a 4-Channel, 104 MSPS Digital Transmit Signal Processor (TSP) AD6623 FEATURES Pin Compatible to the AD6622 18-Bit Parallel Digital IF Output Real or Interleaved Complex 18-Bit Bidirectional Parallel Digital IF Input/Output Allows Cascade of Chips for Additional Channels Clipped or Wrapped Over Range Two’s Complement or Offset Binary Output Four Independent Digital Transmitters in Single Package RAM Coefficient Filter (RCF) Programmable IF and Modulation for Each Channel Programmable Interpolating RAM Coefficient Filter p/4-DQPSK Differential Phase Encoder 3p/8-PSK Linear Encoder 8-PSK Linear Encoder Programmable GMSK Look-Up Table Programmable QPSK Look-Up Table All-Pass Phase Equalizer Programmable Fine Scaler Programmable Power Ramp Unit High Speed CIC Interpolating Filter Digital Resampling for Noninteger Interpolation Rates NCO Frequency Translation Spurious Performance Better than –100 dBc Separate 3-Wire Serial Data Input for Each Channel Bidirectional Serial Clocks and Frames Microprocessor Control 2.5 V CMOS Core, 3.3 V Outputs, 5 V Inputs JTAG Boundary Scan APPLICATIONS Cellular/PCS Base Stations Micro/Pico Cell Base Stations Wireless Local Loop Base Stations Multicarrier, Multimode Digital Transmit GSM, EDGE, IS136, PHS, IS95, TDS CDMA, UMTS, CDMA2000 Phased Array Beam Forming Antennas Software Defined Radio Tuning Resolution Better than 0.025 Hz Real or Complex Outputs FUNCTIONAL BLOCK DIAGRAM NCO = NUMERICALLY CONTROLLED OSCILLATOR/TUNER SDINA DATA SDFIA SPORT SDFOA I RAM COEFFICIENT Q FILTER SCALER I AND Q POWER RAMP CIC5 FILTER Q I I RAM COEFFICIENT Q FILTER SCALER I AND Q POWER RAMP CIC5 FILTER Q I RCIC2 FILTER Q RCIC2 FILTER Q QIN CHAN A NCO IN [17–0] SCLKA SDINB SDFIB DATA SPORT SDFOB I I CHAN B NCO SYNC 4 SCLKB SUMMATION SDINC DATA SDFIC SPORT I I RAM COEFFICIENT Q FILTER SCALER AND Q POWER RAMP I RAM COEFFICIENT Q FILTER SCALER I AND Q POWER RAMP SDFOC I CIC5 FILTER Q I RCIC2 FILTER Q CHAN C NCO SCLKC OEN SDIND SDFID DATA SPORT SDFOD I CIC5 FILTER Q I RCIC2 FILTER Q QOUT CHAN D NCO OUT [17:0] SCLKD MICROPORT JTAG TDL TDO TMS TCK TRST D[7:0] DS DTACK RW MODE A[2:0] CS CLK RESET REV. 0 Information furnished by Analog Devices is believed to be accurate and reliable. However, no responsibility is assumed by Analog Devices for its use, nor for any infringements of patents or other rights of third parties that may result from its use. No license is granted by implication or otherwise under any patent or patent rights of Analog Devices. One Technology Way, P.O. Box 9106, Norwood, MA 02062-9106, U.S.A. Tel: 781/329-4700 www.analog.com Fax: 781/326-8703 © Analog Devices, Inc., 2002 AD6623 TABLE OF CONTENTS FEATURES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 APPLICATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 FUNCTIONAL BLOCK DIAGRAM . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 PRODUCT DESCRIPTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 FUNCTIONAL OVERVIEW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 RECOMMENDED OPERATING CONDITIONS . . . . . . . . . . . . . . . . 4 ELECTRICAL CHARACTERISTICS . . . . . . . . . . . . . . . . . . . . . . . . . . 4 LOGIC INPUTS (5 V TOLERANT) . . . . . . . . . . . . . . . . . . . . . . . . . 4 LOGIC OUTPUTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 IDD SUPPLY CURRENT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 POWER DISSIPATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 GENERAL TIMING CHARACTERISTICS . . . . . . . . . . . . . . . . . . . . . 5 MICROPROCESSOR PORT TIMING CHARACTERISTICS . . . . . . . . 6 MICROPROCESSOR PORT, MODE INM (MODE = 0) . . . . . . . . . 6 MICROPROCESSOR PORT, MOTOROLA (MODE = 1) . . . . . . . . 6 TIMING DIAGRAMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7–9 ABSOLUTE MAXIMUM RATINGS . . . . . . . . . . . . . . . . . . . . . . . . . . 10 THERMAL CHARACTERISTICS . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 EXPLANATION OF TEST LEVELS . . . . . . . . . . . . . . . . . . . . . . . . . . 10 ORDERING GUIDE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 ESD SENSITIVITY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 PIN CONFIGURATION – 128-Lead MQFP . . . . . . . . . . . . . . . . . . . . 11 128 PIN FUNCTION DESCRIPTION . . . . . . . . . . . . . . . . . . . . . . . 12 PIN CONFIGURATION – 196-Lead BGA . . . . . . . . . . . . . . . . . . . . . . 13 196 PIN FUNCTION DESCRIPTION . . . . . . . . . . . . . . . . . . . . . . . 14 POWER SUPPLY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 INPUTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 CONTROL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 MICROPORT CONTROL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 OUTPUTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 JTAG AND BIST . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 SERIAL DATA PORT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 Serial Master Mode (SCS = 0) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 Serial Slave Mode (SCS = 1) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 Self-Framing Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 External Framing Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 Serial Port Cascade Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 Serial Data Format . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 PROGRAMMABLE RAM COEFFICIENT FILTER (RCF) . . . . . . . . . 16 OVERVIEW OF THE RCF BLOCKS . . . . . . . . . . . . . . . . . . . . . . . . 16 INTERPOLATING FIR FILTER . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 RCF CONTROL REGISTERS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 PSK MODULATOR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 /4-DQSPK MODULATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 8-PSK MODULATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 3/8-8-PSK MODULATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 MSK Look-Up Table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 GMSK Look-Up Table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 QPSK Look-Up Table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 PHASE EQUALIZER . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 SCALE AND RAMP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 FINE SCALING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 RCF POWER RAMPING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 CASCADED INTERGRATOR COMB (CIC) INTERPOLATING FILTERS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 CIC Scaling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 CIC5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 rCIC2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 NUMERICALLY CONTROLLED OSCILLATOR/TUNER (NCO) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 Phase Dither . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 Amplitude Dither . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 Phase Offset . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 NCO Frequency Update and Phase Offset Update Hold-Off Counters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 NCO Control Scale . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 SUMMATION BLOCK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 SYNCHRONIZATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 Start . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 Start with No Sync . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 Start with Soft Sync . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 Start with Pin Sync . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 Hop with Soft Sync . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 Hop with Pin Sync . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 Beam with Soft Sync . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 Beam with Pin Sync . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 JTAG INTERFACE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 SCALING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 Multicarrier Scaling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 Single Carrier Scaling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 MICROPORT INTERFACE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 MicroPort Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 EXTERNAL MEMORY MAP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 Intel Nonmultiplexed Mode (INM) . . . . . . . . . . . . . . . . . . . . . . . . . . 30 Motorola Nonmultiplexed Mode (MNM) . . . . . . . . . . . . . . . . . . . . . 30 External Address 7 Upper Address Register (UAR) . . . . . . . . . . . . . . 30 External Address 6 Lower Address Register (LAR) . . . . . . . . . . . . . . 30 External Address 5 Sync . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 External Address 4 Sleep . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 External Address 3:0 (Data Bytes) . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 INTERNAL CONTROL REGISTERS AND ON-CHIP RAM . . . . . . . . . 31 AD6623 and AD6622 Compatibility Common Function Registers (not associated with a particular channel) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 Channel Function Registers (0x1XX = Ch. A, 0x2XX = Ch. B, 0x3XX = Ch. C, 0x4XX = Ch. D) . . . . . . . . . . . 31 (0x000) Summation Mode Control . . . . . . . . . . . . . . . . . . . . . . . . . . 33 (0x001) Sync Mode Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 (0x002) BIST Counter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 (0x003) BIST Result . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 Channel Function Registers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 (0xn00) Start Update Hold-Off Counter . . . . . . . . . . . . . . . . . . . . . . 34 (0xn01) NCO Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 (0xn02) NCO Frequency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 (0xn03) NCO Frequency Update Hold-Off Counter . . . . . . . . . . . . . 34 (0xn04) NCO Phase Offset . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 (0xn05) NCO Phase Offset Update Hold-Off Counter . . . . . . . . . . . . 34 (0xn06) CIC Scale . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 (0xn07) CIC2 Decimation – 1 (MCIC2 – 1) . . . . . . . . . . . . . . . . . . . . . 34 (0xn08) CIC2 Interpolation – 1 (LCIC2 – 1) . . . . . . . . . . . . . . . . . . . . 34 (0xn09) CIC5 Interpolation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 (0xn0A) Number of RCF Coefficients – 1 . . . . . . . . . . . . . . . . . . . . . 34 (0xn0B) RCF Coefficient Offset . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 (0xn0C) Channel Mode Control 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 (0xn0D) Channel Mode Control 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 (0xn0E) Fine Scale Factor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 (0xn0F) RCF Time Slot Hold-Off Counter . . . . . . . . . . . . . . . . . . . . 35 (0xn10–0xn11) RCF Phase Equalizer Coefficients . . . . . . . . . . . . . . . 35 (0xn12–0xn15) FIR-PSK Magnitudes . . . . . . . . . . . . . . . . . . . . . . . . 35 (0xn16) Serial Port Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 (0xn17) Power Ramp Length 0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 (0xn18) Power Ramp Length 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 (0xn19) Power Ramp Rest Time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 (0xn20–0xn1F) Unused . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 (0xn20–0xn3F) Data Memory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 (0xn40–0xn17F) Power Ramp Coefficient Memory . . . . . . . . . . . . . . 35 Pseudocode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 Write Pseudocode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 Read Pseudocode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 APPLICATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 USING THE AD6623 TO PROCESS UMTS CARRIERS . . . . . . . . 36 DIGITAL-TO-ANALOG CONVERTER (DAC) SELECTION . . . . . . . 36 MULTIPLE TSP OPERATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 Determining the Number of TSPs to Use . . . . . . . . . . . . . . . . . . . . 36 Programming Mulitple TSPs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 Driving Multiple TSP Serial Ports . . . . . . . . . . . . . . . . . . . . . . . . . 37 THERMAL MANAGEMENT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 PACKAGE OUTLINE DIMENSIONS . . . . . . . . . . . . . . . . . . . . . . . . . 39 –2– REV. 0 AD6623 PRODUCT DESCRIPTION The AD6623 is a 4-channel Transmit Signal Processor (TSP) that creates high bandwidth data for Transmit Digital-to-Analog Converters (TxDACs) from baseband data provided by a Digital Signal Processor (DSP). Modern TxDACs have achieved sufficiently high sampling rates, analog bandwidth, and dynamic range to create the first Intermediate Frequency (IF) directly. The AD6623 synthesizes multicarrier and multistandard digital signals to drive these TxDACs. The RAM-based architecture allows easy reconfiguration for multimode applications. Modulation, pulseshaping and anti-imaging filters, static equalization, and tuning functions are combined in a single, cost-effective device. Digital IF signal processing provides repeatable manufacturing, higher accuracy, and more flexibility than comparable high dynamic range analog designs. The AD6623 has four identical digital TSPs complete with synchronization circuitry and cascadable wideband channel summation. AD6623 is pin compatible to AD6622 and can operate in AD6622compatible control register mode. The AD6623 utilizes a 3.3 V I/O power supply and a 2.5 V core power supply. All I/O pins are 5 V tolerant. All control registers and coefficient values are programmed through a generic microprocessor interface. Intel and Motorola microprocessor bus modes are supported. All inputs and outputs are LVCMOS compatible. FUNCTIONAL OVERVIEW Each TSP has five cascaded signal processing elements: a programmable interpolating RAM Coefficient Filter (RCF), a programmable Scale and Power Ramp, a programmable fifth order Cascaded Integrator Comb (CIC5) interpolating filter, a flexible second order Resampling Cascaded Integrator Comb filter (rCIC2), and a Numerically Controlled Oscillator/Tuner (NCO). The outputs of the four TSPs are summed and scaled on-chip. In multicarrier wideband transmitters, a bidirectional bus allows the Parallel (wideband) IF Input/Output to drive a second DAC. In this operational mode two AD6623 channels drive one DAC and the other two AD6623 channels drive a second DAC. Multiple AD6623s may be combined by driving the INOUT[17:0] of the succeeding with the OUT[17:0] of the preceding chip. The REV. 0 INOUT[17:0] can alternatively be masked off by software to allow preceding AD6623’s outputs to be ignored. Each channel accepts input data from independent serial ports that may be connected directly to the serial port of Digital Signal Processor (DSP) chips. The RCF implements any one of the following functions: Interpolating Finite Impulse Response (FIR) filter, /4-DQPSK modulator, 8-PSK modulator, or 3 /8-8-PSK modulator, GMSK modulator, and QPSK modulator. Each AD6623 channel can be dynamically switched between the GMSK modulation mode and the 3 /8-8-PSK modulation mode in order to support the GSM/EDGE standard. The RCF also implements an Allpass Phase Equalizer (APE) which meets the requirements of IS-95-A/B standard (CDMA transmission). The programmable Scale and Power Ramp block allows power ramping on a time-slot basis as specified for some air-interface standards (e.g., GSM, EDGE). A fine scaling unit at the programmable FIR filter output allows an easy signal amplitude level adjustment on time slot basis. The CIC5 provides integer rate interpolation from 1 to 32 and coarse anti-image filtering. The rCIC2 provides fractional rate interpolation from 1 to 4096 in steps of 1/512. The wide range of interpolation factors in each CIC filter stage and a highly flexible resampler incorporated into rCIC2 makes the AD6623 useful for creating both narrowband and wideband carriers in a high-speed sample stream. The high resolution 32-bit NCO allows flexibility in frequency planning and supports both digital and analog air interface standards. The high speed NCO tunes the interpolated complex signal from the rCIC2 to an IF channel. The result may be real or complex. Multicarrier phase synchronization pins and phase offset registers allow intelligent management of the relative phase of independent RF channels. This capability supports the requirements for phased array antenna architectures and management of the wideband peak/power ratio to minimize clipping at the DAC. The wideband Output Ports can deliver real or complex data. Complex words are interleaved into real (I) and imaginary (Q) parts at half the master clock rate. –3– AD6623 RECOMMENDED OPERATING CONDITIONS Parameter Test Level Min AD6623 Typ Max Unit VDD VDDIO