Advance Data Sheet December 2000 T8301 Internet Protocol Telephone Phone-On-A-Chip™ IP Solution DSP 1 Introduction Lucent Technologies’ Phone-On-A-Chip IP Solution is a highly integrated set of IC chips that form the basic building blocks for an internet protocol telephone (IPT), residing on a local area network (LAN). The IPT presently consists of two ICs—the T8301 (IPT_DSP) and the T8302 (IPT_ARM*). The T8301 provides the audio processing engine for voice compression and decompression, speakerphone echo cancellation, digital-to-analog and analog-to-digital converters, low-pass filters, and amplifiers to drive standard business telephone handsets and speakerphone hardware. The general-purpose processor chip T8302 controls system I/O (Ethernet, USB, IrDA, etc.) and provides general telephone control features (LED control, keypad button scanning, LCD module interface, etc.). A block diagram of the T8301 can be found in Figure 3 on page 8. Since the DSP1627 is an integral part of the T8301, we will refer to the DSP1627 Digital Signal Processor Data Sheet throughout this discussion. 1.1 Features ■ Dual-port RAM, 6K x 16 (zero wait-state at 80 MHz). ■ Internal SRAM, 16K x 16 (single wait-state at 80 MHz). ■ 16-bit analog-to-digital converter. ■ Programmable gain amplifier on audio input. ■ Fixed gain differential microphone input. ■ Analog input SRAM buffer, 512 x 16. ■ Timed DMA for analog input SRAM. ■ Two 16-bit digital-to-analog converters. ■ Independent simultaneous speaker and handset outputs. ■ Two integrated differential speaker driver outputs. ■ Two analog output SRAM buffers, 512 x 16 each. ■ Two timed DMA outputs for simultaneous handset and speaker audio output. ■ Low-pass filtering on audio inputs and outputs. ■ Serial I/O interface. ■ General-purpose timer counter. ■ Bit I/O interface. ■ JTAG test and debugging control. ■ DSP1627 core with bit manipulation unit. ■ Implementation in 0.35 µm, 5 V silicon technology. ■ DSP clock speeds up to 80 MHz. ■ Packaged in 100-pin TQFP. ■ Instruction ROM, 32K x 16 (zero wait-state at 80 MHz). * ARM is a registered trademark of Advanced RISC Machines Limited. T8301 Internet Protocol Telephone Phone-On-A-Chip IP Solution DSP Advance Data Sheet December 2000 Table of Contents Contents Page 1 Introduction .............................................................. 1 1.1 Features ............................................................ 1 2 Pin Information ........................................................ 3 2.1 T8301 100-Pin TQFP Pin Diagram ................... 3 2.2 Pinout Information ............................................. 4 3 Overview .................................................................. 7 4 DSP1600 Core ........................................................ 9 4.1 Bit Manipulation Unit (BMU) .............................. 9 4.2 Timer ................................................................. 9 4.3 Clock PLL Control ............................................. 9 4.4 Bit Input/Output (BIO) ...................................... 10 4.5 Serial Input/Output (SIO) ................................. 10 4.6 Interrupts and Traps ........................................ 10 4.7 Power Management ........................................ 11 4.8 External Memory Interface (EMI) .................... 11 4.9 T8301 Memory Mapping ................................. 11 4.10 Y Space Memory Map ................................... 15 5 Audio Input/Output Circuitry .................................. 17 5.1 Analog Audio Input Channels .......................... 17 5.2 Programmable Gain Amplifier (PGA) .............. 17 5.3 Analog Audio Output Channels ....................... 18 5.4 Tone Ringer ..................................................... 18 5.5 Audio Codec Block .......................................... 20 5.6 Audio Codec Control Registers ....................... 21 6 DMA Input/Output Channels .................................. 23 6.1 DMA Operation ................................................ 23 6.2 DMA Registers ................................................ 23 7 Hardware Compander ........................................... 26 8 Electrical Specifications ......................................... 28 8.1 Operating Range Specifications ...................... 28 8.2 Analog and Codec Specifications .................... 28 8.3 Crystal Specification ........................................ 29 9 JTAG and Hardware Development System (HDS) ................................. 30 9.1 TMODE Control for JCS/Boundary-Scan Operation ........................................................ 30 9.1.1 Mode 7 Operation (TMODE = 7) ............ 30 9.1.2 Mode 6 operation (TMODE = 6) ............ 30 9.2 The Principle of Boundary-Scan Architecture ..................................................... 30 9.2.1 Boundary-Scan Instruction Register ................................ 32 Figures Page Figure 1. T8301 TQFP Pin Diagram ........................... 3 Figure 2. DSP/ARM Interface Block Diagram ............. 7 Figure 3. T8301 Block Diagram .................................. 8 Figure 4. Crystal Oscillator ......................................... 9 Figure 5. Audio Codec Block Diagram ..................... 20 Figure 6. Hardware Compander Block Diagram ....... 27 Figure 7. Boundary-Scan Architecture ..................... 31 2 Tables Page Table 1. Pin Description ............................................. 4 Table 2. SIO Interface Signals ..................................10 Table 3. DSP1627 INT0N and INT1N ......................11 Table 4. T8301 Instruction/Coefficient Memory Map ..............................................13 Table 5. T8301 Memory-Mapped Peripherals ..........14 Table 6. Data Memory Area: I/O, Register, and Memory ................................15 Table 7. Programmable Gain Amplifier Maximum ....17 Table 8. Tone Ringer Control Register (trc_reg) ......18 Table 9. Tone Ringer Amplitude Control Encoding ........................................19 Table 10. Tone Ringer Frequency Encoding ............19 Table 11. aioc_reg Analog Audio I/O Control ...........21 Table 12. Audio Codec Clock Control Register (aclkc_reg) .....................22 Table 13. Audio Clock Encoding ..............................22 Table 14. DMA Control Register dmac_reg ..............24 Table 15. DMA Starting Address Register setadr_reg ..................................24 Table 16. DMA Transfer Count Register setcnt_reg ..................................24 Table 17. DMA Address Increment Register adrinc_reg ..................................25 Table 18. DMA Transfer Decrement Register cntdec_reg ................................................25 Table 19. config_compander Register ......................26 Table 20. write_linear Register .................................26 Table 21. write_companded Register .......................26 Table 22. read_linear Register .................................26 Table 23. read_companded Register .......................26 Table 24. Operating Range Specifications ...............28 Table 25. AINAN Specifications ...............................28 Table 26. AINCP, AINCN Specifications ..................28 Table 27. AOUTA Specifications ..............................28 Table 28. Speaker#1, Speaker#2 Specifications ......29 Table 29. Digital Low-Pass Filters Specifications .....29 Table 30. Digital-to-Analog Converter Specifications ...........................................29 Table 31. Analog-to-Digital Converter Specifications ...........................................29 Table 32. Boundary-Scan Pin Functions ..................32 Table 33. Debug Mode ..............................................32 Table 34. Boundary-Scan Instruction Register .........32 Table 35. Boundary-Scan Register Description .......33 Lucent Technologies Inc. T8301 Internet Protocol Telephone Phone-On-A-Chip IP Solution DSP Advance Data Sheet December 2000 2 Pin Information 80 1 70 10 T8301 (100-pin TQFP) 60 50 AINCP AINCN AINAN VSSGB VDDGB STI1 STO1 STCK CK8KHZ VDD VSS D0 D1 D2 D3 D4 D5 D6 D7 VDD VSS D8 D9 D10 D11 VDD A11 A10 A9 A8 A7 A6 A5 A4 VSS VDD A3 A2 A1 A0 I_CSN M_CSN X_CSN RWN VSS VDD D15 D14 D13 D12 40 20 30 VDD BIO0 BIO1 BIO2 BIO3 INT0N INT1N STOPN DI1 VSS VDD DO1 SYNC IOLD IOCK VSS VDDPLL CKI1 CKI2 VSSPLL A15 A14 A13 A12 VSS 90 100 VSS CKO CK2MHZ TDI TMS TCK TDO RESETN TMODEN2 TMODEN1 TMODEN0 TRSTN VDDGB VSSGB SVDD SPKDRV2A SPKDRV2B SVSS SVSS SPKDRV1B SPKDRV1A SVDD VDDA AOUTA GNDA 2.1 T8301 100-Pin TQFP Pin Diagram 5-8211(F) Figure 1. T8301 TQFP Pin Diagram Lucent Technologies Inc. 3 T8301 Internet Protocol Telephone Phone-On-A-Chip IP Solution DSP Advance Data Sheet December 2000 2 Pin Information (continued) 2.2 Pinout Information In the following table, reference 1 refers to sections in the T8301 data sheet (this data sheet) and reference 2 refers to sections in the DSP1627 data sheet. Table 1. Pin Description Pin # 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 4 Name VDD BIO0 BIO1 BIO2 BIO3 INT0N INT1N STOPN DI1 VSS VDD DO1 SYNC IOLD IOCK VSS VDDPLL CKI1 CKI2 VSSPLL A15 A14 A13 A12 VSS VDD A11 A10 A9 A8 A7 A6 A5 A4 VSS VDD A3 Description — BIT I/O 0 BIT I/O 1 BIT I/O 2 BIT I/O 3 DSP interrupt 0, active -low DSP interrupt 1, active -low Controls the internal processor clock, active -ow Serial input/output unit (SIO) data in — — Serial input/output unit (SIO) data out Serial input/output unit (SIO) sync Serial input/output unit (SIO) input load/output load Serial input/output unit (SIO) input clock/output clock — OSC and PLL VDD XTL1 input/CMOS clock XTL2 input/CMOS clock OSC and PLL VSS EMI address bus 15 EMI address bus 14 EMI address bus 13 EMI address bus 12 — — EMI address bus 11 EMI address bus 10 EMI address bus 9 EMI address bus 8 EMI address bus 7 EMI address bus 6 EMI address bus 5 EMI address bus 4 — — EMI address bus 3 Reference 1 — 4.4 4.4 4.4 4.4 4.6 4.6 — 4.5 — — 4.5 4.5 4.5 4.5 — — 4.3 4.3 — 4.9 4.9 4.9 4.9 — — 4.9 4.9 4.9 4.9 4.9 4.9 4.9 4.9 — — 4.9 Reference 2 — 4.9 4.9 4.9 4.9 4.3 4.3 4.13 4.7 — — 4.7 4.7 4.7 4.7 — — 4.12 4.12 — — 4.5, 6.2 4.5, 6.2 4.5, 6.2 — — 4.5, 6.2 4.5, 6.2 4.5, 6.2 4.5, 6.2 4.5, 6.2 4.5, 6.2 4.5, 6.2 4.5, 6.2 — — 4.5, 6.2 Lucent Technologies Inc. Advance Data Sheet December 2000 T8301 Internet Protocol Telephone Phone-On-A-Chip IP Solution DSP 2 Pin Information (continued) Table 1. Pin Description (continued) Pin # 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 Name A2 A1 A0 I_CSN M_CSN X_CSN RWN VSS VDD D15 D14 D13 D12 D11 D10 D9 D8 VSS VDD D7 D6 D5 D4 D3 D2 D1 D0 VSS VDD CK8KHZ STCK STO1 STI1 VDDGB VSSGB AINAN AINCN AINCP GNDA AOUTA VDDA Description EMI address bus 2 EMI address bus 1 EMI address bus 0 ARM interrupt chip select, active-low ARM memory chip select, active-low External memory chip select, active-low Read/write, active-low — — EMI data bus 15 EMI data bus 14 EMI data bus 13 EMI data bus 12 EMI data bus 11 EMI data bus 10 EMI data bus 9 EMI data bus 8 — — EMI data bus 7 EMI data bus 6 EMI data bus 5 EMI data bus 4 EMI data bus 3 EMI data bus 2 EMI data bus 1 EMI data bus 0 — — Test clock Serial test clock* Serial test out 1* Serial test in 1* — — (Handset) single-ended microphone input (Speakerphone) microphone differential input negative (Speakerphone) microphone differential input positive — (Handset) single-ended speaker output — Reference 1 4.9 4.9 4.9 4.10, Table 6 4.10, Table 6 4.10, Table 6 — — — 3 3 3 3 3 3 3 3 — — 3 3 3 3 3 3 3 3 — — 9.1 — — — — — 5.1 5.1 5.1 — 5.3 — Reference 2 4.5, 6.2 4.5, 6.2 4.5, 6.2 — — — 4.5, 6.2 — — 4.5, 6.2 4.5, 6.2 4.5, 6.2 4.5, 6.2 4.5, 6.2 4.5, 6.2 4.5, 6.2 4.5, 6.2 — — 4.5, 6.2 4.5, 6.2 4.5, 6.2 4.5, 6.2 4.5, 6.2 4.5, 6.2 4.5, 6.2 4.5, 6.2 — — — — — — — — — — — — — — * Leave open, this is for test purposes only. Lucent Technologies Inc. 5 T8301 Internet Protocol Telephone Phone-On-A-Chip IP Solution DSP Advance Data Sheet December 2000 2 Pin Information (continued) Table 1. Pin Description (continued) Pin # 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 6 Name SVDD SPKDRV1A SPKDRV1B SVSS SVSS SPKDRV2B SPKDRV2A SVDD VSSGB VDDGB TRSTN TMODEN0 TMODEN1 TMODEN2 RESETN TDO TCK TMS TDI CK2MHZ CKO VSS Description — (Speakerphone) speaker#1 differential output driver A (Speakerphone) speaker#1 differential output driver B — — (Speakerphone) speaker#2 differential output driver B (Speakerphone) speaker#2 differential output driver A — — — JTAG test reset input, active-low Test mode 0 Test mode 1 Test mode 2 Chip reset, active-low JTAG test data out JTAG test clock JTAG mode select JTAG test data in Clock out DSP clock out — Reference 1 — 5.3 5.3 — — 5.3 5.3 — — — 9.2 9.1 9.1 9.1 — 9.2 9.2 9.2 9.2 9.1 9.1 — Reference 2 — — — — — — — — — — — — — — 10.2 6.6 6.6 6.6 6.6 — 4.12 — Lucent Technologies Inc. T8301 Internet Protocol Telephone Phone-On-A-Chip IP Solution DSP Advance Data Sheet December 2000 3 Overview The T8301 (DSP) interfaces with the T8302 (ARM) to form the basic building blocks for an internet protocol telephone (IPT), residing on a local area network (LAN); see Figure 2. At the heart of the T8301 integrated circuit is the Lucent Technologies Microelectronics Group DSP1627 digital signal processor core. The DSP1627’s high-performance (80 MIPS) and single-cycle multiply accumulate instruction provide excellent support for execution of voice compression/decompression and echo cancellation algorithms. The DSP1627 core and the analog audio circuitry included on the T8301 IC provide a low-cost silicon solution for the IP exchange telephone’s audio requirements. A block diagram of the T8301 integrated circuit is shown in Figure 3. The DSP1627 core contains the DSP1600 core processor, bit manipulation unit (BMU), dual-port RAM (DPRAM), instruction/coefficient ROM (IROM), bit I/O (BIO), serial I/O (SIO), timer, clock PLL control, vectored interrupts and traps, power management, external memory interface (EMI) with wait-state control, and a JTAG interface with integral hardware development system support. The DSP1627 peripherals communicate with the DSP1627 core through the (D-IDB bus), which is 16 bits wide. The DSP1627 core’s Harvard architecture allows efficient memory utilization by supporting separate instruction (XDB, XAB) and data (YDB, YAB) address spaces. The dual-port RAM (DPRAM) is connected to both address and data buses XDB, YDB, XAB, and YAB, while the instruction ROM is only connected to the XDB and XAB memory bus. The external memory interface provides a mechanism to access I/O devices and memories that are not part of the core DSP1627 hardware. For a complete description of the DSP1627 core and its peripherals, refer to the DSP1627 Digital Signal Processor Data Sheet. A brief description of the functionality of the DSP1627 is provided in the following section. Where necessary, comments are made which reflect differences between the operation of the DSP1627 and the T8301. Please refer to the DSP1627 data sheet for further explanation. X = LEAVE OPEN IF UNUSED 12.288 MHz CLOCK SOURCE CKI1 CKI2 CKO CK8KHZ CK2MHZ HANDSET SPEAKER AND MIC AOUTA SPEAKERPHONE SPEAKER AND MIC SPKDRV1A SPKDRV1B AINCP AINCN HEADSET SPEAKER SPKDRV2A SPKDRV2B AINAN OPTIONAL CLOCK RESOURCES RESETN OPTIONAL EXTERNAL SERIAL CODEC DO1 DI1 IOCK IOLD SYNC OPTIONAL BIT INPUT OUTPUT BIO0 BIO1 BIO2 BIO3 ATE ANALOG TEST PINS TEST MODE SELECT PINS DSP_INT0N M_CS DSP_MCSN I_CSN DSP_ICSN RWN DSP_RWN A0—A11 DSP_A0—DSP_A11 D0—D15 DSP_D0—DSP_D15 DSP STCK STO1 STI1 TMODEN0 TMODEN1 TMODEN2 RESETN INT0N X_CSN TRSTN TDO TDI TC TMS OPTIONAL (MEMORY) DEVICE ON 12K Y DATA BUS ARM BOUNDARY SCAN AND/OR JTAG INT1N Figure 2. DSP/ARM Interface Block Diagram Lucent Technologies Inc. 7 Advance Data Sheet December 2000 T8301 Internet Protocol Telephone Phone-On-A-Chip IP Solution DSP 3 Overview (continued) STCK TONE RINGER 1.5 dB STO1 STI1 trc_reg 2.5 Vp-p SPKDRV1A + SPKDRV1B 12 dB SPKDRV2A SPKDRV2B 12 dB AINCN DMA OUTPUT DMAS COUNTER aioc ADDRESS act1 AOUTA AINCP AUDIO CODEC BLOCK + 30 dB – PGA 0—21 dB IN 3 dB STEPS AINAN act2 OUTPUT DMAH COUNTER AUDIO CLOCK GENERATOR ADDRESS aclkc INPUT DMA COUNTER CKO RWN ADDRESS DECODE CLOCK/PLL & POWER powerc D[15:0] A[15:0] RWN I/O ERAMLO ERAMHI EROM pllc CKI1 DSP1627 CORE EXTERNAL MEMORY INTERFACE mwait TDI ioc JTAG jtag JCON IROM 32K x 16 DPRAM 6K x 16 YDB TMS HDS Trace OR TIMER timerc YAB DATA BUS Breakpoint INT0N srta tdms sdx(in) BYPASS INT1N SIO sdx(out) ID TCK M_CSN INTERNAL SRAM 16K x 16 ‘OR’ CK2MHz TDO I_CSN AIN SRAM BUFFER 512 x 16 STOPN TRSTN A(15:0) dmac reg DMAINT TMODEN2 CKI2 AOUTA SRAM BUFFER 512 x 16 X_CSN TMODEN1 CK8kHz D(15:0) ADDRESS TMODEN0 RESETN SOUT SRAM BUFFER 512 x 16 I N T E R R U P T XAB XDB INSTRUCTION/ COEFFICIENT BUS BMU sioc aa0 saddx IOLD IOCK SYNC DO1 DI1 aa1 DSP 1600 CORE ar0 BIO ar1 sbit ar2 cbit BIO[3:0] ar3 D-IDB timer 0 5-8210 (F) Figure 3. T8301 Block Diagram 8 Lucent Technologies Inc. T8301 Internet Protocol Telephone Phone-On-A-Chip IP Solution DSP Advance Data Sheet December 2000 4 DSP1600 Core The discussions in this section pertain to circuitry that is inside of the dotted outline in Figure 3. For additional resources, please refer to the DSP1627 Digital Signal Processor Data Sheet. The DSP1600 core includes a data arithmetic unit, memory addressing units, cache, and a control section. In combination, these elements support a diverse instruction set for implementing users’ algorithms. 4.1 Bit Manipulation Unit (BMU) The BMU provides extensive bit manipulation capabilities that increase the DSP1627’s efficiency in processing data. 4.2 Timer The DSP1627 core contains a programmable interrupt timer that can be configured to count over a wide range of frequencies. This timer provides flexibility in timing events. 4.3 Clock PLL Control The DSP1627 powers up with the input clock (CKI1/CKI2 in the T8301 IC) as the source for the processor clock. An on-chip clock synthesizer (PLL) can also be used to generate the system clock for the DSP1627, which will run at a frequency multiple of the input clock. The clock synthesizer is deselected and powered down on reset. For lowpower operation, an internally generated slow clock can be used to drive the DSP1627. If both the clock synthesizer and the internally generated slow clock are selected, the slow clock will drive the DSP1627; however, the synthesizer will continue to run. The clock synthesizer and other programmable clock sources are discussed in the DSP1627 data sheet. The use of these programmable clock sources for power management is also discussed in the DSP1627 data sheet. Board designers should refer to the section on VDDA and VSSA connections in the data sheet for specific connection and filtering requirements on the clock synthesizer power and ground leads. CKI1 TO PLL LOAD CAPACITOR 12,288 kHz CRYSTAL OSCILLATOR ÷768 16 kHz TO CODECS CKI2 LOAD CAPACITOR Note: The 12,288 KHz is required as shown. Variations from this crystal frequency will cause detrimental effects to speech quality. Figure 4. Crystal Oscillator Lucent Technologies Inc. 9 T8301 Internet Protocol Telephone Phone-On-A-Chip IP Solution DSP Advance Data Sheet December 2000 4 DSP1600 Core (continued) 4.4 Bit Input/Output (BIO) The BIO provides convenient and efficient monitoring and control of four individually configurable pins. When configured as outputs, the pins can be individually set, cleared, or toggled. When configured as inputs, individual pins or combinations of pins can be tested for patterns. Flags returned by the BIO mesh seamlessly with conditional instructions. Although the DSP1627 has eight BIOs available, the T8301 makes the four low-order BIOs available on pins. 4.5 Serial Input/Output (SIO) The serial I/O interface (SIO) of the T8301 closely follows the serial interface of the DSP1627. The T8301 multiplexes certain DSP1627 SIO pins and eliminates some others to reduce the total pin count. Hysteresis input buffers are used for the SIO clocks on this device (IOLD, IOCK, and SYNC). The table below shows the signals that comprise the T8301 SIO interface. Table 2. SIO Interface Signals Symbol DI1 Type I Serial data in 1. DO1 O Serial data out 1. IOLD* I/O Input/output load for SIO 1. IOCK I/O Input/output clock for SIO 1. SYNC I/O Sync for SIO 1 and 2. Function * IOLD is comprised of the ILD1 and the OLD1 signals from the DSP1627 core tied together. By default, the IOLD signal is an input, which corresponds to the two DSP1627 load signals configured as passive. However, input load 1 (ILD1) may be configured as active, which then configures the IOLD signal as an output. In this case, the internal input load 1 (ILD1) drives the output load signal (OLD1.) IOCK is analogous to IOLD. Input clock 1 can be configured as an output, which would then drive IOCK and OCK1. If the PLL is enabled, care should be taken if using IOCK as an output since there may be an unacceptable amount of jitter on the clock. The SYNC signal is intended to provide synchronization of the serial bus with an external 8 kHz frame clock. When SYNC is configured as an input and asserted, the SIO load counter is reset and IOLD is asserted (if configured as an output). For typical applications, the SIO will be configured to have SYNC and IOCK as inputs and IOLD as an output (from the DSP1627 core). In this configuration, there are thirty-two 8-bit (sixteen 16-bit) time slots for each SIO channel and SYNC provides the 8 kHz SIO frame timing. The timing relationship for this configuration can be found in the DSP1627 data sheet. 4.6 Interrupts and Traps The DSP1627 supports prioritized, vectored interrupts, and a trap. There are eight internal hardware sources for program interrupt and two external interrupt pins. Additionally, there is a trap signal from the hardware development system (HDS). Each of the sources has a unique vector address and priority assigned to it. Refer to the DSP1627 data sheet for more information. The use of the two external DSP1627 core interrupts is shown in Table 3 and in Figure 2. 10 Lucent Technologies Inc. T8301 Internet Protocol Telephone Phone-On-A-Chip IP Solution DSP Advance Data Sheet December 2000 4 DSP1600 Core (continued) INT0N is dedicated to the ARM DCC interrupts in the DSP’s mask ROM. INT1N is internally ORed with the DMA interrupt. The DSP’s mask ROM for INT1 is dedicated to DMA servicing. It is recommended that INT1N float (internal pull up on pin). Table 3. DSP1627 INT0N and INT1N Interrupt Function INT1N Interrupt from DMA block or external interrupt 1, active-low. INT0N External interrupt input 0, active-low. Interrupt Priority 4 (higher) 2 4.7 Power Management There are three different power management control mechanisms: the power control register (POWERC), the stop pin (STOPN), and the AWAIT bit in the ALF register. Refer to the DSP1627 data sheet for more information concerning these registers and their usage. 4.8 External Memory Interface (EMI) The T8301 external memory interface is used to access the non-DSP1627 core features provided in the T8301 integrated circuit. The external memory interface is also used to access off-chip resources such as the interprocessor communication memories contained in the IPT_ARM integrated circuit. The T8301 external memory interface requires one wait-state to access on-chip resources and two wait-states to access 15 ns or faster off-chip resources when operating at 80 MHz. 4.9 T8301 Memory Mapping The T8301 contains various types of memory modules, all with varying characteristics, aside from their memory map location. As a Harvard architecture, the device has two address/data buses; these are referred to as X and Y. The X system is used for program instructions and data, and the Y system is typically for data and memory mapped I/O. Memory is 16 bits wide. The DSP1627 can vary the X bus memory map based on the logic levels on two signals: EXM and LOWPR. However, the T8301 has EXM tied low internally, reducing the possibilities to two. The two memory maps are the DSP1627’s MAP1 and MAP3. LOWPR is software controllable. When using the DSP1627 software tools (with JCS i.e., JTAG communications system) the tools will configure LOWPR automatically based upon the link time compile options of the .if file. Map1 is the default map. The basic difference of the two maps is the type of memory at the reset vector (0x0000). MAP1 has ROM at 0x0000, and MAP3 has RAM at 0x0000. The Y map is fixed. The T8301 is a masked ROM-coded device and contains no flash memory. MAP 1 is typically used for production, and Map 3 is typically used for code development. When used in conjunction to the T8302 ARM embedded processor, the ARM will be required to pass all code and data to the DSP's ram at power up reset. A hardware/software protocol must be instituted to allow the ARM to successfully load code into the DSP. Note: All X memory references are MAP 3. ■ Internal ROM, IROM—32K x 16: — Responds only to the X data bus, the X memory location is 0x4000—0xBFFF. This block will operate with zero wait-states. Lucent Technologies Inc. 11 T8301 Internet Protocol Telephone Phone-On-A-Chip IP Solution DSP Advance Data Sheet December 2000 4 DSP1600 Core (continued) ■ Dual-port (core) RAM, DPRAM—6K x 16: — This block is a true dual-port memory and is accessible simultaneously by both the X and Y bus system. Two locations can be either read or written in the same instruction execution. This memory block resides at locations 0x0000—0x17FF on both the X and Y maps. This block will operate with zero wait states. The DPRAM contains 6K x 16-bit words of zero wait-state memory, which is organized into six banks of 1K x 16-bit words. Each bank has separate ports to the instruction/coefficient and data memory spaces. Dual accesses to both memory spaces in separate banks incur no wait-states; however, accesses to the same bank from both spaces will add one wait-state to the total access time. ■ Internal SRAM, ISRAM—16K x 16: — Although this is a dual-port RAM, there is only one bus system to the RAM itself. The X and Y bus is multiplexed before the RAM and is actually addressed via the external memory interface (EMI). Two locations can be either read or written in the same instruction execution, but will require two clock cycles. The X memory location is at 0xC000—0xFFFF and the Y memory location is at 0x8000—0xBFFF, and also at 0xC000— 0xFFFF. (Referred to as mirrored. A write to 0x8000 on the Y map will also write to 0xC000). There is only one block of 16K; however, it appears twice on the Y map. There is one wait-state required for both the X and Y bus to access this RAM. ■ External SRAM, XSRAM—12K x 16: — Responds only to the Y data bus. The T8301 generates a chip select called X_CSN (active-low), pin 43. It uses the EMI to generate the address and data. There is one wait-state required for both the X and Y bus to access this RAM. 12 Lucent Technologies Inc. T8301 Internet Protocol Telephone Phone-On-A-Chip IP Solution DSP Advance Data Sheet December 2000 4 DSP1600 Core (continued) Table 4. T8301 Instruction/Coefficient Memory Map Address 0x0000 0x0800 0x1000 0x1800 0x2000 0x2800 0x3000 0x3800 0x4000 0x4800 0x5000 0x5800 0x6000 0x6800 0x7000 0x7800 0x8000 0x8800 0x9000 0x9800 0xA000 0xA800 0xB000 0xB800 0xC000 0xC800 0xD000 0xD800 0xE000 0xE800 0xF000 0xF800 X MAP1 X MAP3 Dual-port RAM 6K (DPRAM) Internal ROM 32K (IROM) Reserved 10K (Instructions and constants) Internal ROM 32K (IROM) (Instructions and constants) Internal SRAM 16K (ISRAM) Dual-port RAM 6K (DPRAM) Reserved 10K Lucent Technologies Inc. Internal SRAM 16K (ISRAM) Address 0x0000 0x0800 0x1000 0x1800 0x2000 0x2800 0x3000 0x3800 0x4000 0x4800 0x5000 0x5800 0x6000 0x6800 0x7000 0x7800 0x8000 0x8800 0x9000 0x9800 0xA000 0xA800 0xB000 0xB800 0xC000 0xC800 0xD000 0xD800 0xE000 0xE800 0xF000 0xF800 Y MAP Dual-port RAM 6K (DPRAM) Reserved 10K I/O and ERAMLO (See Table 5.) ERAMLO (See Table 5.) ERAMLO External chip select X_CSN (external SRAM) 12K Internal SRAM 16K (1K—16K block mirrored) (ISRAM) Internal SRAM 16K (1K—16K block mirrored) (ISRAM) 13 T8301 Internet Protocol Telephone Phone-On-A-Chip IP Solution DSP Advance Data Sheet December 2000 4 DSP1600 Core (continued) Table 5. T8301 Memory-Mapped Peripherals Address 0x4000 0x4100 0x4200 0x4300 0x4400 0x4500 0x4600 0x4700 0x4800 0x4900 0x4A00 0x4B00 0x4C00 0x4D00 0x4E00 0x4F00 14 Y MAP Analog I/O devices (I_CSN) DCC control interface Audio input, SRAM (512) read only Handset audio output, SRAM (512) write only Speaker audio output, SRAM (512) write only M_CSN ARM-to-DSP Buffer (1K) M_CSN DSP-to-ARM Buffer (1K) Lucent Technologies Inc. T8301 Internet Protocol Telephone Phone-On-A-Chip IP Solution DSP Advance Data Sheet December 2000 4 DSP1600 Core (continued) 4.10 Y Space Memory Map The table below shows the Y space memory map. This is the area can be addressed via the DSP1627’s R0, R1, R2, and R3 registers, and also by direct (Y-based) addressing. Table 6. Data Memory Area: I/O, Register, and Memory Address R/W DSP CS Function 0x0:0x17FF 0x4000 R/W R/W Internal I/O Internal RAM aioc_reg 0x4001 0x4002 0x4003 R/W R/W R/W I/O I/O I/O 0x4008 0x4010 0x4014 R/W R/W R/W I/O I/O I/O 0x4015 R/W I/O 0x4016 R/W I/O 0x4017 R/W I/O 0x4018 R/W I/O 0x4019 R/W I/O 0x401A R/W I/O Ox401B R/W I/O 0x401C R/W I/O 0x401D R/W I/O 0x401E R/W I/O 0x401F R/W I/O 0x4040 R/W I/O 0x4041 0x4041 0x4042 W R W I/O I/O I/O Lucent Technologies Inc. Description — Analog audio I/O control register, see Table 11. act1_reg Audio codec test register 1. act2_reg Audio codec test register 2. aclkc_reg Audio codec clock control register, see Table 12. trc_reg Tone ringer control register, see Table 8. dmac_reg DMA control register, see Table 14. AINsetadr_reg Audio in DMA starting address register, see Table 15. AINsetcnt_reg Audio in DMA transfer count registers, see Table 16. AINadrinc_reg Audio in DMA address increment registers, see Table 17. AINcntdec_reg Audio in DMA transfer count decrement register, see Table 18. HNDsetadr_reg Handset DMA starting address register, see Table 15. HNDsetcnt_reg Handset DMA transfer count register, see Table 16. HNDadrinc_reg Handset DMA address increment register, (see Table 17). HNDcntdec_reg Handset DMA transfer count decrement register, see Table 18. SPKsetadr_reg Speaker DMA starting address register, see Table 15. SPKsetcnt_reg Speaker DMA transfer count register, see Table 16. SPKadrinc_reg Speaker DMA address increment register, see Table 17. SPKcntdec_reg Speaker DMA transfer count decrement register, see Table 18. config_compander Compander configuration register, see Table 19. write_companded Write companded value register, see Table 21. read_linear Read linear value register, see Table 22. write_linear Write linear value register, see Table 20. Size (words) 6K 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 15 T8301 Internet Protocol Telephone Phone-On-A-Chip IP Solution DSP Advance Data Sheet December 2000 4 DSP1600 Core (continued) Table 6. Data Memory Area: I/O, Register, and Memory (continued) Address R/W DSP CS 0x4042 R I/O 0x4100:0x4107 R/W 0x4108:x410B R/W ERAMLO ERAMLO 0x410C:0x410F R/W ERAMLO 0x4110:0x41FF R/W 0x4200:0x43FF R 0x4400:0x45FF W 0x4600:0x47FF W 0x4800:0x4BFF R ERAMLO ERAMLO ERAMLO ERAMLO ERAMLO 0x4C00:0x4FFF ERAMLO W 0x5000:0x7FFF R/W 0x8000:0xBFFF R/W 16 ERAMLO ERAMHI Function Description read_companded Read companded value register, see Table 23. I_CSN External chip select to access token registers. I_CSN External chip select to access ARM interrupt register. I_CSN External chip select to access DSP interrupt register. I_CSN Reserved. AIN SRAM Audio input SRAM buffer. AOUTA SRAM Handset audio out SRAM buffer. SOUT SRAM Speaker audio out SRAM buffer. M_CSN External chip select to access ARM to DSP RAM (in the T8302 IPT_ARM chip). M_CSN External chip select to access DSP to ARM RAM (in the T8302 IPT_ARM chip). X_CSN External spare chip select. ISRAM Internal SRAM. Size (words) 1 8 4 4 240 512 512 512 1K 1K 12K 16K Lucent Technologies Inc. T8301 Internet Protocol Telephone Phone-On-A-Chip IP Solution DSP Advance Data Sheet December 2000 5 Audio Input/Output Circuitry The discussions in this section pertain to circuitry that is outside of the dotted outline in Figure 3 on page 8. 