Setting this bit high resets all SPI registers to default values. Bit 4 Reserved This bit should always be set high. Bits 3:0 These bits should always mirror bits 4:7 to avoid ambiguity in bit ordering. Address 0x00: chip_port_config Bit ordering and SPI reset are controlled by this register. Bit order can be selected as MSB to LSB (MSB first) Figure 41. Instruction/Address Phase Figure 42. 2-Byte Transfer Figure 43. N-Byte Transfer Rev 0.5.1 Preliminary Page 20 KAD5610P Address 0x02: burst_end If a series of sequential registers are to be set, burst mode can improve throughput by eliminating redundant addressing. In 3-wire SPI mode the burst is ended by pulling the CSB pin high. If the device is operated in 2-wire mode the CSB pin is not available. In that case, setting the burst_end address determines the end of the transfer. During a write operation, the user must be cautious to transmit the correct number of bytes based on the starting and ending addresses. Bits 7:0 Burst End Address This register value determines the ending address of the burst data. DUT Information Address 0x08: chip_id Address 0x09: chip_version The generic die identifier and a revision number, respectively, can be read from these two registers. Indexed DUT Configuration/Control Address 0x10: device_index_A A common SPI map, which can accommodate single-channel or multi-channel devices, is used for all Kenet ADC products. Certain configuration commands (identified as Indexed in the SPI map) can be executed on a per-converter basis. This register determines which converter is being addressed for an Indexed command. It is important to note that only a single converter can be addressed at a time. This register defaults to 00h, indicating that no ADC is addressed. Address 0x20: offset_coarse Address 0x21: offset_fine Parameter 0x20[7:0] 0x21[7:0] Coarse Offset Fine Offset Steps 256 256 –Full Scale (0x80) -24.0mV -1.7mV Mid–Scale (0x00) 0.0mV 0.0mV +Full Scale (0x7F) +23.8mV +1.7mV Nominal Step Size 187.5µV 13.3µV Table 7. Offset Adjustments Address 0x22: gain_coarse Address 0x23: gain_medium Address 0x24: gain_fine Gain of each ADC core can be adjusted in coarse, medium and fine steps. Coarse gain is a 4-bit adjustment while medium and fine are 8-bit. The data format is twos complement for all three registers. The default value of each register will be the result of the self-calibration after initial power-up. If a register is to be incremented or decremented, the user should first read the register value then write the incremented or decremented value back to the same register. Parameter 0x22[3:0] Coarse Gain Steps 16 –Full Scale (0x08) -11.2% Mid–Scale (0x00) 0.0% +Full Scale (0x07) +9.8% Nominal Step Size 1.4% Table 8. Coarse Gain Adjustment The input offset of each ADC core can be adjusted in fine and coarse steps. Both adjustments are made via an 8-bit word as detailed in Table 7. The data format is twos complement. Parameter 0x23[7:0] Medium Gain 0x24[7:0] Fine Gain Steps 256 256 The default value of each register will be the result of the self-calibration after initial power-up. If a register is to be incremented or decremented, the user should first read the register value then write the incremented or decremented value back to the same register. –Full Scale (0x80) -10.56% -1.06% Mid–Scale (0x00) 0.0% 0.0% +Full Scale (0x7F) +10.48% +1.05% Nominal Step Size 0.0825% 0.00825% Table 9. Medium and Fine Gain Adjustments Address 0x25: modes Two distinct reduced power modes can be selected. By default, the tri-level NAPSLP pin can select normal Rev 0.5.1 Preliminary Page 21 KAD5610P operation, nap or sleep modes (refer to Nap/Sleep section). This functionality can be overridden and controlled through the SPI. This is an indexed function when controlled from the SPI, but a global function when driven from the pin. This register is not changed by a Soft Reset. Value 0x25[2:0] Power Down Mode 000 Pin Control 001 Normal Operation 010 Nap Mode 100 Sleep Mode Table 10. Power Down Control Global DUT Configuration/Control Address 0x70: skew_diff The value in the skew_diff register adjusts the timing skew between the two ADCs cores. The nominal range and resolution of this adjustment are given in Table 