TAMBIENT IV IV IV 2.25 3.0 –40 2.5 3.3 +25 2.75 3.6 +70 V V °C Max Unit 5.0 +0.8 10 10 V V µA µA pF ELECTRICAL CHARACTERISTICS Parameter (Conditions) Temp Test Level Min LOGIC INPUTS (5 V TOLERANT) Logic Compatibility Logic “1” Voltage Logic “0” Voltage Logic “1” Current Logic “0” Current Input Capacitance Full Full Full Full Full 25°C IV IV IV IV V 2.0 –0.3 LOGIC OUTPUTS Logic Compatibility Logic “1” Voltage (IOH = 0.25 mA) Logic “0” Voltage (IOL = 0.25 mA) Full Full Full IV IV 2.0 IDD SUPPLY CURRENT CLK = 104 MHz, VDD = 2.75 V* CLK = 104 MHz, VDDIO = 3.6 V* GSM Example: CORE I/O IS-136 Example: CORE I/O WBCDMA Example Sleep Mode POWER DISSIPATION GSM Example IS-136 Example WBCDMA Example Sleep Mode Full Full Typ 3.3 V CMOS 1 0 4 IV IV V 3.3 V CMOS/TTL VDD – 0.2 0.2 Full V IV 422 193 232 56 207 55 Tbd Tbd Full V V V IV 740 700 Tbd Tbd V 0.4 TBD* V V TBD mA mA mA mA mA mA mA mA TBD mW mW mW mW *This specification denotes an absolute maximum supply current for the device. The conditions include all channels active, minimum interpolation in both CIC stages, and maximum switching of input data. In an actual application the power will be less. See the Thermal Management section of the data sheet for further details. –4– REV. 0 AD6623 GENERAL TIMING CHARACTERISTICS1, 2 Parameter (Conditions) Temp Test Level Min CLK Timing Requirements: CLK Period tCLK CLK Width Low tCLKL tCLKH CLK Width High Full Full Full I IV IV 9.6 3 3 RESET Timing Requirement: tRESL RESET Width Low Full I 30.0 ns Input Data Timing Requirements: tSI INOUT[17:0], QIN to ↑CLK Setup Time tHI INOUT[17:0], QIN to ↑CLK Hold Time Full Full IV IV 1 2 ns ns Output Data Timing Characteristics: tDO ↑CLK to OUT[17:0], INOUT[17:0], QOUT Output Delay Time tDZO OEN HIGH to OUT[17:0] Active Full Full IV IV 2 3 SYNC Timing Requirements: SYNC(0, 1, 2, 3) to ↑CLK Setup Time tSS tHS SYNC(0, 1, 2, 3) to ↑CLK Hold Time Full Full IV IV 1 2 Full Full IV IV 4 5 10.5 13 ns ns Full IV 3.5 9 ns Full IV 4 10 ns Full Full Full IV IV IV 1.7 0 0.5 3.5 ns ns ns Full Full Full Full Full IV IV IV IV IV 2 0 2 0 0.5 3 ns ns ns ns ns Full Full Full IV IV IV 3.5 3.5 ns ns ns Full Full Full IV IV IV 1 2.5 4 ns ns ns Full Full Full Full Full IV IV IV IV IV 2 1 1 2.5 10 Master Mode Serial Port Timing Requirements (SCS = 0): Switching Characteristics3 tDSCLK1 ↑CLK to ↑SCLK Delay (divide by 1) ↑CLK to ↑SCLK Delay (for any other divisor) tDSCLKH tDSCLKL ↑CLK to ↓SCLK Delay (divide by 2 or even number) ↓CLK to ↓SCLK Delay tDSCLKLL (divide by 3 or odd number) Channel is Self-Framing SDIN to ↑SCLK Setup Time tSSDI0 tHSDI0 SDIN to ↑SCLK Hold Time tDSFO0A ↑SCLK to SDFO Delay Channel is External-Framing SDFI to ↑SCLK Setup Time tSSFI0 tHSFI0 SDFI to ↑SCLK Hold Time SDIN to ↑SCLK Setup Time tSSDI0 tHSDI0 SDIN to ↑SCLK Hold Time tDSFO0B ↑SCLK to SDFO Delay Slave Mode Serial Port Timing Requirements (SCS = 1): Switching Characteristics3 tSCLK SCLK Period tSCLKL SCLK Low Time tSCLKH SCLK High Time Channel is Self-Framing SDIN to ↑SCLK Setup Time tSSDH tHSDH SDIN to ↑SCLK Hold Time ↑SCLK to SDFO Delay tDSFO1 Channel is External-Framing tSSFI1 SDFI to ↑ SCLK Setup Time SDFI to ↑SCLK Hold Time tHSFI1 tSSDI1 SDIN to ↑SCLK Setup Time tHSDI1 SDIN to ↑SCLK Hold Time tDSFO1 ↓SCLK to SDFO Delay NOTES 1 All Timing Specifications valid over VDD range of 2.375 V to 2.675 V and VDDIO range of 3.0 V to 3.6 V. 2 CLOAD = 40 pF on all outputs (unless otherwise specified). 3 The timing parameters for SCLK, SDIN, SDFI, SDFO, and SYNC apply to all four channels (A, B, C, and D). Specifications subject to change without notice. REV. 0 –5– AD6623AS Typ Max 0.5 × tCLK 6 7.5 Unit ns ns ns ns ns ns ns 2 tCLK 10 ns ns ns ns ns AD6623 MICROPROCESSOR PORT TIMING CHARACTERISTICS1, 2 Temp Test Level Min MODE INM Write Timing: tSC Control3 to ↑CLK Setup Time Control3 to ↑CLK Hold Time tHC WR(RW) to RDY(DTACK) Hold Time tHWR tSAM Address/Data to WR(RW) Setup Time Address/Data to RDY(DTACK) Hold Time tHAM WR(RW) to RDY(DTACK) Delay tDRDY tACC WR(RW) to RDY(DTACK) High Delay Full Full Full Full Full Full Full IV IV IV IV IV IV IV 4.5 2.0 8.0 3.0 2.0 4.0 4 × tCLK MODE INM Read Timing: Control3 to ↑CLK Setup Time tSC tHC Control3 to ↑CLK Hold Time tSAM Address to RD(DS) Setup Time Address to Data Hold Time tHAM tZOZ Data Three-State Delay tDD RDY(DTACK) to Data Delay RD(DS) to RDY(DTACK) Delay tDRDY tACC RD(DS) to RDY(DTACK) High Delay Full Full Full Full Full Full Full Full IV IV IV IV IV IV IV IV 4.5 2.0 3.0 2.0 4.0 8 × tCLK Parameter (Conditions) AD6623AS Typ Max Unit MICROPROCESSOR PORT, MODE INM (MODE = 0) 5 × tCLK 9 × tCLK ns ns ns ns ns ns ns 10 × tCLK 13 × tCLK ns ns ns ns ns ns ns ns 5 × tCLK ns ns ns ns ns ns ns ns MICROPROCESSOR PORT, MOTOROLA (MODE = 1) MODE MNM Write Timing: tSC Control3 to ↑CLK Setup Time tHC Control3 to ↑CLK Hold Time tHDS DS(RD) to DTACK(RDY) Hold Time RW(WR) to DTACK(RDY) Hold Time tHRW tSAM Address/Data to RW(WR) Setup Time tHAM Address/Data to RW(WR) Hold Time DS(RD) to DTACK(RDY) Delay tDDTACK tACC RW(WR) to DTACK(RDY) Low Delay Full Full Full Full Full Full IV IV IV IV IV IV 4.5 2.0 8.0 8.0 3.0 2.0 Full IV 4 × tCLK MODE MNM Read Timing: tSC Control3 to ↑CLK Setup Time Control3 to ↑CLK Hold Time tHC tHDS DS(RD) to DTACK(RDY) Hold Time tSAM Address to DS(RD) Setup Time Address to Data Hold Time tHAM tZD Data Three-State Delay tDD DTACK(RDY) to Data Delay DS(RD) to DTACK(RDY) Delay tDDTACK tACC DS(RD) to DTACK(RDY) Low Delay Full Full Full Full Full Full Full Full Full IV IV IV IV IV IV IV IV IV 4.0 2.0 8.0 3.0 2.0 8 × tCLK 9 × tCLK 10 × tCLK 13 × tCLK ns ns ns ns ns ns ns ns ns NOTES 1 All Timing Specifications valid over VDD range of 2.375 V to 2.675 V and VDDIO range of 3.0 V to 3.6 V. 2 CLOAD = 40 pF on all outputs (unless otherwise specified). 3 Specification pertains to control signals: RW, (WR), DS, (RD), CS. Specifications subject to change without notice. –6– REV. 0 AD6623 TIMING DIAGRAMS tCLK tCLKL CLK tCLKH RESET INOUT[17:0] OUT[17:0] QOUT tRESL tDO tZO tZO OEN Figure 4. RESET Timing Requirements Figure 1. Parallel Output Switching Characteristics CLK CLK tDSCLKH tSI t HI tSCLKH INOUT[17:0] SCLK QIN tSCLKL Figure 2. Wideband Input Timing Figure 5. SCLK Switching Characteristics (Divide by 1) CLK CLK tDSCLKH tSS tHS tDSCLKL SCLK SYNC Figure 6. SCLK Switching Characteristic (Divide by 2 or EVEN Integer) Figure 3. SYNC Timing Inputs CLK tDSCLKH tDSCLKLL SCLK Figure 7. SCLK Switching Characteristic (Divide by 3 or ODD Integer) REV. 0 –7– AD6623 SCLK tDSFO0A SDFO tSSDI0 SDIN tHSDI0 DATAn Figure 8. Serial Port Timing, Master Mode (SCS = 0), Channel is Self-Framing SCLK tDSFO1 SDFO tSSDI1 SDIN tHSDI1 DATAn Figure 9. Serial Port Timing, Slave Mode (SCS = 1), Channel is Self-Framing nCLKs SCLK tDSFO0B SDFO tHSFI0 tSSFI0 SDFI tSSDI0 SDIN tHSDI0 DATAn Figure 10. Serial Port Timing, Master Mode (SCS = 0), Channel is External-Framing nCLKs SCLK tDSFO1 SDFO tSSFI1 tHSFI1 SDFI tSSDI1 SDIN tHSDI1 DATAn Figure 11. Serial Port Timing, Slave Mode (SCS = 1), Channel is External-Framing –8– REV. 0 AD6623 TIMING DIAGRAMS—INM MICROPORT MODE TIMING DIAGRAMS—MNM MICROPORT MODE CLK CLK tHC tSC RD (DS) tHC tHDS DS (RD) tHWR tSC tHRW WR (RW) RW (WR) CS CS tHAM tSAM A[2:0] A[2:0] VALID ADDRESS tHAM tSAM D[7:0] tHAM tSAM VALID ADDRESS tHAM tSAM VALID DATA D[7:0] VALID DATA tDRDY RDY (DTACK) tDDTACK DTACK (RDY) tACC tACC NOTES 1. tACC ACCESS TIME DEPENDS ON THE ADDRESS ACCESSED. ACCESS TIME IS MEASURED FROM FE OF WR TO THE RE OF RDY. 2. tACC REQUIRES A MAXIMUM 9 CLK PERIODS. NOTES 1. tACC ACCESS TIME DEPENDS ON THE ADDRESS ACCESSED. ACCESS TIME IS MEASURED FROM FE OF DS TO THE FE OF DTACK. 2. tACC REQUIRES A MAXIMUM 9 CLK PERIODS. Figure 12. INM Microport Write Timing Requirements Figure 14. MNM Microport Write Timing Requirements CLK CLK tHC tHC tSC tSC RD (DS) DS (RD) WR (RW) RW (WR) CS CS tSAM tSAM A[2:0] VALID ADDRESS tDD tZD D[7:0] tHAM VALID ADDRESS A[2:0] tDD tZD tZD VALID DATA D[7:0] tACC NOTES 1. tACC ACCESS TIME DEPENDS ON THE ADDRESS ACCESSED. ACCESS TIME IS MEASURED FROM FE OF DS TO THE FE OF DTACK. 2. tACC REQUIRES A MAXIMUM 13 CLK PERIODS. NOTES 1. tACC ACCESS TIME DEPENDS ON THE ADDRESS ACCESSED. ACCESS TIME IS MEASURED FROM FE OF WR TO THE RE OF RDY. 2. tACC REQUIRES A MAXIMUM OF 13 CLK PERIODS AND APPLIES TO A[2:0] = 7, 6, 5, 3, 2, 1 REV. 0 tZD tDDTACK DTACK (RDY) tACC Figure 13. INM Microport Read Timing Requirements tHAM VALID DATA tDRDY RDY (DTACK) tHDS Figure 15. Motorola Microport Read Timing Requirements –9– AD6623 ABSOLUTE MAXIMUM RATINGS* THERMAL CHARACTERISTICS Supply Voltage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.6 V Input Voltage . . . . . . . . . . . . . . –0.3 V to +5 V (5 V Tolerant) Output Voltage Swing . . . . . . . . . . –0.3 V to VDDIO + 0.3 V Load Capacitance . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200 pF Junction Temperature Under Bias . . . . . . . . . . . . . . . . . 125°C Storage Temperature Range . . . . . . . . . . . . –65°C to +150°C Lead Temperature (5 sec) . . . . . . . . . . . . . . . . . . . . . . . 280°C 128-Lead MQFP: JA = 33°C/W, no airflow JA = 27°C/W, 200 lfpm airflow JA = 24°C/W, 400 lfpm airflow *Stresses greater than those listed above may cause permanent damage to the device. These are stress ratings only; functional operation of the devices at these or any other conditions greater than those indicated in the operational sections of this specification is not implied. Exposure to absolute maximum rating conditions for extended periods may affect device reliability. 196-Lead BGA: JA = 26.3°C/W, no airflow JA = 22°C/W, 200 lfpm airflow Thermal measurements made in the horizontal position on a 2-layer board. EXPLANATION OF TEST LEVELS I. 100% Production Tested II. 100% Production Tested at 25°C, and Sample Tested at Specified Temperatures III. Sample Tested Only IV. Parameter Guaranteed by Design and Analysis V. Parameter is Typical Value Only ORDERING GUIDE Model Temperature Range Package Description Package Option AD6623AS AD6623ABC AD6623S/PCB AD6623BC/PCB –40°C to +70°C (Ambient) –40°C to +85°C (Ambient) 128-Lead MQFP (Metric Quad Flatpack) 196-Lead BGA (Ball Grid Array) MQFP Evaluation Board with AD6623 and Software BGA Evaluation Board with AD6623 and Software S-128A BC-196 CAUTION ESD (electrostatic discharge) sensitive device. Electrostatic charges as high as 4000 V readily accumulate on the human body and test equipment and can discharge without detection. Although the AD6623 features proprietary ESD protection circuitry, permanent damage may occur on devices subjected to high-energy electrostatic discharges. Therefore, proper ESD precautions are recommended to avoid performance degradation or loss of functionality. –10– WARNING! ESD SENSITIVE DEVICE REV. 0 REV. 0 GND 21 –11– D6 37 GND 38 GND 35 GND 36 D7 33 GND 34 GND 32 GND 31 QOUT 30 OUT17 29 OUT16 28 OUT15 27 VDDIO 26 OUT14 25 OUT13 24 OUT12 23 SDFIC 117 OUT11 22 GND 116 GND 20 SDFIB 115 GND 19 SDFOB 114 OUT10 18 OUT9 17 OUT8 16 OUT7 15 VDDIO 14 OUT6 13 OUT5 12 OUT4 11 OUT3 10 GND 9 OUT2 8 OUT1 7 OUT0 6 GND 5 GND 4 GND 3 OEN 2 GND 1 65 GND 67 CLK 66 VDD 69 SYNC2 68 GND 71 INOUT17 70 QIN 73 SYNC3 72 GND 75 INOUT15 74 INOUT16 78 VDDIO 77 INOUT13 76 INOUT14 80 INOUT11 79 INOUT12 82 INOUT9 81 INOUT10 84 GND 83 GND 86 INOUT8 85 GND 88 INOUT6 87 INOUT7 90 VDDIO 89 INOUT5 92 INOUT3 91 INOUT4 94 INOUT1 93 INOUT2 97 INOUT0 96 GND 95 GND 100 TRST 99 GND 98 GND 102 GND 101 TCK AD6623 PIN CONFIGURATION 128-Lead MQFP GND 103 64 GND VDD 104 63 SYNC1 SDFIA 105 62 SYNC0 TMS 106 61 RESET TDO 107 TDI 108 60 CS SCLKA 109 59 VDD 58 A0 VDDIO 110 57 A1 SDFOA 111 SDINA 112 56 A2 SCLKB 113 55 MODE 54 GND AD6623 53 GND TOP VIEW (Not to Scale) 52 GND 51 RW(WR) SDINB 118 50 DTACK(RDY) SCLKC 119 49 DS(RD) 48 D0 SDFOC 120 SDINC 121 47 VDD 46 D1 VDDIO 122 45 D2 SCLKD 123 44 D3 SDFOD 124 SDIND 125 43 D4 42 GND SDFID 126 VDD 127 41 VDDIO 40 D5 GND 128 39 GND AD6623 128 PIN FUNCTION DESCRIPTIONS Pin Number Mnemonic Type Description GND OEN1 P I Ground Connection Active High Output Enable Pin OUT[17:0] VDD VDDIO QOUT O/T P P O/T D[7:0] DS (RD) DTACK (RDY) I/O/T I Parallel Output Data 2.5 V Supply 3.3 V Supply When HIGH indicates Q Output Data (Complex Output Mode) Bidirectional Microport Data INM Mode: Read Signal, MNM Mode: Data Strobe Signal 51 55 RW (WR) MODE I I 56, 57, 58 60 61 62 63 67 69 70 71, 74–77, 79–82, 86–89, 91–94, 97 A[2:0] CS RESET2 SYNC01 SYNC11 CLK1 SYNC21 QIN1 INOUT[17:0]1 I I I I I I I I I/O 73 100 101 105 106 107 108 109 111 112 113 114 115 117 118 119 120 121 123 124 125 126 SYNC31 TRST2 TCK1 SDFIA TMS2 TDO TDI1 SCLKA SDFOA SDINA1 SCLKB SDFOB SDFIB SDFIC SDINB1 SCLKC SDFOC SDINC1 SCLKD SDFOD SDIND1 SDFID I I I I I O I I/O O I I/O O I I I I/O O I I/O O I I 1, 3–5, 9, 19–21, 31, 32, 34–36, 38, 39, 42, 52–54, 64–65, 68, 72, 83–85, 95, 96, 98, 99, 102, 103, 116, 128 2 29, 28, 27, 25, 24, 23, 22, 18, 17, 16, 15, 13, 12, 11, 10, 8, 7, 6 47, 59, 66, 104, 127 14, 26, 41, 78, 90, 110, 122 30 33, 37, 40, 43, 44, 45, 46, 48 49 50 O Acknowledgment of a Completed Transaction (Signals when µP Port Is Ready for an Access) Open Drain, Must Be Pulled Up Externally Active HIGH Read, Active Low Write Sets Microport Mode: MODE = 1, MNM Mode; MODE = 0, INM Mode Microport Address Bus Chip Select, Active low enable for µP Access Active Low Reset Pin SYNC Signal for Synchronizing Multiple AD6623s SYNC Signal for Synchronizing Multiple AD6623s Input Clock SYNC Signal for Synchronizing Multiple AD6623s When HIGH indicates Q input data (Complex Input Mode) Wideband Input/Output Data (Allows Cascade of Multiple AD6623 Chips In a System) SYNC Signal for Synchronizing Multiple AD6623s Test Reset Pin Test Clock Input Serial Data Frame Input—Channel A Test Mode Select Test Data Output Test Data Input Bidirectional Serial Clock—Channel A Serial Data Frame Sync Output—Channel A Serial Data Input—Channel A Bidirectional Serial Clock—Channel B Serial Data Frame Sync Output—Channel B Serial Data Frame Input —Channel B Serial Data Frame Input—Channel C Serial Data Input—Channel B Bidirectional Serial Clock—Channel C Serial Data Frame Sync Output—Channel C Serial Data Input—Channel C Bidirectional Serial Clock—Channel D Serial Data Frame Sync Outpu—Channel D Serial Data Input—Channel D Serial Data Frame Input—Channel D NOTES 1 Pins with a Pull-Down resistor of nominal 70 kΩ. 2 Pins with a Pull-Up resistor of nominal 70 kΩ. –12– REV. 0 AD6623 PIN CONFIGURATION 196-Lead BGA TOP VIEW 1 2 3 4 5 6 7 8 9 10 11 12 13 14 A B C D E F G 15mm sq. H J K L BALL LEGEND M I/O GROUND N CORE POWER RING POWER P 1.0mm 4 5 6 7 8 9 10 11 12 SDFID SDINC SDINB SDFOB SCLKB SCLKA TDO SDFIA TCK B OEN SDIND SDFOC SDFIC SDINA TDI C SDFOD SKLKD SKLKC SDFIB SDFOA 1 A D 2 3 NC E IN4 VDD VDDIO VDD VDDIO VDD OUT6 VDD GND GND GND GND VDDIO IN3 IN5 IN7 VDDIO GND GND GND GND VDD IN6 IN8 IN9 VDD GND GND GND GND VDDIO IN11 IN10 VDDIO GND GND GND GND VDD IN12 IN14 IN13 VDD VDDIO VDD VDDIO VDD VDDIO IN16 IN17 IN15 D4 D1 DTACK (RDY) MODE (ALE) D5 D3 D0 RW(WR) G OUT8 OUT4 H OUT9 OUT10 J OUT11 OUT13 K OUT14 OUT17 L OUT16 OUT15 QOUT OUT12 D7 N D6 SYNC3 QIN D2 DS(RD) NC = NO CONNECT REV. 0 IN1 IN2 VDDIO OUT4 NC IN0 TMS OUT3 OUT5 P NC OUT1 F M 14 TRST OUT0 OUT2 13 –13– A1 RESET SYNC2 A0 SYNC0 A2 CS SYNC1 CLK NC AD6623 196-PIN FUNCTION DESCRIPTIONS Mnemonic Type Function P P G 2.5 V Supply 3.3 V IO Supply Ground I/O I I I I I I I I I I I I A Input Data (Mantissa) When HIGH Indicates Q Input Data (Complex Input Mode) Active LOW Reset Pin Input Clock All Sync Pins Go to All Four Output Channels All Sync Pins Go to All Four Output Channels All Sync Pins Go to All Four Output Channels All Sync Pins Go to All Four Output Channels Serial Data Input—Channel A Serial Data Input—Channel B Serial Data Input—Channel C Serial Data Input—Channel D Active LOW Chip Select I/O I/O I/O I/O O O O O I I I I I Bidirectional Serial Clock—Channel A Bidirectional Serial Clock—Channel B Bidirectional Serial Clock—Channel C Bidirectional Serial Clock—Channel D Serial Data Frame Sync Output—Channel A Serial Data Frame Sync Output—Channel B Serial Data Frame Sync Output—Channel C Serial Data Frame Sync Output—Channel D Serial Data Frame Input—Channel A Serial Data Frame Input—Channel B Serial Data Frame Input—Channel C Serial Data Frame Input—Channel D Active High Output Enable Pin I/O/T I I O/T I I Bidirectional Microport Data Microport Address Bus Active Low Data Strobe (Active Low Read) Active Low Data Acknowledge (Microport Status Bit) Read Write (Active Low Write) Intel or Motorola Mode Select O O Wideband Output Data When HIGH Indicates Q Output Data (Complex Output Mode) I I I O/T I Test Reset Pin (Active Low) Test Clock Input Test Mode Select Input Test Data Output Test Data Input POWER SUPPLY VDD VDDIO GND INPUTS INOUT[17:0]1 QIN1 RESET2 CLK1 SYNC01 SYNC11 SYNC21 SYNC31 SDINA1 SDINB1 SDINC1 SDIND1 CS CONTROL SCLKA SCLKB SCLKC SCLKD SDFOA SDFOB SDFOC SDFOD SDFIA SDFIB SDFIC SDFID OEN1 MICROPORT CONTROL D[7:0] A[2:0] DS (RD) DTACK (RDY)2 RW (WR) MODE OUTPUTS OUT[17:0] QOUT JTAG AND BIST TRST2 