5.1 Analog Audio Input Channels The T8301 contains analog interfaces designed to support a 150 Ω handset as well as an additional microphone and two speakers. The T8301 integrated circuit contains two audio analog inputs. There is a single-ended input (AINAN) to be connected to a standard business telephone handset receiver. There is a differential input (AINCP, AINCN) to be connected to a microphone. This provides the T8301 with the input circuitry to implement a speakerphone. The differential input is directly connected to a 30 dB amplifier. The input select multiplexer routes AINAN or the output of the fixed 30 dB amplifier to a programmable gain amplifier (PGA). The programmable gain amplifier is adjustable from 0 dB to 21 dB in 3 dB steps. The signal output from the programmable gain amplifier is then routed to the audio codec block to be digitized. Each of the input signals AINAN, AINCP, and AINCN are ac-coupled to their T8301 inputs by a 0.2 µF capacitor. The maximum signal input to the codec is 2.5 Vp-p. If the user sets the amplification to a value that would produce a larger signal than 2.5 Vp-p, the audio codec will saturate and clip the input waveform. The maximum input signal from the handset or from the microphone that can be supported for each gain setting is listed in Table 7. Since the microphone amplifier has a maximum specified signal of 40 mV, the maximum microphone input is not supported for PGA settings of 0 dB and 3 dB. 5.2 Programmable Gain Amplifier (PGA) The programmable gain amplifier is using the PGAS[2:0] bits of the aioc_reg (see Table 11 on page 21). The settable gain values and their tolerances are shown below as well as the maximum allowed input signal voltage from each of the input signals. Inputs greater than these values will saturate the input codec and produce clipped waveforms. Table 7. Programmable Gain Amplifier Maximum Bit Code 000 001 010 011 100 101 110 111 Lucent Technologies Inc. Gain 0 dB ± 0.5 dB 3 dB ± 0.5 dB 6 dB ± 0.5 dB 9 dB ± 0.5 dB 12 dB ± 0.5 dB 15 dB ± 1.0 dB 18 dB ± 1.0 dB 21 dB ± 1.5 dB AINAN 2.500 Vp-p 1.770 Vp-p 1.250 Vp-p 0.844 Vp-p 0.625 Vp-p 0.442 Vp-p 0.313 Vp-p 0.221 Vp-p Max Input Signal AINCN, AINCP Not supported Not supported 40.0 mVp-p 28.3 mVp-p 20.0 mVp-p 14.2 mVp-p 10.0 mVp-p 7.1 mVp-p 17 Advance Data Sheet December 2000 T8301 Internet Protocol Telephone Phone-On-A-Chip IP Solution DSP 5 Audio Input/Output Circuitry (continued) 5.3 Analog Audio Output Channels The T8301 contains two independent analog audio output ports. There is a single-ended output signal, AOUTA, that can be connected to the speaker of a standard 150 Ω business telephone handset or to a differential speaker driver SPKDRV2. Differential speaker driver, SPKDRV1, can be set up to ring the phone by adding in the tone ringer output into its audio path. Differential speaker driver output pins (SPKDRV1A, SPKDRV1B and SPKDRV2A, SPKDRV2B) should be connected to 45 Ω speakers. Both outputs receive their analog signals from the audio codec block, which converts the two digital input streams to analog signals. The maximum signal from the codec is 2.5 Vp-p. The AOUTA signal has a maximum 2.5 Vp-p signal swing. It should maintain a midlevel bias to prevent load noises when the driver is re-enabled. The speaker outputs (SPKDRV1A, SPKDRV1B and SPKDRV2A, SPKDRV2B) each have 3 Vp-p signal swing. Since these outputs are of opposite polarity, the differential signal output is 6 Vp-p. This is a 6 dB effective amplification of the codec output signal. The signals should be biased such that, when power is re-enabled, no audible noises occur. The differential speaker output driver does not have to produce a full 6 Vp-p signal without distortion. Signals above 5 Vp-p measured from SPKDRVxA to SPKDRVxB may be in the nonlinear range of the differential amplifier and exhibit a flattening or clipping characteristic at the output. AOUTA is ac coupled to the handset speaker using a 2 µF capacitor. The speaker driver outputs (SPKDRV1A, SPKDRV1B and SPKDRV2A, SPKDRV2B) are direct coupled to 45 Ω speakers. 5.4 Tone Ringer The T8301 analog circuitry contains a tone ringer generator. When this circuit is powered up and enabled, the ringing tone output is added to the current analog speaker signal and output through the differential speaker driver. Custom tones may be generated by modifying the T8301 firmware. Table 8. Tone Ringer Control Register (trc_reg) Tone Ringer Control Register (trc_reg) Address (0x4008) Bit # 15:13 12 11:8 7:0 Name RSVD TR_EN TR_AC[3:0] TR_FC[7:0] Bit # 15:13 12 Name RSVD TR_EN 11:8 7:0 TR_AC[3:0] TR_FC[7:0] 18 Description Reserved. Tone ringer output enable. If 1, the tone ringer’s output is added into the speaker output path. If 0, the tone ringer’s output is disconnected from the speaker output path. Tone ringer amplitude control, see Table 9. Tone ringer frequency control, see Table 10. (The tone ringer frequencies are listed in hex format). Lucent Technologies Inc. T8301 Internet Protocol Telephone Phone-On-A-Chip IP Solution DSP Advance Data Sheet December 2000 5 Audio Input/Output Circuitry (continued) Table 9. Tone Ringer Amplitude Control Encoding Bit# TR_AC[3:0] 0000 0001 0010 0011 0100 0101 0110 0111 1000 1001 1010 1011 1100 1101 1110 1111 Volts Out (p-p) Silent 0.023 V 0.032 V 0.045 V 0.063 V 0.087 V 0.120 V 0.170 V 0.230 V 0.330 V 0.460 V 0.620 V 0.880 V 1.250 V 1.770 V 2.500 V dB Relative to Maximum Level dc to midvoltage reference –40.60 –37.74 –34.88 –32.02 –29.16 –26.30 –23.44 –20.58 –17.72 –14.86 –12.00 –9.00 –6.00 –3.00 0 Tolerance (dB from Nominal) Not applicable ±0.75 ±0.75 ±0.75 ±0.50 ±0.50 ±0.50 ±0.25 ±0.25 ±0.25 ±0.25 ±0.25 ±0.25 ±0.25 ±0.25 Not applicable Table 10. Tone Ringer Frequency Encoding Hz 24,000 16,000 12,000 9,600 8,000 6,857 6,000 5,333 4,800 4,364 4,000 3,692 3,429 3,200 3,000 2,824 2,667 2,526 2,400 2,286 2,182 Hex 3F 1F 0F 87 43 A1 D0 E8 F4 7A 3D 1E 8F C7 63 B1 58 2C 16 0B 05 Hz 1,067 1,043 1,021 1,000 979.5 960.0 941.2 923.1 905.6 888.9 872.7 857.1 842.1 827.6 813.6 800.0 786.9 774.2 761.9 750.0 738.5 Lucent Technologies Inc. Hex 1D 0E 07 03 81 C0 60 30 98 4C 26 93 49 24 92 C9 64 B2 D9 EC 76 Tone Ringer Frequency Encoding Hz Hex Hz Hex 545.5 35 366.4 6C 539.3 9A 363.6 36 533.3 4D 360.9 1B 527.5 A6 358.2 8D 521.7 D3 355.5 C6 516.1 69 352.9 E3 510.6 34 350.4 F1 505.3 1A 347.8 78 500.0 0D 345.3 3C 494.8 86 342.8 9E 489.8 C3 340.4 CF 484.9 E1 338.0 E7 480.0 F0 335.7 73 475.2 F8 333.3 39 470.6 7C 331.0 9C 466.0 BE 328.8 CE 461.5 DF 326.5 67 457.1 6F 324.3 33 452.8 B7 322.1 19 448.6 DB 320.0 8C 444.4 ED 317.9 46 Hz 277.5 275.9 274.3 272.7 271.2 269.7 268.2 266.7 265.2 263.7 262.3 260.9 259.5 258.1 256.7 255.3 253.9 252.6 251.3 250.0 248.7 Hex 2A 95 CA E5 72 B9 DC EE 77 BB DD 6E 37 9B CD E6 F3 79 BC DE EF Hz 223.3 222.2 221.2 220.2 219.2 218.2 217.2 216.2 215.3 214.3 213.3 212.4 211.5 210.5 209.6 208.7 207.8 206.9 206.0 205.1 204.3 Hex C5 62 31 18 0C 06 83 C1 E0 70 B8 5C AE 57 AB 55 AA D5 EA F5 FA 19 Advance Data Sheet December 2000 T8301 Internet Protocol Telephone Phone-On-A-Chip IP Solution DSP 5 Audio Input/Output Circuitry (continued) Table 10. Tone Ringer Frequency Encoding (continued) Hz 2,087 2,000 1,920 1,846 1,778 1,714 1,655 1,600 1,548 1,500 1,455 1,412 1,371 1,333 1,297 1,263 1,231 1,200 1,171 1,143 1,116 1,091 Hex 02 01 80 40 20 10 88 C4 E2 71 38 1C 8E 47 23 91 48 A4 D2 E9 74 3A Hz 727.3 716.4 705.9 695.7 685.7 676.1 666.7 657.5 648.6 640.0 631.6 623.4 615.4 607.6 600.0 592.6 585.4 578.3 571.4 564.7 558.2 551.7 Hex 3B 9D 4E 27 13 09 04 82 41 A0 50 A8 D4 6A B5 DA 6D B6 5B AD D6 6B Tone Ringer Frequency Encoding Hz Hex Hz Hex 440.4 F6 315.8 A3 436.4 7B 313.7 D1 432.4 BD 311.7 68 428.6 5E 309.7 B4 424.8 AF 307.7 5A 421.1 D7 305.7 2D 417.4 EB 303.8 96 413.8 75 301.9 4B 410.3 BA 300.0 25 406.8 5D 298.1 12 403.4 2E 296.3 89 400.0 17 294.5 44 396.7 8B 292.7 A2 393.4 45 290.9 51 390.2 22 289.2 28 387.1 11 287.4 94 384.0 08 285.7 4A 380.9 84 284.0 A5 377.9 C2 282.4 52 375.0 61 280.7 A9 372.1 B0 279.1 54 369.2 D8 — — Hz 247.4 246.2 244.9 243.7 242.4 241.2 240.0 238.8 237.6 236.5 235.3 234.1 233.0 231.9 230.8 229.7 228.6 227.5 226.4 225.4 224.3 — Hex F7 FB FD 7E BF 5F 2F 97 CB 65 32 99 CC 66 B3 59 AC 56 2B 15 8A — Hz 203.4 202.5 201.7 200.8 200.0 199.2 198.3 197.5 196.7 195.9 195.1 194.3 193.5 192.8 192.0 191.2 190.5 189.7 189.0 188.2 — — Hex 7D 3E 9F 4F A7 53 29 14 0A 85 42 21 90 C8 E4 F2 F9 FC FE FF — — 5.5 Audio Codec Block The T8301 contains a 16-bit analog-to-digital converter and two 16-bit digital-to-audio converters. These converters each contain appropriate antialiasing or smoothing filters. A block diagram of the audio codec block is shown below. . RC FILTER TO HANDSET OUTPUT OR SPEAKER DRIVER 2 DAC 16-bit 1 Mbit/s LOW-PASS RC FILTER DAC 16-bit RCV INTRP x4 SDM x 16 64 kS/s SDM x 16 1 Mbit/s FROM AIN MUX LOW-PASS RC FILTER ADC 16-bit FROM DMAS RCV LPF x2 16 kS/s 8 kS/s RCV INTRP x4 RCV LPF x2 FROM DMAH 8 kS/s 16 kS/s SINC3 DECM / 64 DMUX LOW-PASS AMUX TO SPEAKER DRIVER XMT BPF /2 16 TO DMAIN 5-8212 (F) Figure 5. Audio Codec Block Diagram 20 Lucent Technologies Inc. T8301 Internet Protocol Telephone Phone-On-A-Chip IP Solution DSP Advance Data Sheet December 2000 5 Audio Input/Output Circuitry (continued) 5.6 Audio Codec Control Registers The analog audio input and output control register (aioc_reg) is used to select the active and enabled inputs and outputs. Through this register the input and output channels can also have the clocks shut down to conserve power. Table 11. aioc_reg Analog Audio I/O Control Analog Audio Input and Output Control Register (aioc_reg): Address (0x4000) Bit # 15 14 13 12 11 10 9:8 Name MPWRD SPKFB HNDFB AINFB SPK2EN OLE RSVD Bit # 7 6 5 4 3 2 1:0 Name RSVD PGAS(2) PGAS(1) PGAS (0) SPKEN AOUTAEN AINSS[1:0] Bit # Name 15 MPWRD Value at Reset 1 14 SPKFB 0* Main powerdown. If 1, powerdown. If 0, powerup. Speaker #1 output filter bypass. If 1, the transmit LPF is bypassed in the speaker path; set the corresponding DMA clock to 16 kHz. If 0, the transmit LPF is enabled in the speaker path. 0* Note: The SOC bits in the audio codec clock control register should also be modified. Handset output filter bypass. If 1, the transmit LPF is bypassed in the handset path; set the corresponding DMA clock to 16 kHz. If 0, the transmit LPF is enabled in the handset path. 13 12 HNDFB AINFB 0* 11 SPK2EN 0 10 OLE 0 9:7 RSVD Description Note: The HOC bits in the audio codec clock control register should also be modified. Analog input filter bypass. If 1, the receive BPF is bypassed in the audio input path; set the corresponding DMA clock to 16 kHz. If 0, the receive BPF is enabled in the audio input path. Note: The AINC bits in the audio codec clock control register should also be modified. Enables speaker #2 output channel. If 1, the speaker’s output driver is enabled. If 0, the output driver for the speaker output channel is disabled. Output limit enable. When set, this bit causes the nominal full-scale output for the analog outputs to be limited to approximately half the normal value of 2.5 Vp-p Setting this bit has no effect on the receive gain. Reserved. * If the BPF is bypassed, output from the decimator must be shifted right by 2 bits (6 dB attenuation) to avoid saturation going into the compander. Similarly, if the LPF is bypassed in the speaker or handset path, input into the interpolator must be shifted left by 2 bits. Lucent Technologies Inc. 21 Advance Data Sheet December 2000 T8301 Internet Protocol Telephone Phone-On-A-Chip IP Solution DSP 5 Audio Input/Output Circuitry (continued) Table 11. aioc_reg Analog Audio I/O Control (continued) Bit # Name 6:4 PGAS[2:0] Value at Reset 000 3 SPKEN 0 2 AOUTAEN 0 1:0 AINSS 00 Description PGA gain select. Selects the gain for the programmable gain amplifier. See Table 7 on page 17 for an explanation of the coding. Enables the speaker output channel. If 1, the speaker’s output driver is enabled. If 0, the output driver for the speaker output channel is disabled. Enables the handset output channel. If 1, the handset output driver is enabled. If 0, the output driver for the handset output channel is disabled. Analog input source select. If 11, reserved. If 10, analog input source is from the microphone (AINCN, AINCP). If 01, analog input source is from the handset (AINAN). If 00, mute (default after reset or powerup). Table 12. Audio Codec Clock Control Register (aclkc_reg) Audio Codec Clock Control Register (aclkc_reg) Address (0x4003) Bit # 15:9 8:6 5:3 2:0 Name RSVD SOC(2:0) HOC(2:0) AINC(2:0) Bit # 15:9 8:6 5:3 2:0 Name RSVD SOC(2:0) HOC(2:0) AINC(2:0) Description Reserved. Please refer to Table 13 for bit field description. Please refer to Table 13 for bit field description. Please refer to Table 13 for bit field description. . Table 13. Audio Clock Encoding Bit Code 000 001 010 011 100 101 110 111 22 Audio Clock Encoding SOC, HOC, AINC Description 0 Hz. The clock for the channel is stopped. 8 kHz clock is used for all audio codes except G.722. 16 kHz clock is used for G.722 (must bypass filters). Reserved. Reserved. Reserved. Reserved. Supplies 1 MHz clock to DMA. Reserved for testing only. Lucent Technologies Inc. Advance Data Sheet December 2000 T8301 Internet Protocol Telephone Phone-On-A-Chip IP Solution DSP 6 DMA Input/Output Channels The discussions in this section pertain to circuitry that is outside of the dotted outline in Figure 3 on page 8. There are three timed DMA transfer blocks, each of which transfers data to/from the audio codec block from/to a 512 x 16-bit SRAM. These SRAMs are two-port devices. One port is connected to the DSP1627 address and data bus, and the other is accessed by the DMA controller. These memories transfer data to/from the audio codec block or AOUTA, AIN, and SOUT. These DMA blocks are capable of transferring a 16-bit word to/from the device’s A/D or D/A at the following rates, which are set up by programming the audio codec clock control register: ■ 8 kHz ■ 16 kHz Each channel initiates a transfer between the audio codec block and its respective SRAM on the rising edge of the selected transfer clock. 6.1 DMA Operation The T8301 has three timed DMA transfer channels. The DSP sets up a DMA channel by writing a starting address and a transfer count into the setadr_reg (see Table 15) and setcnt_reg (see Table 16). The DSP then sets the channel’s GO bit in the dmac_reg (see Table 14). When the DMA finishes its current transfer operation, indicated by the BSY bit in the dmac_reg going low, the DMA will transfer the contents of the setadr_reg (see Table 15) to the adrinc_reg (see Table 17) and the cntdec_reg (see Table 18) respectively. The GO bit will be reset to zero and the BSY bit will be set to one, in the dmac_reg on completion of this transfer. When the rising edge of the transfer clock is detected, the DMA controller will transfer a single word to/from memory and the audio codec block. The DMA channel will then increment its address pointer adrinc_reg and decrement its counter cntdec_reg. At the completion of the number of transfers written into the transfer counter (cntdec_reg = 0), the DMA block will set its ION bit in the dmac_reg to 1 and reset its BSY bit to zero. If its IEN bit is set, an interrupt to the DSP will occur. If the DSP has set the GO bit which indicates that it has set up a new transfer or if the DSP responds (sets up a new transfer count and re-enables transfers) before the next rising edge of the transfer clock, data can be continuously transferred at the clocked rate. If the DSP is reading or writing to the memory that a timed DMA is transferring to/from, the DMA can be delayed by a clock cycle to allow the DSP to finish its access. 6.2 DMA Registers Each DMA channel has the following four registers: ■ Starting address register ■ Transfer count register ■ Working address increment register (read only) ■ Working count decrement register (read only) In addition, there is a control and status register that supports all three DMA channels. Lucent Technologies Inc. 23 Advance Data Sheet December 2000 T8301 Internet Protocol Telephone Phone-On-A-Chip IP Solution DSP 6 DMA Input/Output Channels (continued) Table 14. DMA Control Register dmac_reg DMA Control Register (dmac_reg) Address (0x4010) Bit # 15 14 13 12 11 10 9 8 Name RSVD IENSPK IENHND IENAIN RSVD IONSPK IONHND IONAIN Bit # 7 6 5 4 3 2 1 0 Name RSVD SPKBSY HNDBSY AINBSY RSVD SPKGO HNDGO AINGO Bit # 15 14 13 12 11 10 Name RSVD IENSPK IENHND IENAIN RSVD IONSPK 9 IONHND 8 IONAIN 7 6 5 4 3 2 RSVD SPKBSY HNDBSY AINBSY RSVD SPKGO 1 HNDGO 0 AINGO Description Reserved. Interrupt enable speaker output channel. Interrupt enable handset output channel. Interrupt enable analog input channel. Reserved. Interrupt on speaker DMA channel. Indicates a transfer has completed. A physical interrupt to the DSP will only occur if the IENSPK bit is also set. The interrupt is cleared by a read operation. Interrupt on handset DMA channel. Indicates a transfer has completed. A physical interrupt to the DSP will only occur if the IENHND bit is also set. The interrupt is cleared by a read operation. Interrupt on analog input DMA channel. Indicates a transfer has completed. A physical interrupt to the DSP will only occur if the IENAIN bit is also set. The interrupt is cleared by a read operation. Reserved. Speaker DMA channel busy (read only). Handset DMA channel busy (read only). Analog input DMA channel busy (read only). Reserved. DMA start. Starts the DMA channel when set to 1, automatically reset to zero when a count of zero is reached by the DMA transfer counter. DMA start. Starts the DMA channel when set to 1, automatically reset to zero when a count of zero is reached by the DMA transfer counter. DMA start. Starts the DMA channel when set to 1, automatically reset to zero when a count of zero is reached by the DMA transfer counter. Table 15. DMA Starting Address Register setadr_reg Set DMA Address Registers [AINsetadr_reg Address (0x4014)] [HNDsetadr_reg Address (0x4018)] [SPKsetadr_reg Address (0x401C)] 15:9 8:0 Bit # RSVD DMA_ADDRESS_SET_UP[8:0] Name Table 16. DMA Transfer Count Register setcnt_reg Set DMA Count Registers [AINsetcnt_reg Address (0x4015)] [HNDsetcnt_reg Address (0x4019)] [SPKsetcnt_reg Address (0x401D)] 15:9 8:0 Bit # RSVD DMA_COUNT_SET_UP[8:0] Name 24 Lucent Technologies Inc. Advance Data Sheet December 2000 T8301 Internet Protocol Telephone Phone-On-A-Chip IP Solution DSP 6 DMA Input/Output Channels (continued) Table 17. DMA Address Increment Register adrinc_reg DMA Address Increment Registers (Read Only) [AINadrinc_reg Address (0x4016)] [HNDadrinc_reg Address (0x401A)] [SPKadrinc_reg Address (0x401E)] 15:9 8:0 Bit # RSVD DMA_ADDRESS[8:0] Name Table 18. DMA Transfer Decrement Register cntdec_reg Bit # Name DMA Count Decrement Registers (Read Only) [AINcntdec_reg Address (0x4017)] [HNDcntdec_reg Address (0x401B)] [SPKcntdec_reg Address (0x401F)] 15:9 8:0 RSVD DMA_COUNT[8:0] Lucent Technologies Inc. 25 Advance Data Sheet December 2000 T8301 Internet Protocol Telephone Phone-On-A-Chip IP Solution DSP 7 Hardware Compander The discussions in this section pertain to circuitry that is outside of the dotted outline in Figure 3 on page 8. The hardware compander performs companded-to-linear and linear-to-companded conversions. This alleviates the DSP from performing the functions in firmware. The compander supports both µ-law and A-law operations. A block diagram of the compander is shown in Figure 6 on page 27. The compander consists of two write-only registers: write_linear and write_companded. A configuration register (config_compander) and two read registers (read _linear and read_ companded) read the results. Config_compand configures the compander for either µ-law or A-law conversion. Upon reset, the register defaults to µ-law. The DSP performs a linear-to-companded conversion by writing the write_linear register and then reading the read-companded buffer. The companded value at the read buffer remains the same until a new linear value is written to the write_linear register. Similarly, companded to linear is done by write_companded then read_linear. Table 19. config_compander Register Compander Configuration Register (config_compander) 15:1 Bit Reserved Field X After Reset 0 µ−Law 1 = µ−Law 0 = A-Law Table 20. write_linear Register write_linear 15:0 Linear value Bit — Table 21. write_companded Register write_companded 15:0 Companded value Bit — Table 22. read_linear Register read_linear 15:0 Linear value Bit — Table 23. read_companded Register read_companded Bit — 26 15:0 Companded value Lucent Technologies Inc. T8301 Internet Protocol Telephone Phone-On-A-Chip IP Solution DSP Advance Data Sheet December 2000 7 Hardware Compander (continued) µ-LAW CONFIG_COMPANDER WRITE_LINEAR DSP_D[15:0] READ_COMPANDED BUFFER COMPANDER COMBINATORIAL LOGIC DSP1627 WRITE_COMPANDED READ_LINEAR BUFFER 5-8209(F) Figure 6. Hardware Compander Block Diagram The Config_compander register configures the compander for either µ-law or A-law conversion. Upon reset, the register defaults to µ-law, see Table 19 on page 26. Lucent Technologies Inc. 27 Advance Data Sheet December 2000 T8301 Internet Protocol Telephone Phone-On-A-Chip IP Solution DSP 8 Electrical Specifications 8.1 Operating Range Specifications Table 24. Operating Range Specifications Parameter Symbol Min Max Unit TA 0 70 °C VDD 4.75 5.25 V P — 900 mW Ambient Temperature Range Operating Supply Voltage Power Consumption 8.2 Analog and Codec Specifications Table 25. AINAN Specifications Parameter Conditions Value ac-coupled with a 0.2 µF capacitor 1 kΩ—3 kΩ With ac-coupled +2.5 Vp-p input signal (max PGA gain) 6 kΩ—12 kΩ Input signals 100 mV—2.5 Vp-p ≤2% PGA set 12 dB ≤20 dBrnC — ≥50 dB Source Impedance* Input Impedance Total Harmonic Distortion Transmit Idle Channel Noise Power Supply Rejection Ratio * Parameter supplied for reference purposes. Table 26. AINCP, AINCN Specifications Parameter Source Impedance* Input Impedance Total Gain Total Harmonic Distortion Conditions Value ac-coupled w/ 0.2 µF capacitor 1 kΩ— 3 kΩ With ac-coupled 40 mVp-p 12 kΩ—20 kΩ — 30 dB ± 1 dB Input signals 1 mV—40 mV ≤2% Total harmonic distortion input signals 1 mV—40 mV ≤2% * Parameter supplied for reference purposes. Table 27. AOUTA Specifications Parameter VOUT VOUT Device impedance Total harmonic distortion (3.0 dBm0) Total harmonic distortion (0.0 dBm0) 28 Conditions 0 dBm0 3.14 dBm0 — –35 dB max (µA limit) 0.0 dB max (µA limit) Value 0.618 Vrms (±0.5 dB) 2.50 Vp-p typical 150 Ω –40 dB max –65 dB max Lucent Technologies Inc. T8301 Internet Protocol Telephone Phone-On-A-Chip IP Solution DSP Advance Data Sheet December 2000 8 Electrical Specifications (continued) Table 28. Speaker#1, Speaker#2 Specifications Parameter VOUT VOUT Device Impedance Total Gain Total Harmonic Distortion Total Harmonic Distortion Conditions 0 dBm0 3.14 dBm0 — — –35 dB max (µA limit) 0.0 dB max (µA limit) Value 1.236 Vrms (±0.5 dB) 5.00 Vp-p typical 45 Ω 6 dB ± 0.25 dB –40 dB max –65 dB max Note: Maximum digital-to-analog converter range = 2.5 V. This translates into a peak-to-peak differential signal of 5.0 V. All signals measured differentially. Table 29. Digital Low-Pass Filters Specifications Parameter Maximum Ripple Pass-Band Minimum Attenuation Conditions 300 Hz ≤ signal frequency ≤ 3.0 kHz 4 kHz Value 3% 30 dB . Table 30. Digital-to-Analog Converter Specifications Parameter Conditions — Full operating range Full operating range — — — Value 16-bit Monotonic TBD 2.5 Vp-p 1.5 LSB Two’s complement Conditions Value — 16-bit Monotonicity Full operating range Monotonic Accuracy Full operating range TBD Max Step-to-step Size — 1.5 LSB Full Scale Input — 2.5 Vp-p Output Code — Two’s complement Range Monotonicity Accuracy Full Scale Output Max Voltage Change 1-bit Change Input Code . Table 31. Analog-to-Digital Converter Specifications Parameter Range 8.3 Crystal Specification See the DSP1627 Digital Signal Processor Data Sheet for further information. Lucent Technologies Inc. 29 T8301 Internet Protocol Telephone Phone-On-A-Chip IP Solution DSP Advance Data Sheet December 2000 9 JTAG and Hardware Development System (HDS) The JTAG block contains logic for implementing the JTAG/IEEE* P1149.1 standard. A four-signal test port provides a mechanism for accessing the DSP1627 core from remote test equipment or a remote hardware development system. The on-chip HDS performs instruction breakpointing and branch tracing at full speed. Using the JTAG port, the breakpointing is set up and the trace history is read back. 9.1 TMODE Control for JCS/Boundary-Scan Operation TMODEN0, TMODEN1, and TMODEN2 are inputs used to determine test mode operation. Of the eight possible combinations, modes 6 and 7 are significant during the development and production phases. 9.1.1 Mode 7 Operation (TMODE = 7) This is the production mode. Internal pull-up resistors (approximately 50 kΩ) will provide the logic level required. The three pins can be left floating (no external resistors are required). In this mode, boundary-scan is active. The CK8KHz (pin 67), the CK2MHz (pin 98), and the CKO (pin 99) are all dormant (high). 9.1.2 Mode 6 Operation (TMODE = 6) The JCS tools (JTAG communications system) are used in this mode. TMODEN0 must be pulled low externally, TMODEN1, and TMODEN2 can both be left floating to enter this mode. The CK8KHz (pin 67), the CK2MHz (pin 98), and the CKO (pin 99) are active. Should the user require access to any or all of the three clocks in production and still require boundary-scan capabilities for production test, a strong (external) pull-down resistor would be required on TMODEN0 (1 kΩ). The production test must be able to pull TMODEN0 high to allow access to the boundary-scan test. After the test is complete, the pin would normally be low (TMODE 6) allowing the clocks to be active. 9.2 The Principle of Boundary-Scan Architecture Each primary input signal and primary output signal is supplemented with a multipurpose memory element called a boundary-scan cell. Cells on device primary inputs are referred to as input cells and cells on primary outputs are referred to as output cells. Input and output is relative to the core logic of the device. At any time, only one register can be connected from TDI to TDO, e.g., the instruction register (IR), BYPASS, boundary-scan, IDENT, or even some appropriate register internal to the core logic; see Figure 7. The selected register is identified by the decoded output of the instruction register. Certain instructions are mandatory, such as EXTEST (boundary-scan register selected), whereas others are optional, such as the IDCODE instruction (IDENT register selected). * IEEE is a registered trademark of The Institute of Electrical and Electronics Engineers, Inc. 30 Lucent Technologies Inc. T8301 Internet Protocol Telephone Phone-On-A-Chip IP Solution DSP Advance Data Sheet December 2000 9 JTAG and Hardware Development System (HDS) (continued) INTERNAL CORE LOGIC TDO TDI BYPASS TEST DATA IN TEST DATA OUT IDENTIFICATION REGISTER INSTRUCTION REGISTER (IR) TEST MODE SELECT TEST CLOCK TMS TAP TCK CONTROLLER TEST RESET (TRSTN) IEEE 1149.1 CHIP ARCHITECTURE Figure 7. Boundary-Scan Architecture Figure 7 shows the following elements: ■ A set of four dedicated test pins, test data in (TDI), test mode select (TMS), test clock (TCK), test data out (TDO), and one optional test pin test reset (TRSTN). These pins are collectively referred to as the test access port (TAP). ■ A boundary-scan cell on each device’s primary input and primary output pin, connected internally to form a serial boundary-scan register (boundary-scan). ■ A finite-state machine TAP controller with inputs TCK and TMS. ■ An n-bit (n = 4) instruction register (IR), holding the current instruction. ■ A 1-bit bypass register (BYPASS). ■ An optional 32-bit identification register (IDENT) capable of being loaded with a permanent device identification code. Lucent Technologies Inc. 31 Advance Data Sheet December 2000 T8301 Internet Protocol Telephone Phone-On-A-Chip IP Solution DSP 9 JTAG and Hardware Development System (HDS) (continued) Access to JTAG (joint test action group) and boundary-scan will be initially provided through a single set of JTAG pins. The pin definitions are as follows. Table 32. Boundary-Scan Pin Functions Pin 94 95 96 97 89 Boundary-Scan TDO (bscan) TCK (bscan) TMS (bscan) TDI (bscan) TRSTN (bscan) Debug TDO (debug) TCK (debug) TMS (debug) TDI (debug) TRSTN (debug) Comments — Pulled high internally Pulled high internally — Pulled high internally Debug mode, or boundary-scan mode is selected via the TMODE pins as shown below. Table 33. Debug Mode Pin Name 90 TMODEN0 91 TMODEN1 92 TMODEN2 Description If 7 = boundary-scan If 6 = debug Comments Pulled high internally Pulled high internally Pulled high internally 9.2.1 Boundary-Scan Instruction Register The boundary-scan instruction register is 4 bits long and the capture value is 0001. Table 34. Boundary-Scan Instruction Register Instruction EXTEST SAMPLE IDCODE BYPASS Binary Code 0000 0001 0101 1111 Description Places the boundary-scan register in EXTEST mode. Places the boundary-scan register in sample mode. Identification code. Places the bypass register in the scan chain. The idcode values are as follows: Version = 0000 (0x0) Part = 0011011101000110 (0x 3746) Manufacturer = 00000011101 (0x1D) 32 Lucent Technologies Inc. T8301 Internet Protocol Telephone Phone-On-A-Chip IP Solution DSP Advance Data Sheet December 2000 9 JTAG and Hardware Development System (HDS) (continued) Table 35. Boundary-Scan Register Description Boundary-Scan Register Bit Pin 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 Pin Name Ball Enabled State Pin Grouping Control Disable Value BIO_E(0) BIO(0) BIO_E(1) BIO(1) BIO_E(2) BIO(2) BIO_E(3) BIO(3) INT0N INT1N STOPN DI1 DO1_E DO1 SYNC_E SYNC IOLD_E IOLD IOCK_E IOCK DSP_A_E A(15) A(14) A(13) A(12) A(11) A(10) A(9) A(8) A(7) A(6) A(5) A(4) A(3) A(2) A(1) A(0) I_CSN M_CSN X_CSN RWN — 2 — 3 — 4 — 5 6 7 8 9 — 12 — 13 — 14 — 15 — 21 22 23 24 27 28 29 30 31 32 33 34 37 38 39 40 41 42 43 44 Controller I/O Controller I/O Controller I/O Controller I/O Input Input Input Input Controller I/O Controller I/O Controller I/O Controller I/O Controller I/O I/O I/O I/O I/O I/O I/O I/O I/O I/O I/O I/O I/O I/O I/O I/O I/O I/O I/O I/O — BIO_E(0) — BIO_E(1) — BIO_E(2) — BIO_E(3) — — — — — DO1_E — SYNC_E — IOLD_E — IOCK_E — A_E A_E A_E A_E A_E A_E A_E A_E A_E A_E A_E A_E A_E A_E A_E A_E A_E A_E A_E A_E — 0 — 0 — 0 — 0 — — — — — 0 — 0 — 0 — 0 — 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 — High Impedance — High Impedance — High Impedance — High Impedance — — — — — High Impedance — High Impedance — High Impedance — High Impedance — High Impedance High Impedance High Impedance High Impedance High Impedance High Impedance High Impedance High Impedance High Impedance High Impedance High Impedance High Impedance High Impedance High Impedance High Impedance High Impedance High Impedance High Impedance High Impedance High Impedance Lucent Technologies Inc. 33 Advance Data Sheet December 2000 T8301 Internet Protocol Telephone Phone-On-A-Chip IP Solution DSP 9 JTAG and Hardware Development System (HDS) (continued) Table 35. Boundary-Scan Register Description (continued) Boundary-Scan Pin Name Register Bit Pin 41 D_E 42 D(15) 43 D(14) 44 D(13) 45 D(12) 46 D(11) 47 D(10) 48 D(9) 49 D(8) 50 D(7) 51 D(6) 52 D(5) 53 D(4) 54 D(3) 55 D(2) 56 D(1) 57 D(0) 58 CLK_E 59 CK8KHZ 60 STCK_E 61 STCK 62 STI_E 63 STI1 64 STO_E 65 STO1 66 RESETN_E 67 RESETN 68 CK2MHZ 69 CKO 34 Ball Enabled State Pin Grouping Control Disable Value — 47 48 49 50 51 52 53 54 57 58 59 60 61 62 63 64 — 67 — 68 — 70 — 69 — 93 98 99 Controller I/O I/O I/O I/O I/O I/O I/O I/O I/O I/O I/O I/O I/O I/O I/O I/O Controller I/O Controller I/O Controller I/O Controller I/O Controller I/O I/O I/O — D_E D_E D_E D_E D_E D_E D_E D_E D_E D_E D_E D_E D_E D_E D_E D_E — CLK_E — STCK_E — STI_E — STO_E — RESETN_E CLK_E CLK_E — 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 — 0 — 0 — 0 — 0 — 0 0 0 — High Impedance High Impedance High Impedance High Impedance High Impedance High Impedance High Impedance High Impedance High Impedance High Impedance High Impedance High Impedance High Impedance High Impedance High Impedance High Impedance — High Impedance — High Impedance — High Impedance — High Impedance — High Impedance High Impedance High Impedance Lucent Technologies Inc. Advance Data Sheet December 2000 T8301 Internet Protocol Telephone Phone-On-A-Chip IP Solution DSP Notes Lucent Technologies Inc. 35 For additional information, contact your Microelectronics Group Account Manager or the following: INTERNET: http://www.lucent.com/micro E-MAIL: [email protected] N. AMERICA: Microelectronics Group, Lucent Technologies Inc., 555 Union Boulevard, Room 30L-15P-BA, Allentown, PA 18109-3286 1-800-372-2447, FAX 610-712-4106 (In CANADA: 1-800-553-2448, FAX 610-712-4106) ASIA PACIFIC: Microelectronics Group, Lucent Technologies Singapore Pte. Ltd., 77 Science Park Drive, #03-18 Cintech III, Singapore 118256 Tel. (65) 778 8833, FAX (65) 777 7495 CHINA: Microelectronics Group, Lucent Technologies (China) Co., Ltd., A-F2, 23/F, Zao Fong Universe Building, 1800 Zhong Shan Xi Road, Shanghai 200233 P. R. China Tel. (86) 21 6440 0468, ext. 325, FAX (86) 21 6440 0652 JAPAN: Microelectronics Group, Lucent Technologies Japan Ltd., 7-18, Higashi-Gotanda 2-chome, Shinagawa-ku, Tokyo 141, Japan Tel. (81) 3 5421 1600, FAX (81) 3 5421 1700 EUROPE: Data Requests: MICROELECTRONICS GROUP DATALINE: Tel. (44) 7000 582 368, FAX (44) 1189 328 148 Technical Inquiries: GERMANY: (49) 89 95086 0 (Munich), UNITED KINGDOM: (44) 1344 865 900 (Ascot), FRANCE: (33) 1 40 83 68 00 (Paris), SWEDEN: (46) 8 594 607 00 (Stockholm), FINLAND: (358) 9 3507670 (Helsinki), ITALY: (39) 02 6608131 (Milan), SPAIN: (34) 1 807 1441 (Madrid) Lucent Technologies Inc. reserves the right to make changes to the product(s) or information contained herein without notice. No liability is assumed as a result of their use or application. No rights under any patent accompany the sale of any such product(s) or information. Phone-On-A-Chip is a trademark of Lucent Technologies Inc. Copyright © 2000 Lucent Technologies Inc. All Rights Reserved Printed in U.S.A. December 2000 DS01-025IPT (Replaces DS00-030IPT-3) Data Sheet March 2000 DSP1627 Digital Signal Processor 1 Features ■ Optimized for digital cellular applications with a bit manipulation unit for higher coding efficiency. ■ On-chip, programmable, PLL clock synthesizer. ■ 14 ns and 11 ns instruction cycle times at 5 V, 10 ns instruction cycle time at 3.0 V, and 20 ns and 12.5 ns instruction cycle times at 2.7 V, respectively. ■ Mask-programmable memory map option: The DSP1627x36 features 36 Kwords on-chip ROM. The DSP1627x32 features 32 Kwords on-chip ROM and access to 16 Kwords external ROM in the same map. Both feature 6 Kwords on-chip, dual-port RAM and a secure option for on-chip ROM. ■ Low power consumption: — <5.5 mW/MIPS typical at 5 V. — <1.5 mW/MIPS typical at 2.7 V. ■ Flexible power management modes: — Standard sleep: 0.5 mW/MIPS at 5 V. 0.12 mW/MIPS at 2.7 V. — Sleep with slow internal clock: 1.4 mW at 5 V. 0.4 mW at 2.7 V. — Hardware STOP (pin halts DSP): <20 µA. ■ Full-speed in-circuit emulation hardware development system on-chip. ■ Supported by DSP1627 software and hardware development tools. 