11. The default value of this register after power-up is 00h. Parameter 0x70[7:0] Differential Skew Steps 256 –Full Scale (0x08) -6.5ps Mid–Scale (0x00) 0.0ps +Full Scale (0x07) +6.5ps Nominal Step Size 51fs Table 11. Differential Skew Adjustment Address 0x71: phase_slip When using a clock divider, it’s not possible to determine the synchronization of the incoming and divided clock phases. This is particularly important when multiple ADCs are used in a time-interleaved system. The phase slip feature allows the rising edge of the divided clock to be advanced by one input clock cycle, as shown in Figure 44. Figure 44. Phase Slip Address 0x72: clock_divide The KAD5610P has a selectable clock divider that can be set to divide by four, two or one (no division). By default, the tri-level CLKDIV pin selects the divisor (refer to Clock Input section). This functionality can be overridden and controlled through the SPI, as shown in Table 12. This register is not changed by a Soft Reset. Value 0x72[2:0] Clock Divider 000 Pin Control 001 Divide by 1 010 Divide by 2 100 Divide by 4 Table 12. Clock Divider Selection Address 0x73: output_mode_A The output_mode_A register controls the physical output format of the data, as well as the logical coding. The KAD5610P can present output data in two physical formats: LVDS or LVCMOS. Additionally, the drive strength in LVDS mode can be set high (3mA) or low (2mA). By default, the tri-level OUTMODE pin selects the mode and drive level (refer to Digital Outputs section). This functionality can be overridden and controlled through the SPI, as shown in Table 13. Data can be coded in three possible formats: two’s complement, Gray code or offset binary. By default, the tri-level OUTFMT pin selects the data format (refer to Data Format section). This functionality can be overridden and controlled through the SPI, as shown in Table 14. This register is not changed by a Soft Reset. Rev 0.5.1 Preliminary Page 22 KAD5610P Value 0x93[7:5] 000 Pin Control 001 LVDS 2mA 010 LVDS 3mA 100 LVCMOS Table 13. Output Mode Control Value 0x93[2:0] Output Format 000 Pin Control 001 Two’s Complement 010 100 ing. A static word can be placed on the output bus, or two different words can alternate. In the alternate mode, Address 0xC0: test_io Bits 7:6 User Test Mode These bits set the test mode to static (0x00) or alternate (0x01) mode. Other values are reserved. The four LSBs in this register (Output Test Mode) determine the test pattern in combination with registers 0xC2 through 0xC5. Refer to Table 15. Value 0xC0[3:0] Output Test Mode Word 1 Word 2 Gray Code 0000 Off Offset Binary 0001 Midscale 0x8000 N/A 0010 Positive Full-Scale 0xFFFF N/A 0011 Negative Full-Scale 0x0000 N/A Address 0x75: config_status 0100 Checkerboard 0xAAAA 0x5555 Bit 6 DLL Range 0101 Reserved N/A N/A This bit sets the DLL operating range to fast (TBD2MSPS to 250MSPS) or slow (40 to TBD1MSPS). 0110 Reserved N/A N/A 0111 One/Zero 0xFFFF 0x0000 1000 User Pattern user_patt1 user_patt2 Table 14. Output Format Control Address 0x74: output_mode_B Bit 4 DDR Enable Setting this bit enables Double Data-Rate mode. The output_mode_B and config_status registers are used in conjunction to enable DDR mode and select the frequency range of the DLL clock generator. The method of setting these options is different from the other registers. Table 15. Output Test Modes Address 0xC2: user_patt1_lsb Address 0xC3: user_patt1_msb These registers define the lower and upper eight bits, respectively, of the first user-defined test word. Address 0xC2: user_patt2_lsb Address 0xC3: user_patt2_msb These registers define the lower and upper eight bits, respectively, of the second user-defined test word. Figure 45. Setting output_mode_B register The procedure for setting output_mode_B is shown in Figure 45. Read the contents of output_mode_B and config_status and XOR them. Then XOR this result with the desired value for output_mode_B and write that XOR result to the register. DUT Test The KAD5610 can produce preset or user defined patterns on the digital outputs to facilitate in-situ testRev 0.5.1 Preliminary Page 23 KAD5610P DUT Test Global DUT Config/Control Indexed DUT Config/Control DUT SPI Config Info SPI Memory Map Addr (Hex) Parameter Name Bit 7 (MSB) 00 01 02 03-07 08 09 10 11-1F 20 21 22 23 24 25 port_config reserved burst_end reserved chip_id chip_version device_index_A reserved offset_coarse offset_fine gain_coarse gain_medium gain_fine modes SDO Active LSB First 26-5F 60-6F reserved reserved 70 71 skew_diff phase_slip 72 clock_divide 73 output_mode_A 74 output_mode_B 75 76-BF C0 config_status reserved test_io C1 C2 C3 C4 C5 C6-FF Reserved user_patt1_lsb user_patt1_msb user_patt2_lsb user_patt2_msb reserved Bit 6 Bit 5 Bit 4 Bit 3 Soft Reset Bit 2 Bit 1 Bit 0 (LSB) Mirror (bit5) Mirror (bit6) Mirror (bit7) Reserved Burst end address [7:0] Reserved Chip ID # Chip Version # Reserved Reserved Coarse Offset Fine Offset ADC01 ADC00 Coarse Gain Reserved Medium Gain Fine Gain Power Down Mode [2:0] 000=Pin Control 001=Normal Operation 010=Nap 100=Sleep other codes=reserved Def. Value Indexed/ (Hex) Global 00h G 00h G Read only Read only 00h G G I cal. value cal. value cal. value cal. value cal. value 00h NOT affected by Soft Reset I I I I I I Reserved Reserved Differential Skew Reserved Clock Divide [2:0] 000=Pin Control 001=divide by 1 010=divide by 2 100=divide by 4 other codes=reserved Output Format [2:0] 000=Pin Control 001=Twos Complement 010=Gray Code 100=Offset Binary other codes=reserved Output Mode [2:0] 000=Pin Control 001=LVDS 2mA 010=LVDS 3mA 100=LVCMOS other codes=reserved DLL Range 0=fast 1=slow DDR Enable XOR Result User Test Mode [2:0] 00=Single 01=Alternate 10=Single Once 11=Alternate Once Reset PN Long Gen B7 B15 B7 B15 B5 B13 B5 B13 B6 B14 B6 B14 Next Clock Edge XOR Result Reserved Reset PN Short Gen 0=Off Output Test Mode [3:0] 1=Midscale Short 2=+FS Short 3=−FS Short 4=Checker Board 5=reserved 6=reserved B4 B12 B4 B12 Reserved B3 B11 B3 B11 Reserved B2 B10 B2 B10 7Fh 00h G G 00h NOT affected by Soft Reset G 00h NOT affected by Soft Reset G 00h NOT affected by Soft Reset Read Only G 00h G 00h 00h 00h 00h 00h G G G G G G 7=One/Zero Word Toggle 8=User Input 9-15=reserved B1 B9 B1 B9 B0 B8 B0 B8 Table 16. SPI Memory Map Rev 0.5.1 Preliminary Page 24 KAD5610P Equivalent Circuits Figure 46. Analog Inputs AVDD Figure 50. LVDS Outputs To Charge Pipeline AVDD CLKP AVDD 11kΩ 18kΩ AVDD 11kΩ Figure 51. CMOS Outputs 18kΩ CLKN Figure 52. VCM_OUT Output Figure 47. Clock Inputs Layout Considerations Split Ground and Power Planes Figure 48. Tri-Level Digital Inputs Data converters operating at high sampling frequencies require extra care in PC board layout. Many complex board designs benefit from isolating the analog and digital sections. Analog supply and ground planes should be laid out under signal and clock inputs. Locate the digital planes under outputs and logic pins. Grounds should be joined under the chip. Clock Input Considerations Use matched transmission lines to the transformer inputs for the analog input and clock signals. Locate transformers and terminations as close to the chip as possible. Figure 49. Digital Inputs Rev 0.5.1 Preliminary Page 25 KAD5610P Exposed Paddle The exposed paddle must be electrically connected to analog ground (AVSS) and should be connected to a large copper plane using numerous vias for optimal thermal performance. Bypass and Filtering Bulk capacitors should have low equivalent series resistance. Tantalum is a good choice. For best performance, keep ceramic bypass capacitors very close to device pins. Longer traces will increase inductance, resulting in diminished dynamic performance and accuracy. Make sure that connections to ground are direct and low impedance. Avoid forming ground loops. LVDS Outputs Output traces and connections must be designed for 50Ω (100Ω differential) characteristic impedance. Keep traces direct and minimize bends where possible. Avoid crossing ground and power plane breaks with signal traces. LVCMOS Outputs Output traces and connections must be designed for 50Ω characteristic impedance. Unused Inputs Standard logic inputs (RESETN, CSB, SCLK, SDIO, SDO) which will not be operated do not require connection to ensure optimal ADC performance. These inputs can be left floating if they are not used. Tri-level inputs (NAPSLP, OUTMODE, OUTFMT, CLKDIV) accept a floating input as a valid state, and therefore should be biased