TCK1 TMS2 TDO TDI1 NOTES 1 Pins with a Pull-Down resistor of nominal 70 kΩ. 2 Pins with a Pull-Up resistors of nominal 70 kΩ. –14– REV. 0 AD6623 SERIAL DATA PORT The AD6623 has four independent Serial Ports (A, B, C, and D), and each accepts data to its own channel (A, B, C, or D) of the device. Each Serial Port has four pins: SCLK (Serial CLocK), SDFO (Serial Data Frame Out), SDFI (Serial Data Frame In), and SDIN (Serial Data INput). SDFI and SDIN are inputs, SDFO is an output, and SCLK is either input or output depending on the state of SCS (Serial Clock Slave: 0xn16, Bit 4). Each channel can be operated either as a Master or Slave channel depending upon SCS. The Serial Port can be self-framing or accept external framing from the SFDI pin or from the previous adjacent channel (0xn16, Bits 7 and 6). SDFO is used to provide a sync signal to the host. The input sample rate is determined by the CLK divided by channel interpolation factor. If the SCLK rate is not an integer multiple of the input sample rate, then the SDFO will continually adjust the period by one SCLK cycle to keep the average SDFO rate equal to the input sample rate. When the channel is in sleep mode, SDFO is held low. The first SDFO is delayed by the channel reset latency after the Channel Reset is removed. The channel reset latency varies dependent on channel configuration. External Framing Mode In master mode, SCLK is created by a programmable internal counter that divides CLK. When the channel is “sleeping,” SCLK is held low. SCLK becomes active on the first rising edge of CLK after Channel sleep is removed (D0 through D3 of external address 4). Once active, the SCLK frequency is determined by the CLK frequency and the SCLK divider, according to the equations below. In this mode Bit 7 of register 0xn16 is set high. The external framing can come from either the SDFI pin (0xn16, Bit 6 = 0) or the previous adjacent channel (0xn16, Bit 6 = 1). In the case of external framing from a previous channel, it uses the internal frame end signal for serial data frame syncing. When in master mode, SDFO and SDFI transition on the positive edge of SCLK, and SDIN is captured on the positive edge of SCLK. When in slave mode, SDFO and SDFI transition on the negative edge of SCLK, and SDIN is captured on the negative edge of SCLK. AD6623 mode: Serial Port Cascade Configuration Serial Master Mode (SCS = 0) fSCLK fCLK = SCLKdivider + 1 (1) AD6622 mode: fSCLK = fCLK 2 × (SCLKdivider + 1) (2) The SCLK divider is a 5-bit unsigned value located at Internal Channel Address 0xn0D (Bits 4–0), where “n” is 1, 2, 3, or 4 for the chosen channel A, B, C, or D, respectively. The user must select the SCLK divider to insure that SCLK is fast enough to accept full input sample words at the input sample rate. See the design example at the end of this section. The maximum SCLK frequency is equal to the CLK when operating in AD6623 mode serial clock master. When operating in AD6622 compatible mode, the maximum SCLK frequency is one-half the CLK. The minimum SCLK frequency is 1/32 of the CLK frequency in AD6623 mode or 1/64 of the CLK frequency when in AD6622 mode. SDFO changes on the positive edge of SCLK when in master mode. SDIN is captured on positive edge when SCLK is in master mode. In this case the SDFO signal from the last channel of the first chip would be programmed to be a serial data frame end (SFE: 0xn16, Bit 5 = 1). This SDFO signal would then be fed as an input for the second cascaded chip’s SDFI pin input. The second chip would be programmed to accept external framing from the SDFI pin (0xn16, Bit 7 = 1, Bit 6 = 0). Serial Data Format The format of data applied to the serial port is determined by the RCF mode selected in Control Register 0xn0C. Below is a table showing the RCF modes and input data format that it sets. Table I. Serial Data Format 0xn0C 0xn0C Bit 6 Bit 5 0xn0C Bit 4 Serial Data Word Length RCF Mode 0 0 0 0 1 0 1 0 1 0 32 FIR /4-DQPSK GMSK MSK 0 0 1 1 0 Serial Slave Mode (SCS = 1) Any of the AD6623 serial ports may be operated in the serial slave mode. In this mode, the selected AD6623 channel requires that an external device such as a DSP to supply the SCLK. This is done to synchronize the serial port to meet an external timing requirement. SDIN is captured on negative edge of SCLK when in slave mode. Self-Framing Mode In this mode Bit 7 of register 0xn16 is set low. The serial data frame output, SDFO, generates a self-framing data request and is pulsed high for one SCLK cycle at the input sample rate. In this mode, the SDFI pin is not used, and the SDFO signal would be programmed to be a serial data frame request (0xn16, Bit 5 = 0). REV. 0 1 1 1 0 1 1 1 0 1 24 (Bit 9 is high) 16 (Bit 9 is low) FIR, compact 8-PSK 3/8-8-PSK QPSK The serial data input, SDIN, accepts 32-bit words as channel input data. The 32-bit word is interpreted as two 16-bit two’s complement quadrature words, I followed by Q, MSB first. This results in linear I and Q data being provided to the RCF. The first bit is shifted into the serial port starting on the next rising edge of SCLK after the SDFO pulse. Figure 16 shows a timing diagram for SCLK master (SCS = 0) and SDFO set for frame request (SFE = 0). –15– AD6623 PROGRAMMABLE RAM COEFFICIENT FILTER (RCF) CLK Each channel has a fully independent RAM Coefficient Filter (RCF). The RCF accepts data from the Serial Port, processes it, and passes the resultant I and Q data to the CIC filter. A variety of processing options may be selected individually or in combination, including PSK and MSK modulation, FIR filtering, all-pass phase equalization, and scaling with arbitrary ramping. See Table III. tSSDI0 SCLK CLKn tDSDFO0A SDFO tSSDI0 SDI tHSDI0 Table III. Data Format Processing Options DATAn Figure 16. Serial Port Switching Characteristics As an example of the Serial Port operation, consider a CLK frequency of 62.208 MHz and a channel interpolation of 2560. In that case, the input sample rate is 24.3 kSPS (62.208 MHz/2560), which is also the SDFO rate. Substituting, fSCLK ≥ 32 ⫻ fSDFO into the equation and solving for SCLKdivider, we find the minimum value for SCLKdivider according to the equation below. SCLKdivider ≤ fCLK 32 × fSFDO (3) Evaluating this equation for our example, SCLKdivider must be less than or equal to 79. Since the SCLKdivider channel register is a 5-bit unsigned number it can only range from 0 to 31. Any value in that range will be valid for this example, but if it is important that the SDFO period is constant, then there is another restriction. For regular frames, the ratio fSCLK/fSDFO must be equal to an integer of 32 or larger. For this example, constant SDFO periods can only be achieved with an SCLK divider of 31 or less. See Table II for usable SCLK divider values and the corresponding SCLK and fSCLK/fSDFO ratio for the example of L = 2560. In conclusion, SDFO rate is determined by the AD6623 CLK rate and the interpolation rate of the channel. The SDFO rate is equal to the channel input rate. The channel interpolation is equal to RCF interpolation times CIC5 interpolation, times CIC2 interpolation: LCRIC 2 L = LRCF × LCIC 5 × MCRIC 2 (4) The SCLK divide ratio is determined by SCLKdivider as shown in the previous equation. The SCLK must be fast enough to input 32 bits of data prior to the next SDFO. Extra SCLKs are ignored by the serial port. Table II. Example of Usable SCLK Divider Values and fSCLK/fSDPO Ratios for L = 2560 SCLKdivisor fSCLK/fSDFO 0 1 3 4 7 9 15 19 31 2560 1280 640 512 320 256 160 128 80 Processing Block Input Data Output Data Interpolating FIR Filter PSK Modulator I and Q 2 or 3 bits per symbol I and Q MSK Modulator 1 bit per symbol QPSK 2 bits per symbol All-pass Phase Equalizer Scale and Ramp I and Q I and Q Unfiltered I and Q: /4-QPSK, 8-PSK, or 3/8-8-PSK Filtered MSK or GSM I and Q Filtered QPSK I and Q I and Q I and Q OVERVIEW OF THE RCF BLOCKS The Serial Port passes data to the RCF with the appropriate format and bit precision for each RCF configuration, see Figure 17. The data may be modulated vectors or unmodulated bits. I and Q vectors are sent directly to the Interpolating Fir Filter. Unmodulated bits may be sent to the PSK Modulator, the Interpolating MSK Modulator, or the Interpolating QPSK Modulator. The PSK Modulator produces unfiltered I and Q vectors at the symbol rate which are then passed through the Interpolating FIR Filter. The Interpolating MSK Modulator and the Interpolating QPSK Modulator produce oversampled, pulse-shaped vectors directly without employing the Interpolating FIR Filter. When possible, the MSK and QPSK modulators are recommended for increased throughput and decreased power consumption compared to Interpolating FIR Filter. In addition, the Interpolating MSK Modulator can realize filters with nonlinear inter-symbol interference, achieving excellent accuracy for GMSK applications. After interpolation, an optional Allpass Phase Equalizer (APE) can be inserted into the signal path. The APE can realize any real, stable, two-pole, two-zero all-pass filter at the RCF’s interpolated rate. This is especially useful to precompensate for nonlinear phase responses of receive filters in terminals, as specified by IS-95. When active, the APE utilizes shared hardware with the interpolating modulators and filter, which may reduce the allowed RCF throughput, inter-symbol interference, or both. See Figure 18. –16– REV. 0 AD6623 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10 < msb, I, lsb > 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 < msb, I, lsb > 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 BIT s D1 D2 D0 8PSK 4 3 2 1 0 BIT SERIAL SYNC m s X D1 D0 QPSK RAMP 4 3 2 1 0 BIT M S X X D0 MSK/GSM 2 1 0 BIT 0 D1 D0 8PSK 1 0 BIT D1 D0 QPSK 0 BIT D0 MSK/GSM 3 2 1 6 0 < msb, Q, lsb > 4 7 5 4 3 2 1 < msb, Q, lsb > < msb, I, lsb > m 9 8 7 6 5 4 3 2 1 0 < msb, Q, lsb > 0 BIT FIR BIT COMPACT FIR BIT COMPACT FIR Figure 17. Data Formats Supported by the AD6623 when SCLK Master (SCS = 0), and SFDO Set for Frame Request (SFE = 0) INTERPOLATING FIR FILTER PSK MODULATOR INTERPOLATING MSK MODULATOR SCALE AND RAMP ALLPASS PHASE EQUALIZER DATA FROM SERIAL PORT DATA TO CIC FILTERS INTERPOLATING QPSK MODULATOR Figure 18. RCF Block Diagram Table IV. FIR Filter Internal Precision Signal x y Notation Minimum Decimal Hexadecimal (h) Maximum Decimal Hexadecimal (h) I and Q Inputs Coefficients Product Sum FIR Output 1.15 1.15 2.18 4.18 1.17 –1.00000 –1.00000 –0.99969 –7.00000 –1.00000 0.999969 0.999969 1.000000 7.999996 0.999992 0.FFFE 0.FFFE 1.00000 7.FFFFC 0.FFFF8 +1.00000 +1.00000 +3.00020 +8.0000 +1.00000 The Scale and Ramp block adjusts the final magnitude of the modulated RCF output. A synchronization pulse from the SYNC0–3 pins or serial words can be used to command this block to ramp down, pause, and ramp up to a new scale factor. The shape of the ramp is stored in RAM, allowing complete sample by sample control at the RCF interpolated rate. This is particularly useful for time division multiplexed standards such as GSM/EDGE. Modulator configurations can be updated while the ramp is quiet, allowing for GSM and EDGE timeslots to be multiplexed together without resetting or reconfiguring the channel. Each of the RCF processing blocks is discussed in greater detail in the following sections. REV. 0 INTERPOLATING FIR FILTER The Interpolating FIR Filter realizes a real, sum-of-products filter on I and Q inputs using a single interleaved Multiply-Accumulator (MAC) running at the CLK rate. The input signal is interpolated by integer factors to produce arbitrary impulse responses up to 256 output samples long. Each bus in the data path carries bipolar two’s complement values. For the purpose of discussion, we will arbitrarily consider the radix point positioned so that the input data ranges from –1 to just below 1. In Figure 19, the data buses are marked x ⫻ y to denote finite precision limitations. A bus marked x ⫻ y has x bits above the radix and y bits below the radix, which implies a range from –17– AD6623 –2x–1 to 2x–1 – 2–y in 2–y steps. The range limits are tabulated in Table IV for each bus. The hexadecimal values are bit-exact and each MSB has negative weight. Note that the Product bus range is limited by result of the multiplication and the two most significant bits are the same except in one case. INPUT 1.15 DMEM 3216 PRODUCT 1.15 OUTPUT 2.18 COEF 1.15 CMEM 25616 The FIR accepts two’s complement I and Q samples from the serial port with a fixed-point resolution of 16 bits each. When the serial port provides data with less precision, the LSBs are padded with zeroes. ACCUMULATOR 4.18 INPUT 1.17 20, 2–1, 2–2, OR 2–3 Figure 19. Interpolating FIR Filter Block Diagram The RCF realizes a FIR filter with optional interpolation. The FIR filter can produce impulse responses up to 256 output samples long. The FIR response may be interpolated up to a factor of 256, although the best filter performance is usually achieved when the RCF interpolation factor (LRCF) is confined to eight or below. The 256 ⫻ 16 coefficient memory (CMEM) can be divided among an arbitrary number of filters, one of which is selected by the Coefficient Offset Pointer (channel address 0x0B). The polyphase implementation is an efficient equivalent to an integer up-sampler followed FIR filter running at the interpolated rate. The AD6623 RCF realizes a sum-of-products filter using a polyphase implementation. This mode is equivalent to an interpolator followed by a FIR filter running at the interpolated rate. In the functional diagram below, the interpolating block increases the rate by the RCF interpolation factor (LRCF) by inserting LRCF–1 zero valued samples between every input sample. The next block is a filter with a finite impulse response length (NRCF) and an impulse response of h[n], where n is an integer from 0 to NRCF–1. The difference equation for Figure 20 is written below, where h[n] is the RCF impulse response, b[n] is the interpolated input sample sequence at point ‘b’ in the diagram above, and c[n] is the output sample sequence at point ‘c’ in Figure 20. fIN LRCF fIN LRCF a b NRCFTAP FIR FILTER h[n] [] c n = ∑ k=0 [] [ ] h n ×b n–k N RCF – 1 ∑ k=0 [] h n × z –1 The Coef-Mem stores up to 256 16-bit filter coefficients. The CoefMem can be accessed through the Microport from 0x800 to 0x8FF above the processing channel’s base internal address, while the channel’s Prog bit is set (external address 4). For AD6622 compatibility, the lower 128 words are also mirrored from 0x080 to 0x0FF above the processing channel’s base internal address, while the Prog bit is set. To avoid start-up transients, the Data-Mem should be cleared before operation. The Prog bit must then be reset to enable channel operation. There is a single Multiply-Accumulator (MAC) on which both the I and Q operations must be interleaved. Two CLK cycles are required for the MAC to multiply each coefficient by an I and Q pair. The MAC is also used for four additional CLK cycles if the All-pass Phase Equalizer is active. The size of the Data-Mem and Coef-Mem combined with the speed of the MAC determine the total number of the taps per phase (TRCF) that may be calculated. TRCF is the number of RCF input samples that influence each RCF output sample. The maximum available TRCF is calculated by the equation below. 