2 Description The DSP1627 is Lucent Technologies Microelectronics Group first digital signal processor offering 100 MIPS operation at 3.0 V and 80 MIPS operation at 2.7 V with a reduction in power consumption. Designed specifically for applications requiring low power dissipation in digital cellular systems, the DSP1627 is a signal-coding device that can be programmed to perform a wide variety of fixed-point signal processing functions. The device is based on the DSP1600 core with a bit manipulation unit for enhanced signal coding efficiency. The DSP1627 includes a mix of peripherals specifically intended to support processingintensive but cost-sensitive applications in the area of digital wireless communications. The DSP1627x36 contains 36 Kwords of internal ROM (IROM), but it doesn’t support the use of IROM and external ROM (EROM) in the same memory map. The DSP1627x32 supports the use of 32 Kwords of IROM with 16 Kwords of EROM in the same map. Both devices contain 6 Kwords of dual-port RAM (DPRAM), which allows simultaneous access to two RAM locations in a single instruction cycle. ■ Mask-programmable clock options: crystal oscillator, small signal, and CMOS. ■ Low-profile TQFP package (1.5 mm) available. ■ Sequenced accesses to X and Y external memory. ■ Object code compatible with the DSP1617. ■ Single-cycle squaring. ■ 16 x 16-bit multiplication and 36-bit accumulation in one instruction cycle. ■ Instruction cache for high-speed, program-efficient, zerooverhead looping. ■ Dual 25 Mbits/s serial I/O ports with multiprocessor capability—16-bit data channel, 8-bit protocol channel. ■ 8-bit parallel host interface: — Supports 8- or 16-bit transfers. — Motorola* or Intel † compatible. ■ 8-bit control I/O interface. ■ 256 memory-mapped I/O ports. ■ IEEE‡ P1149.1 test port (JTAG boundary scan). The on-chip clock synthesizer can be driven by an external clock whose frequency is a fraction of the instruction rate. * Motorola is a registered trademark of Motorola, Inc. † Intel is a registered trademark of Intel Corp. ‡ IEEE is a registered trademark of The Institute of Electrical and Electronics Engineers, Inc. The device is packaged in a 100-pin BQFP or a 100-pin TQFP and is available with 14 ns and 11 ns instruction cycle times at 5 V, 10 ns instruction cycle times at 3.0 V, and 20 ns and 12.5 ns instruction cycle times at 2.7 V, respectively. The DSP1627 is object code compatible with the DSP1617, while providing more memory and architectural enhancements including an on-chip clock synthesizer and an 8-bit parallel host interface for hardware flexibility. The DSP1627 supports 2.7 V, 3.0 V, and 5 V operation and flexible power management modes required for portable cellular terminals. Several control mechanisms achieve lowpower operation, including a STOP pin for placing the DSP into a fully static, halted state and a programmable power control register used to power down unused on-chip I/O units. These power management modes allow for trade-offs between power reduction and wake-up latency requirements. During system standby, power consumption is reduced to less than 20 µA. Data Sheet March 2000 DSP1627 Digital Signal Processor Table of Contents Contents Page Features.............................................................. 1 Description .......................................................... 1 Pin Information.................................................... 3 Hardware Architecture ........................................ 7 4.1 DSP1627 Architectural Overview ............. 7 4.2 DSP1600 Core Architectural Overview .. 10 4.3 Interrupts and Trap ................................. 11 4.4 Memory Maps and Wait-States .............. 16 4.5 External Memory Interface (EMI)............ 18 4.6 Bit Manipulation Unit (BMU) ................... 19 4.7 Serial I/O Units (SIOs) ............................ 19 4.8 Parallel Host Interface (PHIF)................. 22 4.9 Bit Input/Output Unit (BIO)...................... 23 4.10 Timer ...................................................... 23 4.11 JTAG Test Port....................................... 24 4.12 Clock Synthesis ...................................... 26 4.13 Power Management ............................... 29 5 Software Architecture ....................................... 36 5.1 Instruction Set......................................... 36 5.2 Register Settings .................................... 45 5.3 Instruction Set Formats .......................... 55 6 Signal Descriptions ........................................... 61 6.1 System Interface..................................... 61 6.2 External Memory Interface ..................... 63 6.3 Serial Interface #1 .................................. 64 6.4 Parallel Host Interface or Serial Interface #2 and Control I/O Interface .... 65 6.5 Control I/O Interface ............................... 65 6.6 JTAG Test Interface ............................... 66 7 Mask-Programmable Options ........................... 67 7.1 Input Clock Options ................................ 67 7.2 Memory Map Options ............................. 67 7.3 ROM Security Options............................ 67 8 Device Characteristics ...................................... 68 8.1 Absolute Maximum Ratings.................... 68 8.2 Handling Precautions ............................. 68 8.3 Recommended Operating Conditions .... 68 8.4 Package Thermal Considerations .......... 69 9 Electrical Characteristics and Requirements .... 70 9.1 Power Dissipation................................... 73 10 Timing Characteristics for 5 V Operation .......... 75 10.1 DSP Clock Generation ........................... 76 10.2 Reset Circuit ........................................... 77 10.3 Reset Synchronization............................ 78 Contents 1 2 3 4 2 11 12 13 14 Page 10.4 JTAG I/O Specifications.......................... 79 10.5 Interrupt .................................................. 80 10.6 Bit Input/Output (BIO) ............................. 81 10.7 External Memory Interface...................... 82 10.8 PHIF Specifications ................................ 86 10.9 Serial I/O Specifications.......................... 92 10.10 Multiprocessor Communication .............. 97 Timing Characteristics for 3.0 V Operation ....... 98 11.1 DSP Clock Generation............................ 99 11.2 Reset Circuit ......................................... 100 11.3 Reset Synchronization.......................... 101 11.4 JTAG I/O Specifications........................ 102 11.5 Interrupt ................................................ 103 11.6 Bit Input/Output (BIO) ........................... 104 11.7 External Memory Interface.................... 105 11.8 PHIF Specifications .............................. 109 11.9 Serial I/O Specifications........................ 115 11.10 Multiprocessor Communication ............ 120 Timing Characteristics for 2.7 V Operation ..... 121 12.1 DSP Clock Generation.......................... 122 12.2 Reset Circuit ......................................... 123 12.3 Reset Synchronization.......................... 124 12.4 JTAG I/O Specifications........................ 125 12.5 Interrupt ................................................ 126 12.6 Bit Input/Output (BIO) ........................... 127 12.7 External Memory Interface.................... 128 12.8 PHIF Specifications .............................. 132 12.9 Serial I/O Specifications........................ 138 12.10 Multiprocessor Communication ............ 143 Crystal Electrical Characteristics and Requirements.................................................. 144 13.1 External Components for the Crystal Oscillator............................................... 144 13.2 Power Dissipation ................................. 144 13.3 LC Network Design for Third Overtone Crystal Circuits...................... 147 13.4 Frequency Accuracy Considerations .... 149 Outline Diagrams ............................................ 152 14.1 100-Pin BQFP (Bumpered Quad Flat Pack).............................................. 152 14.2 100-Pin TQFP (Thin Quad Flat Pack)... 153 Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor VSS VSS DO1 90 SYNC1 OLD1 OCK1 ICK1 ILD1 VSS DI1 OBE1 IBF1 VDD DB15 100 DB14 DB13 DB12 DB11 VSS DB10 DB9 DB8 DB7 10 DB6 DB5 VDD 3 Pin Information VDD SADD1 DB4 DOEN1 DB3 PIN #1 IDENTIFIER ZONE DB2 DB1 OCK2/PCSN DO2/PSTAT DB0 IO SYNC2/PBSEL ILD2/PIDS 20 OLD2/PODS ERAMHI 80 VDD IBF2/PIBF OBE2/POBE ERAMLO ICK2/PB0 EROM DI2/PB1 RWN VSS VSS DSP1627 EXM DOEN2/PB2 AB15 AB14 SADD2/PB3 VDD VDD IOBIT0/PB4 30 IOBIT1/PB5 AB13 70 AB12 AB11 IOBIT2/PB6 IOBIT3/PB7 VEC1/IOBIT6 AB7 VEC0/IOBIT7 VSS VSS VSSA CKI2 CKI VDDA TDI TDO TMS VDD TCK CKO RSTB TRAP STOP VSS IACK INT0 INT1 AB0 AB1 AB2 AB3 AB4 AB5 VDD AB6 60 VEC2/IOBIT5 AB8 50 VEC3/IOBIT4 AB9 40 AB10 5-4218 (F).b Figure 1. DSP1627 BQFP Pin Diagram Lucent Technologies Inc. 3 Data Sheet March 2000 DSP1627 Digital Signal Processor VSS SYNC1 DO1 OLD1 OCK1 ICK1 ILD1 DI1 VSS IBF1 OBE1 VDD DB15 DB14 80 1 DB13 DB12 DB11 VSS DB10 DB9 DB8 DB7 DB5 DB6 90 VSS 100 VDD 3 Pin Information (continued) VDD DB4 SADD1 DB3 DOEN1 DB2 OCK2/PCSN DB1 DO2/PSTAT DB0 70 IO ILD2/PIDS ERAMHI OLD2/PODS VDD ERAMLO SYNC2/PBSEL IBF2/PIBF OBE2/POBE 10 EROM ICK2/PB0 RWN DI2/PB1 VSS VSS DSP1627 EXM DOEN2/PB2 AB15 SADD2/PB3 AB14 60 VDD VDD IOBIT0/PB4 AB13 IOBIT1/PB5 AB12 AB11 IOBIT2/PB6 IOBIT3/PB7 20 VEC2/IOBIT5 AB8 VEC1/IOBIT6 AB7 VEC0/IOBIT7 VSS VSS VSSA CKI2 CKI VDDA TDI TDO TMS VDD TCK CKO RSTB TRAP STOP IACK VSS INT0 AB0 INT1 AB1 AB2 30 AB3 AB4 AB5 VDD AB6 50 VEC3/IOBIT4 AB9 40 AB10 5-4219 (F).b Figure 2. DSP1627 TQFP Pin Diagram 4 Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 3 Pin Information (continued) Functional descriptions of pins 1—100 are found in Section 6, Signal Descriptions. The functionality of pins 61 and 62 (TQFP pins 48 and 49) are mask-programmable (see Section 7, Mask-Programmable Options). Input levels on all I and I/O type pins are designed to remain at full CMOS levels when not driven by the DSP. Table 1. Pin Descriptions BQFP Pin 1, 2, 3, 4, 5, 7, 8, 9, 10, 11, 12, 15, 16, 17, 18, 19 20 21 23 24 25 27 28, 29, 31, 32, 33, 34, 35, 36, 37, 40, 41, 42, 43, 44, 45, 46 47 48 50 51 52 53 54 56 57 58 59 TQFP Pin 88, 89, 90, 91, 92, 94, 95, 96, 97, 98, 99, 2, 3, 4, 5, 6 7 8 10 11 12 14 15, 16, 18, 19, 20, 21, 22, 23, 24, 27, 28, 29, 30, 31, 32, 33 34 35 37 38 39 40 41 43 44 45 46 61 62 65 66 67 68 48 49 52 53 54 55 Symbol DB[15:0] Type Name/Function I/O* External Memory Data Bus DB[15:0]. IO ERAMHI ERAMLO EROM RWN EXM AB[15:0] O† O† O† O† O† I O* Data Address 0x4000 to 0x40FF I/O Enable. Data Address 0x8000 to 0xFFFF External RAM Enable. Data Address 0x4100 to 0x7FFF External RAM Enable. Program Address External ROM Enable. Read/Write Not. External ROM Enable. External Memory Address Bus 15—0. INT1 INT0 IACK STOP TRAP RSTB CKO TCK TMS TDO TDI I I O* I I/O* I O† I I‡ O§ I‡ CKI** CKI2** VEC0/IOBIT7 VEC1/IOBIT6 VEC2/IOBIT5 VEC3/IOBIT4 I I I/O* I/O* I/O* I/O* Vectored Interrupt 1. Vectored Interrupt 0. Interrupt Acknowledge. STOP Input Clock. Nonmaskable Program Trap/Breakpoint Indication. Reset Bar. Processor Clock Output. JTAG Text Clock. JTAG Test Mode Select. JTAG Test Data Output. JTAG Test Data Input. Mask-Programmable Input Clock Option CMOS Small Crystal Signal Oscillator XLO, 10 pF capacitor to VSS CKI VAC VCM XHI, 10 pF capacitor to VSS VSSA Vectored Interrupt Indication 0/Status/Control Bit 7. Vectored Interrupt Indication 1/Status/Control Bit 6. Vectored Interrupt Indication 2/Status/Control Bit 5. Vectored Interrupt Indication 3/Status/Control Bit 4. CMOS CKI Open * 3-states when RSTB = 0, or by JTAG control. † 3-states when RSTB = 0 and INT0 = 1. Output = 1 when RSTB = 0 and INT0 = 0, except CKO which is free-running. ‡ Pull-up devices on input. § 3-states by JTAG control. ** See Section 7, Mask-Programmable Options. †† For SIO multiprocessor applications, add 5 kΩ external pull-up resistors to SADD1 and/or SADD2 for proper initialization. Lucent Technologies Inc. 5 Data Sheet March 2000 DSP1627 Digital Signal Processor 3 Pin Information (continued) Functional descriptions of pins 1—100 are found in Section 6, Signal Descriptions. Table 1. Pin Descriptions (continued) BQFP Pin 69 70 71 72 74 75 77 78 79 80 81 82 83 84 85 86 87 90 91 92 93 94 95 96 98 99 6, 15, 26, 38, 49, 64, 76, 89, 97 14, 22, 30, 39, 55, 73, 88, 100 60 63 TQFP Pin Symbol 56 IOBIT3/PB7 57 IOBIT2/PB6 58 IOBIT1/PB5 59 IOBIT0/PB4 61 SADD2/PB3†† 62 DOEN2/PB2 64 DI2/PB1 65 ICK2/PB0 66 OBE2/POBE 67 IBF2/PIBF 68 OLD2/PODS 69 ILD2/PIDS 70 SYNC2/PBSEL 71 DO2/PSTAT 72 OCK2/PCSN 73 DOEN1 74 SADD1†† 77 SYNC1 78 DO1 79 OLD1 80 OCK1 81 ICK1 82 ILD1 83 DI1 85 IBF1 86 OBE1 93, 1, 13, VSS 25, 36, 51, 63, 76, 84 100, 9, 17, VDD 26, 42, 60, 75, 87 47 VDDA 50 VSSA Type I/O* I/O* I/O* I/O* I/O* I/O* I/O* I/O* O* O* I/O* I/O* I/O* I/O* I/O* I/O* I/O* I/O* O* I/O* I/O* I/O* I/O* I O* O* P Name/Function Status/Control Bit 3/PHIF Data Bus Bit 7. Status/Control Bit 2/PHIF Data Bus Bit 6. Status/Control Bit 1/PHIF Data Bus Bit 5. Status/Control Bit 0/PHIF Data Bus Bit 4. SIO2 Multiprocessor Address/PHIF Data Bus Bit 3. SIO2 Data Output Enable/PHIF Data Bus Bit 2. SIO2 Data Input/PHIF Data Bus Bit 1. SIO2 Input Clock/PHIF Data Bus Bit 0. SIO2 Output Buffer Empty/PHIF Output Buffer Empty. SIO2 Input Buffer Full/PHIF Input Buffer Full. SIO2 Output Load/PHIF Output Data Strobe. SIO2 Input Load/PHIF Input Data Strobe. SIO2 Multiprocessor Synchronization/PHIF Byte Select. SIO2 Data Output/PHIF Status Register Select. SIO2 Output Clock/PHIF Chip Select Not. SIO1 Data Output Enable. SIO1 Multiprocessor Address. SIO1 Multiprocessor Synchronization. SIO1 Data Output. SIO1 Output Load. SIO1 Output Clock. SIO1 Input Clock. SIO1 Input Load. SIO1 Data Input. SIO1 Input Buffer Full. SIO1 Output Buffer Empty. Ground. P Power Supply. P P Analog Power Supply. Analog Ground. * 3-states when RSTB = 0, or by JTAG control. † 3-states when RSTB = 0 and INT0 = 1. Output = 1 when RSTB = 0 and INT0 = 0. § Pull-up devices on input. ‡ 3-states by JTAG control. ** See Section 7, Mask-Programmable Options. †† For SIO multiprocessor applications, add 5 kΩ external pull-up resistors to SADD1 and/or SADD2 for proper initialization. 6 Lucent Technologies Inc. Data Sheet March 2000 4 Hardware Architecture The DSP1627 device is a 16-bit, fixed-point programmable digital signal processor (DSP). The DSP1627 consists of a DSP1600 core together with on-chip memory and peripherals. Added architectural features give the DSP1627 high program efficiency for signal coding applications. 4.1 DSP1627 Architectural Overview Figure 3 shows a block diagram of the DSP1627. The following modules make up the DSP1627. DSP1600 Core The DSP1600 core is the heart of the DSP1627 chip. The core contains data and address arithmetic units, and control for on-chip memory and peripherals. The core provides support for external memory wait-states and onchip, dual-port RAM and features vectored interrupts and a trap mechanism. Dual-Port RAM (DPRAM) This module contains six banks of zero wait-state memory. Each bank consists of 1K 16-bit words and has separate address and data ports to the instruction/coefficient and data memory spaces. A program can reference memory from either space. The DSP1600 core automatically performs the required multiplexing. If references to both ports of a single bank are made simultaneously, the DSP1600 core automatically inserts a wait-state and performs the data port access first, followed by the instruction/coefficient port access. A program can be downloaded from slow, off-chip memory into DPRAM, and then executed without wait-states. DPRAM is also useful for improving convolution performance in cases where the coefficients are adaptive. Since DPRAM can be downloaded through the JTAG port, full-speed remote in-circuit emulation is possible. DPRAM can also be used for downloading self-test code via the JTAG port. Read-Only Memory (ROM) The DSP1627x36 contains 36K 16-bit words of zero wait-state mask-programmable ROM for program and fixed coefficients. Similarly, the DSP1627x32 has 32K 16-bit words of ROM and access to 16 Kwords of external ROM. External Memory Multiplexer (EMUX) The EMUX is used to connect the DSP1627 to external memory and I/O devices. It supports read/write operations from/to instruction/coefficient memory (X memory space) and data memory (Y memory space). The DSP1600 core automatically controls the EMUX. Instruc- Lucent Technologies Inc. DSP1627 Digital Signal Processor tions can transparently reference external memory from either set of internal buses. A sequencer allows a single instruction to access both the X and the Y external memory spaces. Clock Synthesis The DSP powers up with a 1X input clock (CKI/CKI2) as the source for the processor clock. An on-chip clock synthesizer (PLL) can also be used to generate the system clock for the DSP, which will run at a frequency multiple of the input clock. The clock synthesizer is deselected and powered down on reset. For low-power operation, an internally generated slow clock can be used to drive the DSP. If both the clock synthesizer and the internally generated slow clock are selected, the slow clock will drive the DSP; however, the synthesizer will continue to run. The clock synthesizer and other programmable clock sources are discussed in Section 4.12. The use of these programmable clock sources for power management is discussed in Section 4.13. Bit Manipulation Unit (BMU) The BMU extends the DSP1600 core instruction set to provide more efficient bit operations on accumulators. The BMU contains logic for barrel shifting, normalization, and bit field insertion/extraction. The unit also contains a set of 36-bit alternate accumulators. The data in the alternate accumulators can be shuffled with the data in the main accumulators. Flags returned by the BMU mesh seamlessly with the DSP1600 conditional instructions. Bit Input/Output (BIO) The BIO provides convenient and efficient monitoring and control of eight individually configurable pins. When configured as outputs, the pins can be individually set, cleared, or toggled. When configured as inputs, individual pins or combinations of pins can be tested for patterns. Flags returned by the BIO mesh seamlessly with conditional instructions. Serial Input/Output Units (SIO and SIO2) SIO and SIO2 offer asynchronous, full-duplex, doublebuffered channels that operate at up to 25 Mbits/s (for 20 ns instruction cycle in a nonmultiprocessor configuration), and easily interface with other Lucent Technologies fixed-point DSPs in a multiple-processor environment. Commercially available codecs and time-division multiplex (TDM) channels can be interfaced to the serial I/O ports with few, if any, additional components. SIO2 is identical to SIO. An 8-bit serial protocol channel may be transmitted in addition to the address of the called processor in multiprocessor mode. This feature is useful for transmitting highlevel framing information or for error detection and correction. SIO2 and BIO are pin-multiplexed with the PHIF. 7 Data Sheet March 2000 DSP1627 Digital Signal Processor 4 Hardware Architecture (continued) DB[15:0] AB[15:0] RWN I/O EXM EROM ERAMHI ERAMLO JTAG BOUNDARY SCAN * EXTERNAL MEMORY INTERFACE & EMUX ioc DUAL-PORT RAM 6K x 16 ROM 36K/32K x 16† YAB YDB XDB XAB BMU aa0 DSP1600 CORE TDI ID * TCK BYPASS * TMS HDS TRST TRACE * ar0 TIMER ar1 timerc ar2 VEC[3:0] OR IOBIT[7:4] timer0 ar3 SIO PHIF DO2 OR PSTAT sdx(OUT) phifc OLD2 OR PODS OCK2 OR PCSN M U X DI1 ICK1 ILD1 PSTAT * OBE2 OR POBE ILD2 OR PIDS JCON * aa1 IDB ICK2 OR PB0 TDO BREAKPOINT * CKI CKI2 CKO RSTB STOP TRAP INT[1:0] IACK SYNC2 OR PBSEL jtag srta powerc pllc SIO2 pdx0(IN) tdms sdx2(OUT) pdx0(OUT) DI2 OR PB1 sdx(IN) BIO sbit IBF2 OR PIBF cbit DOEN2 OR PB2 SADD2 OR PB3 srta2 IBF1 DO1 OCK1 OLD1 OBE1 tdms2 sioc sdx2(IN) saddx SYNC1 SADD1 DOEN1 sioc2 IO BIT[3:0] OR PB[7:4] saddx2 5-4142 (F).f * These registers are accessible through the pins only. † 36K x 16 for the DSP1627x36; 32K x 16 for the DSP1627x32. Figure 3. DSP1627 Block Diagram 8 Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 4 Hardware Architecture (continued) Table 2. DSP1627 Block Diagram Legend Symbol aa<0—1> ar<0—3> BIO BMU BREAKPOINT BYPASS cbit EMUX HDS ID IDB ioc JCON jtag pdx0(in) pdx0(out) PHIF phifc pllc powerc PSTAT ROM saddx saddx2 sbit sdx(in) sdx2(in) sdx(out) sdx2(out) SIO SIO2 sioc sioc2 srta srta2 tdms tdms2 TIMER timer0 timerc TRACE XAB XDB YAB YDB Lucent Technologies Inc. Name Alternate Accumulators. Auxiliary BMU Registers. Bit Input/Output Unit. Bit Manipulation Unit. Four Instruction Breakpoint Registers. JTAG Bypass Register. Control Register for BIO. External Memory Multiplexer. Hardware Development System. JTAG Device Identification Register. Internal Data Bus. I/O Configuration Register. JTAG Configuration Registers. 16-bit Serial/Parallel Register. Parallel Data Transmit Input Register 0. Parallel Data Transmit Output Register 0. Parallel Host Interface. Parallel Host Interface Control Register. Phase-Locked Loop Control Register. Power Control Register. Parallel Host Interface Status Register. Internal ROM (36 Kwords for DSP1627x36, 32 Kwords for DSP1627x32). Multiprocessor Protocol Register. Multiprocessor Protocol Register for SIO2. Status Register for BIO. Serial Data Transmit Input Register. Serial Data Transmit Input Register for SIO2. Serial Data Transmit Output Register. Serial Data Transmit Output Register for SIO2. Serial Input/Output Unit. Serial Input/Output Unit #2. Serial I/O Control Register. Serial I/O Control Register for SIO2. Serial Receive/Transmit Address Register. Serial Receive/Transmit Address Register for SIO2. Serial I/O Time-division Multiplex Signal Control Register. Serial I/O Time-division Multiplex Signal Control Register for SIO2. Programmable Timer. Timer Running Count Register. Timer Control Register. Program Discontinuity Trace Buffer. Program Memory Address Bus. Program Memory Data Bus. Data Memory Address Bus. Data Memory Data Bus. 9 Data Sheet March 2000 DSP1627 Digital Signal Processor 4 Hardware Architecture (continued) Parallel Host Interface (PHIF) The PHIF is a passive, 8-bit parallel port which can interface to an 8-bit bus containing other Lucent Technologies DSPs (e.g., DSP1620, DSP1627, DSP1628, DSP1629, DSP1611, DSP1616, DSP1617, DSP1618), microprocessors, or peripheral I/O devices. The PHIF port supports either Motorola or Intel protocols, as well as 8-bit or 16-bit transfers, configured in software. The port data rate depends upon the instruction cycle rate. A 25 ns instruction cycle allows the PHIF to support data rates up to 11.85 Mbytes/s, assuming the external host device can transfer 1 byte of data in 25 ns. The PHIF is accessed in two basic modes: 8-bit or 16-bit mode. In 16-bit mode, the host determines an access of the high or low byte. In 8-bit mode, only the low byte is accessed. Software-programmable features allow for a glueless host interface to microprocessors (see Section 4.8, Parallel Host Interface). In systems with multiple processors, the processors may be configured such that any processor reaching a breakpoint will cause all the other processors to be trapped (see Section 4.3, Interrupts and Trap). Pin Multiplexing In order to allow flexible device interfacing while maintaining a low package pin count, the DSP1627 multiplexes 16 package pins between BIO, PHIF, VEC[3:0], and SIO2. Upon reset, the vectored interrupt indication signals, VEC[3:0], are connected to the package pins while IOBIT[4:7] are disconnected. Setting bit 12, EBIOH, of the ioc register connects IOBIT[4:7] to the package pins and disconnects VEC[3:0]. Upon reset, the parallel host interface (PHIF) is connected to the package pins while the second serial port (SIO2) and IOBIT[3:0] are disconnected. Setting bit 10, ESIO2, of the ioc register connects the SIO2 and IOBIT[3:0] and disconnects the PHIF. Timer Power Management The timer can be used to provide an interrupt at the expiration of a programmed interval. The interrupt may be single or repetitive. More than nine orders of magnitude of interval selection are provided. The timer may be stopped and restarted at any time. Many applications, such as portable cellular terminals, require programmable sleep modes for power management. There are three different control mechanisms for achieving low-power operation: the powerc control register, the STOP pin, and the AWAIT bit in the alf register. The AWAIT bit in the alf register allows the processor to go into a power-saving standby mode until an interrupt occurs. The powerc register configures various powersaving modes by controlling internal clocks and peripheral I/O units. The STOP pin controls the internal processor clock. The various power management options may be chosen based on power consumption and/or wake-up latency requirements. Hardware Development System (HDS) Module The on-chip HDS performs instruction breakpointing and branch tracing at full speed without additional offchip hardware. Using the JTAG port, the breakpointing is set up, and the trace history is read back. The port works in conjunction with the HDS code in the on-chip ROM and the hardware and software in a remote computer. The HDS code must be linked to the user's application code and reside in the first 4 Kwords of ROM. The on-chip HDS cannot be used with the secure ROM masking option (see Section 7.3, ROM Security Options). Four hardware breakpoints can be set on instruction addresses. A counter can be preset with the number of breakpoints to receive before trapping the core. Breakpoints can be set in interrupt service routines. Alternately, the counter can be preset with the number of cache instructions to execute before trapping the core. Every time the program branches instead of executing the next sequential instruction, the addresses of the instructions executed before and after the branch are caught in circular memory. The memory contains the last four pairs of program discontinuities for hardware tracing. 