according to the desired functionality. Definitions Analog Input Bandwidth is the analog input frequency at which the spectral output power at the fundamental frequency (as determined by FFT analysis) is reduced by 3dB from its full-scale low-frequency value. This is also referred to as Full Power Bandwidth. Aperture Delay or Sampling Delay is the time required after the rise of the clock input for the sampling switch to open, at which time the signal is held for conversion. Aperture Jitter is the RMS variation in aperture delay for a set of samples. Rev 0.5.1 Preliminary Clock Duty Cycle is the ratio of the time the clock wave is at logic high to the total time of one clock period. Differential Non-Linearity (DNL) is the deviation of any code width from an ideal 1 LSB step. Effective Number of Bits (ENOB) is an alternate method of specifying Signal to Noise-and-Distortion Ratio (SINAD). In dB, it is calculated as: ENOB = (SINAD-1.76) / 6.02 Gain Error is the ratio of the difference between the voltages that cause the lowest and highest code transitions to the full-scale voltage (less 2 LSB). It is typically expressed in percent. Integral Non-Linearity (INL) is the deviation of each individual code from a line drawn from negative fullscale (1/2 LSB below the first code transition) through positive full-scale (1/2 LSB above the last code transition). The deviation of any given code from this line is measured from the center of that code. Least Significant Bit (LSB) is the bit that has the smallest value or weight in a digital word. Its value in terms of input voltage is VFS/(2N-1) where N is the resolution in bits. Missing Codes are output codes that are skipped and will never appear at the ADC output. These codes cannot be reached with any input value. Most Significant Bit (MSB) is the bit that has the largest value or weight. Pipeline Delay is the number of clock cycles between the initiation of a conversion and the appearance at the output pins of the data. Power Supply Rejection Ratio (PSRR) is the ratio of a change in input voltage necessary to correct a change in output code that results from a change in power supply voltage. Signal to Noise-and-Distortion (SINAD) is the ratio of the RMS signal amplitude to the RMS value of the sum of all other spectral components below one half the clock frequency, including harmonics but excluding DC. Signal-to-Noise Ratio (without Harmonics) is the ratio of the RMS signal amplitude to the sum of all other spectral components below one-half the sampling frequency, excluding harmonics and DC. SNR and SINAD are either given in units of dBc (dB to carrier) when the absolute power of the fundamental is used as the reference, or dBFS (dB to full scale) when the converter’s full-scale input power is used as the reference. Page 26 KAD5610P Spurious-Free-Dynamic Range (SFDR) is the ratio of the RMS signal amplitude to the RMS value of the peak spurious spectral component. The peak spurious spectral component may or may not be a harmonic. Two-Tone SFDR is the ratio of the RMS value of the lowest power input tone to the RMS value of the peak spurious component, which may or may not be an IMD product. Outline Dimensions Figure 53. 72QFN Dimensions Rev 0.5.1 Preliminary Page 27 KAD5610P Ordering Guide RoHS The KAD5610P is compliant with EU directive 2002/95/EC regarding the Restriction of Hazardous Substances (RoHS). Contact Kenet for a materials declaration for this product. Model Speed Package Temp. Range KAD5610P-25Q72 250MSPS 72-QFN -40°C to +85°C KAD5610P-21Q72 210MSPS 72-QFN -40°C to +85°C KAD5610P-17Q72 170MSPS 72-QFN -40°C to +85°C KAD5610P-12Q72 125MSPS 72-QFN -40°C to +85°C Revision History 14-May-07: Rev 0.1 Updated to new format 21-Jun-07: Rev 0.2 Errata Updated 13-Aug-07: Rev 0.3 Content/specification updates 07-Dec-07: Rev 0.4 Content/specification updates 21-Feb-08: Rev 0.5 New Pinout, Updated specifications, added functional descriptions 25-Feb-08: Rev 0.5.1 Added skew_diff SPI register description (p. 22) Preliminary Datasheet This datasheet contains preliminary technical data, which is subject to change without notice. Contact Kenet prior to initiating design activity using this product. Rev 0.5.1 Preliminary Page 28