256 fCLK – 2 × APE (7) TRCF ≤ least of 16, floor , floor 2 × f L RCF SDO c (5) This difference equation can be described by the transfer function from point ‘b’ to ‘c’ as: H bc ( z ) = The Data-Mem stores the most recent 16 I and Q pairs for a total of 32 words. The size of the Data-Mem limits the RCF impulse response to 16 ⫻ LRCF output samples. When the data words from the Serial Port have fewer than 16 bits, the LSBs are padded with zeroes. The Data-Mem can be accessed through the Microport from 0x20 to 0x5F above the processing channel’s base internal address, while the channel’s Prog bit is set (external address 4). In order to avoid start-up transients, the Data-Mem should be cleared before operation. The Prog bit must then be reset to enable normal operation. fIN LRCF Figure 20. RCF Interpolation N RCF – 1 of reducing by a factor of LRCF both the time needed to calculate an output and the required data memory (DMEM). The price of these benefits is that the user must place the coefficients into the coefficient memory (CMEM) indexed by the interpolation phase. The process of selecting the coefficients and placing them into the CMEM is broken into three steps shown below. (6) The actual implementation of this filter uses a polyphase decomposition to skip the multiply-accumulates when b[n–k] is zero. Compared to the diagram above, this implementation has the benefits The impulse response length at the output of the RCF is determined by the product of the number of interfering input samples (TRCF) and the RCF interpolation factor (LRCF), as shown by equation (8) below. The values of NRCF and TRCF are programmed into control registers. LRCF is not a control register, but NRCF and TRCF must be set so that LRCF is an integer. If the integer interpolation by the RCF results in an inconvenient sample rate at the output of the RCF, the desired output rate can usually be achieved by selecting non-integer interpolation in the resampling CIC2 filter. –18– N RCF = TRCF × LRCF (8) REV. 0 AD6623 Table V. RCF Control Registers Channel Address Bit Width 0x0A 0x0B 0x0C 16 8 10 0x0D 8 0x0E 16 0x0F 18 0x110 0x111 0x112 0x113 0x114 0x115 0x116 16 16 16 16 16 16 8 REV. 0 Description 15–8: NRCF –1 B; 7–0: NRCF –1 A 7–0: ORCF 9: Ch. A Compact FIR Input Word Length 0: 16 bits–8 I followed by 8 Q 1: 24 bits–12 I followed by 12 Q 8: Ch. A RCF PRBS Enable 7: Ch A RCF PRBS Length 0: 15 1: 8,388,607 6–4: Ch. A RCF Mode Select 000 = FIR 001 = p/4-DQPSK Modulator 010 = GMSK Look-Up Table 011 = MSK Look-Up Table 100 = FIR compact mode 101 = 8-PSK 110 = 3p/8-8PSK Modulator 111 = QPSK Look-Up Table 3–0: Ch. A RCF Taps per Phase 7–6: RCF Coarse Scale (g): 00 = 0 dB 01 = –6 dB 10 = –12 dB 11 = –18 dB 5: Ch. A Allpass Ph. Eq. Enable 4–0: Serial Clock Divider (1, ..., 32) 15–2: Ch. A Unsigned Scale Factor 1–0: Reserved 17–16: Ch. A Time Slot Sync Select 00: Sync0 (See 0x001 Time Slot) 01: Sync1 10: Sync2 11: Sync3 15–0: Ch. A RCF Scale Hold-Off Counter 1) Ramp Down (if Ramp is enabled) 2) Update Scale and Mode 3) Ramp Up (if Ramp is enabled) 15–0: Ch. A RCF Phase EQ Coef1 15–0: Ch. A RCF Phase EQ Coef2 15–0: Ch. A RCF MPSK Magnitude 0 15–0: Ch. A RCF MPSK Magnitude 1 15–0: Ch. A RCF MPSK Magnitude 2 15–0: Ch. A RCF MPSK Magnitude 3 7: Reserved 6: Ch. A Serial Data Frame Select 0: Serial Data Frame Request 1: Serial Data Frame End Channel Address 0x117 0x118 0x119 0x11A–0x11F 0x120–0x13F 0x140–0x17F Bit Width Description 6 6 5 16 16 0x180–0x1FF 16 5: Ch. A External SDFI Select 0: Internal SDFI 1: External SDFI 4: Ch. A SCLK Slave Select 0: Master 1: Slave 3: Ch. A Serial Fine Scale Enable 2: Ch. A Serial Time Slot Sync Enable (ignored in FIR mode) 1: Ch. A Ramp Interpolation Enable 0: Ch. A Ramp Enable 5–0: Ch. A Mode 0 Ramp Length, R0–1 5–0: Ch. A Mode 1 Ramp Length, R1–1 4–0: Ch. A Ramp Rest Time, Q Reserved 15–0: Ch. A Data Memory 15–14: Reserved 13–0: Ch. A Power Ramp Memory 15–0: Ch. A Coefficient Memory This address is mirrored at 0x900–0x97F and contiguously extended at 0x980–0x9FF PSK MODULATOR The PSK Modulator is an AD6623 extension feature that is only available when the control register bit 0x000:7 is high. The PSK Modulator creates 32-bit complex inputs to the Interpolating FIR Filter from two or three data bits captured by the serial port. The FIR Filter operates exactly as if the 32bit word came directly from the serial port. There are three PSK modulation options to choose from: /4-DQPSK, 8-PSK, and 3/8-8-PSK. Every symbol of any of these modulations can be represented by one of the 16 phases shown in Figure 21. 0 Figure 21. 16-Phase Modulations –19– AD6623 All of these phase locations are represented in rectangular coordinates by only four unique magnitudes in the positive and negative directions. These four values are read from four channel registers that are programmed according to the following table, which gives the generic formulas and a specific example. The example is notable because it is only 0.046 dB below full-scale and the 16-bit quantization is so benign at that magnitude, that the rms error is better than –122 dBc. It is also worth noting that because none of the phases are aligned with the axes, magnitudes slightly beyond 0.16 dB above full-scale are achievable. SERIAL QPSK MAPPER [1:0] SPH PHASE [3:0] [3:0] RPH [3:0] 2 Figure 22. QPSK Mapper The Sph word is calculated by the QPSK Mapper according to the following truth table. Table VI. Program Registers Table VIII. QPSK Mapper Truth Table Channel Register Magnitude M Magnitude E 0x7F53 Serial [1:0] Sph [3:0] 0x12 0x13 0x14 0x15 M 3 cos(p/16) M 3 cos(3p/16) M 3 cos(5p/16) M 3 cos(7p/16) 0x7CE1 0x69DE 0x46BD 0x18D7 00b 01b 11b 10b 0 4 8 12 Using the four channel registers from the preceding table, the PSK Modulator assembles the 16 phases according to Table VII. Table VII. PSK Modulator Phase Phase I Value Q Value 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0x12 0x13 0x14 0x15 –0x15 –0x14 –0x13 –0x12 –0x12 –0x13 –0x14 –0x15 0x15 0x14 0x13 0x12 0x15 0x14 0x13 0x12 0x12 0x13 0x14 0x15 –0x15 –0x14 –0x13 –0x12 –0x12 –0x13 –0x14 –0x15 8-PSK Modulation IS-136+ compliant 8-PSK modulation is selected by setting the channel register 0x0C: 6–4 to 101b. The Phase word is calculated according to the following diagram. The three LSBs of the serial input word update the payload bits once per symbol. SERIAL [2:0] PHASE [3:0] Figure 23. 8-PSK Mapper The Phase word is calculated by the 8-PSK Mapper according to the following truth table: Table IX. 8-PSK Mapper Truth Table The following three sections show how the phase values are created for each PSK modulation mode. /4-DQPSK Modulation 8PSK MAPPER IS-136 compliant /4-DQPSK modulation is selected by setting the channel register 0x0C: 6–4 to 001b. The phase word is calculated according to the following diagram. The two LSBs of the serial input word update the payload bits once per symbol. The QPSK Mapper creates a data dependent static phase word (Sph) which is added to a time dependent rotating phase word (Rph). The Rph starts at zero when the RCF is reset or switches modes via a sync pulse. Otherwise, the Rph increments by two on every symbol. Serial [2:0] Sph [3:0] 111b 011b 010b 000b 001b 101b 100b 110b 0 2 4 6 8 10 12 14 3 /8-8-PSK Modulation EDGE compliant 3 /8-8-PSK modulation is selected by setting the channel register 0x0C: 6–4 to 110b. The phase word is calculated according to the following diagram. The three LSBs of the serial input word update the payload bits once per symbol. The 8-PSK Mapper creates a data-dependent static phase word (Sph) which is added to a time-dependent rotating phase word (Rph). The 8-PSK Mapper operates exactly as described in the preceding 8-PSK Modulation section. The Rph starts at zero when the RCF is reset or switches modes via a sync pulse. Otherwise, the Rph increments by three on every symbol. –20– REV. 0 AD6623 SERIAL 8PSK MAPPER [2:0] SPH PHASE [3:0] [3:0] (2 –15) ⫻ 2, or 0.00006103515625. The register values, in hexadecimal, and the corresponding coefficient weight from positive full-scale through zero to negative full-scale is illustrated in Table X. RPH [3:0] Table X. Coefficient Weights 3 Figure 24. 3 π/8-8-PSK Mapper MSK Look-Up Table The MSK Look-Up Table mode for the RCF is selected in Control Register 0x10C. In the MSK Mode, the RCF performs arbitrary pulse-shaping based on four symbols of impulse response. For the MSK Mode (3, [16] bits), the serial input format is 11 bits of scaling (MSB first) followed by 1 bit of data. The 11 bits can be used to scale the input data. GMSK Look-Up Table The GMSK Look-Up Table mode for the RCF is selected in Control Register 0x10C. In the GMSK Mode, the RCF performs arbitrary pulse-shaping based on four symbols of impulse response. For the GMSK Mode (3, [16] bits), the serial input format is 11 bits of scaling (MSB first) followed by 1 bit of data. The 11 bits can be used to scale the input data. The IS-95 Standard includes a phase equalizer after matched filtering at the baseband transmit side of a base station. This filter pre-distorts the transmitted signal at the base station in order to compensate for the distortion introduced to the received signal by the analog baseband filtering in a handset. The AD6623 includes this functionality in the form of an Infinite Impulse Response (IIR) all-pass filter in the RCF. This Phase Equalizer pre-distort filter has the following transfer function: X(z) Y ( z ) 1 + b1z + b2z 2 = X ( z ) z 2 + b1z + b2 Z–1 (9) Z–1 b1 b2 Z–1 Z–1 Y(z) Figure 25. Second Order All-Pass IIR Filter The Phase Equalizer is enabled/disabled in Control Register 0xn0D Bit 5. The coefficients b1 and b2 are located in Control Registers 0xn10 and 0xn11 respectively The format for b1 and b2 is two’s complement fractional binary with a range of (–2, 2). With one bit for sign at most significant bit position there are 15 bits for magnitude. The value of one bit is REV. 0 0x7FFF .. 0x0001 0x0000 0xFFFF .. 0x8001 0x8000 1.999938964844 0.00006103515625 0 –0.00006103515625 –1.999938964844 –2 Table XI. b1 and b2 Coefficients Oversampling b0 b1 b2 4 5 6 8 1 1 1 1 –1.45514 –1.56195 –1.63412 –1.72513 0.57832 0.64526 0.69414 0.76047 SCALE AND RAMP Scale factors can be range from 0 to [CHF]–1/[CHF] with a resolution 1/[CHF] FINE SCALING PHASE EQUALIZER H( z ) = Coefficient Weight Table XI shows the recommended b1 and b2 coefficients for the respective oversampling rate. QPSK Look-Up Table The QPSK Filter mode for the RCF is selected in Control Register 0xX0C. In the QPSK Mode, the RCF performs baseband linear pulse-shaping based on filter impulse response up to 12 symbols. For the QPSK Mode (3, [16]bits), the serial input format is 13 bits of scaling (MSB first) followed by one bit I and then one bit Q. The 13 bits can be used to scale the input data. Register Value AD6623 allows fine scaling of the RCF output signal. A scale factor of 12 to 14 bits is available through the Microport. The Microport fine scale factor is located in Channel Register 0xn0E. RCF POWER RAMPING The output of the RCF will be multiplied by a 14-bit ramping profile before entering the CIC filters. It is a RAM programmable engine that starts indexing through ramping coefficients when the RAMP bit works its way through the chain. It will then count a programmable number of samples and then RAMP down in reverse order. This will allow the ramping values to update at a modest rate relative to the DAC and still contain the spectral leakage associated with the ramping. A user should provide through the MicroPort the ramping coefficient values, the number of samples to ramp up, and one bit to define the air-interface standard. The programmable power ramp up/down unit allows power ramping on time-slot basis as specified from some wireless transmission technologies (e.g, TDMA). CASCADED INTEGRATOR COMB (CIC) INTERPOLATING FILTERS The I and Q outputs of the RCF stage are interpolated by two cascaded integrator comb (CIC) filters. The CIC section is separated into three discrete blocks: a fifth order filter (CIC5), a second order resampling filter (rCIC2), and a scaling block (CIC Scaling). The CIC5 and rCIC2 blocks each exhibit a gain that changes with respect to their rate change factors, LrCIC2, MrCIC2 and LCIC5. The product of these gains must be compensated for in a shared CIC Scaling block and can be done to within 6 dB The remaining compensation can come from the RCF (in the form of coefficient scaling) or the fine scaling unit. –21– AD6623 baseband, but internal registers peak in response to various dynamic inputs. As long as LCIC5 is confined to 32 or less, there is no possibility of overflow at any register. CIC Scaling The scale factor SCIC is a programmable unsigned integer between 4 and 32. This is a combined scaler for the CIC5 and rCIC2 stages. The overall gain of the CIC section is given by the equation below 4 CIC _ Gain = LCIC 5 × LrCIC 2 × 2–S CIC L × f sin π CIC 5 fCIC 5 1 CIC5( f ) = LCIC 5 f sin π f CIC 5 (10) CIC5 The first CIC filter stage, the CIC5, is a fifth order interpolating cascaded integrator comb whose impulse response is completely defined by its interpolation factor, LCIC5. The value LCIC5–1 can be independently programmed for each channel at location 0xn09. 2–SCIC LCIC5 CIC_SCALE CIC5 5 (16) The pass band droop of CIC5 should be calculated using this equation and can be compensated for in the RCF stage. The gain should be calculated from the CIC scaling section above. As an example, consider an input from the RCF whose bandwidth is 0.141 of the RCF output rate, centered at baseband. Interpolation by a factor of five reveals five images, as shown below. LrCIC2 MrCIC2 rCIC2 10 Figure 26. CIC5 –10 While this control register is 8 bits wide, LCIC5 should be confined to the range from 1 to 32 to avoid the possibility of internal overflow for full scale inputs. The output rate of this stage is given by the equation below. –50 dB fCIC 2 = fCIC 5 × LCIC 5 –30 (11) –70 –90 The transfer function of the CIC5 is given by the following equations with respect to the CIC5 output sample rate, fsamp5. 1 – z –L CIC 5 CIC 5( z ) = –1 1– z –110 5 –130 (12) The SCIC value can be independently programmed for each channel at Control Register 0xn06. SCIC may be safely calculated according to equation (13) below to ensure the net gain through the CIC stages. SCIC serves to frame which bits of the CIC output are transferred to the NCO stage. This results in controlling the data out of the CIC stages in 6 dB Increments. For the best dynamic range, SCIC should be set to the smallest value possible (lowest attenuation) without creating an overflow condition. This can be safely accomplished using the equation below. To ensure the CIC output data is in range, equation (13) must always be met. The maximum total interpolation rate may be limited by the amount of scaling available. ( ) –150 –3 –2 –1 0 1 2 3 Figure 27. Interpolation Images The CIC5 rejects each of the undesired images while passing the image at baseband. The images of a pure tone at channel center (DC) are nulled perfectly, but as the bandwidth increases the rejection is diminished. The lower band edge of the first image always has the least rejection. In this example, the CIC5 is interpolating by a factor of five and the input signal has a bandwidth of 0.141 of the RCF output sample rate. The plot below shows –110 dBc rejection of the lower band edge of the first image. All other image frequencies have better rejection. 10 –10 (13) –30 0 ≤ SCIC ≤ 58 (14) –50 dB SCIC ≥ ceil 4 × log 2 ( LCIC 5 ) + log 2 ( LCIC 2 ) This polynomial fraction can be completely reduced as follows demonstrating a finite impulse response with perfect phase linearity for all values of LCIC5. –70 –90 –110 k LCIC 5 –1 j2π LCIC 5 –1 CIC5( z ) = ∑ z – k = ∑ z –1 – e LCIC 5 k=0 k =1 5 5 (15) –130 –150 –3 The frequency response of the CIC5 can be expressed as follows. The initial 1/LCIC5 factor normalizes for the increased rate, which is appropriate when the samples are destined for a DAC with a zero order hold output. The maximum gain is LCIC54 at –2 –1 0 1 2 3 Figure 28. –110 dBc Rejection –22– REV. 0 AD6623 Table XII lists maximum bandwidth that will be rejected to various levels for CIC5 interpolation factors from 1 to 32. The example above corresponds to the listing in the –110 dB column and the LCIC5 = 5 row. It is worth noting here that the rejection of the CIC5 improves as the interpolation factor increases. Table XIII. Maximum Permissible L RC1C2 Values Table XII. Max Bandwidth of Rejection for L CIS Values –110 dB –100 dB –90 dB –80 dB –70 dB Full 0.101 0.126 0.136 0.136 0.143 0.144 0.145 0.146 0.146 0.147 0.147 0.147 0.147 0.148 0.148 0.148 0.148 0.148 0.148 0.148 0.148 0.148 0.148 0.148 0.148 0.148 0.148 0.148 0.148 0.148 0.148 Full 0.127 0.159 0.170 0.175 0.178 0.179 0.180 0.181 0.182 0.182 0.182 0.183 0.183 0.183 0.183 0.183 0.183 0.183 0.184 0.184 0.184 0.184 0.184 0.184 0.184 0.184 0.184 0.184 0.184 0.184 0.184 Full 0.160 0.198 0.211 0.217 0.220 0.222 0.224 0.224 0.225 0.226 0.226 0.226 0.226 0.227 0.227 0.227 0.227 0.227 0.227 0.227 0.227 0.227 0.227 0.227 0.227 0.227 0.227 0.227 0.227 0.227 0.228 Full 0.203 0.246 0.262 0.269 0.272 0.275 0.276 0.277 0.278 0.278 0.279 0.279 0.279 0.280 0.280 0.280 0.280 0.280 0.280 0.280 0.280 0.280 0.280 0.281 0.281 0.281 0.281 0.281 0.281 0.281 0.281 Full 0.256 0.307 0.325 0.333 0.337 0.340 0.341 0.342 0.343 0.344 0.344 0.345 0.345 0.345 0.345 0.346 0.346 0.346 0.346 0.346 0.346 0.346 0.346 0.346 0.346 0.346 0.346 0.346 0.346 0.346 0.346 Maximum Allowed LrCIC2 32 31 24–30 23 1–22 4096 1162 30 – 4 log 2 ( L CIC 3 ) ] LrCIC 2 = 2[ 3836 4096 Two parameters determine the rate change in this block. They are the interpolation factor (LrCIC2, 12 bits) and the decimation factor (M, 9 bits). When combined, and subject to the maximum values of LrCIC2, the total rate change can be any fraction in the form of: RrCIC2 = L ≥ 1, 1 ≤ L ≤ 4096 , 1 ≤ M ≤ 512 M (17) The only constraint is that the ratio L/M must be greater than or equal to one. This implies that the rCIC2 has a net interpolation of 1 or more. Resampling is implemented by apparently increasing the input sample rate by the factor L, using zero stuffing for the new data samples. Following the resampler is a second order cascaded integrator comb filter. Filter characteristics are determined only by the fractional rate change (L/M). The filter can produce output signals at the full CLK rate of the AD6623. The output rate of this stage is given by the equation below. fout = LrCIC 2 frCIC 2 MrCIC 2 (18) Both LrCIC2 and MrCIC2 are unsigned integers. The interpolation rate (LrCIC2) may be from 1 to 4096 and the decimation (MrCIC2) may be between 1 and 512. The stage can be bypassed by setting the L and M to 1. The transfer function of the rCIC2 is given by the following equations with respect to the rCIC2 output sample rate, fout. 1 – z –L rCIC 2 rCIC 2( z ) = –1 1– z rCIC2 The rCIC2 filter is a second order Cascaded Resampling Integrator Comb filter whose impulse response is completely defined by its rate change factors, LrCIC2 and MrCIC2. The resampler is implemented using a unique technique that does not require a high-speed clock (but it’s still completely jitter-free), thus simplifying the design and saving power. The resampler allows for noninteger relationships between the input data rate and the master clock. This allows easier implementation of systems that are either multimode or require a master clock that is not a multiple of the input data rate to be used. The value of LrCIC2 is limited from 1 to 4096 subject to some restrictions based upon the chosen LCIC5 value as shown in Table XIII. Chosen LCIC5 Value 2 (19) The frequency response of the rCIC2 can be expressed as follows. The maximum gain is LrCIC2 at baseband. The initial MrCIC2/LrCIC2 factor normalizes for the increased rate, which is appropriate when the samples are destined for a DAC with a zero order hold output. rCIC 2( f ) = MrCIC 2 LrCIC 2 LrCIC 2 × f sin π fout f sin π fout 2 (20) The pass-band droop of CIC5 should be calculated using this equation and can be compensated for in the RCF stage. The gain should be calculated from the CIC scaling section above. The values MrCIC2–1, LrCIC2–1 can be independently programmed for each channel at locations 0xn07, 0xn08. While these control registers are nine bits and 12 bits wide respectively, MrCIC2–1 and REV. 0 –23– AD6623 LrCIC2–1 should be confined to the ranges shown by Table XIII according to the interpolation factor of the CIC5. Exceeding the recommended guidelines may result in overflow for input sequences at or near full scale. While relatively large ratios of LrCIC2/MrCIC2 allow for the larger overall interpolations with minimal power consumption, LrCIC2/MrCIC2 should be minimized to achieve the best overall image rejection. As an example, consider an input from the CIC5 whose bandwidth is 0.0033 of the CIC5 rate, centered at baseband. Interpolation by a factor of five reveals five images, as shown below. 