10 4.2 DSP1600 Core Architectural Overview Figure 4 shows a block diagram of the DSP1600 core. System Cache and Control Section (SYS) This section of the core contains a 15-word cache memory and controls the instruction sequencing. It handles vectored interrupts and traps, and also provides decoding for registers outside of the DSP1600 core. SYS stretches the processor cycle if wait-states are required (wait-states are programmable for external memory accesses). SYS sequences downloading via JTAG of selftest programs to on-chip, dual-port RAM. The cache loop iteration count can be specified at run time under program control as well as at assembly time. Lucent Technologies Inc. Data Sheet March 2000 4 Hardware Architecture (continued) Data Arithmetic Unit (DAU) The data arithmetic unit (DAU) contains a 16 x 16-bit parallel multiplier that generates a full 32-bit product in one instruction cycle. The product can be accumulated with one of two 36-bit accumulators. The accumulator data can be directly loaded from, or stored to, memory in two 16-bit words with optional saturation on overflow. The arithmetic logic unit (ALU) supports a full set of arithmetic and logical operations on either 16- or 32-bit data. A standard set of flags can be tested for conditional ALU operations, branches, and subroutine calls. This procedure allows the processor to perform as a powerful 16- or 32-bit microprocessor for logical and control applications. The available instruction set is fully compatible with the DSP1617 instruction set. See Section 5.1 for more information on the instruction set. The user also has access to two additional DAU registers. The psw register contains status information from the DAU (see Table 26, Processor Status Word Register). The arithmetic control register, auc, is used to configure some of the features of the DAU (see Table 27) including single-cycle squaring. The auc register alignment field supports an arithmetic shift left by one and left or right by two. The auc register is cleared by reset. The counters c0 to c2 are signed, 8 bits wide, and may be used to count events such as the number of times the program has executed a sequence of code. They are controlled by the conditional instructions and provide a convenient method of program looping. Y Space Address Arithmetic Unit (YAAU) The YAAU supports high-speed, register-indirect, compound, and direct addressing of data (Y) memory. Four general-purpose, 16-bit registers, r0 to r3, are available in the YAAU. These registers can be used to supply the read or write addresses for Y space data. The YAAU also decodes the 16-bit data memory address and outputs individual memory enables for the data access. The YAAU can address the six 1 Kword banks of onchip DPRAM or three external data memory segments. Up to 48 Kwords of off-chip RAM are addressable, with 16K addresses reserved for internal RAM. Two 16-bit registers, rb and re, allow zero-overhead modulo addressing of data for efficient filter implementations. Two 16-bit signed registers, j and k, are used to hold user-defined postmodification increments. Fixed increments of +1, –1, and +2 are also available. Four compound-addressing modes are provided to make read/write operations more efficient. Lucent Technologies Inc. DSP1627 Digital Signal Processor The YAAU allows direct (or indexed) addressing of data memory. In direct addressing, the 16-bit base register (ybase) supplies the 11 most significant bits of the address. The direct data instruction supplies the remaining 5 bits to form an address to Y memory space and also specifies one of 16 registers for the source or destination. X Space Address Arithmetic Unit (XAAU) The XAAU supports high-speed, register-indirect, instruction/coefficient memory addressing with postmodification of the register. The 16-bit pt register is used for addressing coefficients. The signed register i holds a user-defined postincrement. A fixed postincrement of +1 is also available. Register PC is the program counter. Registers pr and pi hold the return address for subroutine calls and interrupts, respectively. The XAAU decodes the 16-bit instruction/coefficient address and produces enable signals for the appropriate X memory segment. The addressable X segments are internal ROM (up to 36 Kwords for the DSP1627x36, up to 32 Kwords for the DSP1627x32), six 1K banks of DPRAM, and external ROM. The locations of these memory segments depend upon the memory map selected (see Table 5). A security mode can be selected by mask option. This prevents unauthorized access to the contents of on-chip ROM (see Section 7, Mask-Programmable Options). 4.3 Interrupts and Trap The DSP1627 supports prioritized, vectored interrupts and a trap. The device has eight internal hardware sources of program interrupt and two external interrupt pins. Additionally, there is a trap pin and a trap signal from the hardware development system (HDS). A software interrupt is available through the icall instruction. The icall instruction is reserved for use by the HDS. Each of these sources of interrupt and trap has a unique vector address and priority assigned to it. DSP16A interrupt compatibility is not maintained. The software interrupt and the traps are always enabled and do not have a corresponding bit in the ins register. Other vectored interrupts are enabled in the inc register (see Table 29, Interrupt Control (inc) Register) and monitored in the ins register (see Table 30, Interrupt Status (ins) Register). When the DSP1627 goes into an interrupt or trap service routine, the IACK pin is asserted. In addition, pins VEC[3:0] encode which interrupt/ trap is being serviced. Table 4 details the encoding used for VEC[3:0]. 11 Data Sheet March 2000 DSP1627 Digital Signal Processor 4 Hardware Architecture (continued) i (16) SYS cloop (7) ins (16) alf (16) inc (16) mwait (16) XDB XAAU MUX ADDER CACHE CONTROL 1 XAB pr (16) pc (16) pi (16) pt (16) IDB BRIDGE YAAU DAU yh (16) x (16) YDB yl (16) j (16) –1, 0, 1, 2 k (16) 16 x 16 MPY 32 p (32) MUX SHIFT (–2, 0, 1, 2) ADDER MUX YAB rb (16) 36 ALU/SHIFT c0 (8) c2 (8) auc (16) a0 (36) a1 (36) re (16) MUX c1 (8) CMP r0 (16) psw (16) r1 (16) r2 (16) 16 ybase (16) r3 (16) EXTRACT/SAT 5-1741 (F).b Figure 4. DSP1600 Core Block Diagram 12 Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 4 Hardware Architecture (continued) Table 3. DSP1600 Core Block Diagram Legend Symbol 16 x 16 MPY a0—a1 alf ALU/SHIFT auc c0—c2 cloop CMP DAU i IDB inc ins j k MUX mwait p PC pi pr psw pt r0—r3 rb re SYS x XAAU XAB XDB YAAU YAB YDB ybase y Name 16-bit x 16-bit Multiplier. Accumulators 0 and 1 (16-bit halves specified as a0, a0l, a1, and a1l)*. AWAIT, LOWPR, Flags. Arithmetic Logic Unit/Shifter. Arithmetic Unit Control. Counters 0—2. Cache Loop Count. Comparator. Digital Arithmetic Unit. Increment Register for the X Address Space. Internal Data Bus. Interrupt Control. Interrupt Status. Increment Register for the Y Address Space. Increment Register for the Y Address Space. Multiplexer. External Memory Wait-states Register. Product Register (16-bit halves specified as p, pl). Program Counter. Program Interrupt Return Register. Program Return Register. Processor Status Word. X Address Space Pointer. Y Address Space Pointers. Modulo Addressing Register (begin address). Modulo Addressing Register (end address). System Cache and Control Section. Multiplier Input Register. X Space Address Arithmetic Unit. X Space Address Bus. X Space Data Bus. Y Space Address Arithmetic Unit. Y Space Address Bus. Y Space Data Bus. Direct Addressing Base Register. DAU Register (16-bit halves specified as y, yl). * F3 ALU instructions with immediates require specifying the high half of the accumulators as a0h and a1h. Lucent Technologies Inc. 13 Data Sheet March 2000 DSP1627 Digital Signal Processor 4 Hardware Architecture (continued) Signaling Interrupt Service Status Interruptibility Five pins of DSP1627 are devoted to signaling interrupt service status. The IACK pin goes high while any interrupt or user trap is being serviced, and goes low when the ireturn instruction from the service routine is issued. Four pins, VEC[3:0], carry a code indicating which of the interrupts or trap is being serviced. Table 4 contains the encodings used by each interrupt. Vectored interrupts are serviced only after the execution of an interruptible instruction. If more than one vectored interrupt is asserted at the same time, the interrupts are serviced sequentially according to their assigned priorities. See Table 4 for the priorities assigned to the vectored interrupts. Interrupt service routines, branch and conditional branch instructions, cache loops, and instructions that only decrement one of the RAM pointers, r0 to r3 (e.g., *r3− −), are not interruptible. A trap is similar to an interrupt, but it gains control of the processor by branching to the trap service routine even when the current instruction is noninterruptible. It may not be possible to return to normal instruction execution from the trap service routine since the machine state cannot always be saved. In particular, program execution cannot be continued from a trapped cache loop or interrupt service routine. While in a trap service routine, another trap is ignored. When set to 1, the status bits in the ins register indicate that an interrupt has occurred. The processor must reach an interruptible state (completion of an interruptible instruction) before an enabled vectored interrupt will be acted on. An interrupt will not be serviced if it is not enabled. Polled interrupt service can be implemented by disabling the interrupt in the inc register and then polling the ins register for the expected event. Vectored Interrupts Tables 29 and 30 show the inc and ins registers. A logic 1 written to any bit of inc enables (or unmasks) the associated interrupt. If the bit is cleared to a logic 0, the interrupt is masked. Note that neither the software interrupt nor traps can be masked. The occurrence of an interrupt that is not masked will cause the program execution to transfer to the memory location pointed to by that interrupt's vector address, assuming no other interrupt is being serviced (see Table 4, Interrupt Vector Table). The occurrence of an interrupt that is masked causes no automatic processor action, but will set the corresponding status bit in the ins register. If a masked interrupt occurs, it is latched in the ins register, but the interrupt is not taken. When unlatched, this latched interrupt will initiate automatic processor interrupt action. See the DSP1611/17/18/27 Digital Signal Processor Information Manual for a more detailed description of the interrupts. 14 Traps due to HDS breakpoints have no effect on either the IACK or VEC[3:0] pins. Instead, they show the interrupt state or interrupt source of the DSP when the trap occurred. Clearing Interrupts The PHIF interrupts (PIBF and POBE) are cleared by reading or writing the parallel host interface data transmit registers pdx0[in] and pdx0[out], respectively. The SIO and SIO2 interrupts (IBF, IBF2, OBE, and OBE2) are cleared by reading or writing, as appropriate, the serial data registers sdx[in], sdx2[in], sdx[out], and sdx2[out]. The JTAG interrupt (JINT) is cleared by reading the jtag register. Three of the vectored interrupts are cleared by writing to the ins register. Writing a 1 to the INT0, INT1, or TIME bits in the ins will cause the corresponding interrupt status bit to be cleared to a logic 0. The status bit for these vectored interrupts is also cleared when the ireturn instruction is executed, leaving set any other vectored interrupts that are pending. Traps The TRAP pin of the DSP1627 is a bidirectional signal. At reset, it is configured as an input to the processor. Asserting the TRAP pin will force a user trap. The trap mechanism is used for two purposes. It can be used by an application to rapidly gain control of the processor for asynchronous time-critical event handling (typically for catastrophic error recovery). It is also used by the HDS for breakpointing and gaining control of the processor. Separate vectors are provided for the user trap (0x46) and the HDS trap (0x3). Traps are not maskable. Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 4 Hardware Architecture (continued) Table 4. Interrupt Vector Table Source No Interrupt Software Interrupt INT0 JINT INT1 TIME IBF2 OBE2 Reserved Reserved Reserved IBF OBE PIBF POBE TRAP from HDS TRAP from User Vector — 0x2 0x1 0x42 0x4 0x10 0x14 0x18 0x1c 0x20 0x24 0x2c 0x30 0x34 0x38 0x3 0x46 Priority — 1 2 3 4 7 8 9 10 11 12 14 15 16 17 18 19 = highest VEC[3:0] 0x0 0x1 0x2 0x8 0x9 0xc 0xd 0xe 0x0 0x1 0x2 0x3 0x4 0x5 0x6 —* 0x7 Issued by — icall pin jtag in pin timer SIO2 in SIO2 out — — — SIO in SIO out PHIF in PHIF out breakpoint, jtag, or pin pin * Traps due to HDS breakpoints have no effect on VEC[3:0] pins. A trap has four cycles of latency. At most, two instructions will execute from the time the trap is received at the pin to when it gains control. An instruction that is executing when a trap occurs is allowed to complete before the trap service routine is entered. (Note that the instruction could be lengthened by wait-states.) During normal program execution, the pi register contains either the address of the next instruction (two-cycle instruction executing) or the address following the next instruction (one-cycle instruction executing). In an interrupt service routine, pi contains the interrupt return address. When a trap occurs during an interrupt service routine, the value of the pi register may be overwritten. Specifically, it is not possible to return to an interrupt service routine from a user trap (0x46) service routine. Continuing program execution when a trap occurs during a cache loop is also not possible. The HDS trap causes circuitry to force the program memory map to MAP1 (with on-chip ROM starting at address 0x0) when the trap is taken. The previous memory map is restored when the trap service routine exits by issuing an ireturn. The map is forced to MAP1 because the HDS code, if present, resides in the on-chip ROM. Using the Lucent Technologies development tools, the TRAP pin may be configured to be an output, or an input vectoring to address 0x3. In a multiprocessor environment, the TRAP pins of all the DSPs present can be tied together. During HDS operations, one DSP is selected by the host software to be the master. The master processor's TRAP pin is configured to be an output. Lucent Technologies Inc. The TRAP pins of the slave processors are configured as inputs. When the master processor reaches a breakpoint, the master's TRAP pin is asserted. The slave processors will respond to their TRAP input by beginning to execute the HDS code. AWAIT Interrupt (Standby or Sleep Mode) Setting the AWAIT bit (bit 15) of the alf register (alf = 0x8000) causes the processor to go into a powersaving standby or sleep mode. Only the minimum circuitry on the chip required to process an incoming interrupt remains active. After the AWAIT bit is set, one additional instruction will be executed before the standby power-saving mode is entered. A PHIF or SIO word transfer will complete if already in progress. The AWAIT bit is reset when the first interrupt occurs. The chip then wakes up and continues executing. Two nop instructions should be programmed after the AWAIT bit is set. The first nop (one cycle) will be executed before sleeping; the second will be executed after the interrupt signal awakens the DSP and before the interrupt service routine is executed. The AWAIT bit should be set from within the cache if the code which is executing resides in external ROM where more than one wait-state has been programmed. This ensures that an interrupt will not disturb the device from completely entering the sleep state. 15 Data Sheet March 2000 DSP1627 Digital Signal Processor 4 Hardware Architecture (continued) For additional power savings, set ioc = 0x0180 and timerc = 0x0040 in addition to setting alf = 0x8000. This will hold the CKO pin low and shut down the timer and prescaler (see Table 38 and Table 31). For a description of the control mechanisms for putting the DSP into low-power modes, see Section 4.13, Power Management. 4.4 Memory Maps and Wait-States The DSP1600 core implements a modified Harvard architecture that has separate on-chip 16-bit address and data buses for the instruction/coefficient (X) and data (Y) memory spaces. Table 5 shows the instruction/coefficient memory space maps for both the DSP1627x36 and DSP1627x32. The differences between the x36 and x32 memory maps can be seen by comparing the respective MAP1 and MAP3. For instance, MAP1 of the x36 provides for 36 Kwords of IROM and 6 Kwords of dual-port RAM (DPRAM), whereas MAP1 of the x32 provides for 32 Kwords of IROM, 6 Kwords of DPRAM, and 16 Kwords of EROM. The DSP1627 provides a multiplexed external bus which accesses external RAM (ERAM) and ROM (EROM). Programmable wait-states are provided for external memory accesses. The instruction/coefficient memory map is configurable to provide application flexibility. Table 6 shows the data memory space, which has one map. Instruction/Coefficient Memory Map Selection In determining which memory map to use, the processor evaluates the state of two parameters. The first is the LOWPR bit (bit 14) of the alf register. The LOWPR bit of the alf register is initialized to 0 automatically at reset. LOWPR controls the starting address in memory assigned to the six 1K banks of dual-port RAM. If LOWPR is low, internal dual-port RAM begins at address 0xC000. If LOWPR is high, internal dual-port RAM begins at address 0x0. LOWPR also moves IROM from 0x0 in MAP1 to 0x4000 in MAP3, and EROM from 0x0 in MAP2 to 0x4000 in MAP4. The second parameter is the value at reset of the EXM pin (pin 27 or pin 14, depending upon the package type). EXM determines whether the internal 36 Kwords ROM (IROM) will be addressable in the memory map. The Lucent Technologies development system tools, together with the on-chip HDS circuitry and the JTAG port, can independently set the memory map. Specifically, during an HDS trap, the memory map is forced to 16 MAP1. The user's map selection is restored when the trap service routine has completed execution. MAP1 MAP1 has the IROM starting at 0x0 and six 1 Kword banks of DPRAM starting at 0xC000. Additionally, MAP1 for the x32 has 16 Kwords of EROM starting at 0x8000. MAP1 is used if DSP1627 has EXM low at reset and the LOWPR parameter is programmed to zero. It is also used during an HDS trap. MAP2 MAP2 differs from MAP1 in that the lowest 48 Kwords reference external ROM (EROM). MAP2 is used if EXM is high at reset, the LOWPR parameter is programmed to zero, and an HDS trap is not in progress. MAP3 MAP3 has the six 1 Kword banks of DPRAM starting at address 0x0. In MAP3 of the x36, the 36 Kwords of IROM start at 0x4000. Similarly, for the x32, 32 Kwords of IROM start at 0x4000. Additionally, MAP3 for the x32 has 16 Kwords of EROM starting at 0xC000. MAP3 is used if EXM is low at reset, the LOWPR bit is programmed to 1, and an HDS trap is not in progress. Note that this map is not available if the secure mask-programmable option has been ordered. MAP4 MAP4 differs from MAP3 in that addresses above 0x4000 reference external ROM (EROM). This map is used if the LOWPR bit is programmed to 1, an HDS trap is not in progress, and, either EXM is high during reset, or the secure mask-programmable option has been ordered. Whenever the chip is reset using the RSTB pin, the default memory map will be MAP1 or MAP2, depending upon the state of the EXM pin at reset. A reset through the HDS will not reinitialize the alf register, so the previous memory map is retained. Boot from External ROM After RSTB goes from low to high, the DSP1627 comes out of reset and fetches an instruction from address zero of the instruction/coefficient space. The physical location of address zero is determined by the memory map in effect. If EXM is high at the rising edge of RSTB, MAP2 is selected. MAP2 has EROM at location zero; thus, program execution begins from external memory. If EXM is high and INT1 is low when RSTB rises, the mwait register defaults to 15 wait-states for all external memory segments. If INT1 is high, the mwait register defaults to 0 wait-states. Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 4 Hardware Architecture (continued) Table 5. Instruction/Coefficient Memory Maps DSP1627x36 X Address AB[0:15] MAP 1* EXM = 0 LOWPR = 0† 0 4K 6K 12K 16K 20K 24K 28K 32K 36K 40K 44K 48K 52K 54K 56K 60K—64K 0x0000 0x1000 0x1800 0x3000 0x4000 0x5000 0x6000 0x7000 0x8000 0x9000 0xA000 0xB000 0xC000 0xD000 0xD800 0xE000 0xFFFF IROM (36K) MAP 2 EXM = 1 LOWPR = 0 EROM (48K) MAP 3‡ EXM = 0 LOWPR = 1 MAP 4 EXM = 1 LOWPR = 1 DPRAM (6K) Reserved (10K) IROM (36K) DPRAM (6K) Reserved (10K) EROM (48K) Reserved (12K) DPRAM (6K) Reserved (10K) DPRAM (6K) Reserved (10K) Reserved (12K) * MAP1 is set automatically during an HDS trap. The user-selected map is restored at the end of the HDS trap service routine. † LOWPR is an alf register bit. The Lucent Technologies development system tools can independently set the memory map. ‡ MAP3 is not available if the secure mask-programmable option is selected. DSP1627x32 X Address AB[0:15] MAP 1* EXM = 0 LOWPR = 0† MAP 2 EXM = 1 LOWPR = 0 MAP 3‡ EXM = 0 LOWPR = 1 MAP 4 EXM = 1 LOWPR = 1 0 4K 6K 12K 16K 20K 24K 28K 32K 36K 40K 44K 48K 52K 54K 56K 60K—64K 0x0000 0x1000 0x1800 0x3000 0x4000 0x5000 0x6000 0x7000 0x8000 0x9000 0xA000 0xB000 0xC000 0xD000 0xD800 0xE000 0xFFFF IROM (32K) EROM (48K) DPRAM (6K) Reserved (10K) IROM (32K) DPRAM (6K) Reserved (10K) EROM (48K) DPRAM (6K) Reserved (10K) EROM (16K) EROM (16K) DPRAM (6K) Reserved (10K) * MAP1 is set automatically during an HDS trap. The user-selected map is restored at the end of the HDS trap service routine. † LOWPR is an alf register bit. The Lucent Technologies development system tools can independently set the memory map. ‡ MAP3 is not available if the secure mask-programmable option is selected. Lucent Technologies Inc. 17 Data Sheet March 2000 DSP1627 Digital Signal Processor 4 Hardware Architecture (continued) 4.5 External Memory Interface (EMI) Data Memory Mapping The external memory interface supports read/write operations from instruction/coefficient memory, data memory, and memory-mapped I/O devices. The DSP1627 provides a 16-bit external address bus, AB[15:0], and a 16-bit external data bus, DB[15:0]. These buses are multiplexed between the internal buses for the instruction/coefficient memory and the data memory. Four external memory segment enables, ERAMLO, IO, ERAMHI, and EROM, select the external memory segment to be addressed. Table 6. Data Memory Map (Not to Scale) Decimal Address 0 6K Address in r0, r1, r2, r3 0x0000 0x1800 Segment DPRAM[1:6] Reserved (10K) 16K 0x4000 IO 16,640 0x4100 ERAMLO 32K 0x8000 ERAMHI 64K – 1 0xFFFF On the data memory side (see Table 6), the six 1K banks of dual-port RAM are located starting at address 0. Addresses from 0x4000 to 0x40FF reference a 256word memory-mapped I/O segment (IO). Addresses from 0x4100 to 0x7FFF reference the low external data RAM segment (ERAMLO). Addresses above 0x8000 reference high external data RAM (ERAMHI). Wait-States The number of wait-states (from 0 to 15) used when accessing each of the four external memory segments (ERAMLO, IO, ERAMHI, and EROM) is programmable in the mwait register (see Table 36). When the program references memory in one of the four external segments, the internal multiplexer is automatically switched to the appropriate set of internal buses, and the associated external enable of ERAMLO, IO, ERAMHI, or EROM is issued. The external memory cycle is automatically stretched by the number of wait-states configured in the appropriate field of the mwait register. 18 If a data memory location with an address between 0x4100 and 0x7FFF is addressed, ERAMLO is asserted low. If one of the 256 external data memory locations, with an address greater than or equal to 0x4000, and less than or equal to 0x40FF, is addressed, IO is asserted low. IO is intended for memory-mapped I/O. If a data memory location with an address greater than or equal to 0x8000 is addressed, ERAMHI is asserted low. When the external instruction/coefficient memory is addressed, EROM is asserted low. The flexibility provided by the programmable options of the external memory interface (see Table 36, mwait Register and Table 38, ioc Register) allows the DSP1627 to interface gluelessly with a variety of commercial memory chips. Each of the four external memory segments, ERAMLO, IO, ERAMHI, and EROM, has a number of wait-states that is programmable (from 0 to 15) by writing to the mwait register. When the program references memory in one of the four external segments, the internal multiplexer is automatically switched to the appropriate set of internal buses, and the associated external enable of ERAMLO, IO, ERAMHI, or EROM is issued. The external memory cycle is automatically stretched by the number of wait-states in the appropriate field of the mwait register. When writing to external memory, the RWN pin goes low for the external cycle. The external data bus, DB[15:0], is driven by the DSP1627 starting halfway through the cycle. The data driven on the external data bus is automatically held after the cycle unless an external read cycle immediately follows. The DSP1627 has one external address bus and one external data bus for both memory spaces. Since some instructions provide the capability of simultaneous access to both X space and Y space, some provision must be made to avoid collisions for external accesses. The DSP1627 has a sequencer that does the external X access first, and then the external Y access, transparently to the programmer. Wait-states are maintained as Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 4 Hardware Architecture (continued) 4.6 Bit Manipulation Unit (BMU) programmed in the mwait register. For example, let two instructions be executed: the first reads a coefficient from EROM and writes data to ERAM; the second reads a coefficient from EROM and reads data from ERAM. The sequencer carries out the following steps at the external memory interface: read EROM, write ERAM, read EROM, and read ERAM. Each step is done in sequential one-instruction cycle steps, assuming zero waitstates are programmed. Note that the number of instruction cycles taken by the two instructions is four. Also, in this case, the write hold time is zero. The BMU interfaces directly to the main accumulators in the DAU providing the following features: The DSP1627 allows writing into external instruction/ coefficient memory. By setting bit 11, WEROM, of the ioc register (see Table 38), writing to (or reading from) data memory or memory-mapped I/O asserts the EROM strobe instead of ERAMLO, IO, or ERAMHI. Therefore, with WEROM set, EROM appears in both Y space (replacing ERAM) and X space, in its normal position. Bit 14 of the ioc register (see Table 38), EXTROM, may be used with WEROM to download to a full 64K of external memory. When WEROM and EXTROM are both asserted, address bit 15 (AB15) is held low, aliasing the upper 32K of external memory into the lower 32K. When an access to internal memory is made, the AB[15:0] bus holds the last valid external memory address. Asserting the RSTB pin low 3-states the AB[15:0] bus. After reset, the AB[15:0] value is undefined. The leading edge of the memory segment enables can be delayed by approximately one-half a CKO period by programming the ioc register (see Table 38). This is used to avoid a situation in which two devices drive the data bus simultaneously. Bits 7, 8, and 13 of the ioc register select the mode of operation for the CKO pin (see Table 38). Available options are a free-running unstretched clock, a wait-stated sequenced clock (runs through two complete cycles during a sequenced external memory access), and a wait-stated clock based on the internal instruction cycle. These clocks drop to the low-speed internal ring oscillator when SLOWCKI is enabled (see 4.13, Power Management). The high-to-low transitions of the wait-stated clock are synchronized to the high-to-low transition of the free-running clock. Also, the CKO pin provides either a continuously high level, a continuously low level, or changes at the rate of the internal processor clock. This last option, only available with the crystal and small-signal input clock options, enables the DSP1627 CKI input buffer to deliver a full-rate clock to other devices while the DSP1627 itself is in one of the low-power modes. Lucent Technologies Inc. ■ Barrel shifting—logical and arithmetic, left and right shift ■ Normalization and extraction of exponent ■ Bit-field extraction and insertion These features increase the efficiency of the DSP in applications such as control or data encoding and decoding. For example, data packing and unpacking, in which short data words are packed into one 16-bit word for more efficient memory storage, is very easy. In addition, the BMU provides two auxiliary accumulators, aa0 and aa1. In one instruction cycle, 36-bit data can be shuffled, or swapped, between one of the main accumulators and one of the alternate accumulators. The ar<0—3> registers are 16-bit registers that control the operations of the BMU. They store a value that determines the amount of shift or the width and offset fields for bit extraction or insertion. Certain operations in the BMU set flags in the DAU psw register and the alf register (see Table 26, Processor Status Word (psw) Register, and Table 35, alf Register). The ar<0—3> registers can also be used as general-purpose registers. The BMU instructions are detailed in Section 5.1. For a thorough description of the BMU, see the DSP1611/17/ 18/27 Digital Signal Processor Information Manual. 4.7 Serial I/O Units (SIOs) The serial I/O ports on the DSP1627 device provide a serial interface to many codecs and signal processors with little, if any, external hardware required. Each highspeed, double-buffered port (sdx and sdx2) supports back-to-back transmissions of data. SIO and SIO2 are identical. The output buffer empty (OBE and OBE2) and input buffer full (IBF and IBF2) flags facilitate the reading and/or writing of each serial I/O port by programor interrupt-driven I/O. There are four selectable active clock speeds. A bit-reversal mode provides compatibility with either the most significant bit (MSB) first or least significant bit (LSB) first serial I/O formats (see Table 22, Serial I/O Control Registers (sioc and sioc2)). A multiprocessor I/O configuration is supported. This feature allows up to eight DSP161X devices to be connected together on an SIO port without requiring external glue logic. 