10 –10 –30 dB –50 –70 –90 –110 –130 –150 –2 –1 0 1 2 Figure 29. CIC5 Interpolation Images The rCIC2 rejects each of the undesired images while passing the image at baseband. The images of a pure tone at channel center (DC) are nulled perfectly, but as the bandwidth increases the rejection is diminished. The lower band edge of the first image always has the least rejection. In this example, the rCIC2 is interpolating by a factor of five and the input signal has a bandwidth of 0.0033 of the CIC5 output sample rate. Figure 30 below shows –110 dBc rejection of the lower band edge of the first image. All other image frequencies have better rejection. –100 dB –90 dB –80 dB –70 dB Full 0.0023 0.0029 0.0032 0.0033 0.0034 0.0034 0.0035 0.0035 0.0035 0.0035 0.0035 0.0035 0.0035 0.0035 0.0035 0.0035 0.0035 0.0035 0.0035 0.0035 0.0035 0.0035 0.0035 0.0035 0.0035 0.0035 0.0035 0.0035 0.0035 0.0035 0.0035 Full 0.0040 0.0052 0.0057 0.0059 0.0060 0.0061 0.0062 0.0062 0.0062 0.0062 0.0062 0.0062 0.0063 0.0063 0.0063 0.0063 0.0063 0.0063 0.0063 0.0063 0.0063 0.0063 0.0063 0.0063 0.0063 0.0063 0.0063 0.0063 0.0063 0.0063 0.0063 Full 0.0072 0.0093 0.0101 0.0105 0.0107 0.0108 0.0109 0.0110 0.0110 0.0110 0.0111 0.0111 0.0111 0.0111 0.0111 0.0111 0.0111 0.0111 0.0111 0.0111 0.0111 0.0111 0.0112 0.0112 0.0112 0.0112 0.0112 0.0112 0.0112 0.0112 0.0112 Full 0.0127 0.0165 0.0179 0.0186 0.0189 0.0192 0.0193 0.0194 0.0195 0.0195 0.0196 0.0196 0.0196 0.0197 0.0197 0.0197 0.0197 0.0197 0.0197 0.0197 0.0197 0.0197 0.0197 0.0198 0.0198 0.0198 0.0198 0.0198 0.0198 0.0198 0.0198 Full 0.0226 0.0292 0.0316 0.0328 0.0334 0.0338 0.0341 0.0343 0.0344 0.0345 0.0346 0.0346 0.0347 0.0347 0.0347 0.0348 0.0348 0.0348 0.0348 0.0348 0.0348 0.0348 0.0348 0.0349 0.0349 0.0349 0.0349 0.0349 0.0349 0.0349 0.0349 Each channel has a fully independent tuner. The tuner accepts data from the CIC filter, tunes it to a digital Intermediate Frequency (IF), and passes the result to a shared summation block. The tuner consists of a 32-bit quadrature NCO and a Quadrature Amplitude Mixer (QAM). The NCO serves as a local oscillator and the QAM translates the interpolated channel data from baseband to the NCO frequency. The worst case spurious signal from the NCO is better than –100 dBc for all output frequencies. The tuner can produce real or complex outputs as requested by the shared summation block. –10 –30 –50 dB –110 dB NUMERICALLY CONTROLLED OSCILLATOR/TUNER (NCO) 10 –70 –90 –110 In the complex mode, the NCO serves as a quadrature local oscillator running at fCLK/2 capable of producing any frequency step between –fCLK/4 and +fCLK/4 with a resolution of fCLK/233 (0.0121 Hz for fCLK = 104 MHz). –130 –150 –3 Table XIV. Maximum Bandwidth of Rejection for LCIC2 Values –2 –1 0 1 2 3 Figure 30. rCIC2 –110 dBc Table XIV lists maximum bandwidth that will be rejected to various levels for CIC2 interpolation factors from 1 to 32. The example above corresponds to the listing in the –110 dB column and the LCIC2 = 5 row. The rejection of the CIC2 improves as the interpolation factor increases. In the real mode, the NCO serves as a quadrature local oscillator running at fCLK capable of producing any frequency step between –fCLK/2 and +fCLK/2 with a resolution of fCLK/232 (0.0242 Hz for fCLK = 104 MHz). The quadrature portion of the output is discarded. Negative frequencies are distinguished from positive –24– REV. 0 AD6623 PHASE OFFSET 16 MICROPROCESSOR INTERFACE 16 NCO FREQUENCY WORD 32 D Q 32 I DATA FROM CIC5 32 D Q 32 32 ON OFF 32 ANGLE TO CARTESIAN CONVERSION PN GEN. I I, Q Q PN GEN. ON OFF Q DATA FROM CIC5 CLK Figure 31. Numerically Controlled Oscillator and QAM Mixer frequencies solely by spectral inversion. The digital IF is calculated using the equation: f IF = f NCO × NCO _ frequency 232 (21) where: NCO_frequency is the value written to 0xn02, fIF is the desired intermediate frequency, and fNCO is fCLK/2 for complex outputs and fCLK for real outputs. Phase Dither The AD6623 provides a phase dither option for improving the spurious performance of the NCO. Phase dither is enabled by writing a “1” to Bit 3 of Channel Register 0xn01. When phase dither is enabled, spurs due to phase truncation in the NCO are randomized. The choice of whether phase dither is used in a system will ultimately be decided by the system goals and the choice of IF frequency. The 18 most significant bits of the phase accumulator are used by the angle to Cartesian conversion. If the NCO frequency has all zeroes below the 18th bit, then phase dither has no effect. If the fraction below the 18th bit is near a 1/2 or 1/3 of the 18th bit, then spurs will accumulate separated from the IF by 1/2 or 1/3 of the CLK frequency. The smaller the denominator of this residual fraction, the larger the spurs due to phase truncation will be. If the phase truncation spurs are unacceptably high for a given frequency, then the phase dither can reduce these at the penalty of a slight elevation in total error energy. If the phase truncation spurs are small, then phase dither will not be effective in reducing them further, but a slight elevation in total error energy will occur. This register allows multiple NCOs to be synchronized to produce sine waves with a known phase relationship. NCO Frequency Update and Phase Offset Update Hold-Off Counters The update of both the NCO frequency and phase offset can be synchronized with internal Hold-Off counters. Both of these counters are 16-bit unsigned integers and are clocked at the master CLK rate. These Hold-Off counters used in conjunction with the frequency or phase offset registers, allow beam forming and frequency hopping. See the Synchronization section of the data sheet for additional details. The NCO phase can also be cleared on Sync (set to 0x0000) by setting Bit 2 of Channel Register 0xn01 high. NCO Control Scale The output of the NCO can be scaled in four steps of 6 dB each via Channel Register 0xn01, Bits 1–0. Table XV show a breakdown of the NCO Control Scale. The NCO always has loss to accommodate the possibility that both the I and Q inputs may reach full-scale simultaneously, resulting in a 3 dB input magnitude. Table XV. NCO Control Scale 0xn01 Bit 1 0xn01 Bit 0 NCO Output Level 0 0 1 1 0 1 0 1 –6 dB (no attenuation) –12 dB attenuation –18 dB attenuation –24 dB attenuation Amplitude Dither SUMMATION BLOCK Amplitude dither can also be used to improve spurious performance of the NCO. Amplitude dither is enabled by writing a “1” to Bit 4 of Channel Register at 0xn01. When enabled, amplitude dither can reduce spurs due to truncation at the input to the QAM. If the entire frequency word is close to a fraction that has a small denominator, the spurs due to amplitude truncation will be large and amplitude dither will spread these spurs effectively. Amplitude dither also will increase the total error energy by approximately 3 dB. For this reason amplitude dither should be used judiciously. The Summation Block of the AD6623 serves to combine the outputs of each channel to create a composite multicarrier signal. The four channels are summed together and the result is then added with the 18-bit Wideband Input Bus (IN[17:0]). The final summation is then driven on the 18-bit Wideband Output Bus (OUT[17:0]) on the rising edge of the high speed clock. If the OEN input is high then this output bus is three-stated. If the OEN input is low then this bus will be driven by the summed data. The OEN is active high to allow the Wideband Output Bus to be connected to other busses without using extra logic. Most other busses (like 374 type registers) require a low output enable, which is opposite of the AD6623 OEN, thus eliminating extra circuitry. Phase Offset The phase offset (Channel Register 0xn04) adds an offset to the phase accumulator of the NCO. This is a 16-bit register that is interpreted as a 16-bit unsigned integer. Phase offset ranges from 0 to nearly 2 radians with a resolution of /32768 radians. REV. 0 The wideband parallel input IN[17:0] is defined as bidirectional, to support dual parallel outputs. Each parallel output produces the sum of two of the four internal TSPs and AD6623 that can –25– AD6623 drive two DACs. Channels are added in pairs (A + B), (C + D) as shown in Figure 32. CHANNELS A+B OUT [17:0] 14-BIT DAC AD6623 CHANNELS IN/OUT C+D [17:0] 14-BIT DAC Figure 32. AD6623 Driving Two DACs The Wideband Output Bus may be interpreted as a two’s complement number or as an offset binary number as defined by bit 1 of the Summation Mode Control Register at address 0x000. When this bit is high, then the Wideband Output is in two’s complement mode and when it is low it is configured for offset binary output data. The MSB (bit 17) of the Wideband Output Bus is typically used as a guard bit for the purpose of clipping the wideband output bus when bit 0 of the Summation Mode Control Register at address 0x000 is high. If clip detection is enabled then bit 17 of the output bus is not used as a data bit. Instead, bit 16 will become the MSB and is connected to the MSB of the DAC. Configuring the DAC in this manner gives the summation block a gain of 0 dB. When clip detection is not enabled and bit 17 is used as a data bit then the summation block will have a gain of –6.02 dB. to output complex data on the Wideband Output Bus. The I data samples would be identified when QOUT is low and the Q data samples would be identified when QOUT is high. The second method of obtaining complex data is to provide a QIN signal that toggles on every rising edge of the CLK. This could be obtained by connecting the QOUT of another AD6623 to QIN as shown in Figure 33. In a cascaded system the QIN of the first AD6623 in the chain would typically be tied high and the QOUT of the first AD6623 would be connected to the QIN of the following part. All AD6623s will synchronize themselves to the QIN input so that the proper samples are always paired and the Wideband Output bus represents valid complex data samples. Table XV shows different parallel input and output data bus formats as a function of QIN and QOUT. Table XV. Valid Output Bus Data Modes QIN Wideband Input IN[17:0] Output Data Type OUT[17:0], QOUT Low High Pulsed Real Zero Complex Real Complex Complex There are two data output modes. The first is offset binary. This mode is used only when driving offset binary DACs. Two’s complement mode may be used in one of two circumstances. The first is when driving a DAC that accepts two’s complement data. The second is when driving another AD6623 in cascade mode. When clipping is enabled, the two’s complement mode output bus will clip to 0x2FFFF for output signals more positive than the output can express and it will clip to 0x3000 for signals more negative than the output can express. In offset binary mode the output bus will clip to 0x3FFFF for output signals more positive than the output can express and it will clip to 0x2000 for signals more negative than the output can express. The Wideband Input is always interpreted as an 18-bit two’s complement number and is typically connected to the Wideband Output Bus of another AD6623 in order to send more than four carriers to a single DAC. The Output Bus of the preceeding AD6623 should be configured in two’s complement mode and clip detection disabled. The 18-bit resolution insures that the noise and spur performance of the wideband data stream does not become the limiting factor as large numbers of carriers are summed. There is a two-clock cycle latency from the Wideband Input Bus to the Wideband Output Bus. This latency may be calibrated out of the system by use of the Start Hold-Off counter. The preceding AD6623 in a cascaded chain can be started two CLK cycles before the following AD6623 is started and the data from each AD6623 will arrive at the DAC on the same clock cycle. In systems where the individual signals are not correlated, this is usually not necessary. The AD6623 is capable of outputting both real and complex data. When in Real mode the QIN input is tied low signaling that all inputs on the Wideband Input Bus are real and that all outputs on the Wideband Output Bus are real. The Wideband Input Bus will be pulled low and no data will be added to the composite signal if this port is unused (not connected). If complex data is desired there are two ways this can be obtained. The first method is to simply set the QIN input of the AD6623 high and to set the Wideband Input Bus low. This allows the AD6623 LOGIC1 QOUT QIN AD6623 LOGIC0 IN [17:0] OUT [17:0] QIN AD6623 IN [17:0] OUT [16:3] 14-BIT DAC Figure 33. Cascade Operation of Two AD6623s SYNCHRONIZATION Three types of synchronization can be achieved with the AD6623. These are Start, Hop, and Beam. Each is described in detail below. The synchronization is accomplished with the use of a shadow register and a Hold-Off counter. See Figure 34 for a simplistic schematic of the NCO shadow register and NCO Frequency Hold-Off counter to understand basic operation. Enabling the clock (AD6623 CLK) for the Hold-Off counter can occur with either a Soft_Sync (via the micro port), or a Pin Sync (via the AD6623 Sync pin, Pin 62). The functions that include shadow registers to allow synchronization include: 1. Start 2. Hop (NCO Frequency) 3. Beam (NCO Phase Offset) Start Refers to the start-up of an individual channel, chip, or multiple chips. If a channel is not used, it should be put in the Sleep Mode to reduce power dissipation. Following a hard reset (low pulse on the AD6623 RESET pin), all channels are placed in the Sleep Mode. Start With No Sync If no synchronization is needed to start multiple channels or multiple AD6623s, the following method should be used to initialize the device. 1. To program a channel, it must first be set to the Program Mode (bit high) and Sleep Mode (bit high) (Ext Address 4). The Program Mode allows programming of data memory and coefficient memory (all other registers are programmable whether in Program Mode or not). Since no synchronization is used, all –26– REV. 0 AD6623 Sync bits are set low (External Address 5). All appropriate control and memory registers (filter) are then loaded. The Start Update Hold-Off Counter (0xn00) should be set to 0. 2. Set the appropriate program and sleep bits low (Ext Address 4). This enables the channel. The channel must have Program and Sleep Mode low to activate a channel. NCO REGISTER NCO SHADOW REGISTER 32 32 D Q D Q NCO PHASE ACCUMULATOR 32 ENA HOLDOFF COUNTER HOP HOLDOFF MICROPROCESSOR INTERFACE 16 D Q 16 D Q A jump from one NCO frequency to a new NCO frequency. This change in frequency can be synchronized via microprocessor control or an external Sync signal as described below. To set the NCO frequency without synchronization the following method should be used. START COUNTER START HOLDOFF 16 Hop D C=1 PL C=0 ENA HOP SYNC 16 D C=1 PL C=0 CLR Set Frequency No Hop CLR Q SLEEP SET ENA START SYNC 1. Set the NCO Frequency Hold-Off counter to 0. 2. Load the appropriate NCO frequency. The new frequency will be immediately loaded to the NCO. Hop with Soft Sync CLK RESET PIN Figure 34. NCO Shadow Register and Hold-Off Counter Start with Soft Sync The AD6623 includes the ability to synchronize channels or chips under microprocessor control. One action to synchronize is the start of channels or chips. The Start Update Hold-Off Counter (0xn00) in conjunction with the Start bit and Sync bit (Ext Address 5) allow this synchronization. Basically the Start Update Hold-Off Counter delays the Start of a channel(s) by its value (number of AD6623 CLKs). The following method is used to synchronize the start of multiple channels via microprocessor control. 1. Set the appropriate channels to sleep mode (a hard reset to the AD6623 Reset pin brings all four channels up in sleep mode). 2. Write the Start Update Hold-Off Counter(s) (0xn00) to the appropriate value (greater than 1 and less than 216–1). If the chip(s) is not initialized, all other registers should be loaded at this step. 3. Write the Start bit and the SyncX(s) bit high (Ext Address 5). 4. This starts the Start Update Hold-Off Counter counting down. The counter is clocked with the AD6623 CLK signal. When it reaches a count of one the sleep bit of the appropriate channel(s) is set low to activate the channel(s). Start with Pin Sync A Sync pin is provided on the AD6623 to provide the most accurate synchronization, especially between multiple AD6623s. Synchronization of start with an external signal is accomplished with the following method. REV. 0 1. Set the appropriate channels to sleep mode (a hard reset to the AD6623 Reset pin brings all four channels up in sleep mode). 2. Write the Start Update Hold-Off Counter(s) (0xn00) to the appropriate value (greater than 1 and less than 216–1). If the chip(s) is not initialized, all other registers should be loaded at this step. 3. Set the Start on Pin Sync bit and the appropriate Sync Pin Enable high (0xn01). 4. When the Sync pin is sampled high by the AD6623 CLK this enables the count down of the Start Update Hold-Off Counter. The counter is clocked with the AD6623 CLK signal. When it reaches a count of one the sleep bit of the appropriate channel(s) is set low to activate the channel(s). The AD6623 includes the ability to synchronize a change in NCO frequency of multiple channels or chips under microprocessor control. The NCO Frequency Hold-Off counter (0xn03) in conjunction with the Hop bit and the Sync bit (Ext Address 5) allow this synchronization. Basically the NCO Frequency Hold-Off counter delays the new frequency from being loaded into the NCO by its value (number of AD6623 CLKs). The following method is used to synchronize a hop in frequency of multiple channels via microprocessor control. 1. Write the NCO Frequency Hold-Off (0xn03) counter to the appropriate value (greater than 1 and less then 216–1). 2. Write the NCO Frequency register(s) to the new desired frequency. 3. Write the hop bit and the Sync(s) bit high (Ext Address 5). 4. This starts the NCO Frequency Hold-Off counter counting down. The counter is clocked with the AD6623 CLK signal. When it reaches a count of one the new frequency is loaded into the NCO. Hop with Pin Sync A Sync pin is provided on the AD6623 to provide the most accurate synchronization, especially between multiple AD6623s. Synchronization of hopping to a new NCO frequency with an external signal is accomplished with the following method. 1. Write the NCO Frequency Hold-Off counter(s) (0xn03) to the appropriate value (greater than 1 and less than 216–1). 2. Write the NCO Frequency register(s) to the new desired frequency. 3. Set the Hop on Pin Sync bit and the appropriate Sync Pin Enable high (0xn01). 4. When the Sync pin is sampled high by the AD6623 CLK this enables the count down of the NCO Frequency Hold-Off counter. The counter is clocked with the AD6623 CLK signal. When it reaches a count of one the new frequency is loaded into the NCO. –27– AD6623 Beam Table XVII. Test Access Port Pins A change in phase for a particular channel and can be synchronized with respect to other channels or AD6623s. This change in phase can be synchronized via microprocessor control or an external Sync signal. To set the amplitude without synchronization the following method should be used. Set Phase No Beam Name Pin Number Description TRST TCK TMS TDI TDO 100 101 106 108 107 Test Access Port Reset Test Clock Test Access Port Mode Select Test Data Input Test Data Output 1. Set the NCO Phase Offset Update Hold-Off Counter (0xn05) to 0. 2. Load the appropriate NCO Phase Offset (0xn04). The NCO Phase Offset will be immediately loaded. Note that TCK and TDI are internally pulled down which is opposite of IEEE Standard 1149.1. These pins may be connected to external pull-up resistors, with the associated additional current draw through the pull-ups, or left unconnected. Beam with Soft Sync The AD6623 supports four op codes are shown in Table XVIII. These instructions set the mode of the JTAG interface. The AD6623 includes the ability to synchronize a change in NCO phase of multiple channels or chips under microprocessor control. The NCO Phase Offset Update Hold-Off Counter in conjunction with the Beam bit and the Sync bit (Ext Address 5) allow this synchronization. Basically the NCO Phase Offset Update HoldOff Counter delays the new phase from being loaded into the NCO/RCF by its value (number of AD6623 CLKs). The following method is used to synchronize a beam in phase of multiple channels via microprocessor control. 1. Write the NCO Phase Offset Update Hold-Off Counter (0xn05) to the appropriate value (greater than 1 and less then 216–1). 2. Write the NCO Phase Offset register(s) to the new desired phase and amplitude. 3. Write the beam bit and the Sync(s) bit high (Ext Address 5). 4. This starts the NCO Phase Offset Update Hold-Off Counter counting down. The counter is clocked with the AD6623 CLK signal. When it reaches a count of one the new phase is loaded into the NCO. Beam with Pin Sync A Sync pin is provided on the AD6623 to provide the most accurate synchronization, especially between multiple AD6623s. Synchronization of beaming to a new NCO Phase Offset with an external signal is accomplished using the following method. 1. Write the NCO Phase Offset Hold-Off (0xn05) counter(s) to the appropriate value (greater than 1 and less than 216–1). 2. Write the NCO Phase Offset register(s) to the new desired phase and amplitude. 3. Set the Beam on Pin Sync bit and the appropriate Sync Pin Enable high (0xn01). 4. When the Sync pin is sampled high by the AD6623 CLK this enables the count down of the NCO Phase Offset Hold-Off counter. The counter is clocked with the AD6623 CLK signal. When it reaches a count of one the new phase is loaded into the NCO registers. JTAG INTERFACE The AD6623 supports a subset of IEEE Standard 1149.1 specification. For additional details of the standard, please see IEEE Standard Test Access Port and Boundary-Scan Architecture, IEEE-1149 publication from IEEE. The AD6623 has five pins associated with the JTAG interface. These pins are used to access the on-chip Test Access Port and are listed in Table XVII. Table XVIII. Op Codes Instruction Op Code IDCODE BYPASS SAMPLE/PRELOAD EXTEST 10 11 01 00 The Vendor Identification Code (Table XIX) can be accessed through the IDCODE instruction and has the following format. Table XIX. Vendor Identification Code MSB Part Version Number 0000 Manufacturer LSB ID Number Mandatory 0010 0111 1000 0000 000 1110 0101 1 A BSDL file for this device is available from Analog Devices, Inc. Contact Analog Devices for more information. SCALING Proper scaling of the wideband output is critical to maximize the spurious and noise performance of the AD6623. A relatively small overflow anywhere in the data path can cause the spurious free dynamic range to drop precipitously. Scaling down the output levels also reduces dynamic range relative to an approximately constant noise floor. A well-balanced scaling plan at each point in the signal path will be rewarded with optimum performance. The scaling plan can be separated into two parts: multicarrier scaling and single-carrier scaling. Multicarrier Scaling An arbitrary number of AD6623s can be cascaded to create a composite digital IF with many carriers. As the number of carriers increases, the peak to RMS ratio of the composite digital IF will increase as well. It is possible and beneficial to limit the peak to RMS ratio through careful frequency planning and controlled phase offsets. Nevertheless, in most cases with a large number of carriers, the worst-case peak is an unlikely event. The AD6623 immediately preceding the DAC can be programmed to clip rather than wrap around (see the Summation Block description). For a large number of carriers, a rare but finite chance of clipping at the AD6623 wideband output will result in superior dynamic range compared to lowering each carrier level until clipping is impossible. This will also be the case for most DACs. Through analysis or experimentation, an optimal output level of individual carriers can be determined for any particular DAC. –28– REV. 0 AD6623 Single-Carrier Scaling Once the optimal power level is determined for each carrier, one must determine the best way to achieve that level. The maximum SNR can be achieved by maximizing the intermediate power level at each processing stage. This can be done by assuming the proper level at the output and working along the following path: Summation, NCO, CIC, Ramp, RCF, and finally, Fine Scaler Unit. The Summation Block is intended to combine multiple carriers with each carrier at least 6 dB below full scale. For this configuration, the AD6623 driving the DAC should have clip detection enabled. OUT17 becomes a clip indicator that reports clipping in both polarities. If the DAC requires offset binary outputs, then the internal offset binary conversion should be enabled as well. Any preceding cascaded AD6623s should disable clip detection and offset binary conversion. The IN17–IN0 of the first AD6623 in the cascade should be grounded. See the Summation Block section for details. In this configuration, intermediate OUT17s will serve as guard bits that allow intermediate sums to exceed full scale. As long as the final output does not exceed 6 dB over full scale, the clip detector will perform correctly. If a single carrier needs to exceed –6 dB full scale, hardwired scaling can be accomplished according to Table XX. This is most useful when the AD6623 is processing a Single Wideband Carrier such as UMTS or CDMA 2000. Table XX. Hardwired Scaling Max. Single Carrier Level Connect to DAC MSB Clip Detect Offset Binary Compensation –12.04 dB –6.02 dB 0 dB +6.02 dB OUT17 OUT16 OUT15 OUT14 N/A ± + only + only Internal Internal 0x08000 0x0C000 The NCO/Tuner is equipped with an output scaler that ranges from –6.02 dB to –24.08 dB below full scale, in 6.02 dB steps. See the NCO/Tuner section for details. The best SNR will be achieved by maximizing the input level to the NCO and using the largest possible NCO attenuation. For example, to achieve an output level –20 dB below full scale, one should set the CIC output level to –1.94 dB below full scale and attenuate by –18.06 dB in the NCO. The CIC is equipped with an output scaler that ranges from 0 dB to –186.64 dB below full scale in 6.02 dB steps. This large attenuation is necessary to compensate for the potentially large gains associated with CIC interpolation. See the CIC section for details. For example to achieve an output level of –1.94 dB below full scale, with a CIC5 interpolation of 27 (114.51 dB gain) and a CIC2 interpolation of 3 (9.54 dB gain), one should set the CIC_Scale to 20 and the Fine Scale Unit output level to –5.59 dB below full scale. (22) –1.94 – 9.54 – 114.51 + 20 × 6.02 = –5.59 The ramp unit when bypassed will have exactly 0 dB of gain and can be ignored. When in use, the gain is dependant on what value is stored in the last valid RMEM location. RMEM words are 14 bits [0–1), so when the value is positive full scale, the gain is about –0.0005 dB; probably neglectable. The RCF coefficients should be normalized to positive full scale. This will yield the greatest dynamic range. The RCF is equipped with an output scaler that ranges from 0 dB to –18.06 dB below full scale in 6.02 dB steps. This attenuation can be used to partially compensate for filter gain in the RCF. For example, if the maximum REV. 0 gain of the RCF coefficients is 11.26 dB, the RCF coarse scale should be set to 2 (12.04 dB). This yields an RCF output level and fine scale input level of –0.78 dB (23) 11.26 – 12.04 = –0.78 The fine scale unit is left to turn a –0.78 dB level into a –5.59 dB level. This requires a gain of –4.81 dB, which corresponds to a 14-bit [0–2] scale value of 1264h. All subsequent rescalings during chip operation should be relative to this maximum. –5.59 – 0.78 = –4.81 (24) –4.81 floor 10 20 × 213 = 1264h (25) Finally, as described in the RCF section, there may be a worst-case peak of a phase that is larger than the channel center gain. In the preceding example, if the worst case to channel center ratio is larger than 4.59 dB (potentially overflowing the RCF), then the RCF_ Coarse_Scale should be reduced by one and the CIC_Scale should be increased by one. In the preceding example, if the worst case to channel center ratio is larger than 5.59 dB (potentially overflowing the RCF and CIC), then the RCF_Coarse_Scale should be reduced by one and the NCO_Output_Scale should be increased by one. MICROPORT INTERFACE The MicroPort interface is the communications port between the AD6623 and the host controller. There are two modes of bus operation: Intel nonmultiplexed mode (INM), and Motorola nonmultiplexed mode (MNM) that is set by hard-wiring the MODE pin to either ground or supply. The mode is selected based on the use of the MicroPort control lines (DS or RD, DTACK or RDY, RW or WR) and the capabilities of the host processor. See the timing diagrams for details on the operation of both modes. The External Memory Map provides data and address registers to read and write the extensive control registers in the Internal Memory Map. The control registers access global chip functions and multiple control functions for each independent channel. MicroPort Control All accesses to the internal registers and memory of the AD6623 are accomplished indirectly through the use of the microprocessor port external registers shown in Table XXI. Accesses to the External Registers are accomplished through the 3 bit address bus (A[2:0]) and the 8-bit data bus (D[7:0]) of the AD6623 (MicroPort). External Address [3:0] provides access to data read from or written to the internal memory (up to 32 bits). External Address [0] is the least significant byte and External Address [3] is the most significant byte. External Address [4] controls the Sleep Mode of each channel. External Address [5] controls the sync status of each channel. External Address [7:6] determines the Internal Address selected and whether this address is incremented after subsequent reads and/or writes to the internal registers. EXTERNAL MEMORY MAP The External Memory Map is used to gain access to the Internal Memory Map described below. External Address [7:6] sets the Internal Address to which subsequent reads or writes will be performed. The top two bits of External Address [7] allow the user to set the address to auto increment after reads, writes, or both. All internal data words have widths that are less than or equal to 32 bits. Accesses to External Address [0] also triggers access to the AD6623’s internal memory map. Thus during writes to the –29– AD6623 Table XXI. External Registers External Address D7 D6 D5 External Data D4 D3 D2 D1 D0 7:UAR 6:LAR 5:Sync 4:Sleep 3:Byte3 2:Byte2 1:Byte1 0:Byte0 Wrinc IA7 – Prog D ID31 ID23 ID15 ID7 Rdinc IA6 Beam Prog C ID30 ID22 ID14 ID6 – IA5 Hop Prog B ID29 ID21 ID13 ID5 – IA4 Start Prog A ID28 ID20 ID12 ID4 IAIO IA2 Sync C Sleep C ID26 ID18 ID10 ID2 IA9 IA1 Sync B Sleep B ID25 ID17 ID9 ID1 IA8 IA0 Sync A Sleep A ID24 ID16 ID8 ID0 internal registers, External Address [0] must be written last to insure all data is transferred. Reads are the opposite in that External Address [0] must be the first data register read (after setting the appropriate internal address) to initiate an internal access. External Address [5:4] reads and writes are transferred immediately to internal control registers. External Address [4] is the sleep register. The sleep bits can be set collectively