19 DSP1627 Digital Signal Processor 4 Hardware Architecture (continued) The serial data may be internally looped back by setting the SIO loopback control bit, SIOLBC, of the ioc register. SIOLBC affects both the SIO and SIO2. The data output signals are wrapped around internally from the output to the input (DO1 to DI1 and DO2 to DI2). To exercise loopback, the SIO clocks (ICK1, ICK2, OCK1, and OCK2) should either all be in the active mode, 16-bit condition, or each pair should be driven from one external source in passive mode. Similarly, pins ILD1 (ILD2) and OLD1 (OLD2) must both be in active mode or tied together and driven from one external frame clock in passive mode. During loopback, DO1, DO2, DI1, DI2, ICK1, ICK2, OCK1, OCK2, ILD1, ILD2, OLD1, OLD2, SADD1, SADD2, SYNC1, SYNC2, DOEN1, and DOEN2 are 3-stated. Setting DODLY = 1 (sioc and sioc2) delays DO by one phase of OCK so that DO changes on the falling edge of OCK instead of the rising edge (DODLY = 0). This reduces the time available for DO to drive DI and to be valid for the rising edge of ICK, but increases the hold time on DO by half a cycle on OCK. Programmable Modes Programmable modes of operation for the SIO and SIO2 are controlled by the serial I/O control registers (sioc and sioc2). These registers, shown in Table 22, are used to set the ports into various configurations. Both input and output operations can be independently configured as either active or passive. When active, the DSP1627 generates load and clock signals. When passive, load and clock signal pins are inputs. Since input and output can be independently configured, each SIO has four different modes of operation. Each of the sioc registers is also used to select the frequency of active clocks for that SIO. Finally, these registers are used to configure the serial I/O data formats. The data can be 8 or 16 bits long, and can also be input/ output MSB first or LSB first. Input and output data formats can be independently configured. Multiprocessor Mode The multiprocessor mode allows up to eight processors (DSP1629, DSP1628, DSP1627, DSP1620, DSP1618, DSP1617, DSP1616, DSP1611) to be connected together to provide data transmission among any of the DSPs in the system. Either SIO port (SIO or SIO2) may be independently used for the multiprocessor mode. The multiprocessor interface is a four-wire interface, consisting of a data channel, an address/protocol channel, a transmit/receive clock, and a sync signal (see Figure 5). The DI1 and DO1 pins of all the DSPs are connected to transmit and receive the data channel. The SADD1 pins of all the DSPs are connected to trans20 Data Sheet March 2000 mit and receive the address/protocol channel. ICK1 and OCK1 should be tied together and driven from one source. The SYNC1 pins of all the DSPs are connected. In the configuration shown in Figure 5, the master DSP (DSP0) generates active SYNC1 and OCK1 signals while the slave DSPs use the SYNC1 and OCK1 signals in passive mode to synchronize operations. In addition, all DSPs must have their ILD1 and OLD1 signals in active mode. While ILD1 and OLD1 are not required externally for multiprocessor operation, they are used internally in the DSP's SIO. Setting the LD field of the master's sioc register to a logic level 1 will ensure that the active generation of SYNC1, ILD1, and OLD1 is derived from OCK1 (see Table 22). With this configuration, all DSPs should use ICK1 (tied to OCK1) in passive mode to avoid conflicts on the clock (CK) line (see the DSP1611/17/18/27 Digital Signal Processor Information Manual for more information). Four registers (per SIO) configure the multiprocessor mode: the time-division multiplexed slot register (tdms or tdms2), the serial receive and transmit address register (srta or srta2), the serial data transmit register (sdx or sdx2), and the multiprocessor serial address/protocol register (saddx or saddx2). Multiprocessor mode requires no external logic and uses a TDM interface with eight 16-bit time slots per frame. The transmission in any time slot consists of 16 bits of serial data in the data channel and 16 bits of address and protocol information in the address/protocol channel. The address information consists of the transmit address field of the srta register of the transmitting device. The address information is transmitted concurrently with the transmission of the first 8 bits of data. The protocol information consists of the transmit protocol field written to the saddx register and is transmitted concurrently with the last 8 bits of data (see Table 25, Multiprocessor Protocol Register). Data is received or recognized by other DSP(s) whose receive address matches the address in the address/protocol channel. Each SIO port has a user-programmable receive address and transmit address associated with it. The transmit and receive addresses are programmed in the srta register. In multiprocessor mode, each device can send data in a unique time slot designated by the tdms register transmit slot field (bits 7—0). The tdms register has a fully decoded transmit slot field in order to allow one DSP1627 device to transmit in more than one time slot. This procedure is useful for multiprocessor systems with less than eight DSP1627 devices when a higher bandwidth is necessary between certain devices in that system. The DSP operating during time slot 0 also drives SYNC1. Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor Each SIO also has a fully decoded transmitting address specified by the srta register transmit address field (bits 7—0). This is used to transmit information regarding the destination(s) of the data. The fully decoded receive address specified by the srta register receive address field (bits 15—8) determines which data will be received. The SIO protocol channel data is controlled via the saddx register. When the saddx register is written, the The SIO2 functions the same as the SIO. Please refer to Pin Multiplexing in Section 4.1 for a description of pin multiplexing of BIO, PHIF, VEC[3:0], and SIO2. DO DI DSP 7 SYNC SADD DO DI ICK OCK DSP 1 SYNC SADD ICK OCK DO DI DSP 0 Using SIO2 DATA CHANNEL SYNC Therefore, to prevent spurious inputs, the address/protocol channel should be pulled up to VDD with a 5 kΩ resistor, or it should be guaranteed that the bus is driven in every time slot. (If the SYNC1 signal is externally generated, then this pull-up is required for correct initialization.) An example use of the protocol channel is to use the top 3 bits of the saddx value as an encoded source address for the DSPs on the multiprocessor bus. This leaves the remaining 5 bits available to convey additional control information, such as whether the associated field is an opcode or data, or whether it is the last word in a transfer, etc. These bits can also be used to transfer parity information about the data. Alternatively, the entire field can be used for data transmission, boosting the bandwidth of the port by 50%. SADD In order to prevent multiple bus drivers, only one DSP can be programmed to transmit in a particular time slot. In addition, it is important to note that the address/protocol channel is 3-stated in any time slot that is not being driven. lower 8 bits contain the 8-bit protocol field. On a read, the high-order 8 bits read from saddx are the most recently received protocol field sent from the transmitting DSP's saddx output register. The low-order 8 bits are read as 0s. ICK OCK 4 Hardware Architecture (continued) 5 kΩ VDD CLOCK ADDRESS/PROTOCOL CHANNEL SYNC SIGNAL 5-4181 (F).a Figure 5. Multiprocessor Communication and Connections Lucent Technologies Inc. 21 Data Sheet March 2000 DSP1627 Digital Signal Processor 4 Hardware Architecture (continued) register. Setting PMODE selects 16-bit transfer mode. An input pin controlled by the host, PBSEL, determines an access of either the high or low bytes. The assertion level of the PBSEL input pin is configurable in software using bit 3 of the phifc register, PBSELF. Table 7 summarizes the port's functionality as controlled by the PSTAT and PBSEL pins and the PBSELF and PMODE fields. 4.8 Parallel Host Interface (PHIF) The DSP1627 has an 8-bit parallel host interface for rapid transfer of data with external devices. This parallel port is passive (data strobes provided by an external device) and supports either Motorola or Intel microcontroller protocols. The PHIF also provides for 8-bit or 16-bit data transfers. As a flexible host interface, it requires little or no glue logic to interface to other devices (e.g., microcontrollers, microprocessors, or another DSP). For 16-bit transfers, if PBSELF is zero, the PIBF and POBE flags are set after the high byte is transferred. If PBSELF is one, the flags are set after the low byte is transferred. In 8-bit mode, only the low byte is accessed, and every completion of an input or output access sets PIBF or POBE. The data path of the PHIF consists of a 16-bit input buffer, pdx0(in), and a 16-bit output buffer, pdx0(out). Two output pins, parallel input buffer full (PIBF) and parallel output buffer empty (POBE), indicate the state of the buffers. In addition, there are two registers used to control and monitor the PHIF's operation: the parallel host interface control register (phifc, see Table 28), and the PHIF status register (PSTAT, see Table 8). The PSTAT register, which reflects the state of the PIBF and POBE flags, can only be read by an external device when the PSTAT input pin is asserted. The phifc register defines the programmable options for this port. Bit 1 of the phifc register, PSTROBE, configures the port to operate either with an Intel protocol where only the chip select (PCSN) and either of the data strobes (PIDS or PODS) are needed to make an access, or with a Motorola protocol where the chip select (PCSN), a data strobe (PDS), and a read/write strobe (PRWN) are needed. PIDS and PODS are negative assertion data strobes while the assertion level of PDS is programmable through bit 2, PSTRB, of the phifc register. Finally, the assertion level of the output pins, PIBF and POBE, is controlled through bit 4, PFLAG. When PFLAG is set low, PIBF and POBE output pins have positive assertion levels. By setting bit 5, PFLAGSEL, the logical OR of PIBF and POBE flags (positive assertion) is seen at the output pin PIBF. By setting bit 7 in phifc, PSOBEF, the polarity of the POBE flag in the status register, PSTAT, can be changed. PSOBEF has no effect on the POBE pin. The function of the pins, PIDS and PODS, is programmable to support both the Intel and Motorola protocols. The pin, PCSN, is an input that, when low, enables PIDS and PODS (or PRWN and PDS, depending on the protocol used). While PCSN is high, the DSP1627 ignores any activity on PIDS and/or PODS. If a DSP1627 is intended to be continuously accessed through the PHIF port, PCSN should be grounded. If PCSN is low and their respective bits in the inc register are set, the assertion of PIDS and PODS by an external device causes the DSP1627 device to recognize an interrupt. Pin Multiplexing Please refer to Pin Multiplexing in Section 4.1 for a description of BIO, PHIF, VEC[3:0], and SIO2 pins. Programmability The parallel host interface can be programmed for 8-bit or 16-bit data transfers using bit 0, PMODE, of the phifc Table 7. PHIF Function (8-bit and 16-bit Modes) PMODE Field 0 (8-bit) 0 0 0 1 (16-bit) 1 1 1 PSTAT Pin 0 0 1 1 0 0 1 1 PBSEL Pin 0 1 0 1 0 1 0 1 PBSELF Field = 0 pdx0 low byte reserved PSTAT reserved pdx0 low byte pdx0 high byte PSTAT reserved PBSELF Field = 1 reserved pdx0 low byte reserved PSTAT pdx0 high byte pdx0 low byte reserved PSTAT Table 8. pstat Register as Seen on PB[7:0] Bit Field 22 7 6 5 4 RESERVED 3 2 1 PIBF 0 POBE Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 4 Hardware Architecture (continued) 4.10 Timer 4.9 Bit Input/Output Unit (BIO) The interrupt timer is composed of the timerc (control) register, the timer0 register, the prescaler, and the counter itself. The timer control register (see Table 31, timerc Register) sets up the operational state of the timer and prescaler. The timer0 register is used to hold the counter reload value (or period register) and to set the initial value of the counter. The prescaler slows the clock to the timer by a number of binary divisors to allow for a wide range of interrupt delay periods. The BIO controls the directions of eight bidirectional control I/O pins, IOBIT[7:0]. If a pin is configured as an output, it can be individually set, cleared, or toggled. If a pin is configured as an input, it can be read and/or tested. The lower half of the sbit register (see Table 33) contains current values (VALUE[7:0]) of the eight bidirectional pins IOBIT[7:0]. The upper half of the sbit register (DIREC[7:0]) controls the direction of each of the pins. A logic 1 configures the corresponding pin as an output; a logic 0 configures it as an input. The upper half of the sbit register is cleared upon reset. The cbit register (see Table 34) contains two 8-bit fields, MODE/MASK[7:0] and DATA/PAT[7:0]. The values of DATA/PAT[7:0] are cleared upon reset. The meaning of a bit in either field depends on whether it has been configured as an input or an output in sbit. If a pin has been configured to be an output, the meanings are MODE and DATA. For an input, the meanings are MASK and PAT (pattern). Table 9 shows the functionality of the MODE/ MASK and DATA/PAT bits based on the direction selected for the associated IOBIT pin. Those bits that have been configured as inputs can be individually tested for 1 or 0. For those inputs that are being tested, there are four flags produced: allt (all true), allf (all false), somet (some true), and somef (some false). These flags can be used for conditional branch or special instructions. The state of these flags can be saved and restored by reading and writing bits 0 to 3 of the alf register (see Table 35). Table 9. BIO Operations DIREC[n]* 1 (Output) 1 (Output) 1 (Output) 1 (Output) 0 (Input) 0 (Input) 0 (Input) 0 (Input) MODE/ MASK[n] 0 0 1 1 0 0 1 1 DATA/ PAT[n] 0 1 0 1 0 1 0 1 Action Clear Set No Change Toggle No Test No Test Test for Zero Test for One * 0 ≤ n ≤ 7. If a BIO pin is switched from being configured as an output to being configured as an input and then back to being configured as an output, the pin retains the previous output value. The counter is a 16-bit down counter that can be loaded with an arbitrary number from software. It counts down to 0 at the clock rate provided by the prescaler. Upon reaching 0 count, a vectored interrupt to program address 0x10 is issued to the DSP1627, providing the interrupt is enabled (bit 8 of inc and ins registers). The counter will then either wait in an inactive state for another command from software, or will automatically repeat the last interrupting period, depending upon the state of the RELOAD bit in the timerc register. When RELOAD is 0, the counter counts down from its initial value to 0, interrupts the DSP1627, and then stops, remaining inactive until another value is written to the timer0 register. Writing to the timer0 register causes both the counter and the period register to be written with the specified 16-bit number. When RELOAD is 1, the counter counts down from its initial value to 0, interrupts the DSP1627, automatically reloads the specified initial value from the period register into the counter, and repeats indefinitely. This provides for either a single timed interrupt event or a regular interrupt clock of arbitrary period. The timer can be stopped and started by software, and can be reloaded with a new period at any time. Its count value, at the time of the read, can also be read by software. Due to pipeline stages, stopping and starting the timer may result in one inaccurate count or prescaled period. When the DSP1627 is reset, the bottom 6 bits of the timerc register and the timer0 register and counter are initialized to 0. This sets the prescaler to CKO/2*, turns off the reload feature, disables timer counting, and initializes the timer to its inactive state. The act of resetting the chip does not cause a timer interrupt. Note that the period register is not initialized on reset. The T0EN bit of the timerc register enables the clock to the timer. When T0EN is a 1, the timer counts down towards 0. When T0EN is a 0, the timer holds its current count. Pin Multiplexing Please refer to Pin Multiplexing in Section 4.1 for a description of BIO, PHIF, VEC[3:0], and SIO2 pins. Lucent Technologies Inc. * Frequency of CKO/2 is equivalent to either CKI/2 for the PLL bypassed or related to CKI by the PLL multiplying factors. See Section 4.12, Clock Synthesis. 23 Data Sheet March 2000 DSP1627 Digital Signal Processor 4 Hardware Architecture (continued) The PRESCALE field of the timerc register selects one of 16 possible clock rates for the timer input clock (see Table 31, timerc Register). Setting the DISABLE bit of the timerc register to a logic 1 shuts down the timer and the prescaler for power savings. Setting the TIMERDIS, bit 4, in the powerc register has the same effect of shutting down the timer. The DISABLE bit and the TIMERDIS bit are cleared by writing a 0 to their respective registers to restore the normal operating mode. 4.11 JTAG Test Port The DSP1627 uses a JTAG/IEEE 1149.1 standard fourwire test port for self-test and hardware emulation. There is no separate TRST input pin. An instruction register, a boundary-scan register, a bypass register, and a device identification register have been implemented. The device identification register coding for the DSP1627 is shown in Table 37. The instruction register (IR) is 4 bits long. The instruction for accessing the device ID is 0xE (1110). The behavior of the instruction register is summarized in Table 10. Cell 0 is the LSB (closest to TDO). The first line shows the cells in the IR that capture from a parallel input in the capture-IR controller state. The second line shows the cells that always load a logic 1 in the capture-IR controller state. The third line shows the cells that always load a logic 0 in the capture-IR controller state. Cell 3 (MSB of IR) is tied to status signal PINT, and cell 2 is tied to status signal JINT. The state of these signals can therefore be captured during capture-IR and shifted out during SHIFT-IR controller states. Boundary-Scan Register All of the chip's inputs and outputs are incorporated in a JTAG scan path shown in Table 11. The types of boundary-scan cells are as follows: ■ I = input cell ■ O = 3-state output cell ■ B = bidirectional (I/O) cell ■ OE = 3-state control cell ■ DC = bidirectional control cell Table 10. JTAG Instruction Register IR Cell #: Parallel Input? Always Logic 1? Always Logic 0? 24 3 Y N N 2 Y N N 1 N N Y 0 N Y N Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 4 Hardware Architecture (continued) Note that the direction of shifting is from TDI to cell 104 to cell 103 . . . to cell 0 of TDO. Table 11. JTAG Boundary-Scan Register Cell 0 1 2 3 4 5 6 7 8 9 10—25 26 27 28—31 32—36 37 38—48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 Type OE O I DC B I O I OE I O I O O B DC B O O I DC B DC B DC B DC B OE O DC B DC B DC B DC Signal Name/Function Controls cells 1, 27—31 CKO RSTB Controls cell 4 TRAP STOP† IACK INT0 Controls cells 6, 10—25, 49, 50, 78, 79 INT1 AB[0:15] EXM RWN EROM, ERAMLO, ERAMHI, IO DB[0:4] Controls cells 32—36, 38—48 DB[5:15] OBE1 IBF1 DI1 Controls cell 53 ILD1 Controls cell 55 ICK1 Controls cell 57 OCK1 Controls cell 59 OLD1 Controls cell 61 DO1 Controls cell 63 SYNC1 Controls cell 65 SADD1 Controls cell 67 DOEN1 Controls cell 69 Cell 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 Type B DC B DC B DC B DC B O O DC B DC B DC B DC B DC B DC B DC B DC B DC B DC B DC B DC B I Signal Name/Function OCK2/PCSN* Controls cell 71 DO2/PSTAT* Controls cell 73 SYNC2/PBSEL* Controls cell 75 ILD2/PIDS* Controls cell 77 OLD2/PODS* IBF2/PIBF* OBE2/POBE* Controls cell 81 ICK2/PB0* Controls cell 83 DI2/PB1* Controls cell 85 DOEN2/PB2* Controls cell 87 SADD2/PB3* Controls cell 89 IOBIT0/PB4* Controls cell 91 IOBIT1/PB5* Controls cell 93 IOBIT2/PB6* Controls cell 95 IOBIT3/PB7* Controls cell 97 VEC3/IOBIT4* Controls cell 99 VEC2/IOBIT5* Controls cell 101 VEC1/IOBIT6* Controls cell 103 VEC0/IOBIT7* CKI‡ * Please refer to Pin Multiplexing in Section 4.1 for a description of pin multiplexing of BIO, PHIF, VEC[3:0], and SIO2. † Note that shifting a zero into this cell in the mode to scan a zero into the chip will disable the processor clocks just as the STOP pin will. ‡ When the JTAG SAMPLE instruction is used, this cell will have a logic one regardless of the state of the pin. Lucent Technologies Inc. 25 Data Sheet March 2000 DSP1627 Digital Signal Processor 4 Hardware Architecture (continued) 4.12 Clock Synthesis SLOWCKI powerc fSLOW CLOCK RING OSCILLATOR CKI INPUT CLOCK fCKI fCKI VCO CLOCK fVCO M U X INTERNAL PROCESSOR CLOCK fINTERNAL CLOCK ÷2 LOCK PLLSEL (FLAG TO INDICATE LOCK CONDITION OF PLL) ÷N PHASE DETECTOR CHARGE PUMP VCO PLLEN pllc LOOP FILTER Nbits[2:0] ÷M PLL/SYNTHESIZER Mbits[4:0] LF[3:0] 5-4520 (F) Figure 6. Clock Source Block Diagram The DSP1627 provides an on-chip, programmable clock synthesizer. Figure 6 is the clock source diagram. The 1X CKI input clock, the output of the synthesizer, or a slow internal ring oscillator can be used as the source for the internal DSP clock. The clock synthesizer is based on a phase-locked loop (PLL), and the terms clock synthesizer and PLL are used interchangeably. On powerup, CKI is used as the clock source for the DSP. This clock is used to generate the internal processor clocks and CKO, where fCKI = fCKO. Setting the appropriate bits in the pllc control register (described in Table 32) will enable the clock synthesizer to become the clock source. The powerc register, which is discussed in Section 4.13, can override the selection to stop clocks or force the use of the slow clock for lowpower operation. 26 PLL Control Signals The input to the PLL comes from one of the three maskprogrammable clock options: CMOS, crystal, or smallsignal. The PLL cannot operate without an external input clock. To use the PLL, the PLL must first be allowed to stabilize and lock to the programmed frequency. After the PLL has locked, the LOCK flag is set and the lock detect circuitry is disabled. The synthesizer can then be used as the clock source. Setting the PLLSEL bit in the pllc register will switch sources from fCKI to fVCO/2 without glitching. It is important to note that the setting of the pllc register must be maintained. Otherwise, the PLL will seek the new set point. Every time the pllc register is written, the LOCK flag is reset. Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 4 Hardware Architecture (continued) The frequency of the PLL output clock, fVCO, is determined by the values loaded into the 3-bit N divider and the 5-bit M divider. When the PLL is selected and locked, the frequency of the internal processor clock is related to the frequency of CKI by the following equations: fVCO = fCKI * M/N fINTERNAL CLOCK = fCKO = fVCO ÷ 2 The frequency of the VCO, fVCO, must fall within the range listed in Table 63. Also note that fVCO must be at least twice fCKI. The coding of the Mbits and Nbits is described as follows: Mbits = M − 2 if (N == 1) Nbits = 0x7 else Nbits = N − 2 where N ranges from 1 to 8 and M ranges from 2 to 20. The loop filter bits LF[3:0] should be programmed according to Table 64. Lucent Technologies Inc. Two other bits in the pllc register control the PLL. Clearing the PLLEN bit powers down the PLL; setting this bit powers up the PLL. Clearing the PLLSEL bit deselects the PLL so that the DSP is clocked by a 1X version of the CKI input; setting the PLLSEL bit selects the PLLgenerated clock for the source of the DSP internal processor clock. The pllc register is cleared on reset and powerup. Therefore, the DSP comes out of reset with the PLL deselected and powered down. M and N should be changed only while the PLL is deselected. The values of M and N should not be changed when powering down or deselecting the PLL. As previously mentioned, the PLL also provides a user flag, LOCK, to indicate when the loop has locked. When this flag is not asserted, the PLL output is unstable. The DSP should not be switched to the PLL-based clock without first checking that the lock flag is set. The lock flag is cleared by writing to the pllc register. When the PLL is deselected, it is necessary to wait for the PLL to relock before the DSP can be switched to the PLLbased clock. Before the input clock is stopped, the PLL should be powered down. Otherwise, the LOCK flag will not be reset and there may be no way to determine if the PLL is stable, once the input clock is applied again. The lock-in time depends on the frequency of operation and the values programmed for M and N (see Table 64). 27 Data Sheet March 2000 DSP1627 Digital Signal Processor 4 Hardware Architecture (continued) PLL Programming Examples The following section of code illustrates how the PLL would be initialized on powerup, assuming the following operating conditions: ■ CKI input frequency = 10 MHz ■ Internal clock and CKO frequency = 50 MHz ■ VCO frequency = 100 MHz ■ Input divide down count N = 2 (Set Nbits[2:0] = 000 to get N = 2, as described in Table 32.) ■ Feedback down count M = 20 (Set Mbits[4:0] = 10010 to get M = 18 + 2 = 20, as described in Table 32.) The device would come out of reset with the PLL disabled and deselected. pllinit: pllc = 0x2912 pllc = 0xA912 call pllwait pllc = 0xE912 goto start pllwait: if lock return goto pllwait /* /* /* /* /* Running CKI input clock at 10 MHz, set up counters in PLL */ Power on PLL, but PLL remains deselected */ Loop to check for LOCK flag assertion */ Select high-speed, PLL clock */ User's code, now running at 50 MHz */ Programming examples which illustrate how to use the PLL with the various power management modes are listed in Section 4.13. Latency The switch between the CKI-based clock and the PLL-based clock is synchronous. This method results in the actual switch taking place several cycles after the PLLSEL bit is changed. During this time, actual code can be executed, but it will be at the previous clock rate. Table 12 shows the latency times for switching between CKI-based and PLLbased clocks. In the example given, the delay to switch to the PLL source is 1—4 CKO cycles and to switch back is 11—31 CKO cycles. Table 12. Latency Times for Switching Between CKI and PLL-Based Clocks Switch to PLL-based clock Switch from PLL-based clock Minimum Latency (Cycles) 1 M/N + 1 Maximum Latency (Cycles) N+2 M + M/N + 1 Frequency Accuracy and Jitter When using the PLL to multiply the input clock frequency up to the instruction clock rate, it is important to realize that although the average frequency of the internal clock and CKO will have about the same relative accuracy as the input clock, noise sources within the DSP will produce jitter on the PLL clock such that each individual clock period will have some error associated with it. The PLL is guaranteed only to have sufficiently low jitter to operate the DSP, and thus, this clock should not be used as an input to jitter-sensitive devices in the system. VDDA and VSSA Connections The PLL has its own power and ground pins, VDDA and VSSA. Additional filtering should be provided for VDDA in the form of a ferrite bead connected from V DDA to VDD and two decoupling capacitors (4.7 µF tantalum in parallel with a 0.01 µF ceramic) from VDDA to VSS. VSSA can be connected directly to the main ground plane. This recommendation is subject to change and may need to be modified for specific applications depending on the characteristics of the supply noise. Note: For devices with the CMOS clock input option, the CKI2 pin should be connected to VSSA. 28 Lucent Technologies Inc. Data Sheet March 2000 4 Hardware Architecture (continued) 4.13 Power Management There are three different control mechanisms for putting the DSP1627 into low-power modes: the powerc control register, the STOP pin, and the AWAIT bit in the alf register. The PLL can also be disabled with the PLLEN bit of the pllc register for more power saving. Powerc Control Register Bits The powerc register has 10 bits that power down various portions of the chip and select the clock source: XTLOFF: Assertion of the XTLOFF bit powers down the crystal oscillator or the small-signal input circuit, disabling the internal processor clock. Assertion of the XTLOFF bit to disable the crystal oscillator also prevents its use as a noninverting buffer. Since the oscillator and the small-signal input circuits take many cycles to stabilize, care must be taken with the turn-on sequence, as described later. SLOWCKI: Assertion of the SLOWCKI bit selects the ring oscillator as the clock source for the internal processor clock instead of CKI or the PLL. When CKI or the PLL is selected, the ring oscillator is powered down. Switching of the clocks is synchronized so that no partial or short clock pulses occur. Two nops should follow the instruction that sets or clears SLOWCKI. NOCK: Assertion of the NOCK bit synchronously turns off the internal processor clock, regardless of whether its source is provided by CKI, the PLL, or the ring oscillator. The NOCK bit can be cleared by resetting the chip with the RSTB pin, or asserting the INT0 or INT1 pins. Two nops should follow the instruction that sets NOCK. The PLL remains running, if enabled, while NOCK is set. INT0EN: This bit allows the INT0 pin to asynchronously clear the NOCK bit, thereby allowing the device to continue program execution from where it left off without any loss of state. No chip reset is required. It is recommended that, when INT0EN is to be used, the INT0 interrupt be disabled in the inc register so that an unintended interrupt does not occur. After the program resumes, the INT0 interrupt in the ins register should be cleared. INT1EN: This bit enables the INT1 pin to be used as the NOCK clear, exactly like INT0EN previously described. DSP1627 Digital Signal Processor SIO1DIS: This is a powerdown signal to the SIO1 I/O unit. It disables the clock input to the unit, thus eliminating any sleep power associated with the SIO1. Since the gating of the clocks may result in incomplete transactions, it is recommended that this option be used in applications where the SIO1 is not used or when reset may be used to reenable the SIO1 unit. Otherwise, the first transaction after reenabling the unit may be corrupted. SIO2DIS: This bit powers down the SIO2 in the same way SIO1DIS powers down the SIO1. PHIFDIS: This is a powerdown signal to the parallel host interface. It disables the clock input to the unit, thus eliminating any sleep power associated with the PHIF. Since the gating of the clocks may result in incomplete transactions, it is recommended that this option be used in applications where the PHIF is not used, or when reset may be used to reenable the PHIF. Otherwise, the first transaction after reenabling the unit may be corrupted. TIMERDIS: This is a timer disable signal which disables the clock input to the timer unit. Its function is identical to the DISABLE field of the timerc control register. Writing a 0 to the TIMERDIS field will continue the timer operation. Figure 7 shows a functional view of the effect of the bits of the powerc register on the clock circuitry. It shows only the high-level operation of each bit. Not shown are the bits that power down the peripheral units. STOP Pin Assertion (active-low) of the STOP pin has the same effect as setting the NOCK bit in the powerc register. The internal processor clock is synchronously disabled until the STOP pin is returned high. Once the STOP pin is returned high, program execution will continue from where it left off without any loss of state. No chip reset is required. The PLL remains running, if enabled, during STOP assertion. The pllc Register Bits The PLLEN bit of the pllc register can be used to power down the clock synthesizer circuitry. Before shutting down the clock synthesizer circuitry, the system clock should be switched to either CKI using the PLLSEL bit of pllc, or to the ring oscillator using the SLOWCKI bit of powerc. The following control bits power down the peripheral I/O units of the DSP. These bits can be used to further reduce the power consumption during standard sleep mode. Lucent Technologies Inc. 29 Data Sheet March 2000 DSP1627 Digital Signal Processor 4 Hardware Architecture (continued) PLLEN XTLOFF OFF CRYSTAL OSCILLATOR, OR SMALL SIGNAL CLOCK CKI2 RING OSCILLATOR PLL DEEP SLEEP ON fSLOW CLOCK fVCO/2 MASK-PROGRAMMABLE OPTION CKI fCKI CMOS INPUT CLOCK SYNC. MUX PLLSEL SLOWCKI STOP HW STOP NOCK SW STOP CLEAR NOCK DEEP SLEEP DISABLE SYNC. GATE RSTB fINTERNAL CLOCK INT0 INTERNAL PROCESSOR CLOCK INT0EN INT1 INT1EN 5-4124 (F).h Notes: The functions in the shaded ovals are bits in the powerc control register. The functions in the nonshaded ovals are bits in the pllc control register. Deep sleep is the state arrived at either by a hardware or software stop of the internal processor clock. The switching of the multiplexers and the synchronous gate is designed so that no partial clocks or glitching will occur. When the deep sleep state is entered with the ring oscillator selected, the internal processor clock is turned off before the ring oscillator is powered down. PLL select is the PLLSEL bit of pllc; PLL powerdown is the PLLEN bit of pllc. Figure 7. Power Management Using the powerc and the pllc Registers 30 Lucent Technologies Inc. Data Sheet March 2000 4 Hardware Architecture (continued) Await Bit of the alf Register Setting the AWAIT bit of the alf register causes the processor to go into the standard sleep state or power-saving standby mode. Operation of the AWAIT bit is the same as in the DSP1610, DSP1611, DSP1616, DSP1617, and DSP1618. In this mode, the minimum circuitry required to process an incoming interrupt remains active, and the PLL remains active if enabled. An interrupt will return the processor to the previous state, and program execution will continue. The action resulting from setting the AWAIT bit and the action resulting from setting bits in the powerc register are mostly independent. As long as the processor is receiving a clock, whether slow or fast, the DSP may be put into standard sleep mode with the AWAIT bit. Once the AWAIT bit is set, the STOP pin can be used to stop and later restart the processor clock, returning to the standard sleep state. If the processor clock is not running, however, the AWAIT bit cannot be set. Power Management Sequencing There are important considerations for sequencing the power management modes. Both the crystal oscillator and the small-signal clock input circuits have start-up delays which must be taken into account, and the PLL requires a delay to reach lock-in. Also, the chip may or may not need to be reset following a return from a lowpower state. Devices with a crystal oscillator or small-signal input clocking option may use the XTLOFF bit in the powerc register to power down the on-chip oscillator or smallsignal circuitry, thereby reducing the power dissipation. When reenabling the oscillator or the small-signal circuitry, it is important to bear in mind that a start-up interval exists during which time the clocks are not stable. Lucent Technologies Inc. DSP1627 Digital Signal Processor Two scenarios exist here: 1. Immediate Turn-Off, Turn-On with RSTB: This scenario applies to situations where the target device is not required to execute any code while the crystal oscillator or small-signal input circuit is powered down and where restart from a reset state can be tolerated. In this case, the processor clock derived from either the oscillator or the small-signal input is running when XTLOFF is asserted. This effectively stops the internal processor clock. When the system chooses to reenable the oscillator or small-signal input, a reset of the device will be required. The reset pulse must be of sufficient duration for the oscillator start-up interval to be satisfied. A similar interval is required for the small-signal input circuit to reach its dc operating point. A minimum reset pulse of 20 ms will be adequate. The falling edge of the reset signal, RSTB, will asynchronously clear the XTLOFF field, thus re-enabling the power to the oscillator or small-signal circuitry. The target DSP will then start execution from a reset state, following the rising edge of RSTB. 2. Running from Slow Clock While XTLOFF Active: The second scenario applies to situations where the device needs to continue execution of its target code when the crystal oscillator or small-signal input is powered down. In this case, the device switches to the slow ring oscillator clock first, by enabling the SLOWCKI field before writing a 1 to the XTLOFF field. Two nops are needed in between the two write operations to the powerc register. The target device will then continue execution of its code at slow speed, while the crystal oscillator or small-signal input clock is turned off. Switching from the slow clock back to the high-speed crystal oscillator clock is then accomplished in three user steps. First, XTLOFF is cleared. Then, a user-programmed routine sets the internal timer to a delay to wait for the crystal's oscillations to become stable. When the timer counts down to zero, the high-speed clock is selected by clearing the SLOWCKI field, either in the timer's interrupt service routine or following a timer polling loop. If PLL operation is desired, then an additional routine is necessary to enable the PLL and wait for it to lock. 31 Data Sheet March 2000 DSP1627 Digital Signal Processor 4 Hardware Architecture (continued) Power Management Examples Without the PLL The following examples show the more significant options for reducing the power dissipation. These are valid only if the pllc register is set to disable and deselect the PLL (PLLEN = 0, PLLSEL = 0). Standard Sleep Mode. This is the standard sleep mode. While the processor is clocked with a high-speed clock, CKI, the alf register's AWAIT bit is set. Peripheral units may be turned off to further reduce the sleep power. powerc = 0X00F0 sleep:a0 = 0x8000 do 1 { alf = a0 nop } nop nop cont: . . . powerc = 0x0 /* /* /* /* /* Turn off peripherals, core running with CKI */ Set alf register in cache loop if running from */ external memory with >1 wait state */ Stop internal processor clock, interrupt circuits */ active */ /* /* /* /* Needed for bedtime execution. Only sleep power */ consumed here until.... interrupt wakes up the device */ User code executes here */ Turn peripheral units back on */ Sleep with Slow Internal Clock. In this case, the ring oscillator is selected to clock the processor before the device is put to sleep. This will reduce the power dissipation while waiting for an interrupt to continue program execution. powerc = 0x40F0 2*nop sleep:a0 = 0x8000 do 1 { alf = a0 nop } nop nop cont: . . . powerc = 0x00F0 2*nop powerc = 0x0000 /* /* /* /* /* /* Turn off peripherals and select slow clock */ Wait for it to take effect */ Set alf register in cache loop if running from */ external memory with >1 wait state */ Stop internal processor clock, interrupt circuits */ active */ /* /* /* /* /* /* Needed for bedtime execution. Reduced sleep power */ consumed here.... Interrupt wakes up the device */ User code executes here */ Select high-speed clock */ Wait for it to take effect */ Turn peripheral units back on */ Note that, in this case, the wake-up latency is determined by the period of the ring oscillator clock. Sleep with Slow Internal Clock and Crystal Oscillator/Small-Signal Disabled. If the target device contains the crystal oscillator or the small-signal clock option, the clock input circuitry can be powered down to further reduce power. In this case, the slow clock must be selected first. powerc = 0x40F0 2*nop powerc = 0xC0F0 sleep:a0 = 0x8000 do 1 { alf = a0 nop } nop nop powerc = 0x40F0 call xtlwait cont: powerc = 0x00F0 2*nop powerc = 0x0000 /* /* /* /* /* /* /* Turn off peripherals and select slow clock */ Wait for it to take effect */ Turn off the crystal oscillator */ Set alf register in cache loop if running from */ external memory with >1 wait state */ Stop internal processor clock, interrupt circuits */ active */ /* /* /* /* /* /* /* Needed for bedtime execution. Reduced sleep power */ consumed here.... Interrupt wakes up the device */ Clear XTLOFF, reenable oscillator/small-signal */ Wait until oscillator/small-signal is stable */ Select high-speed clock */ Wait for it to take effect */ Turn peripheral units back on */ Note that, in this case, the wake-up latency is dominated by the crystal oscillator or small-signal start-up period. 32 Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 4 Hardware Architecture (continued) Software Stop. In this case, all internal clocking is disabled. INT0, INT1, or RSTB may be used to reenable the clocks. If the device uses the crystal oscillator or small-signal clock option, the power management must be done in correct sequence. powerc = 0x4000 2*nop powerc = 0xD000 inc = NOINT0 sopor:powerc = 0xF000 3*nop cont: powerc = 0x4000 call waitxtl powerc = 0x0 2*nop ins = 0x0010 /* /* /* /* /* /* /* /* /* /* /* /* /* /* SLOWCKI asserted */ Wait for it to take effect */ XTLOFF asserted if applicable and INT0EN asserted */ Disable the INT0 interrupt */ NOCK asserted, all clocks stop */ Minimum switching power consumed here */ Some nops will be needed */ INT0 pin clears the NOCK field, clocking resumes */ INT0EN cleared and XTLOFF cleared, if applicable*/ Wait for the crystal oscillator/small-signal to */ stabilize, if applicable*/ Clear SLOWCKI field, back to high speed */ Wait for it to take effect */ Clear the INT0 status bit */ In this case also, the wake-up latency is dominated by the crystal oscillator or small-signal start-up period. The previous examples do not provide an exhaustive list of options available to the user. Many different clocking possibilities exist for which the target device may be programmed, depending on: ■ The clock source to the processor. ■ Whether the user chooses to power down the peripheral units. ■ The operational state of the crystal oscillator/small-signal clock input, powered or unpowered. ■ Whether the internal processor clock is disabled through hardware or software. ■ The combination of power management modes the user chooses. ■ Whether or not the PLL is enabled. An example subroutine for xtlwait follows: xtlwait: loop1: timer0 = 0x2710 timerc = 0x0010 inc = 0x0000 a0 = ins a0 = a0 & 0x0100 if eq goto loop1 ins = 0x0100 return Lucent Technologies Inc. /* /* /* /* /* /* /* /* Load a count of 10,000 into the timer Start the timer with a PRESCALE of two Disable the interrupts Poll the ins register Check bit 8 (TIME) of the ins register Loop if the bit is not set Clear the TIME interrupt bit Return to the main program */ */ */ */ */ */ */ */ 33 DSP1627 Digital Signal Processor Data Sheet March 2000 4 Hardware Architecture (continued) Power Management Examples with the PLL The following examples show the more significant options for reducing power dissipation if operation with the PLL clock synthesizer is desired. Standard Sleep Mode, PLL Running. This mode would be entered in the same manner as without the PLL. While the input to the clock synthesizer, CKI, remains running, the alf register's AWAIT bit is set. The PLL will continue to run and dissipate power. Peripheral units may be turned off to further reduce the sleep power. powerc = 0x00F0 sleep:a0 = 0x8000 do 1 { alf = a0 nop } nop nop cont: . . . powerc = 0x0 /* /* /* /* /* Turn off peripherals, core running with PLL */ Set alf register in cache loop if running from */ external memory with >1 wait state */ Stop internal processor clock, interrupt circuits */ active */ /* /* /* /* Needed for bedtime execution. Only sleep power plus PLL */ power consumed here.... Interrupt wakes up the device */ User code executes here */ Turn peripheral units back on */ Sleep with Slow Internal Clock, PLL Running. In this case, the ring oscillator is selected to clock the processor before the device is put to sleep. This will reduce power dissipation while waiting for an interrupt to continue program execution. powerc = 0x40F0 2*nop sleep:a0 = 0x8000 do 1 { alf = a0 nop } nop nop cont: . . . powerc = 0x00F0 2*nop powerc = 0x0000 34 /* /* /* /* /* /* Turn off peripherals and select slow clock */ Wait for slow clock to take effect */ Set alf register in cache loop if running from */ external memory with >1 wait state */ Stop internal processor clock, interrupt circuits */ active */ /* /* /* /* /* /* /* Needed for bedtime execution. Reduced sleep power, PLL */ power, and ring oscillator power consumed here... */ Interrupt wakes up the device */ User code executes here */ Select high-speed PLL based clock */ Wait for it to take effect */ Turn peripheral units back on */ Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 4 Hardware Architecture (continued) Sleep with Slow Internal Clock and Crystal Oscillator/Small-Signal Disabled, PLL Disabled. If the target device contains the crystal oscillator or the small-signal clock option, the clock input circuitry can be powered down to further reduce power. In this case, the slow clock must be selected first, and then the PLL must be disabled, since the PLL cannot run without the clock input circuitry being active. powerc = 0x40F0 2*nop pllc = 0x29F2 powerc = 0xC0F0 sleep:a0 = 0x8000 do 1 { alf = a0 nop } nop nop powerc = 0x40F0 call xtlwait pllc = 0xE9F2 call pllwait cont: powerc = 0x00F0 2*nop powerc = 0x0000 /* /* /* /* /* /* /* /* Turn off peripherals and select slow clock */ Wait for slow clock to take effect */ Disable PLL (assume N = 1,M = 20, LF = 1001) */ Disable crystal oscillator */ Set alf register in cache loop if running from */ external memory with >1 wait state */ Stop internal processor clock, interrupt circuits */ active */ /* /* /* /* /* /* /* /* /* Needed for bedtime execution. Reduced sleep power consumed here.... Interrupt wakes up device */ Clear XTLOFF, leave PLL disabled */ Wait until crystal oscillator/small-signal is stable */ Enable PLL, continue to run off slow clock */ Loop to check for LOCK flag assertion */ Select high-speed PLL based clock */ Wait for it to take effect */ Turn peripherals back on */ Software Stop, PLL Disabled. In this case, all internal clocking is disabled. INT0, INT1, or RSTB may be used to reenable the clocks. If the device uses the crystal oscillator or small-signal clock option, the power management must be done in the correct sequence, with the PLL being disabled before shutting down the clock input buffer. powerc = 0x4000 2*nop pllc = 0x29F2 powerc = 0xD000 sopor:powerc = 0xF000 3*nop cont: powerc = 0x4000 call xtlwait pllc = 0xE9F2 call pllwait powerc = 0x0 2*nop ins = 0x0010 Lucent Technologies Inc. /* /* /* /* /* /* /* /* /* /* /* /* /* /* /* /* /* SLOWCKI asserted */ Wait for slow clock to take effect */ Disable PLL (assume N = 1, M = 20, LF = 1001) */ XTLOFF asserted, if applicable and INT0EN asserted */ NOCK asserted, all clocks stop */ Minimum switching power consumed here */ Some nops will be needed */ INT0 pin clears NOCK field, clocking resumes */ INTOEN cleared and XTLOFF cleared, if applicable */ Wait until crystal oscillator/small-signal is stable */ if applicable */ Enable PLL, continue to run off slow clock */ Loop to check for LOCK flag assertion */ Select high-speed PLL based clock */ Wait for it to take effect */ Clear the INT0 status bit */ 35 Data Sheet March 2000 DSP1627 Digital Signal Processor 5 Software Architecture Multiply/ALU Instructions 5.1 Instruction Set Note that the function statements and transfer statements in Table 13 are chosen independently. Any function statement (F1) can be combined with any transfer statement to form a valid multiply/ALU instruction. If either statement is not required, a single statement from either column also constitutes a valid instruction. The number of cycles to execute the instruction is a function of the transfer column. (An instruction with no transfer statement executes in one instruction cycle.) Whenever PC, pt, or rM is used in the instruction and points to external memory, the programmed number of wait-states must be added to the instruction cycle count. All multiply/ALU instructions require one word of program memory. The no-operation (nop) instruction is a specialcase encoding of a multiply/ALU instruction and executes in one cycle. The assembly-language representation of a nop is either nop or a single semicolon. The DSP1627 processor has seven types of instructions: multiply/ALU, special function, control, F3 ALU, BMU, cache, and data move. The multiply/ALU instructions are the primary instructions used to implement signal processing algorithms. Statements from this group can be combined to generate multiply/accumulate, logical, and other ALU functions and to transfer data between memory and registers in the data arithmetic unit. The special function instructions can be conditionally executed based on flags from the previous ALU or BMU operation, the condition of one of the counters, or the value of a pseudorandom bit in the DSP1627 device. Special function instructions perform shift, round, and complement functions. The F3 ALU instructions enrich the operations available on accumulators. The BMU instructions provide high-performance bit manipulation. The control instructions implement the goto and call commands. Control instructions can also be executed conditionally. Cache instructions are used to implement low-overhead loops, conserve program memory, and decrease the execution time of certain multiply/ALU instructions. Data move instructions are used to transfer data between memory and registers or between accumulators and registers. See the DSP1611/17/18/27 Digital Signal Processor Information Manual for a detailed description of the instruction set. The following operators are used in describing the instruction set: ■ * 16 x 16-bit –> 32-bit multiplication or register-indirect addressing when used as a prefix to an address register or denotes direct addressing when used as a prefix to an immediate ■ + 36-bit addition† ■ – 36-bit subtraction† ■ >> Arithmetic right shift ■ >>> Logical right shift ■ << ■ <<< Logical left shift ■ | 36-bit bitwise OR† ■ & 36-bit bitwise AND† ■ ^ 36-bit bitwise EXCLUSIVE OR† ■ : Compound address swapping, accumulator shuffling ■ ~ One's complement A single-cycle squaring function is provided in DSP1627. By setting the X = Y = bit in the auc register, any instruction that loads the high half of the y register also loads the x register with the same value. A subsequent instruction to multiply the x register and y register results in the square of the value being placed in the p register. The instruction a0 = p p = x*y y = *r0++ with the X = Y = bit set to one will read the value pointed to by r0, load it to both x and y, multiply the previously fetched value of x and y, and transfer the previous product to a0. A table of values pointed to by r0 can thus be squared in a pipeline with one instruction cycle per each value. Multiply/ALU instructions that use x = X transfer statements (such as a0 = p p = x*y y = *r0++ x = *pt++) are not recommended for squaring because pt will be incremented even though x is not loaded from the value pointed to by pt. Also, the same conflict wait occurrences from reading the same bank of internal memory or reading from external memory apply, since the X space fetch occurs (even though its value is not used). Arithmetic left shift † These are 36-bit operations. One operand is 36-bit data in an accumulator; the other operand may be 16, 32, or 36 bits. 36 Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 5 Software Architecture (continued) Table 13. Multiply/ALU Instructions Function Statement p=x*y aD = p p=x*y aD = aS + p p=x*y aD = aS – p p=x*y aD = p aD = aS + p aD = aS – p aD = y aD = aS + y aD = aS – y aD = aS & y aD = aS | y aD = aS ^ y aS – y aS & y Transfer Statement† y=Y x=X y = aT x=X y[l] = Y aT[l] = Y x=Y Y Y = y[l] Y = aT[l] Z:y x=X Z:y[l] Z:aT[l] Cycles (Out/In Cache)‡ 2/1 2/1 1/1 1/1 1/1 1/1 2/2 2/2 2/2 2/2 2/2 † The l in [ ] is an optional argument that specifies the low 16 bits of aT or y. ‡ Add cycles for: 1. When an external memory access is made in X or Y space and wait-states are programmed, add the number of wait-states. 2. If an X space access and a Y space access are made to the same bank of DPRAM in one instruction, add one cycle. Note: For transfer statements when loading the upper half of an accumulator, the lower half is cleared if the corresponding CLR bit in the auc register is zero. auc is cleared by reset. Table 14. Replacement Table for Multiply/ALU Instructions Replace aD, aS, aT X Y Z Value a0, a1 *pt++, *pt++i Meaning One of two DAU accumulators. X memory space location pointed to by pt. pt is postmodified by +1 and i, respectively. *rM, *rM++, *rM--, rM++j RAM location pointed to by rM (M = 0, 1, 2, 3). rM is postmodified by 0, +1, –1, or j, respectively. *rMzp, *rMpz, *rMm2, *rMjk Read/Write compound addressing. rM (M = 0, 1, 2, 3) is used twice. First, postmodified by 0, +1, –1, or j, respectively; and, second, postmodified by +1, 0, +2, or k, respectively. Lucent Technologies Inc. 37 Data Sheet March 2000 DSP1627 Digital Signal Processor 5 Software Architecture (continued) Special Function Instructions All forms of the special function require one word of program memory and execute in one instruction cycle. (If PC points to external memory, add programmed wait-states.) aD = aS >> 1 } aD = aS >> 4 aD = aS >> 8 aD = aS >> 16 Arithmetic right shift (sign preserved) of 36-bit accumulators aD = aS — Load destination accumulator from source accumulator aD = –aS — 2's complement aD = ~aS* — 1's complement aD = rnd(aS) — Round upper 20 bits of accumulator aDh = aSh + 1 — Increment upper half of accumulator (lower half cleared) aD = aS + 1 — Increment accumulator aD = y — Load accumulator with 32-bit y register value with sign extend aD = p — Load accumulator with 32-bit p register value with sign extend aD = aS << 1 } aD = aS << 4 aD = aS << 8 aD = aS << 16 Arithmetic left shift (sign not preserved) of the lower 32 bits of accumulators (upper 4 bits are sign-bit-extended from bit 31 at the completion of the shift) The above special functions can be conditionally executed, as in: if CON instruction and with an event counter ifc CON instruction which means: if CON is true then c1 = c1 + 1 instruction c2 = c1 else c1 = c1 + 1 The above special function statements can be executed unconditionally by writing them directly, e.g., a0 = a1. Table 15. Replacement Table for Special Function Instructions Replace aD aS CON Value a0, a1 Meaning One of two DAU accumulators. mi, pl, eq, ne, gt, le, lvs, lvc, mvs, mvc, c0ge, See Table 17 for definitions of mnemonics. c0lt, c1ge, c1lt, heads, tails, true, false, allt, allf, somet, somef, oddp, evenp, mns1, nmns1, npint, njint, lock * This function is not available for the DSP16A. 38 Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 5 Software Architecture (continued) Control Instructions All control instructions executed unconditionally execute in two cycles, except icall which takes three cycles. Control instructions executed conditionally execute in three instruction cycles. (If PC, pt, or pr point to external memory, add programmed wait-states.) Control instructions executed unconditionally require one word of program memory, while control instructions executed conditionally require two words. Control instructions cannot be executed from the cache. goto JA† goto pt call JA† call pt icall‡ return ireturn (goto pr) (goto pi) † The goto JA and call JA instructions should not be placed in the last or next-to-last instruction before the boundary of a 4 Kwords page. If the goto or call is placed there, the program counter will have incremented to the next page and the jump will be to the next page, rather than to the desired current page. ‡ The icall instruction is reserved for development system use. The above control instructions, with the exception of ireturn and icall, can be conditionally executed. For example: if le goto 0x0345 Table 16. Replacement Table for Control Instructions Replace CON JA Value mi, pl, eq, ne, gt, le, nlvs, lvc, mvs, mvc, c0ge, c0lt, c1ge, c1lt, heads, tails, true, false, allt, allf, somet, somef, oddp, evenp, mns1, nmns1, npint, njint, lock 12-bit value Lucent Technologies Inc. Meaning See Table 17 for definitions of mnemonics. Least significant 12 bits of absolute address within the same 4 Kwords memory section. 39 Data Sheet March 2000 DSP1627 Digital Signal Processor 5 Software Architecture (continued) Conditional Mnemonics (Flags) Table 17 lists mnemonics used in conditional execution of special function and control instructions. Table 17. DSP1627 Conditional Mnemonics Test pl eq gt lvs mvs c0ge c1ge heads true allt somet oddp mns1 npint lock Meaning Result is nonnegative (sign bit is bit 35). ≥ 0 Result is equal to 0. = 0 Result is greater than 0. > 0 Logical overflow set.* Mathematical overflow set.† Counter 0 greater than or equal to 0. Counter 1 greater than or equal to 0. Pseudorandom sequence bit set. The condition is always satisfied in an if instruction. All True, all BIO input bits tested compared successfully. Some True, some BIO input bits tested compared successfully. Odd Parity, from BMU operation. Minus 1, result of BMU operation. Not PINT, used by hardware development system. The PLL has achieved lock and is stable. Test mi ne le lvc mvc c0lt c1lt tails false allf somef evenp nmns1 njint Meaning Result is negative. < 0 Result is not equal to 0. ≠ 0 Result is less than or equal to 0. ≤ 0 Logical overflow clear. Mathematical overflow clear. Counter 0 less than 0. Counter 1 less than 0. Pseudorandom sequence bit clear. The condition is never satisfied in an if instruction. All False, no BIO input bits tested compared successfully. Some False, some BIO input bits tested did not compare successfully. Even Parity, from BMU operation. Not Minus 1, result of BMU operation. Not JINT, used by hardware development system. * Result is not representable in the 36-bit accumulators (36-bit overflow). † Bits 35—31 are not the same (32-bit overflow). Notes: Testing the state of the counters (c0 or c1) automatically increments the counter by one. The heads or tails condition is determined by a randomly set or cleared bit, respectively. The bit is randomly set with a probability of 0.5. A random rounding function can be implemented with either heads or tails. The random bit is generated by a ten-stage pseudorandom sequence generator (PSG) that is updated after either a heads or tails test. The pseudorandom sequence may be reset by writing any value to the pi register, except during an interrupt service routine (ISR). While in an ISR, writing to the pi register updates the register and does not reset the PSG. If not in an ISR, writing to the pi register resets the PSG. (The pi register is updated, but will be written with the contents of the PC on the next instruction.) Interrupts must be disabled when writing to the pi register. If an interrupt is taken after the pi write, but before pi is updated with the PC value, the ireturn instruction will not return to the correct location. If the RAND bit in the auc register is set, however, writing the pi register never resets the PSG. 40 Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 5 Software Architecture (continued) F3 ALU Instructions These instructions are implemented in the DSP1600 core. They allow accumulator two-operand operations with either another accumulator, the p register, or a 16-bit immediate operand (IM16). The result is placed in a destination accumulator that can be independently specified. All operations are done with the full 36 bits. For the accumulator with accumulator operations, both inputs are 36 bits. For the accumulator with p register operations, the p register is sign-extended into bits 35—32 before the operation. For the accumulator high with immediate operations, the immediate is sign-extended into bits 35—32 and the lower bits 15—0 are filled with zeros, except for the AND operation, for which they are filled with ones. These conventions allow the user to do operations with 32-bit immediates by programming two consecutive 16-bit immediate operations. The F3 ALU instructions are shown in Table 18. Table 18. F3 ALU Instructions F3 ALU Instructions† Cachable (One-Cycle) aD = aS + aT aD = aS – aT aD = aS & aT aD = aS | aT aD =aS ^ aT aS – aT aS & aT aD = aS + p aD = aS – p aD = aS & p aD = aS | p aD = aS ^ p aS – p aS & p Not Cachable (Two-Cycle)‡ aD = aSh + IM16 aD = aSh – IM16 aD = aSh & IM16 aD = aSh | IM16 aD = aSh ^ IM16 aSh – IM16 aSh & IM16 aD = aSl + IM16 aD = aSl – IM16 aD = aSl & IM16 aD = aSl | IM16 aD = aSl ^ IM16 aSl – IM16 aSl & IM16 Note: The F3 ALU instructions that do not have a destination accumulator are used to set flags for conditional operations, i.e., bit test operations. † If PC points to external memory, add programmed wait-states. ‡ The h and l are required notation in these instructions. F4 BMU Instructions The bit manipulation unit in the DSP1627 provides a set of efficient bit manipulation operations on accumulators. It contains four auxiliary registers, ar<0—3> (arM, M = 0, 1, 2, 3), two alternate accumulators (aa0—aa1), which can be shuffled with the working set, and four flags (oddp, evenp, mns1, and nmns1). The flags are testable by conditional instructions and can be read and written via bits 4—7 of the alf register. The BMU also sets the LMI, LEQ, LLV, and LMV flags in the psw register. ■ LMI = 1 if negative (i.e., bit 35 = 1) ■ LEQ = 1 if zero (i.e., bits 35—0 are 0) ■ LLV = 1 if (a) 36-bit overflow, or if (b) illegal shift on field width/offset condition ■ LMV = 1 if bits 31—35 are not the same (32-bit overflow) The BMU instructions and cycle times follow. (If PC points to external memory, add programmed wait-states.) All BMU instructions require 1 word of program memory unless otherwise noted. Please refer to the DSP1611/17/18/ 27 Digital Signal Processor Information Manual for further discussion of the BMU instructions. Lucent Technologies Inc. 41 Data Sheet March 2000 DSP1627 Digital Signal Processor 5 Software Architecture (continued) ■ Barrel Shifter: aD = aS >> IM16 Arithmetic right shift by immediate (36-bit, sign filled in); 2-cycle, 2-word. aD = aS >> arM Arithmetic right shift by arM (36-bit, sign filled in); 1-cycle. aD = aS >> aS Arithmetic right shift by aS (36-bit, sign filled in); 2-cycle. aD = aS >>> IM16 Logical right shift by immediate (32-bit shift, 0s filled in); 2-cycle, 2-word. aD = aS >>> arM Logical right shift by arM (32-bit shift, 0s filled in); 1-cycle. aD = aS >>> aS Logical right shift by aS (32-bit shift, 0s filled in); 2-cycle. aD = aS << IM16 Arithmetic left shift† by immediate (36-bit shift, 0s filled in); 2-cycle, 2-word. aD = aS << arM Arithmetic left shift† by arM (36-bit shift, 0s filled in); 1-cycle. aD = aS << aS Arithmetic left shift† by aS (36-bit shift, 0s filled in); 2-cycle. aD = aS <<< IM16 Logical left shift by immediate (36-bit shift, 0s filled in); 2-cycle, 2-word. aD = aS <<< arM Logical left shift by arM (36-bit shift, 0s filled in); 1-cycle. aD = aS <<< aS Logical left shift by aS (36-bit shift, 0s filled in); 2-cycle. † Not the same as the special function arithmetic left shift. Here, the guard bits in the destination accumulator are shifted into, not sign-extended. ■ ■ Normalization and Exponent Computation: aD = exp(aS) Detect the number of redundant sign bits in accumulator; 1-cycle. aD = norm(aS, arM) Normalize aS with respect to bit 31, with exponent in arM; 1-cycle. Bit Field Extraction and Insertion: aD = extracts(aS, IM16) Extraction with sign extension, field specified as immediate; 2-cycle, 2-word. aD = extracts(aS, arM) Extraction with sign extension, field specified in arM; 1-cycle. aD = extractz(aS, IM16) Extraction with zero extension, field specified as immediate; 2-cycle, 2-word. aD = extractz(aS, arM) Extraction with zero extension, field specified in arM; 1-cycle. aD = insert(aS, IM16) Bit field insertion, field specified as immediate; 2-cycle, 2-word. aD = insert(aS, arM) Bit field insertion, field specified in arM; 2-cycle. Note: The bit field to be inserted or extracted is specified as follows. The width (in bits) of the field is the upper byte of the operand (immediate or arM), and the offset from the LSB is in the lower byte. ■ Alternate Accumulator Set: aD = aS:aa0 Shuffle accumulators with alternate accumulator 0 (aa0); 1-cycle. aD = aS:aa1 Shuffle accumulators with alternate accumulator 1 (aa1); 1-cycle. Note: The alternate accumulator gets what was in aS. aD gets what was in the alternate accumulator. Table 19. Replacement Table for F3 ALU Instructions and F4 BMU Instructions Replace aD, aT, aS IM16 arM 42 Value a0 or a1 immediate ar<0—3> Meaning One of the two accumulators. 16-bit data, sign-, zero-, or one-extended as appropriate. One of the auxiliary BMU registers. Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 5 Software Architecture (continued) Cache Instructions Cache instructions require one word of program memory. The do instruction executes in one instruction cycle, and the redo instruction executes in two instruction cycles. (If PC points to external memory, add programmed waitstates.) Control instructions and long immediate values cannot be stored inside the cache. The instruction formats are as follows: ■ do K { ■ instr1 ■ instr2 ■ . ■ . ■ . ■ instrN ■ } ■ redo K Table 20. Replacement Table for Cache Instructions Replace K Instruction Encoding cloop† N 1 to 127 1 to 15 Meaning Number of times the instructions are to be executed taken from bits 0—6 of the cloop register. Number of times the instructions to be executed is encoded in the instruction. 1 to 15 instructions can be included. † The assembly-language statement, do cloop (or redo cloop), is used to specify that the number of iterations is to be taken from the cloop register. K is encoded as 0 in the instruction encoding to select cloop. When the cache is used to execute a block of instructions, the cycle timings of the instructions are as follows: 1. In the first pass, the instructions are fetched from program memory and the cycle times are the normal out-ofcache values, except for the last instruction in the block of NI instructions. This instruction executes in two cycles. 2. During pass two through pass K – 1, each instruction is fetched from cache and the in-cache timings apply. 3. During the last (Kth) pass, the block of instructions is fetched from cache and the in-cache timings apply, except that the timing of the last instruction is the same as if it were out-of-cache. 4. If any of the instructions access external memory, programmed wait-states must be added to the cycle counts. The redo instruction treats the instructions currently in the cache memory as another loop to be executed K times. Using the redo instruction, instructions are reexecuted from the cache without reloading the cache. The number of iterations, K, for a do or redo can be set at run time by first moving the number of iterations into the cloop register (7 bits unsigned), and then issuing the do cloop or redo cloop. At the completion of the loop, the value of cloop is decremented to 0; hence, cloop needs to be written before each do cloop or redo cloop. Lucent Technologies Inc. 43 Data Sheet March 2000 DSP1627 Digital Signal Processor 5 Software Architecture (continued) Data Move Instructions Data move instructions normally execute in two instruction cycles. (If PC or rM point to external memory, any programmed wait-states must be added. In addition, if PC and rM point to the same bank of DPRAM, then one cycle must be added.) Immediate data move instructions require two words of program memory; all other data move instructions require only one word. The only exception to these statements is a special case immediate load (short immediate) instruction. If a YAAU register is loaded with a 9-bit short immediate value, the instruction requires only one word of memory and executes in one instruction cycle. All data move instructions, except those doing long immediate loads, can be executed from within the cache. The data move instructions are as follows: ■ R = IM16 ■ aT[l] = R ■ SR = IM9 ■ Y=R ■ R=Y ■ Z:R ■ R = aS[l] ■ DR = *(OFFSET) ■ *(OFFSET) = DR Table 21. Replacement Table for Data Move Instructions Replace R DR aS, aT Y Value Meaning Any of the registers in Table 51 — r<0—3>, a0[l], a1[l], y[l], p, pl, x, Subset of registers accessible with direct addressing. pt, pr, psw a0, a1 High half of accumulator. Same as in multiply/ALU instructions. *rM, *rM++, *rM--, *rM++j Z IM16 IM9 OFFSET *rMzp, *rMpz, *rMm2, *rMjk 16-bit value 9-bit value 5-bit value from instruction 11-bit value in base register SR r<0—3>, rb, re, j, k Same as in multiply/ALU instructions. Long immediate data. Short immediate data for YAAU registers. Value in bits [15:5] of ybase register form the 11 most significant bits of the base address. The 5-bit offset is concatenated to this to form a 16-bit address. Subset of registers for short immediate. Notes: sioc, sioc2, tdms, tdms2, srta, and srta2 registers are not readable. When signed registers less than 16 bits wide (c0, c1, c2) are read, their contents are sign-extended to 16 bits. When unsigned registers less than 16 bits wide are read, their contents are zero-extended to 16 bits. Loading an accumulator with a data move instruction does not affect the flags. 44 Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 5 Software Architecture (continued) 5.2 Register Settings Tables 22 through 38 describe the programmable registers of the DSP1627 device. Table 40 describes the register settings after reset. Note that the following abbreviations are used in the tables: ■ x = don't care ■ R = read only ■ W = read/write The reserved (RSVD) bits in the tables should always be written with zeros to make the program compatible with future chip versions. Table 22. Serial I/O Control Registers sioc Bit Field 10 DODLY 9 LD Field DODLY Value 0 1 LD 0 1 00 01 10 11 0 1 0 1 0 1 0 1 0 1 0 1 0 1 CLK MSB OLD ILD OCK ICK OLEN ILEN 8 7 6 MSB CLK 5 OLD 4 ILD 3 OCK 2 ICK 1 OLEN 0 ILEN Description DO changes on the rising edge of OCK. DO changes on the falling edge of OCK. This delay in driving DO increases the hold time on DO by half a cycle of OCK. In active mode, ILD1 and/or OLD1 = ICK1/16, active SYNC1 = ICK1/[128/256*]. In active mode, ILD1 and/or OLD1 = OCK1/16, active SYNC1 = OCK1/[128/256*]. Active clock = CKI/2 (1X). Active clock = CKI/6 (1X). Active clock = CKI/8 (1X). Active clock = CKI/10 (1X). LSB first. MSB first. OLD1 is an input (passive mode). OLD1 is an output (active mode). ILD1 is an input (passive mode). ILD1 is an output (active mode). OCK1 is an input (passive mode). OCK1 is an output (active mode). ICK1 is an input (passive mode). ICK1 is an output (active mode). 16-bit output. 8-bit output. 16-bit input. 8-bit input. * See tdms register, SYNC field. sioc2‡ Bit Field 10 DODLY2 9 LD2 8 7 CLK2 6 MSB2 5 OLD2 4 ILD2 3 OCK2 2 ICK2 1 OLEN2 0 ILEN2 † See tdms register, SYNC field. ‡ The bit definitions of the sioc2 register are identical to the sioc register bit definitions. Lucent Technologies Inc. 45 Data Sheet March 2000 DSP1627 Digital Signal Processor 5 Software Architecture (continued) Table 23. Time-Division Multiplex Slot Registers tdms Bit Field 9 SYNCSP Field SYNCSP† 8 MODE 7 Value 0‡ 1 MODE TRANSMIT SLOT SYNC 0 1 1xxxxxx x1xxxxx xx1xxxx xxx1xxx xxxx1xx xxxxx1x xxxxxx1 1 0 6 5 4 3 TRANSMIT SLOT 2 1 0 SYNC Description 0 *. SYNC1 = ICK1/128 if LD = SYNC1 = OCK1/128 if LD = 1*. SYNC1 = ICK1/256 if LD = 0*. SYNC1 = OCK1/256 if LD = 1*. Multiprocessor mode off; DOEN1 is an input (passive mode). Multiprocessor mode on; DOEN1 is an output (active mode). Transmit slot 7. Transmit slot 6. Transmit slot 5. Transmit slot 4. Transmit slot 3. Transmit slot 2. Transmit slot 1. Transmit slot 0, SYNC1 is an output (active mode). SYNC1 is an input (passive mode). * See sioc register, LD field. ‡ Select this mode when in multiprocessor mode. tdms2§ Bit Field 9 SYNCSP2† 8 MODE2 7 6 5 4 3 TRANSMIT SLOT2 2 1 0 SYNC2 † See sioc register, LD field. ‡ Select this mode when in multiprocessor mode. § The tdms2 register bit definitions are identical to the tdms register bit definitions. 46 Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 5 Software Architecture (continued) Table 24. Serial Receive/Transmit Address Registers srta Bit Field 15 14 13 12 11 10 RECEIVE ADDRESS Field RECEIVE ADDRESS TRANSMIT ADDRESS 9 8 Value 1xxxxxxx x1xxxxxx xx1xxxxx xxx1xxxx xxxx1xxx xxxxx1xx xxxxxx1x xxxxxxx1 1xxxxxxx x1xxxxxx xx1xxxxx xxx1xxxx xxxx1xxx xxxxx1xx xxxxxx1x xxxxxxx1 7 6 5 4 3 2 1 TRANSMIT ADDRESS 0 Description Receive address 7. Receive address 6. Receive address 5. Receive address 4. Receive address 3. Receive address 2. Receive address 1. Receive address 0. Transmit address 7. Transmit address 6. Transmit address 5. Transmit address 4. Transmit address 3. Transmit address 2. Transmit address 1. Transmit address 0. srta2† Bit Field 15 14 13 12 11 10 9 RECEIVE ADDRESS2 8 7 6 5 4 3 2 1 TRANSMIT ADDRESS2 0 † The srta2 field definitions are identical to the srta register field definitions. Table 25. Multiprocessor Protocol Registers saddx Bit Field Write Read saddx2 15—8 X Read Protocol Field [7:0] 7—0 Write Protocol Field [7:0] 0 ‡ Bit Field Write Read 15—8 X Read Protocol2 Field [7:0] 7—0 Write Protocol2 Field [7:0] 0 ‡ The saddx2 field definitions are identical to the saddx register field definitions. Lucent Technologies Inc. 47 Data Sheet March 2000 DSP1627 Digital Signal Processor 5 Software Architecture (continued) Table 26. Processor Status Word (psw) Register Bit Field 15 14 13 DAU FLAGS Field DAU 12 11 X Value Wxxx xWxx xxWx xxxW W Wxxx xWxx xxWx xxxW W Wxxx xWxx xxWx xxxW FLAGS* a1[V] a1[35:32] a0[V] a0[35:32] 10 X 9 a1[V] 8 7 6 5 a1[35:32] 4 a0[V] 3 2 1 0 a0[35:32] Description LMI — logical minus when set (bit 35 = 1). LEQ — logical equal when set (bit [35:0] = 0). LLV — logical overflow when set. LMV — mathematical overflow when set. Accumulator 1 (a1) overflow when set. Accumulator 1 (a1) bit 35. Accumulator 1 (a1) bit 34. Accumulator 1 (a1) bit 33. Accumulator 1 (a1) bit 32. Accumulator 0 (a0) overflow when set. Accumulator 0 (a0) bit 35. Accumulator 0 (a0) bit 34. Accumulator 0 (a0) bit 33. Accumulator 0 (a0) bit 32. * The DAU flags can be set by either BMU or DAU operations. Table 27. Arithmetic Unit Control (auc) Register† Bit Field 8 RAND 7 X=Y= 6 Field RAND Value 0 X=Y= 1 0 1 CLR SAT ALIGN 1xx x1x xx1 1x x1 00 01 10 11 5 CLR 4 3 2 SAT 1 0 ALIGN Description Pseudorandom sequence generator (PSG) reset by writing the pi register only outside an interrupt service routine. PSG never reset by writing the pi register. Normal operation. All instructions which load the high half of the y register also load the x register, allowing single-cycle squaring with p = x * y. Clearing yl is disabled (enabled when 0). Clearing a1l is disabled (enabled when 0). Clearing a0l is disabled (enabled when 0). a1 saturation on overflow is disabled (enabled when 0). a0 saturation on overflow is disabled (enabled when 0). a0, a1 ← p. a0, a1 ← p/4. a0, a1 ← p x 4 (and zeros written to the two LSBs). a0, a1 ← p x 2 (and zero written to the LSB). † The auc is 9 bits [8:0]. The upper 7 bits [15:9] are always zero when read and should always be written with zeros to make the program compatible with future chip versions. The auc register is cleared at reset. 48 Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 5 Software Architecture (continued) Table 28. Parallel Host Interface Control (phifc) Register Bit Field 15—7 RSVD 6 PSOBEF Field Value PMODE 0 1 0 1 0 1 0 1 PSTROBE PSTRB PBSELF PFLAG 0 1 0 1 PFLAGSEL PSOBEF 0 1 5 PFLAGSEL 4 PFLAG 3 PBSELF 2 PSTRB 1 PSTROBE 0 PMODE Description 8-bit data transfers. 16-bit data transfers. Intel protocol: PIDS and PODS data strobes. Motorola protocol: PRWN and PDS data strobes. When PSTROBE = 1, PODS pin (PDS) active-low. When PSTROBE = 1, PODS pin (PDS) active-high. In either mode, PBSEL pin = 0 → pdx0 low byte. See Table 7. If PMODE = 0, PBSEL pin = 1 → pdx0 low byte. If PMODE = 1, PBSEL pin = 0 → pdx0 high byte. PIBF and POBE pins active-high. PIBF and POBE pins active-low. Normal. PIBF flag ORed with POBE flag and output on PIBF pin; POBE pin unchanged (output buffer empty). Normal. POBE flag as read through PSTAT register is active-low. Table 29. Interrupt Control (inc) Register Bit Field 15 JINT* 14—11 RSVD 10 OBE2 9 IBF2 8 TIME 7—6 RSVD 5—4 INT[1:0] 3 PIBF 2 POBE 1 OBE 0 IBF * JINT is a JTAG interrupt and is controlled by the HDS. It may be made unmaskable by the Lucent Technologies development system tools. Encoding: A 0 disables an interrupt; a 1 enables an interrupt. Table 30. Interrupt Status (ins) Register Bit Field 15 JINT 14—11 RSVD 10 OBE2 9 IBF2 8 TIME 7—6 RSVD 5—4 INT[1:0] 3 PIBF 2 POBE 1 OBE 0 IBF Encoding: A 0 indicates no interrupt. A 1 indicates an interrupt has been recognized and is pending or being serviced. If a 1 is written to bits 4, 5, or 8 of ins, the corresponding interrupt is cleared. Lucent Technologies Inc. 49 Data Sheet March 2000 DSP1627 Digital Signal Processor 5 Software Architecture (continued) Table 31. timerc Register Bit Field 15—7 RSVD Field DISABLE RELOAD T0EN PRESCALE 6 DISABLE Value 0 1 0 1 0 1 — 5 RELOAD 4 T0EN 3—0 PRESCALE Description Timer enabled. Timer and prescaler disabled. The period register and timer0 are not reset. Timer stops after counting down to 0. Timer automatically reloads and repeats indefinitely. Timer holds current count. Timer counts down to 0. See table below. PRESCALE Field PRESCALE Frequency of Timer Interrupts CKO/2 CKO/4 CKO/8 CKO/16 CKO/32 CKO/64 CKO/128 CKO/256 0000 0001 0010 0011 0100 0101 0110 0111 PRESCALE 1000 1001 1010 1011 1100 1101 1110 1111 Frequency of Timer Interrupts CKO/512 CKO/1024 CKO/2048 CKO/4096 CKO/8192 CKO/16384 CKO/32768 CKO/65536 Table 32. Phase-Locked Loop Control (pllc) Register Bit Field 15 PLLEN Field PLLEN PLLSEL ICP SEL5V LF[3:0] Nbits[2:0] Mbits[4:0] 50 14 PLLSEL Value 0 1 0 1 — 0 1 — — — 13 ICP 12 SEL5V 11—8 LF[3:0] 7—5 Nbits[2:0] 4—0 Mbits[4:0] Description PLL powered down. PLL powered up. DSP internal clock taken directly from CKI. DSP internal clock taken from PLL. Charge pump current selection (see Table 64 for proper value). 3 V operation (see Table 64 for proper value). 5 V operation (see Table 64 for proper value). Loop filter setting (see Table 64 for proper value). Encodes N, 1 ≤ N ≤ 8, where N = Nbits[2:0] + 2, unless Nbits[2:0] = 111, then N = 1. Encodes M, 2 ≤ M ≤ 20, where M = Mbits[4:0] + 2, fINTERNAL CLOCK = fCKI x (M/(2N)). Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 5 Software Architecture (continued) Table 33. sbit Register Bit Field 15 14 Field DIREC VALUE 13 12 11 DIREC[7:0] Value 1xxxxxxx x1xxxxxx xx1xxxxx xxx1xxxx xxxx1xxx xxxxx1xx xxxxxx1x xxxxxxx1 Rxxxxxxx xRxxxxxx xxRxxxxx xxxRxxxx xxxxRxxx xxxxxRxx xxxxxxRx xxxxxxxR 10 9 8 7 6 5 4 3 VALUE[7:0] 2 1 0 Description IOBIT7 is an output (input when 0). IOBIT6 is an output (input when 0). IOBIT5 is an output (input when 0). IOBIT4 is an output (input when 0). IOBIT3 is an output (input when 0). IOBIT2 is an output (input when 0). IOBIT1 is an output (input when 0). IOBIT0 is an output (input when 0). Reads the current value of IOBIT7. Reads the current value of IOBIT6. Reads the current value of IOBIT5. Reads the current value of IOBIT4. Reads the current value of IOBIT3. Reads the current value of IOBIT2. Reads the current value of IOBIT1. Reads the current value of IOBIT0. Table 34. cbit Register Bit Field 15 14 13 12 MODE/MASK[7:4] DIREC[n]* 1 (Output) 1 (Output) 1 (Output) 1 (Output) 0 (Input) 0 (Input) 0 (Input) 0 (Input) 11 10 9 8 MODE/MASK[3:0] MODE/MASK[n] 0 0 1 1 0 0 1 1 7 6 5 4 DATA/PAT[7:4] DATA/PAT[n] 0 1 0 1 0 1 0 1 3 2 1 0 DATA/PAT[3:0] Action Clear Set No Change Toggle No Test No Test Test for Zero Test for One * 0 ≤ n ≤ 7. Lucent Technologies Inc. 51 Data Sheet March 2000 DSP1627 Digital Signal Processor 5 Software Architecture (continued) Table 35. alf Register Bit Field 15 AWAIT Field AWAIT Value 1 0 1 0 — LOWPR FLAGS 14 LOWPR 13—0 FLAGS Action Power-saving standby mode or standard sleep enabled. Normal operation. The internal DPRAM is addressed beginning at 0x0000 in X space. The internal DPRAM is addressed beginning at 0xc000 in X space. See table below. Bit 13—8 7 6 5 4 3 2 1 0 Flag Reserved nmns1 mns1 evenp oddp somef somet allf allt Use — NOT-MINUS-ONE from BMU MINUS-ONE from BMU EVEN PARITY from BMU ODD PARITY from BMU SOME FALSE from BIO SOME TRUE from BIO ALL FALSE from BIO ALL TRUE from BIO Table 36. mwait Register Bit Field 15—12 EROM[3:0] 11—8 ERAMHI[3:0] 7—4 IO[3:0] 3—0 ERAMLO[3:0] If the EXM pin is high and the INT1 is low upon reset, the mwait register is initialized to all 1s (15 wait-states for all external memory). Otherwise, the mwait register is initialized to all 0s (0 wait-states) upon reset. Table 37. DSP1627 32-Bit JTAG ID Register Bit Field 31 RESERVED Field RESERVED SECURE CLOCK ROMCODE Value 0 0 1 01 10 11 — PART ID ROMCODE Letter Value 52 30 SECURE 27—19 ROMCODE 18—12 PART ID 11—0 0x03B Mask-Programmable Features — Nonsecure ROM option. Secure ROM option. Small-signal input clock option. Crystal oscillator input clock option. CMOS level input clock option. Users ROMCODE ID: The ROMCODE ID is the 9-bit binary value of the following expression: (20 x value for first letter) + (value of second letter), where the values of the letters are in the following table. For example, ROMCODE GK is (20 x 6) + (9) = 129 or 0 1000 0001. DSP1627x36 with 36K IROM and no EROM in MAP1 or MAP3. DSP1627x32 with 32K IROM and 16K EROM in MAP1 and MAP3. 0x1C 0x2C A 0 29—28 CLOCK B 1 C 2 D 3 E 4 F 5 G 6 H 7 J 8 K 9 L M N P R S T U W Y 10 11 12 13 14 15 16 17 18 19 Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 5 Software Architecture (continued) Table 38. ioc Register* Bit Field 15 RSVD 14 13 12 EXTROM CKO2 EBIOH 11 WEROM 10 ESIO2 9 SIOLBC 8—7 CKO[1:0] 6—4 RSVD 3—0 DENB[3:0] * The field definitions for the ioc register are different from the DSP1610. ioc Fields ioc Field EXTROM CKO2 EBIOH WEROM ESIO2 SIOLBC CKO[1:0] DENB3 DENB2 DENB1 DENB0 CKO2 Description If 1, sets AB15 low during external memory accesses when WEROM = 1. CKO configuration (see below). If 1, enables high half of BIO, IOBIT[4:7], and disables VEC[3:0] from pins. If 1, allows writing into external program (X) memory. If 1, enables SIO2 and low half of BIO, and disables PHIF from pins. If 1, DO1 and DO2 looped back to DI1 and DI2. CKO configuration (see below). If 1, delay EROM. If 1, delay ERAMHI. If 1, delay IO. If 1, delay ERAMLO. CKO1 CKO0 0 0 0 0 0 1 1X CKI CKI/(1 + W) 0 1 0 1 0 1 1 1 0 0 1 0 1 0 CKI CKI/(1 + W) 1 1 1 1 0 1 CKO Output Description PLL — CKI x M/(2N) Free-running clock. CKI x (M/(2N)) / [1 + W] Wait-stated clock.*, † 1 Held high.*, †, ‡ 0 Held low. CKI Output of CKI buffer. CKI x (M/(2N)) / [1 + W] Sequenced, wait-stated clock.*, †, ‡, § Reserved Reserved * The phase of CKI is synchronized by the rising edge of RSTB. † When SLOWCKI is enabled in the powerc register, these options reflect the low-speed internal ring oscillator. ‡ The wait-stated clock reflects the internal instruction cycle and may be stretched based on the mwait register setting (see Table 36). During sequenced external memory accesses, it completes one cycle. § The sequenced wait-stated clock completes two cycles during a sequenced external memory access and may be stretched based on the mwait register setting (see Table 36). Lucent Technologies Inc. 53 Data Sheet March 2000 DSP1627 Digital Signal Processor 5 Software Architecture (continued) Table 39. powerc Register The powerc register configures various power management modes. Bit Field 15 14 13 12 11 10 9—8 7 6 5 4 3—0 XTLOFF SLOWCKI NOCK INT0EN RSVD INT1EN RSVD SIO1DIS SIO2DIS PHIFDIS TIMERDIS RSVD Note: The reserved (RSVD) bits should always be written with zeros to make the program compatible with future chip versions. powerc fields Field XTLOFF SLOWCKI NOCK INT0EN INT1EN SIO1DIS SIO2DIS PHIFDIS TIMERDIS Description 1 = powerdown crystal oscillator or small-signal clock input. 1 = select ring oscillator clock (internal slow clock). 1 = disable internal processor clock. 1 = INT0 clears NOCK field. 1 = INT1 clears NOCK field. 1 = disable SIO1. 1 = disable SIO2. 1 = disable PHIF. 1 = disable timer. A • indicates that this bit is unknown on powerup reset and unaffected on subsequent reset. An S indicates that this bit shadows the PC. P indicates the value on an input pin, i.e., the bit in the register reflects the value on the corresponding input pin. Table 40. Register Settings After Reset Register Bits 15—0 Register Bits 15—0 r0 r1 r2 r3 j k rb re pt pr pi i p pl x y yl auc psw c0 c1 c2 sioc srta sdx tdms phifc pdx0 ybase •••••••••••••••• •••••••••••••••• •••••••••••••••• •••••••••••••••• •••••••••••••••• •••••••••••••••• 0000000000000000 0000000000000000 •••••••••••••••• •••••••••••••••• SSSSSSSSSSSSSSSS •••••••••••••••• •••••••••••••••• •••••••••••••••• •••••••••••••••• •••••••••••••••• •••••••••••••••• 0000000000000000 ••••00•••••••••• •••••••••••••••• ••••••••••••••• •••••••••••••••• ••••••0000000000 •••••••••••••••• •••••••••••••••• ••••••0000000000 0000000000000000 0000000000000000 •••••••••••••••• inc ins sdx2 saddx cloop mwait saddx2 sioc2 cbit sbit ioc jtag 0000000000000000 0000010000000110 •••••••••••••••• •••••••••••••••• 000000000••••••• 0000000000000000† •••••••••••••••• ••••••0000000000 •••••••••••••••• 00000000PPPPPPPP 0000000000000000 •••••••••••••••• a0 a0l a1 a1l timerc timer0 tdms2 srta2 powerc pllc ar0 ar1 ar2 ar3 •••••••••••••••• •••••••••••••••• •••••••••••••••• •••••••••••••••• ••••••••00000000 0000000000000000 ••••••0000000000 •••••••••••••••• 0000000000000000 0000000000000000 •••••••••••••••• •••••••••••••••• •••••••••••••••• •••••••••••••••• alf 00000000•••••••• † If EXM is high and INT1 is low when RSTB goes high, mwait will contain all ones instead of all zeros. 54 Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 5 Software Architecture (continued) 5.3 Instruction Set Formats This section defines the hardware-level encoding of the DSP1627 device instructions. Multiply/ALU Instructions Format 1: Multiply/ALU Read/Write Group Field Bit 15 T 13 14 12 D 10 11 S 9 F1 8 7 6 5 X 4 5 X 4 5 X 4 5 X 4 Y 3 2 1 0 1 0 1 0 Format 1a: Multiply/ALU Read/Write Group Field Bit 15 T 13 14 12 aT 10 11 S 9 F1 8 7 6 Y 3 2 Format 2: Multiply/ALU Read/Write Group Field Bit 15 T 13 14 12 D 10 11 S 9 F1 8 7 6 Y 3 2 Format 2a: Multiply/ALU Read/Write Group Field Bit 15 T 13 14 12 aT 10 11 S 9 F1 8 7 6 Y 3 2 1 0 3 CON 2 1 0 aT 0 1 2 1 0 Special Function Instructions Format 3: F2 ALU Special Functions Field Bit 15 T 13 14 12 11 D 10 S 9 F2 11 D S F3 Immediate Operand (IM16) 10 9 8 7 6 8 7 6 5 4 Format 3a: F3 ALU Operations Field Bit T 15 14 13 12 SRC2 5 4 3 Format 3b: BMU Operations Field Bit T 15 14 Lucent Technologies Inc. 13 12 11 D S F4[3—1] Immediate Operand (IM16) 10 9 8 7 6 0 F4[0] 5 4 AR 3 2 1 0 55 Data Sheet March 2000 DSP1627 Digital Signal Processor 5 Software Architecture (continued) Control Instructions Format 4: Branch Direct Group Field Bit T 15 14 JA 13 12 11 10 9 10 B 9 8 7 6 5 4 3 6 Reserved 5 4 3 2 1 0 2 1 0 0 1 0 1 0 Format 5: Branch Indirect Group Field Bit 15 14 T 13 12 11 8 7 Format 6: Conditional Branch Qualifier/Software Interrupt (icall) Field Bit 15 14 T 13 12 11 SI 10 Reserved 8 7 9 6 5 4 CON 3 2 6 5 4 3 Note: A branch instruction immediately follows except for a software interrupt (icall). Data Move Instructions Format 7: Data Move Group Field Bit 15 14 T 13 12 11 aT 10 R 9 8 7 Y/Z 2 Format 8: Data Move (immediate operand—2 words) Field Bit T 15 14 13 D 12 11 R Immediate Operand (IM16) 9 8 7 6 5 10 Reserved 4 3 2 1 0 1 0 0 0 Format 9: Short Immediate Group Field Bit 15 14 T 13 I 12 11 10 9 9 9 8 7 Short Immediate Operand (IM9) 6 5 4 3 2 Format 9a: Direct Addressing Field Bit 15 14 T 13 12 11 R/W 10 T 13 12 11 10 DR 8 7 6 1 5 8 7 6 5 4 OFFSET 3 2 1 4 K 3 Cache Instructions Format 10: Do/Redo Field Bit 56 15 14 NI 2 1 Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 5 Software Architecture (continued) Table 43. aT Field Specifies transfer accumulator. Field Descriptions aT 0 1 Table 41. T Field Specifies the type of instruction. T 0000x 00010 00011 00100 00101 00110 00111 01000 01000 01001 01001 01010 01011 01011 01100 01101 01110 01111 1000x 10010 10011 10100 10101 10110 10111 11000 11000 11001 11010 11011 11100 11101 11110 11110 11111 Operation goto JA Short imm j, k, rb, re Short imm r0, r1, r2, r3 Y = a1[l] F1 Z : aT[l] F1 Y F1 aT[l] = Y F1 Bit 0 = 0, aT = R Bit 0 = 1, aTl = R Bit 10 = 0, R = a0 Bit 10 = 1, R = a0l R = IM16 Bit 10 = 0, R = a1 Bit 10 = 1, R = a1l Y=R Z:R do, redo R=Y call JA ifc CON F2 if CON F2 Y = y[l] F1 Z : y[l] F1 x=Y F1 y[l] = Y F1 Bit 0 = 0, branch indirect Bit 0 = 1, F3 ALU y = a0 x = X F1 Cond. branch qualifier y = a1 x = X F1 Y = a0[l] F1 Z:y x=X F1 Bit 5 = 0, F4 ALU (BMU) Bit 5 = 1, direct addressing y=Yx=X F1 Format 4 9 9 1 2a 1 1a 7 7 7 7 8 7 7 7 7 10 7 4 3 3 1 2 1 1 5 3a 1 6 1 1 2 3b 9a 1 Register Accumulator 1 Accumulator 0 Table 44. S Field Specifies a source accumulator. S 0 1 Register Accumulator 0 Accumulator 1 Table 45. F1 Field Specifies the multiply/ALU function. F1 0000 0001 0010 0011 0100 0101 0110 0111 1000 1001 1010 1011 1100 1101 1110 1111 Operation aD = pp = x * y aD = aS + pp = x * y p=x*y aD = aS – pp = x * y aD = p aD = aS + p nop aD = aS – p aD = aS | y aD = aS ^ y aS & y aS – y aD = y aD = aS + y aD = aS & y aD = aS – y Table 46. X Field Specifies the addressing of ROM data in two-operand multiply/ALU instructions. Specifies the high or low half of an accumulator or the y register in one-operand multiply/ALU instructions. Table 42. D Field Operation Two-Operand Multiply/ALU 0 *pt++ Specifies a destination accumulator. 1 D 0 1 Lucent Technologies Inc. Register Accumulator 0 Accumulator 1 X 0 1 *pt++i One-Operand Multiply/ALU aTl, yl aTh, yh 57 Data Sheet March 2000 DSP1627 Digital Signal Processor 5 Software Architecture (continued) Table 49. F2 Field Specifies the special function to be performed. Table 47. Y Field Specifies the form of register indirect addressing with postmodification. Y 0000 Operation *r0 0001 *r0++ *r0-- 0010 0011 0100 0101 0110 0111 1000 1001 1010 1011 *r0++j *r1 *r1++ *r1-*r1++j *r2 *r2++ *r2-- 1100 *r2++j *r3 1101 *r3++ 1110 *r3-*r3++j 1111 Table 48. Z Field Specifies the form of register indirect compound addressing with postmodification. 58 Z 0000 Operation *r0zp 0001 *r0pz 0010 *r0m2 0011 *r0jk 0100 *r1zp 0101 *r1pz 0110 *r1m2 0111 *r1jk 1000 *r2zp 1001 *r2pz 1010 *r2m2 1011 *r2jk 1100 *r3zp 1101 *r3pz 1110 *r3m2 1111 *r3jk F2 0000 0001 0010 0011 0100 0101 0110 0111 1000 1001 1010 1011 1100 1101 1110 1111 Operation aD = aS >> 1 aD = aS << 1 aD = aS >> 4 aD = aS << 4 aD = aS >> 8 aD = aS << 8 aD = aS >> 16 aD = aS << 16 aD = p aDh = aSh + 1 aD = ~aS aD = rnd(aS) aD = y aD = aS + 1 aD = aS aD = – aS Table 50. CON Field Specifies the condition for special functions and conditional control instructions. CON 00000 00001 00010 00011 00100 00101 00110 00111 01000 01001 01010 01011 01100 01101 Condition mi pl eq ne lvs lvc mvs mvc heads tails c0ge c0lt c1ge c1lt Other codes Reserved CON 01110 01111 10000 10001 10010 10011 10100 10101 10110 10111 11000 11001 11010 11011 11100 — Condition true false gt le allt allf somet somef oddp evenp mns1 nmns1 npint njint lock — Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 5 Software Architecture (continued) Table 52. B Field Table 51. R Field Specifies the type of branch instruction (except software interrupt). Specifies the register for data move instructions. R 000000 000001 000010 000011 000100 000101 000110 000111 001000 001001 001010 001011 001100 001101 001110 001111 010000 010001 010010 010011 010100 010101 010110 010111 011000 011001 011010 011011 011100 011101 011110 011111 Register r0 r1 r2 r3 j k rb re pt pr pi i p pl pllc Reserved x y yl auc psw c0 c1 c2 sioc srta sdx tdms phifc pdx0 Reserved ybase R 100000 100001 100010 100011 100100 100101 100110 100111 101000 101001 101010 101011 101100 101101 101110 101111 110000 110001 110010 110011 110100 110101 110110 110111 111000 111001 111010 111011 111100 111101 111110 111111 Register inc ins sdx2 saddx cloop mwait saddx2 sioc2 cbit sbit ioc jtag Reserved Reserved Reserved Reserved a0 a0l a1 a1l timerc timer0 tdms2 srta2 powerc Reserved ar0 ar1 ar2 ar3 Reserved alf B 000 001 010 011 1xx Operation return ireturn goto pt call pt Reserved Table 53. DR Field DR Value 0000 0001 0010 0011 0100 0101 0110 0111 1000 1001 1010 1011 1100 1101 1110 1111 Register r0 r1 r2 r3 a0 a0l a1 a1l y yl p pl x pt pr psw Table 54. I Field Specifies a register for short immediate data move instructions. I 00 01 10 11 Register r0/j r1/k r2/rb r3/re Table 55. SI Field Specifies when the conditional branch qualifier instruction should be interpreted as a software interrupt instruction. SI 0 1 Lucent Technologies Inc. Operation Not a software interrupt Software interrupt 59 Data Sheet March 2000 DSP1627 Digital Signal Processor 5 Software Architecture (continued) NI Field Number of instructions to be loaded into the cache. Zero implies redo operation. K Field Number of times the NI instructions in cache are to be executed. Zero specifies use of value in cloop register. JA Field 12-bit jump address. R/W Field A zero specifies a write, *(O) = DR. A one specifies a read, DR = *(O). Table 56. F3 Field Specifies the operation in an F3 ALU instruction. F3 1000 1001 1010 1011 1101 1110 1111 aD = aS[h, l] aD = aS[h, l] aS[h, l] aS[h, l] aD = aS[h, l] aD = aS[h, l] aD = aS[h, l] Operation | {aT, IM16, p} ^ {aT, IM16, p} & {aT, IM16, p} – {aT, IM16, p} + {aT, IM16, p} & {aT, IM16, p} – {aT, IM16, p} Table 58. BMU Encodings F4 0000 0001 0000 0001 1000 1001 1000 1001 1100 1101 1100 1101 0000 0001 1110 0010 1110 0010 1110 1010 0111 0111 AR 00xx 00xx 10xx 10xx 0000 0000 1000 1000 0000 0000 1000 1000 1100 11xx 0000 00xx 0100 01xx 1000 10xx 0000 0001 Operation aD = aS >> arM aD = aS << arM aD = aS >>> arM aD = aS <<< arM aD = aS >> aS aD = aS << aS aD = aS >>> aS aD = aS <<< aS aD = aS >> IM16 aD = aS << IM16 aD = aS >>> IM16 aD = aS <<< IM16 aD = exp(aS) aD = norm(aS, arM) aD = extracts(aS, IM16) aD = extracts(aS, arM) aD = extractz(aS, IM16) aD = extractz(aS, arM) aD = insert(aS, IM16) aD = insert(aS, arM) aD = aS:aa0 aD = aS:aa1 Note: xx encodes the auxiliary register to be used. 00 (ar0), 01(ar1), 10 (ar2), or 11(ar3). Table 57. SRC2 Field Specifies operands in an F3 ALU instruction. SRC2 00 10 01 11 60 Operands aSl, IM16 aSh, IM16 aS, aT aS, p Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 6 Signal Descriptions AB[15:0] 16 DB[15:0] 16 RWN EXTERNAL MEMORY INTERFACE EXM 2 4 EROM ERAMHI IO RSTB CKO CKI2 CKI STOP INT[1:0] VEC[3:0] OR IOBIT[4:7] IACK TRAP SYSTEM INTERFACE OR CONTROL I/O INTERFACE ERAMLO PSTAT OR DO2 PODS OR OLD2 PCSN OR OCK2 DO1 POBE OR OBE2 DSP1627 PBSEL OR SYNC2 OLD1 PB0 OR ICK2 OCK1 OBE1 SERIAL INTERFACE #1 PIDS OR ILD2 PARALLEL HOST INTERFACE OR SERIAL INTERFACE #2 AND CONTROL I/O INTERFACE PB1 OR DI2 DI1 ILD1 PIBF OR IBF2 PB2 OR DOEN2 ICK1 PB3 OR SADD2 IBF1 4 SYNC1 PB[7:4] OR IOBIT[3:O] SADD1 JTAG TEST INTERFACE DOEN1 TDI TDO TCK TMS 5-4006 (C) Figure 8. DSP1627 Pinout by Interface Figure 8 shows the pinout for the DSP1627. The signals can be separated into five interfaces as shown. These interfaces and the signals that comprise them are described below. 6.1 System Interface The system interface consists of the clock, interrupt, and reset signals for the processor. RSTB Reset: Negative assertion. A high-to-low transition causes the processor to enter the reset state. The auc, powerc, sioc, sioc2, phifc, pdx0, tdms, tdms2, timerc, timer0, sbit (upper byte), inc, ins (except OBE, OBE2, Lucent Technologies Inc. and PODS status bits set), alf (upper 2 bits, AWAIT and LOWPR), ioc, rb, and re registers are cleared. The mwait register is initialized to all 0s (zero wait-states) unless the EXM pin is high and the INT1 pin is low. In that case, the mwait register is initialized to all 1s (15 wait-states). Reset clears IACK, VEC[3:0]/IOBIT[4:7], IBF, and IBF2. The DAU condition flags are not affected by reset. IOBIT[7:0] are initialized as inputs. If any of the IOBIT pins are switched to outputs (by writing sbit), their initial value will be logic zero (see Table 40, Register Settings After Reset). Upon negation of the signal, the processor begins execution at location 0x0000 in the active memory map (see Section 4.4, Memory Maps and Wait-States). 