by the address. The sleep bits can be cleared by operation of start syncs (described below). External Address [5] is the sync register. These bits are write only. There are three types of syncs: start, hop, and beam. Each of these can be sent to any or all of the four channels. For example, a write of X0010100 would issue a start sync to channel C only. A write of X1101111 would issue a beam sync and a hop sync to all channels. The internal address bus is 12 bits wide and the internal data bus is 32 bits wide. External address 7 is the UAR (Upper Address Register) and stores the upper four bits of the address space in UAR[3:0]. UAR[7:6] define the auto-increment feature. If Bit 6 is high, the internal address is incremented after an internal read. If Bit 7 is high, the internal address is incremented after an internal write. If both bits are high, the internal address in incremented after either a write or a read. This feature is designed for sequential access to internal locations. External address 6 is the LAR (Lower Address Register) and stores the lower 8 bits of the internal address. External addresses 3 through 0 store the 32 bits of the internal data. All internal accesses are two clock cycles long. Writing to an internal location with a data width of 16 bits is achieved by first writing the upper four bits of the address to bits 3 through 0 of the UAR (bits 7 and 6 of the UAR are written to determine whether or not the auto increment feature is enabled). The LAR is then written with the lower eight bits of the internal address (it doesn’t matter if the LAR is written before the UAR as long as both are written before the internal access). Since the data width of the internal address is 16 bits, only data register 1 and data register 0 are needed. Data register 1 must be written first because the write to data register 0 triggers the internal access. Data register 0 must always be the last register written to initiate the internal write. Reading from the MicroPort is accomplished in a similar manner. The internal address is first written. A read from data register 0 activates the internal read, thus register 0 must always be read first to initiate an internal read. This provides the 8 LSBs of the internal read through the MicroPort (D[7:0]). Additional bytes are then read by changing the external address (A[2:0]) and performing additional reads. If data register 3 (or any other) is read before data register 0, incorrect data will be read. Data register 0 must be read first in order to transfer data from the core memory to the external memory locations. Once the data register is read, the remaining locations may be examined in any order. IAII IA3 Sync D Sleep D ID27 ID19 ID11 ID3 Access to the external registers of Table XX is accomplished in one of two modes using the CS, DS(RD), RW(WR), and DTACK(RDY) inputs. The access modes are Intel Nonmultiplexed mode and Motorola Nonmultiplexed mode. These modes are controlled by the MODE input (MODE = 0 for INM, MODE = 1 for MNM). Intel Nonmultiplexed Mode (INM) MODE must be tied low to operate the AD6623 MicroPort in INM mode. The access type is controlled by the user with the chip select (CS), read (RD), and write (WR) inputs. The ready (RDY) signal is produced by the MicroPort to communicate to the user the MicroPort is ready for an access. RDY goes low at the start of the access and is released when the internal cycle is complete. See the timing diagrams for both the read and write modes in the Specifications. Motorola Nonmultiplexed Mode (MNM) MODE must be tied high to operate the AD6623 MicroPort in MNM mode. The access type is controlled by the user with the chip select (CS), data strobe (DS), and read/write (RW) inputs. The data acknowledge (DTACK) signal is produced by the MicroPort to acknowledge the completion of an access to the user. DTACK goes low when an internal access is complete and then will return high after DS is deasserted. See the timing diagrams for both the read and write modes in the Specifications. The DTACK(RDY) pin is configured as an open drain so that multiple devices may be tied together at the microprocessor/ microcontroller without contention. The MicroPort of the AD6623 allows for multiple accesses while CS is held low (CS can be tied permanently low if the MicroPort is not shared with additional devices). The user can access multiple locations by pulsing the RW(WR) or DS(RD) lines and changing the contents of the external three bit address bus (A[2:0]). External Address 7 Upper Address Register (UAR) Sets the four most significant bits of the internal address, effectively selecting channels 1, 2, 3, or 4 (D2:D0). The autoincrement of read and write are also set (D7:D6). External Address 6 Lower Address Register (LAR) Sets the internal address 8 LSBs (D7:D0). External Address 5 Sync This register is write only. Bits in this address control the synchronization of the AD6623 channels. If the user intends to bring up channels with no synchronization requirements then all bits of this register should be written low. Two types of sync signals are available with the AD6623. The first is Soft Sync. Soft Sync is software synchronization enabled through the MicroPort. The second synchronization method is Pin Sync. Pin Sync is enabled by a signal –30– REV. 0 AD6623 applied to the Sync pin (Pin 62). See the Synchronization section for detailed explanations of the different modes. INTERNAL COUNTER REGISTERS AND ON-CHIP RAM AD6623 and AD6622 Compatibility External Address 4 Sleep The AD6623 functionality is a superset of the AD6622 functionality. The AD6623 is pin-compatible with AD6622. Bits in this register determine how the chip is programmed and enables the channels. The program bits (D7:D4) must be set high to allow programming of CMEM and DMEM for each channel. Sleep bits (D3:D0) are used to activate or sleep channels. These can be used manually by the user to bring up a channel by simply writing the required channel high. These bits can also be used in conjunction with the Start and Sync signals available in External Address 5 to synchronize the channels. See the Synchronization section for a detailed explanation of different modes. AD6622 compatibility is selected when bit 7 of Internal Control Register 0x000 is low. In this state, all AD6623 extended control registers are cleared. While in the AD6622 mode the unused AD6623 pins are three-stated. Listed below is the mapping of internal AD6623 registers. AD6622 compatibility is selected by setting 0x000:7 low. In this state, all AD6623 extended control registers are cleared. Registers marked as “Reserved” must be written low. External Address 3:0 (Data Bytes) These registers return or accept the data to be accessed for a read or write to internal addresses Common Function Registers (not associated with a particular channel) Internal Address 0x000 0x001 0x002 0x003 Bit 7 6–5 4 3 2 1 0 7 6 5 4 3 2 1 0 23–0 15–0 AD6622 Compatible Description 1 AD6623 Extensions Description AD6623 Extension = 11 No Change Wideband Input Disable3 Dual Output Enable3 No Change No Change No Change No Change No Change No Change No Change No Change No Change No Change No Change BIST Counter1, 3 BIST Value (read only) AD6623 Extension = 0 Reserved Reserved Reserved Reserved Offset Binary Outputs1 Clip Wideband I/O3 First Sync Only3 Beam on Pin Sync3 Hop on Pin Sync3 Start on Pin Sync3 Ch. D Sync0 Pin Enable3 Ch. C Sync0 Pin Enable3 Ch. B Sync0 Pin Enable3 Ch. A Sync0 Pin Enable3 Unused Unused Channel Function Registers (0x1XX = Ch. A, 0x2XX = Ch. B, 0x3XX = Ch. C, 0x4XX = Ch. D) Internal Address Bit AD6622 Compatible Description AD6623 Extensions Description 0x100 17–16 Unused 15–0 7–5 4 3 2 1–0 Ch. A Start Hold–Off Counter3 Reserved Ch. A NCO Amplitude Dither Enable Ch. A NCO Phase Dither Enable Ch. A NCO Clear Phase Accumulator on Sync Ch. A NCO Scale 00: –6 dB 01: –12 dB 10: –18 dB 11: –24 dB Ch. A NCO Frequency Value3 Unused Ch. A Start Sync Select3 00: Sync0 (See 0x001) 01: Sync1 10: Sync2 11: Sync3 No Change No Change No Change No Change No Change No Change No Change No Change No Change No Change No Change Ch. A Hop Sync Select3 00: Sync0 (See 0x001 Hop) 01: Sync1 10: Sync2 11: Sync3 0x101 0x102 0x103 REV. 0 31–0 17–16 –31– AD6623 Channel Function Registers (continued) Internal Address 0x104 0x105 0x106 0x107 0x108 0x109 0x10A 0x10B 0x10C Bit AD6622 Compatible Description AD6623 Extensions Description 3 15–0 15–0 17–16 Ch. A NCO Frequency Update Hold–Off Counter Ch. A NCO Phase Offset2 Reserved 15–0 7–5 4–0 8–0 11–8 7–0 7–0 15–8 7 6–0 7 6–0 15–10 9 Ch. A NCO Phase Offset Update Hold–off Counter3 Reserved Ch. A CIC Scale, SCIC Reserved Reserved Ch. A C1C2 Interpolation, L2 –1 Ch. A C1C5 Interpolation, L5 –1 Reserved Reserved Ch. A RCF TapsA, NRCF – 1 (7 bits) 1 Reserved Ch. A RCF Coefficient Offset, ORCF (7 bits) 1 Unused Unused 8 7 Unused Ch. A PRBS Length3 0: 15 1: 8,388,607 Ch. A RCF PRBS Enable Ch. A RCF Mode Select (1 of 2)3 Ch. A RCF Mode Select (2 of 2)3 00: FIR 01: FIR 10: QPSK 11: MSK 6 5 4 0x10D 3–0 7–6 0x10E 5 4–0 15 Ch. A RCF (Taps per Phase3) –1 Ch. A RCF Coarse Scale (a) 00: 0 dB 01: –6 dB 10: –12 dB 11: –18 dB Ch. A RCF Phase EQ Enable Ch. A Serial Clock Divisor (2, 4, …64) Ch. A Fine Scale Factor Enable 0x10F 14–2 1–0 17–16 Ch. A RCF Unsigned Scale Factor 2 Reserved Unused 15–0 Ch. A RCF Scale Hold–Off Counter1 –32– No Change No Change2 Ch. A Phase Sync Select3 00: Sync0 (See 0x001 Beam) 01: Sync1 10: Sync2 11: Sync3 No Change No Change No Change Ch. A CIC2 Decimation, M2 –1 Ch. A CCI2 Interpolation, L2 –1, extended No Change No Change Ch. A RCF TapsB, NRCF – 1 (8 bits)3 Ch. A RCF TapsA, NRCF – 1 (new MSB)2 No Change2 Ch. A RCF Coef Offset, ORCF (new MSB)2 No Change2 Reserved Ch. A Compact FIR Input Word Length 0: 16 bits—8 I followed by 8 Q 1: 24 bits—12 I followed by 12 Q Ch. A RCF PRBS Enable Ch. A RCF PRBS Length3 0: 15 1: 8,388,607 Ch. A RCF Mode Select (1 of 3)2 Ch. A RCF Mode Select (2 of 3)2 Ch. A RCF Mode Select (3 of 3)2 000: FIR 001: /4–DQPSK 010: GMSK 011: MSK 100: FIR, Compact Input Resolution 101: 8–PSK 110: 3π/8–8PSK 111: QPSK No Change2 No Change2 No Change Ch. A Serial Clock Divisor (1, 2,…32) Ch. A Unsigned Scale Factor2 This is extended to allow values in the range (0–2). No Change2 Reserved Ch. A Time Slot Sync Select 00: Sync0 (See 0x001 Time Slot) 01: Sync1 10: Sync2 11: Sync3 The counter is unchanged, but instead of just scale update, when the counter hits one, the following sequence is initiated: 1. Ramp Down (if Ramp is enabled) REV. 0 AD6623 Channel Function Registers (continued) Internal Address 0x110 0x111 0x112 0x113 0x114 0x115 0x116 0x117 0x118 0x119 0x11A–11F 0x120–13F 0x140–17F 0x180–1FF Bit AD6622 Compatible Description 15–0 15–0 15–0 15–0 15–0 15–0 7–6 Ch. A RCF Phase EQ Coef1 Ch. A RCF Phase EQ Coef2 Unused Unused Unused Unused Unused 5 Unused 4 Unused 3 2 Unused Unused 1 0 5–0 4–0 4–0 Unused Unused Unused Unused Unused 15–0 15–14 13–0 15–0 Unused Ch. A Data RAM Unused Unused Ch. A Coefficient RAM AD6623 Extensions Description 2. Update RCF Mode Select registers marked with “2”. 3. Ramp Up (if Ramp is enabled) No Change No Change Ch. A RCF FIR–PSK Magnitude 0 Ch. A RCF FIR–PSK Magnitude 1 Ch. A RCF FIR–PSK Magnitude 2 Ch. A RCF FIR–PSK Magnitude 3 Ch. A Serial Data Frame Input Select 0x: Internal Frame Request 10: External SDFI Pad 11: Previous Channel’s Frame End Ch. A Serial Data Frame Output Select 0: Serial Data Frame Request 1: Serial Data Frame End Ch. A Serial Clock Slave (SCS) SCS = 0: Master Mode (SCLK is an output) SCS = 1: Slave Mode (SCLK is an input) Ch. A Serial Fine Scale Enable1 Ch. A Serial Time Slot Sync Enable (ignored in FIR mode) Ch. A Ramp Interpolation Enable Ch. A Ramp Enable Ch. A Mode 0 Ramp Length, R0–1 Ch. A Mode 1 Ramp Length, R1–1 Ch. A Ramp Rest Time, Q (No inputs requested during rest time.) No Change No Change No Change Ch. A Ramp RAM No Change This address is mirrored at 0x900–0x97F and contiguously extended at 0x980–0x9FF NOTES 1 Clear on RESET. 2 These bits update after a Start or a Beam Sync. See CR 0x10F. 3 Allows dynamic updates. (0x000) Summation Mode Control Controls features in the summation block of the AD6623. Bit 5–6: Reserved. Bit 4: Low: Wideband Input Enabled. High: Wideband Input Disabled. Bit 3: Low: Dual Output Disabled. High: Dual Output Enabled. Bit 2: Reserved. Bit 1: Low: Output data will be in two’s complement. High: Output data will be in offset binary. Bit 0: Low: Over-range will wrap. High: Over-range will clip to full scale. (0x001) Sync Mode Control Bit 7: Bit 6: Bit 5: REV. 0 Ignores all but the first pin sync. Beam on pin Sync. Hop on pin Sync. Bit 4: High enables the count down of the Start Hold-Off Counter. The counter is clocked with the AD6623 CLK signal. When it reaches a count of one the Sleep bit of the appropriate channel(s) is set low to activate the channel(s). Bit 3–0: High enables synchronization of these channels. See the Synchronization section of the data sheet for detailed explanation. (0x002) BIST Counter Sets the length, in CLK cycles, of the built-in self test. (0x003) BIST Result A read-only register containing the result after a self test. Must be compared to a known good result for a given setup to determine pass/fail. –33– AD6623 Channel Function Registers (0xn06) CIC Scale The following registers are channel-specific. ‘0xn’ denotes that these values are represented as hexadecimal numbers; ‘n’ represents the specified channel. Valid channels are n = 1, 2, 3, and 4. Bits 4–0: Sets the CIC scaling per the equation below. (0xn00) Start Update Hold-Off Counter The Start Update Hold-Off Counter is used to synchronize start–up of AD6623 channels and can be used to synchronize multiple chips. The Start Update Hold-Off Counter is clocked by the AD6623 CLK (master clock). (0xn01) NCO Control Set the NCO scaling per Table XXI. Table XXI. NCO Control (0xn01) Bit 2: Bit 1 Bit 0 NCO Output Level 0 0 1 1 0 1 0 1 –6 dB (no attenuation) –12 dB attenuation –18 dB attenuation –24 dB attenuation High enables NCO phase dither. Bit 4: High enables NCO amplitude dither. Bit 7–5: Reserved and should be written low. This register sets the number of RCF Coefficients and is limited to a maximum of 256. The programmed value is the number of RCF Coefficients – 1. There is an A register and a B register at this memory location. Value A is used when the RCF is operating in mode 0 and value B is used when in mode 1. The RCF mode bit of interest here is bit 6 of address 0xn0C. (0xn0B) RCF Coefficient Offset This register sets the offset for RCF Coefficients and is normally set to 0. It can be viewed as a pointer which selects the portion of the CMEM used when computing the RCF filter. This allows multiple filters to be stored in the Coefficient memory space, selecting the appropriate filter by setting the offset. (0xn0C) Channel Mode Control 1 Bit 9: (0xn03) NCO Frequency Update Hold-Off Counter See the Synchronization section for detailed explanation. If no synchronization is required, this register should be set to 0. Bit 17–16: The Hop Sync Select bits are used to set which sync pin will initiate a hop sequence. The Hold-Off Counter is used to synchronize the change of NCO frequencies. (0xn04) NCO Phase Offset Bit 8: Bit 7: Bits 6–4: This register is a 16-bit unsigned integer that is added to the phase accumulator of the NCO. This allows phase synchronization of multiple channels of the AD6623(s). The NCO Phase Offset contains a shadow register for synchronization purposes. The shadow can be read back directly, the NCO Phase Offset cannot. See the Synchronization section for details. (0xn05) NCO Phase Offset Update Hold-Off Counter See the Synchronization section for a detailed explanation. If no synchronization is required, this register should be set to 0. Bit 17–16: The Phase Sync Select bits are used to set which sync pin will initiate a phase sync sequence. The Hold-Off Counter is used to synchronize the change of NCO phases. This register is used to set the interpolation in the CIC2 filter. The value written to this register is the interpolation minus one. The CIC2 interpolation can range from 1 to 4096. LrCIC2 must be chosen equal to or larger than MrCIC2 and both must be chosen such that a suitable CIC2 Scalar can be chosen. For more details the CIC2 section should be consulted. (0xn0A) Number of RCF Coefficients – 1 This register is a 32-bit unsigned integer that sets the NCO Frequency. The NCO Frequency contains a shadow register for synchronization purposes. The shadow can be read back directly, the NCO Frequency cannot. f NCOFREQUENCY = 232 × CHANNEL (26) CLK Bit 15–0: (0xn08) CIC2 Interpolation – 1 (L CIC2 – 1) This register sets the interpolation rate for the CIC5 filter stage (unsigned integer). The programmed value is the CIC5 Interpolation – 1. Maximum interpolation is limited by the CIC scaling available (See the CIC section). (0xn02) NCO Frequency Bit 15–0: This register is used to set the decimation in the CIC2 filter. The value written to this register is the decimation minus one. The CIC2 decimation can range from 1 to 512 depending upon the interpolation of the CIC2. There is no timing error associated with this decimation. See the CIC2 section for further details. (0xn09) CIC5 Interpolation – 1 High clears the NCO phase accumulator to 0 on either a Soft Sync or Pin Sync (see Synchronization for details). Bit 3: (27) (0xn07) CIC2 Decimation – 1 (M CIC2 – 1) Bit 17–16: The Start Sync Select bits are used to set which sync pin will initiate a start sequence. Bit 1:0 )) See the CIC section for details. See the Synchronization section for detailed explanation. If no synchronization is required, this register should be set to 0. Bit 15–0: ( ( 4 CIC _ Scale = ceil × log 2 LCIC 5 × LCIC 2 High, selecting compact FIR mode results in 24-bit serial word length (12 I followed by 12 Q). When low, selecting compact FIR mode results in 16-bit serial word length (8 I followed by 8 Q). High enables RCF Pseudo-Random Input Select. High selects a Pseudo-Random sequence length of 8,388,607. Low selects a Pseudo-Random Sequence length of 15. Sets the channel input format as shown in Table XXII. Table XXII. Channel Inputs Bit 6 Bit 5 Bit 4 Input Mode 0 0 0 0 1 1 1 1 0 0 1 1 0 0 1 1 0 1 0 1 0 1 0 1 FIR /4-DQPSK GSM MSK Compact FIR 8PSK 3/8-8PSK QPSK –34– REV. 0 AD6623 Bit 6 Bits 3–0: Can be set through the serial port (see section on serial word formats). (0xn17) Power Ramp Length 0 Sets (NRCF/LRCF) –1 (0xn18) Power Ramp Length 1 (0xn0D) Channel Mode Control 2 Bits 7–6: Sets the RCF Coarse Scale as shown in Table XXIII. This is the length of the ramp for mode 0, minus one. This is the length of the ramp for mode 1, minus one. Setting this to zero disables dual ramps. (0xn19) Power Ramp Rest Time This is the number of RCF output samples to rest for between a ramp down and a ramp up. Table XXIII. RCF Coarse Scale Bit 7 Bit 6 RCF Coarse Scale (dB) 0 0 1 1 0 1 0 1 0 –6 –12 –18 (0xn1A–0xn1F) Unused (0xn20–0xn3F) Data Memory This group of registers contain the RCF Filter Data. See the RCF section for additional details. (0xn40–0xn17F) Power Ramp Coefficient Memory Bit 5: High enables the RCF phase equalizer. Bits 4–0: Sets the serial clock divider (SDIV) that determines the serial clock frequency based on the following equation. fSCLK CLK = SDIV + 1 (28) (0xn0E) Fine Scale Factor Bits 15–2: Bits 1–0: Sets the RCF Fine Scale Factor as an unsigned number representing the values (0,2). This register is shadowed for synchronization purposes. The shadow can be read back directly, the Fine Scale Factor can not. Reserved. (0xn0F) RCF Time Slot Hold-Off Counter Bits 17–16: The Time Slot Sync Select bits are used to set which sync pin will initiate a time slot sync sequence. Bits 15–0: The Hold-Off Counter is used to synchronize the change of RCF Fine Scale. See the Synchronization section for a detailed explanation. If no synchronization is required, this register should be set to 0. (0xn10–0xn11) RCF Phase Equalizer Coefficients See the RCF section for details. (0xn12–0xn15) FIR-PSK Magnitudes See the RCF section for details. (0xn16) Serial Port Setup Bits 7–6: Serial Data Frame Start Select Title XXIV. Serial Port Setup Bit 7 Bit 6 Serial Data Frame Start 0 1 1 X 0 1 Internal Frame Request External SDFI Pad Previous Channel’s Frame End Bit 5: High means SDFO is a frame end, low means SDFO is a frame request. Bit 4: High selects serial slave mode. SCLK is an input in serial slave mode. Bit 3: High enables Fine Scaling through the Serial Port (not available in FIR Mode). Bit 2: High enables Serial Time Slot Syncs (not available in FIR Mode). Bit 1: High enables Power Ramp coefficient interpolation. Bit 0: High enables the Power Ramp. REV. 0 This group of registers contain the Power Ramp Coefficients. See the Power Ramp section for additional details. (0xn80–0xn1FF) Coefficient Memory This group of registers contain the RCF Filter Coefficients. See the RCF section for additional details. PSEUDOCODE Write Pseudocode Void Write_Micro(ext_address, int data); Main() { /* This code shows the programming of the NCO frequency register using the Write_Micro function defined above. The variable address is the External Address A[2:0] and data is the value to be placed in the external interface register. Internal Address = 0x102, channel 1 */ /*Holding registers for NCO byte wide access data*/ int d3, d2, d1, d0; /*NCO frequency word (32 bits wide)*/ NCO_FREQ=0x1BEFEFFF; /*write Chan */ Write_Micro(7, 0x01); /*write Addr */ Write_Micro(6,0x02); /*write Byte 3*/ d3=(NCO_FREQ & 0xFF02Y∞00)>>24; Write_Micro(3,d3); /*write Byte 2*/ d2=(NCO_FREQ & 0xFF0000)>>16; Write_Micro(2,d2); /*write Byte 1*/ d1=(NCO_FREQ & 0xFF00)>>8; Write_Micro(1,d1); /*write Byte 0, Byte 0 is written last and causes an internal write to occur*/ d0=NCO_FREQ & 0xFF; Write_Micro(0,d0); } –35– AD6623 Read Pseudocode MULTIPLE TSP OPERATION Void Read_Micro(ext_address); Main() { /* This code shows the reading of the NCO frequency register using the Read_Micro function defined above. The variable address is the External Address A[2:0] Each of the four Transmit Signal Processors (TSPs) of the AD6623 can adequately reject the interpolation images of narrow bandwidth carriers such as AMPS, IS-136, GSM, EDGE, and PHS. Wider bandwidth carriers such as IS-95 and IMT2000 require a coordinated effort of multiple processing channels. This section demonstrates how to coordinate multiple TSPs to create wider bandwidth channels without sacrificing image rejection. As an example, a UMTS carrier is modulated using four TSP channels (an entire AD6623). The same principles can be applied to different designs using more or fewer TSPs. This section does not explore techniques for using multiple TSPs to solve problems other than Serial Port or RCF throughput. Internal Address = 0x102, channel 1 */ /*Holding registers for NCO byte wide access data*/ int d3, d2, d1, d0; /*NCO frequency word (32 bits wide)*/ /*write Chan */ Write_Micro(7, 0x01); /*write Addr*/ Write_Micro(6,0x02); /*read Byte 0, all data is moved from the Internal Registers to the interface registers on this access, thus Byte 0 must be accessed first for the other Bytes to be valid*/ d0=Read_Micro(0) & 0xFF; /*read Byte 1*/ d1=Read_Micro(1) & 0xFF; /*read Byte 2*/ d2=Read_Micro(2) & 0xFF; /*read Byte 0 */ d3=Read_Micro(3) & 0xFF; } Designing filter coefficients and control settings for de-interleaved TSPs is no harder than designing a filter for a single TSP. For example, if four TSPs are to be used, simply divide the input data rate by four and generate the filter as normal. For any design, a better filter can always be realized by incrementing the number of TSPs to be used. When it is time to program the TSPs, only two small differences must be programmed. First, each channel is configured with exactly the same filter, scalers, modes and NCO frequency. Since each channel receives data at one-quarter the data rate and in a staggered fashion, the Start Hold-Off Counters must also be staggered (see “Programming Multiple TSPs” section). Second, the phase offset of each NCO must be set to match the demultiplexed ratio (in this example). Thus the phase offset should be set to 90 degrees (16384 which is one-quarter of a 16-bit register). Determining the Number of TSPs to Use APPLICATIONS The AD6623 provides considerable flexibility for the control of the synchronization, relative phasing, and scaling of the individual channel inputs. Implementation of a multichannel transmitter invariably begins with an analysis of the output spectrum that must be generated. USING THE AD6623 TO PROCESS UMTS CARRIERS The AD6623 may be used to process two UMTS carriers, each with an output oversampling rate of 24⫻ (i.e., 92.16 MSPS). The AD6623 configuration used to accomplish this consists of using two processing channels in parallel to process each UMTS carrier. Please refer to the Technical Note: Processing Two UMTS Carriers with 24⫻ Oversampling Using the AD6623. DIGITAL TO ANALOG CONVERTER (DAC) SELECTION The selection of a high-performance DAC depends on a number of factors. The dynamic range of the DAC must be considered from a noise and spectral purity perspective. The 14-bit AD9772A is the best choice for overall bandwidth, noise, and spectral purity. In order to minimize the complexity of the analog interpolation filter which must follow the DAC, the sample rate of the master clock is generally set to at least three times the maximum analog frequency of interest. In the case where a 15 MHz band of interest is to be up-converted to RF, the lowest frequency might be 5 MHz and the upper band edge at 20 MHz (offset from dc to afford the best image reject filter after the first digital IF). The minimum sample rate would be set to 65 MSPS. Consideration must also be given to data rate of the incoming data stream, interpolation factors, and the clock rate of the DSP. There are three limitations of a single TSP that can be overcome by deinterleaving an input stream into multiple TSPs: Serial Port bandwidth, the time restriction to the RCF impulse response length (NRCF), and the DMEM restriction to NRCF. If the input sample rate is faster than the Serial Port can accept data, the data can be de-interleaved into multiple Serial Ports. Recalling from the Serial Port description, the SCLK frequency (fSCLK) is determined by the equation below. To minimize the number of processing channels, SCLKdivider should be set as low as possible to get the highest fSCLK- that the serial data source can accept. fSCLK = fCLK SCLKdivider + 1 (29) A minimum of 32 SCLK cycles are required to accept an input sample, so the minimum number of TSPs (NTSP) due to limited Serial Port bandwidth is a function of the input sample rate (fIN), as shown in the equation below. 32 × fIN NTSP ≥ ceil fSCLK (30) For example for a UMTS system, we will assume fCLK = 76.8 MHz, and the serial data source can drive data at 38.4 Mbps (SCLKdivider = 0). To achieve fIN = 3.84 MHz, the minimum NTSP is 3 with a Serial Clock fSCLK = 52 MHz which is a limitation of the Serial Port (This is TSP channels, not TSP ICs). Multiple TSPs are also required if the RCF does not have enough time or DMEM space to calculate the required RCF filter. Recalling the maximum NTAPS– equation from the RCF description, are three restrictions to the RCF impulse response length, NRCF. –36– REV. 0 AD6623 Time Restriction CMEM Restriction 1 N RCF ≤ min , 16 × LRCF 256 2 (31) Programming Multiple TSPs Configuring the TSPs for de-interleaved operation is straight forward. All the Channel Registers and the CMEM of each TSP are programmed identically, except the Start Hold-Off Counters and NCO Phase Offset. DMEM Restriction where: L = LRCF × LCIC 5 × Scaling must be considered as normal with an interpolation factor of L, to guarantee no overflow in the RCF, CIC, or NCOs. The output level at the summation port should be calculated using an interpolation factor of L/NTSP. LCIC 2 N × fCLK = TSP MCIC 2 fIN (32) De-interleaving the input data into multiple TSPs extends the time restriction and may possibly extend the DMEM restriction, but will not extend the CMEM restriction. Deinterleaving the input stream to multiple TSPs divides the input sample rate to each TSP by the number of TSPs used (NTSP). To keep the output rate fixed, L must be increased by a factor of NCH, which extends the time restriction. This increase in L may be achieved by increasing any one or more of LRCF, LCIC5, or LCIC2 within their normal limits. Achieving a larger L by increasing LRCF instead of LCIC5 or LCIC2 will relieves the DMEM restriction as well. In a UMTS example, NTSP = 4, fCLK = 76.8 MHz, and fIN = 3.84 MHz, resulting in L = 80. Factoring L into LRCF = 10, LCIC = 8, and LCIC2 = 1 results in a maximum NRCF = 40 due to the time restriction. Figure 37 shows an example RCF impulse response which has a frequency response as shown in Figure 38 from 0 Hz to 7.68 MHz (fIN ⫻ LRCF/NTSP). The composite RCF and CIC frequency response is shown in Figure 38, on the same frequency scale. This figure demonstrates a good approximation to a root-raised-cosine with a roll-off factor of 0.22, a passband ripple of 0.1 dB, and a stopband ripple better than –70 dB until the lobe of the first image which peaks at –60 dB about 7.68 MHz from the carrier center. This lobe could be reduced by shifting more of the interpolation towards the RCF, but that would sacrifice near in performance. As shown, the first image can be easily rejected by an analog filter further up the signal path. In order to separate the input timing to each TSP, the Hold-Off Counters must be used to start each TSP successively in response to a common Start SYNC. The Start SYNC may originate from the SYNC pin or the MicroPort. Each subsequent TSP must have a Hold-Off Counter value L/NTSP larger than its predecessor’s. If the TSPs are located on cascaded AD6623s, the Hold-Off Counters of the upstream device should be incremented by an additional one. In the UMTS example, L = 80 and NTSP = 4, so to respond as quickly as possible to a Start SYNC, the Hold-Off Counter values should be 1, 21, 41, and 61. Driving Multiple TSP Serial Ports When configured properly, the AD6623 will drive each SDFO out of phase. Each new piece of data should be driven only into the TSP that pulses its SDFO pin at that time. In the UMTS example in Figure 35, L = 80 and NTSP = 4, so each serial port need only accept every fourth input sample. Each serial port is shifting at peak capacity, so sample 1, 2, and 3 begin shifting into Serial Ports B, C, and D before sample 0 is completed into Serial Port A. SDFOA SDFOB SDFOC 32 RAM COEF FILTER SDFOD 32 3.84 MCPS DATA RE-FORMATTER 32 RAM COEF FILTER 32 RAM COEF FILTER CIC NCO Q 76.8MSPS I CIC NCO Q 76.8 MSAMPLES/SEC SUMMATION BLOCK 76.8MSPS I CIC NCO Q 76.8MSPS I CIC NCO Q COMLEX SIGNAL 32 BITS (16, I, 16 Q) REAL OR IMAGINARY SIGNAL Figure 36. Summation Block REV. 0 7 76.8MSPS 9.6MSPS 0.96 MCPS 6 3 I 9.6MSPS 0.96 MCPS 32 RAM COEF FILTER 5 Figure 35. UMTS Example 9.6MSPS 0.96 MCPS 4 1 2 9.6MSPS 0.96 MCPS 0 –37– DAC AD6623 1.0 10 0 –10 CIC ROLLOFF –20 0.5 dBc MAGNITUDE –30 –40 –50 –60 0.0 IDEAL FILTER –70 COMPOSITE AD6623 RESPONSE –80 –90 –100 0 5 10 15 20 25 30 35 40 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 COEFFICIENT kHz Figure 39. RCF and CIL, Frequency Response for WBCDMA Figure 37. Typical Impulse Response for WBCDMA 10 THERMAL MANAGEMENT The power dissipation of the AD6623 is primarily determined by three factors: the clock rate, the number of channels active, and the distribution of interpolation rates. The faster the clock rate the more power dissipated by the CMOS structures of the AD6623 and the more channels active the higher the overall power of the chip. Low interpolation rates in the CIC stages (CIC5, CIC2) results in higher power dissipation. All these factors should be analyzed as each application has different thermal requirements. 0 –10 –20 dBc –30 –40 –50 –60 –70 The AD6623 128-Lead MQFP is specially designed to provide excellent thermal performance. To achieve the best performance the power and ground leads should be connected directly to planes on the PC board. This provides the best thermal transfer from the AD6623 to the PC board. –80 –90 –100 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 kHz Figure 38. RAM Coefficient Filter, Frequency Response for WBCDMA –38– REV. 0 AD6623 OUTLINE DIMENSIONS Dimensions shown in inches and (mm). 128-Lead Terminal Metric Quad Flatpack (MQFP) 0.685 (17.40) 0.677 (17.20) 0.669 (17.00) 0.555 (14.10) 0.551 (14.00) 0.547 (13.90) 0.134 (3.40) MAX 0.041 (1.03) 0.035 (0.88) 0.031 (0.78) 128 1 103 102 SEATING PLANE 0.791 (20.10) 0.787 (20.00) 0.783 (19.90) TOP VIEW (PINS DOWN) 0.921 (23.40) 0.913 (23.20) 0.906 (23.00) 0.003 (0.08) MAX 0.010 (0.25) MIN 38 39 0.110 (2.80) 0.106 (2.70) 0.102 (2.60) 65 64 0.020 (0.50) BSC* 0.011 (0.27) 0.009 (0.22) 0.007 (0.17) * THE ACTUAL POSITION OF EACH LEAD IS WITHIN 0.00315 (0.08) FROM ITS IDEAL POSITION WHEN MEASURED IN THE LATERAL DIRECTION. CENTER FIGURES ARE TYPICAL UNLESS OTHERWISE NOTED 196-Lead Ball Grid Array (BGA) 0.594 (15.10) 0.591 (15.00) SQ 0.587 (14.90) 0.039 (1.00) BSC BALL PITCH 14 12 10 8 6 4 2 13 11 9 7 5 3 1 A B C D E F G H J K L M N P BALL A1 INDICATOR 0.512 (13.0) BSC SQ TOP VIEW 0.059 (1.50) MAX 0.039 (1.00) BSC BALL PITCH BOTTOM VIEW DETAIL A 0.039 (1.00) 0.033 (0.85) SEATING PLANE 0.021 (0.53) 0.017 (0.43) 0.008 (0.20) 0.028 (0.70) COPLANARITY 0.024 (0.60) 0.020 (0.50) BALL DIAMETER DETAIL A NOTES 1. CONTROLLING DIMENSIONS ARE IN MILLIMETERS 2. THE ACTUAL POSITION OF THE BALL GRID IS WITHIN 0.008 (0.20) OF ITS IDEAL POSITION RELATIVE TO THE PACKAGE EDGES 3. THE ACTUAL POSITION OF EACH BALL IS WITHIN 0.004 (0.10) OF ITS IDEAL POSITION TO THE BALL GRID 4. CENTER FIGURES ARE NOMINAL DIMENSIONS REV. 0 –39– –40– PRINTED IN U.S.A. C02768–0–1/02(0)