61 Data Sheet March 2000 DSP1627 Digital Signal Processor 6 Signal Descriptions (continued) ■ A free-running output clock that runs at the CKI rate, independent of the powerc register setting. This option is only available with the crystal and small-signal clock options. When the PLL is selected, the CKO frequency equals the input CKI frequency regardless of how the PLL is programmed. ■ A logic 0. ■ A logic 1. CKI Input Clock: A mask-programmable option selects one of three possible input buffers for the CKI pin (see Section 7, Mask-Programmable Options, and Table 1, Pin Descriptions). The internal CKI from the output of the selected input buffer can then drive the internal processor clock directly (1X) or drive the on-chip PLL (see Section 4.13). The PLL allows the CKI input clock to be at a lower frequency than the internal processor clock. CKI2 Input Clock 2: Used with mask-programmable input clock options which require an external crystal or small signal differential across CKI and CKI2 (see Table 1, Pin Descriptions). When the CMOS option is selected, this pin should be tied to VSSA. STOP INT[1:0] Processor Interrupts 0 and 1: Positive assertion. Hardware interrupt inputs to the DSP1627. Each is enabled via the inc register. When enabled and asserted, each cause the processor to vector to the memory location described in Table 4. INT1 is used in conjunction with EXM to select the desired reset initialization of the mwait register (see Table 36). When both INT0 and RSTB are asserted, all output and bidirectional pins (except TDO, which 3-states by JTAG control) are put in a 3-state condition. Stop Input Clock: Negative assertion. A high-to-low transition synchronously stops all of the internal processor clocks leaving the processor in a defined state. Returning the pin high will synchronously restart the processor clocks to continue program execution from where it left off without any loss of state. This hardware feature has the same effect as setting the NOCK bit in the powerc register (see Table 39). VEC[3:0] CKO Interrupt Acknowledge: Positive assertion. IACK signals when an interrupt is being serviced by the DSP1627. IACK remains asserted while in an interrupt service routine, and is cleared when the ireturn instruction is executed. Clock Out: Buffered output clock with options programmable via the ioc register (see Table 38). The selectable CKO options (see Tables 38 and 29) are as follows: ■ A free-running output clock at the frequency of the internal processor clock; runs at the internal ring oscillator frequency when SLOWCKI is enabled. ■ A wait-stated clock based on the internal instruction cycle; runs at the internal ring oscillator frequency when SLOWCKI is enabled. ■ A sequenced, wait-stated clock based on the EMI sequencer cycle; runs at the internal ring oscillator frequency when SLOWCKI is enabled. 62 Interrupt Output Vector: These four pins indicate which interrupt is currently being serviced by the device. Table 4 shows the code associated with each interrupt condition. VEC[3:0] are multiplexed with IOBIT[4:7]. IACK TRAP Trap Signal: Positive assertion. When asserted, the processor is put into the trap condition, which normally causes a branch to the location 0x0046. The hardware development system (HDS) can configure the trap pin to cause an HDS trap, which causes a branch to location 0x0003. Although normally an input, the pin can be configured as an output by the HDS. As an output, the pin can be used to signal an HDS breakpoint in a multiple processor environment. Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 6 Signal Descriptions (continued) EROM 6.2 External Memory Interface External ROM Enable Signal: Negative assertion. When asserted, the signal indicates an access to external program memory (see Table 5, Instruction/Coefficient Memory Maps). This signal's leading edge can be delayed via the ioc register (see Table 38). The external memory interface is used to interface the DSP1627 to external memory and I/O devices. It supports read/write operations from/to program and data memory spaces. The interface supports four external memory segments. Each external memory segment can have an independent number of software-programmable wait-states. One hardware address is decoded, and an enable line is provided, to allow glueless I/O interfacing. AB[15:0] External Memory Address Bus: Output only. This 16-bit bus supplies the address for read or write operations to the external memory or I/O. During external memory accesses, AB[15:0] retain the value of the last valid external access. DB[15:0] External Memory Data Bus: This 16-bit bidirectional data bus is used for read or write operations to the external memory or I/O. RWN Read/Write Not: When a logic 1, the pin indicates that the memory access is a read operation. When a logic 0, the memory access is a write operation. ERAMHI External RAM High Enable Signal: Negative assertion. When asserted, the signal indicates an access to external data memory addresses 0x8000 through 0xFFFF (see Table 6, Data Memory Map). This signal's leading edge can be delayed via the ioc register (see Table 38). ERAMLO External RAM Low Enable Signal: Negative assertion. When asserted, the signal indicates an access to external data memory addresses 0x4100 through 0x7FFF (see Table 6, Data Memory Map). This signal's leading edge can be delayed via the ioc register (see Table 38). IO External I/O Enable Signal: Negative assertion. When asserted, the signal indicates an access to external data memory addresses 0x4000 through 0x40FF (see Table 6, Data Memory Map). This memory segment is intended for memory-mapped I/O. This signal's leading edge can be delayed via the ioc register (see Table 38). EXM External Memory Select: Input only. This signal is latched into the device on the rising edge of RSTB. The value of EXM latched in determines whether the internal ROM is addressable in the instruction/coefficient memory map. If EXM is low, internal ROM is addressable. If EXM is high, only external ROM is addressable in the instruction/coefficient memory map (see Table 5, Instruction/Coefficient Memory Maps). EXM chooses between MAP1 or MAP2 and between MAP3 or MAP4. Lucent Technologies Inc. 63 Data Sheet March 2000 DSP1627 Digital Signal Processor 6 Signal Descriptions (continued) OCK1 6.3 Serial Interface #1 Output Clock: The clock for serial output data. In active mode, OCK1 is an output; in passive mode, OCK1 is an input, according to the sioc register OCK field (see Table 22). Input has typically 0.7 V hysteresis. The serial interface pins implement a full-featured synchronous/asynchronous serial I/O channel. In addition, several pins offer a glueless TDM interface for multiprocessing communication applications (see Figure 5, Multiprocessor Communications and Connections). OLD1 Data Input: Serial data is latched on the rising edge of ICK1, either LSB or MSB first, according to the sioc register MSB field (see Table 22). Output Load: The clock for loading the output shift register, osr, from the output buffer sdx[out]. A falling edge of OLD1 indicates the beginning of a serial output word. In active mode, OLD1 is an output; in passive, OLD1 is an input, according to the sioc register OLD field (see Table 22). Input has typically 0.7 V hysteresis. ICK1 OBE1 Input Clock: The clock for serial input data. In active mode, ICK1 is an output; in passive mode, ICK1 is an input, according to the sioc register ICK field (see Table 22). Input has typically 0.7 V hysteresis. Output Buffer Empty: Positive assertion. OBE1 is asserted when the output buffer, sdx[out], is emptied (moved to the output shift register for transmission). It is cleared with a write to the buffer, as in sdx = a0. OBE1 is also set by asserting RSTB. DI1 ILD1 SADD1 Input Load: The clock for loading the input buffer, sdx[in], from the input shift register isr. A falling edge of ILD1 indicates the beginning of a serial input word. In active mode, ILD1 is an output; in passive mode, ILD1 is an input, according to the sioc register ILD field (see Table 22). Input has typically 0.7 V hysteresis. IBF1 Input Buffer Full: Positive assertion. IBF1 is asserted when the input buffer, sdx[in], is filled. IBF1 is negated by a read of the buffer, as in a0 = sdx. IBF1 is also negated by asserting RSTB. Serial Address: Negative assertion. A 16-bit serial bit stream typically used for addressing during multiprocessor communication between multiple DSP16xx devices. In multiprocessor mode, SADD1 is an output when the tdms time slot dictates a serial transmission; otherwise, it is an input. Both the source and destination DSP can be identified in the transmission. SADD1 is always an output when not in multiprocessor mode and can be used as a second 16-bit serial output. See the DSP1611/17/18/27 Digital Signal Processor Information Manual for additional information. SADD1 is 3-stated when DOEN1 is high. When used on a bus, SADD1 should be pulled high through a 5 kΩ resistor. DO1 SYNC1 Data Output: The serial data output from the output shift register (osr), either LSB or MSB first (according to the sioc register MSB field). DO1 changes on the rising edges of OCK1. DO1 is 3-stated when DOEN1 is high. DOEN1 Data Output Enable: Negative assertion. An input when not in the multiprocessor mode. DO1 and SADD1 are enabled only if DOEN1 is low. DOEN1 is bidirectional when in the multiprocessor mode (tdms register MODE field set). In the multiprocessor mode, DOEN1 indicates a valid time slot for a serial output. 64 Multiprocessor Synchronization: Typically used in the multiprocessor mode, a falling edge of SYNC1 indicates the first word (time slot 0) of a TDM I/O stream and causes the resynchronization of the active ILD1 and OLD1 generators. SYNC1 is an output when the tdms register SYNC field is set (i.e., selects the master DSP and uses time slot 0 for transmit). As an input, SYNC1 must be tied low unless part of a TDM interface. When used as an output, SYNC1 = [ILD1/OLD1]/8 or 16, depending on the setting of the SYNCSP field of the tdms register. When configured as described above, SYNC1 can be used to generate a slow clock for SIO operations. Input has typically 0.7 V hysteresis. Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 6 Signal Descriptions (continued) PODS 6.4 Parallel Host Interface or Serial Interface #2 and Control I/O Interface Parallel Output Data Strobe: An input pin, software configurable to support both Intel and Motorola protocols. This interface pin multiplexes a parallel host interface with a second serial I/O interface and a 4-bit I/O interface. The interface selection is made by writing the ESIO2 bit in the ioc register (see Table 38 and Section 4.1). The functions and signals for the second SIO correspond exactly with those in SIO #1. Therefore, the pin descriptions below discuss only PHIF and BIO pin functionality. In Intel mode: Negative assertion. When PODS is pulled low by an external device, the DSP1627 places the contents of the parallel output register, pdx0, onto the PB bus. PB[7:0] PIBF Parallel I/O Data Bus: This 8-bit bidirectional bus is used to input data to, or output data from, the PHIF. Parallel Input Buffer Full: An output pin with positive assertion; configurable in software. This flag is cleared after reset, indicating an empty input buffer pdx0[in]. Note that PB[3:0] are pin multiplexed with SIO2 functionality, and PB[7:4] are pin multiplexed with BIO unit pins IOBIT[3:0] (see Section 4.1). PCSN Peripheral Chip Select Not: Negative assertion. PCSN is an input. While PCSN is low, the data strobes PIDS and PODS are enabled. While PCSN is high, the DSP1627 ignores any activity on PIDS and PODS. PBSEL Peripheral Byte Select: An input pin, configurable in software. Selects the high or low byte of pdx0 available for host accesses. PSTAT Peripheral Status Select: PSTAT is an input. When a logic 0, the PHIF will output the pdx0[out] register on the PB bus. When a logic 1, the PHIF will output the contents of the PSTAT register on PB[7:0]. PIDS Parallel Input Data Strobe: An input pin, software configurable to support both Intel and Motorola protocols. In Intel mode: Negative assertion. PIDS is pulled low by an external device to indicate that data is available on the PB bus. The DSP latches data on the PB bus on the rising edge (low-to-high transition) of PIDS or PCSN, whichever comes first. In Motorola mode: PIDS(PRWN*) functions as a read/ write strobe. The external device sets PIDS(PRWN*) to a logic 0 to indicate that data is available on the PB bus (write operation by the external device). A logic 1 on PIDS(PRWN*) indicates an external read operation by the external device. Lucent Technologies Inc. In Motorola mode: Software-configurable assertion level. The external device uses PODS(PDS*) as its data strobe for both read and write operations. PIBF is set immediately after the rising edge of PIDS or PCSN, indicating that data has been latched into the pdx0[in] register. When the DSP1627 reads the contents of this register, emptying the buffer, the flag is cleared. Configured in software, PIBF may become the logical OR of the PIBF and POBE flags. POBE Parallel Output Buffer Empty: An output pin with positive assertion; configurable in software. This flag is set after reset, indicating an empty output buffer pdx0[out]. POBE is set immediately after the rising edge of PODS or PCSN, indicating that the data in pdx0[out] has been driven onto the PB bus. When the DSP1627 writes to pdx0[out], filling the buffer, this flag is cleared. 6.5 Control I/O Interface This interface is used for status and control operations provided by the bit I/O unit of the DSP1627. It is pin multiplexed with the PHIF and VEC[3:0] pins (see Section 4.1). Setting the ESIO2 and EBIOH bits in the ioc register provides a full 8-bit BIO interface at the associated pins. IOBIT[7:0] I/O Bits [7:0]: Each of these bits can be independently configured as either an input or an output. As outputs, they can be independently set, toggled, or cleared. As inputs, they can be tested independently or in combinations for various data patterns. * Motorola mode signal name. 65 Data Sheet March 2000 DSP1627 Digital Signal Processor 6 Signal Descriptions (continued) TDO 6.6 JTAG Test Interface Test Data Output: JTAG serial output signal. Serialscanned data and status bits are output on this pin. The JTAG test interface has features that allow programs and data to be downloaded into the DSP via four pins. This provides extensive test and diagnostic capability. In addition, internal circuitry allows the device to be controlled through the JTAG port to provide on-chip in-circuit emulation. Lucent Technologies provides hardware and software tools to interface to the on-chip HDS via the JTAG port. Note: The DSP1627 provides all JTAG/IEEE 1149.1 standard test capabilities including boundary scan. See the DSP1611/17/18/27 Digital Signal Processor Information Manual for additional information on the JTAG test interface. TMS Test Mode Select: JTAG mode control signal that, when combined with TCK, controls the scan operations. This pin has an internal pull-up resistor. TCK Test Clock: JTAG serial shift clock. This signal clocks all data into the port through TDI, and out of the port through TDO, and controls the port by latching the TMS signal inside the state-machine controller. TDI Test Data Input: JTAG serial input signal. All serialscanned data and instructions are input on this pin. This pin has an internal pull-up resistor. 66 Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 7 Mask-Programmable Options The DSP1627 contains a ROM that is mask-programmable. The selection of several programmable features is made when a custom ROM is encoded. These features select the input clock options, the instruction/coefficient memory map option, and the hardware emulation or ROM security option, as summarized in Table 59. Table 59. DSP1627 ROM Options Features Input Clock Memory Map ROM Security Options CMOS Level Small Signal Crystal DSP1627x36 DSP1627x32 Nonsecure Secure Comments 2.7 V, 3.0 V, and 5.0 V. 2.7 V, 3.0 V, and 5.0 V. 2.7 V, 3.0 V, and 5.0 V. 36 Kwords IROM, no EROM in MAP1 or MAP3. 32 Kwords IROM, 16 Kwords EROM in MAP1 and MAP3. Specify and link 1627hds.v#*, allows emulation. Specify and link crc16.v#†, no emulation capability. * 1627hds.v# (# indicates the current version number) is the relocatable HDS object code. It uses approximately 140 words and must reside in the first 4 Kwords of ROM. † crc16.v# is the cyclic redundancy check object code. It uses approximately 75 words and must reside in the first 4 Kwords of ROM. See the DSP1600 Support Tools Manual for detailed information. 7.1 Input Clock Options For all input options, the input clock CKI can run at some fraction of the internal clock frequency by setting the PLL multiplication factors appropriately (see Section 4.12, Clock Synthesis). When the PLL is bypassed, the input clock CKI frequency is the internal clock frequency. If the mask option for using an external crystal is chosen, the internal oscillator may be used as a noninverting input buffer by supplying a CMOS level to the CKI pin and leaving the CKI2 pin open. 7.2 Memory Map Options The DSP1627 offers a DSP1627x36 or a DSP1627x32 where the difference is in the instruction/coefficient memory maps. The DSP1627x36 contains 36 Kwords of internal ROM (IROM), but it doesn’t support the use of IROM and external ROM (EROM) in the same memory map. The DSP1627x32 supports the use of only 32 Kwords of IROM with 16 Kwords of EROM in the same memory map. See Section 4.4 Memory Maps and Wait-States for further description. 7.3 ROM Security Options The DSP1600 Hardware Development System (HDS) provides on-chip in-circuit emulation and requires that the relocatable HDS code be linked to the application code. This code's object file is called 1627hds.v#, where # is a unique version identifier. Refer to the DSP1627-ST software tools release for more specific information. If on-chip in-circuit emulation is desired, a nonsecure ROM must be chosen. If ROM security is desired with the DSP1627, the HDS cannot be used. To provide testing of the internal ROM contents on a secure ROM device, a cyclic redundancy check (CRC) program is called by and linked with the user's source code. The CRC code resides in the first 4 Kwords of ROM. See the DSP1600 Support Tools Manual for more detailed information. Lucent Technologies Inc. 67 DSP1627 Digital Signal Processor Data Sheet March 2000 8 Device Characteristics 8.1 Absolute Maximum Ratings Stresses in excess of the absolute maximum ratings can cause permanent damage to the device. These are absolute stress ratings only. Functional operation of the device is not implied at these or any other conditions in excess of those given in the operational sections of the data sheet. Exposure to absolute maximum ratings for extended periods can adversely affect device reliability. External leads can be bonded and soldered safely at temperatures of up to 300 °C. Voltage Range on VDD with Respect to Ground Using Devices Designed for 5 V Operation .............–0.5 V to +7 V Voltage Range on VDD with Respect to Ground Using Devices Designed for 3 V Operation ..........–0.5 V to +4.6 V Voltage Range on Any Pin ............................................................................................ .VSS – 0.5 V to VDD + 0.5 V Power Dissipation................................................................................................................................................ 1 W Ambient Temperature Range ......................................................................................................... –40 °C to +85 °C Storage Temperature Range .................................................................................................................... –65 °C to +150 °C 8.2 Handling Precautions All MOS devices must be handled with certain precautions to avoid damage due to the accumulation of static charge. Although input protection circuitry has been incorporated into the devices to minimize the effect of this static buildup, proper precautions should be taken to avoid exposure to electrostatic discharge during handling and mounting. Lucent Technologies employs a human-body model for ESD susceptibility testing. Since the failure voltage of electronic devices is dependent on the current, voltage, and hence, the resistance and capacitance, it is important that standard values be employed to establish a reference by which to compare test data. Values of 100 pF and 1500 Ω are the most common and are the values used in the Lucent Technologies human-body model test circuit. The breakdown voltage for the DSP1627 is greater than 2000 V. 8.3 Recommended Operating Conditions Table 60. Recommended Operating Conditions Maximum Instruction Rate (MIPS) Device Speed 50 20 ns 80 12.5 ns 100 10 ns 70 14 ns 90 11 ns Input Clock Package Supply Voltage Ambient Temperature VDD (V) TA (°°C) Min Max Min Max CMOS, small-signal, BQFP 2.7 3.3 –40 85 crystal or TQFP CMOS, small-signal, BQFP 2.7 3.3 –40 85 crystal or TQFP CMOS, small-signal, BQFP 3.0 3.6 –40 85 crystal or TQFP CMOS, small-signal, BQFP 4.75 5.25 –40 85 crystal or TQFP CMOS, small-signal, BQFP 4.75 5.25 –40 85 crystal or TQFP The ratio of the instruction cycle rate to the input clock frequency is 1:1 without the PLL (referred to as 1X operation) and M/(2N) with the PLL selected (see Section 4.12). Device speeds greater than 50 MIPS do not support 1X operation; use the PLL. 68 Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 8 Device Characteristics (continued) 8.4 Package Thermal Considerations The recommended operating temperature specified above is based on the maximum power, package type, and maximum junction temperature. The following equations describe the relationship between these parameters. If the applications' maximum power is less than the worst-case value, this relationship determines a higher maximum ambient temperature or the maximum temperature measured at top dead center of the package. TA = TJ – P x ΘJA TTDC = TJ – P x ΘJ-TDC where TA is the still-air ambient temperature and TTDC is the temperature measured by a thermocouple at the top dead center of the package. Maximum Junction Temperature (TJ) in 100-Pin BQFP ................................................................................. 125 °C 100-pin BQFP Maximum Thermal Resistance in Still-Air-Ambient (ΘJA) ..................................................... 55 °C/W 100-pin BQFP Maximum Thermal Resistance, Junction to Top Dead Center (ΘJ-TDC) ............................... 12 °C/W Maximum Junction Temperature (TJ) in 100-Pin TQFP ................................................................................. 125 °C 100-pin TQFP Maximum Thermal Resistance in Still-Air-Ambient (ΘJA) ..................................................... 30 °C/W 100-pin TQFP Maximum Thermal Resistance, Junction to Top Dead Center (ΘJ-TDC) ................................. 6 °C/W WARNING: Due to package thermal constraints, proper precautions in the user's application should be taken to avoid exceeding the maximum junction temperature of 125 °C. Otherwise, the device will be affected adversely. Lucent Technologies Inc. 69 Data Sheet March 2000 DSP1627 Digital Signal Processor 9 Electrical Characteristics and Requirements The following electrical characteristics are preliminary and are subject to change. Electrical characteristics refer to the behavior of the device under specified conditions. Electrical requirements refer to conditions imposed on the user for proper operation of the device. The parameters below are valid for the conditions described in Section 8.3, Recommended Operating Conditions. Table 61. Electrical Characteristics and Requirements Parameter Input Voltage: Low High Input Current (except TMS, TDI): Low (VIL = 0 V, VDD = 5.25 V) High (VIH = 5.25 V, VDD = 5.25 V) Input Current (TMS, TDI): Low (VIL = 0 V, VDD = 5.25 V) High (VIH = 5.25 V, VDD = 5.25 V) Output Low Voltage: Low (IOL = 2.0 mA) Low (IOL = 50 µA) Output High Voltage: High (IOH = –2.0 mA) High (IOH = –50 µA) Output 3-State Current: Low (VDD = 5.25 V, VIL = 0 V) High (VDD = 5.25 V, VIH = 5.25 V) Input Capacitance Symbol Min Max Unit VIL VIH –0.3 0.7 * VDD 0.3 * VDD VDD + 0.3 V V IIL IIH –5 — — 5 µA µA IIL IIH –100 — — 5 µA µA VOL VOL — — 0.4 0.2 V V VOH VOH VDD – 0.7 VDD – 0.2 — — V V IOZL IOZH CI –10 — — — 10 5 µA µA pF Table 62. Electrical Requirements for Mask-Programmable Input Clock Options Parameter CKI CMOS Level Input Voltage: Low High Small-signal Peak-to-peak Voltage* (on CKI) Small-signal Input Duty Cycle† Small-signal Input Voltage Range (pins: CKI, CKI2) Small-signal Buffer Frequency Range Frequency Range of Fundamental Mode or Overtone Crystal Series Resistance of Fundamental Mode or Overtone Crystal (pins: CKI, CKI2) Mutual Capacitance of Crystal (includes board stray capacitance) Symbol Min Max Unit VIL VIH Vpp –0.3 0.7 * VDD 0.6 0.3 * VDD VDD + 0.3 — V V V DCyc Vin 45 0.2 * VDD 55 0.6 * VDD % V fss fX — 5 35 25 MHz MHz RS — 40 Ω C0 — 7 pF * The small-signal buffer must be used in single-ended mode where an ac waveform (sine or square) is applied to CKI and a dc voltage approximately equal to the average value of CKI is applied to CKI2, as shown in the figure below. The maximum allowable ripple on CKI2 is 100 mV. † Duty cycle for a sine wave is defined as the percentage of time during each clock cycle that the voltage on CKI exceeds the voltage on CKI2. CKI CKI2 70 Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 9 Electrical Characteristics and Requirements (continued) Additional Electrical Requirements with Crystal Option: See Section 13, Crystal Electrical Characteristics and Requirements. Table 63. PLL Electrical Specifications, VCO Frequency Ranges Parameter VCO frequency range (VDD = 3 V ± 10%)* VCO frequency range (VDD = 3.0 V – 3.6 V)* VCO frequency range (VDD = 5 V ± 5%)* Input Jitter at CKI Symbol fVCO fVCO fVCO — Min 50 50 70 — Max 160 200 180 200 Unit MHz MHz MHz ps-rms * The M and N counter values in the pllc register must be set so that the VCO will operate in the appropriate range (see Table 63). Choose the lowest value of N and then the appropriate value of M for fINTERNAL CLOCK = fCKI x (M/(2N)) = fVCO/2. Table 64. PLL Electrical Specifications and pllc Register Settings M VDD pllc13 (ICP) 23—24 21—22 19—20 16—18 12—15 8—11 2—7 19—20 17—18 16 14—15 12—13 10—11 8—9 7 5—6 2—4 2.7 V – 3.6 V 2.7 V – 3.6 V 2.7 V – 3.6 V 2.7 V – 3.6 V 2.7 V – 3.6 V 2.7 V – 3.6 V 2.7 V – 3.6 V 5 V ± 5% 5 V ± 5% 5 V ± 5% 5 V ± 5% 5 V ± 5% 5 V ± 5% 5 V ± 5% 5 V ± 5% 5 V ± 5% 5 V ± 5% 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 pllc12 (SEL5V) 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 pllc[11:8] (LF[3:0]) 1011 1010 1001 1000 0111 0110 0100 1110 1101 1100 1011 1010 1001 1000 0111 0110 0101 Typical Lock-in Time (µ µs)* 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 * Lock-in time represents the time following assertion of the PLLEN bit of the pllc register during which the PLL output clock is unstable. The DSP must operate from the 1X CKI input clock or from the slow ring oscillator while the PLL is locking. Completion of the lock-in interval is indicated by assertion of the LOCK flag. Lucent Technologies Inc. 71 Data Sheet March 2000 DSP1627 Digital Signal Processor 9 Electrical Characteristics and Requirements (continued) VDD VOH (V) VDD – 0.1 DEVICE UNDER TEST VDD – 0.2 VOH IOH VDD – 0.3 VDD – 0.4 0 5 10 15 20 25 30 35 40 45 50 IOH (mA) 5-4007 (F).a Figure 9. Plot of VOH vs. IOH Under Typical Operating Conditions 0.4 VOL (V) 0.3 DEVICE UNDER TEST 0.2 VOL IOL 0.1 0 0 5 10 15 20 25 30 35 40 45 50 I OL (mA) 5-4008 (F).b Figure 10. Plot of VOL vs. IOL Under Typical Operating Conditions 72 Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 9 Electrical Characteristics and Requirements (continued) 9.1 Power Dissipation Power dissipation is highly dependent on DSP program activity and the frequency of operation. The typical power dissipation listed is for a selected application. The following electrical characteristics are preliminary and are subject to change. Table 65. Power Dissipation and Wake-Up Latency Operating Mode (Unused Inputs at VDD or VSS) Normal Operation ioc = 0x0180 PLL Disabled CKI & CKO = 40 MHz CMOS Crystal Oscillator Small Signal CKI & CKO = 0 MHz CMOS Small Signal Normal Operation ioc = 0x0180 PLL Enabled pllc = 0xFC0E CKI = 10 MHz CKO = 40 MHz CMOS Crystal Oscillator Small Signal Standard Sleep, External Interrupt alf[15] = 1, ioc = 0x0180 PLL Disabled During Sleep CMOS Crystal Oscillator Small Signal Standard Sleep, External Interrupt alf[15] = 1, ioc = 0x0180 PLL Enabled During Sleep CMOS Crystal Oscillator Small Signal Sleep with Slow Internal Clock Crystal/Small Signal Enabled powerc[15:14] = 01, alf[15] = 1, ioc = 0x0180 PLL Disabled During Sleep CMOS Crystal Oscillator Small Signal Typical Power Dissipation (mW) Wake-Up Latency I/O Units ON I/O Units OFF (PLL Not Used (PLL Used powerc[7:4] = 0x0 powerc[7:4] = 0xf During Wake State) During Wake State) 5V 3V 5V 3V 5V 3V 5V 3V 220 241 223 74 80 76 214 235 217 72 78 74 — — — — — — 0.19 3.0 0.067 1.1 0.19 3.0 0.067 1.1 — — — — 228 77 222 75 249 83 243 81 231 78 225 77 Power Management Modes CKO = 40 MHz — — — — — — 25.2 46.2 28.0 8.4 14.0 9.8 17.8 38.8 20.8 5.6 12.0 7.2 3T* 3T* 3T* 3T* + tL† 3T* + tL† 3T* + tL† 33.2 54.0 36.0 10.9 17.1 12.4 25.8 46.0 28.8 7.5 14.0 9.2 — — — 3T* 3T* 3T* 1.4 21.9 3.9 0.4 6.2 2.1 1.1 21.8 3.8 0.3 6.1 2.0 1.5 µs 1.5 µs 1.5 µs 5.0 µs 5.0 µs 5.0 µs 1.5 µs + tL 5.0 µs + tL 1.5 µs + tL 5.0 µs + tL 1.5 µs + tL 5.0 µs + tL * T = CKI clock cycle for 1X input clock option or T = CKI clock cycle divided by M/(2N) for PLL clock option (see Section 4.12). † tL = PLL lock time (see Table 64). Lucent Technologies Inc. 73 Data Sheet March 2000 DSP1627 Digital Signal Processor 9 Electrical Characteristics and Requirements (continued) Table 65. Power Dissipation and Wake-Up Latency (continued) Operating Mode (Unused inputs at VDD or VSS) Typical Power Dissipation (mW) I/O Units ON powerc[7:4] = 0x0 5V 3V I/O Units OFF powerc[7:4] = 0xf 5V 3V Wake-Up Latency (PLL Not Used During Wake State) 5V 3V (PLL Used During Wake State) 5V 3V Sleep with Slow Internal Clock Crystal/Small Signal Enabled powerc[15:14] = 01, alf[15] = 1, ioc = 0x0180 PLL Enabled During Sleep CMOS Crystal Oscillator Small Signal Sleep with Slow Internal Clock Crystal/Small Signal Disabled powerc[15:14] = 11, alf[15] = 1, ioc = 0x0180 PLL Disabled During Sleep 8.3 27.5 10.0 3.0 9.9 4.5 7.5 24.5 10.0 2.7 8.8 4.0 — — — Crystal Oscillator Small Signal 0.67 0.67 0.24 0.24 0.56 0.56 0.16 0.16 20 ms 20 µs 20 µs + tL† 20 µs + tL† 0.19 0.067 0.19 0.067 3T* 3T* + tL† 0.19 0.19 0.067 0.067 0.19 0.19 0.067 0.067 20 ms 20 µs 20 µs +tL† 20 µs + tL† 0.19 20.0 3.0 0.067 6.0 1.1 0.19 20.0 3.0 0.067 6.0 1.1 3T* 3T* 3T* — — — 5.6 25.6 8.6 2.4 8.4 3.5 5.6 25.6 8.6 2.4 8.4 3.5 3T* 3T* 3T* 3T* 3T* 3T* Software Stop powerc[15:12] = 0011 PLL Disabled During STOP CMOS Software Stop powerc[15:12] = 1111 PLL Disabled During STOP Crystal Oscillator Small Signal Hardware Stop (STOP = VSS) powerc[15:12] = 0000 PLL Disabled During STOP CMOS Crystal Oscillator Small Signal Hardware Stop (STOP = VSS) powerc[15:12] = 0000 PLL Enabled During STOP CMOS Crystal Oscillator Small Signal 1.5 µs 1.5 µs 1.5 µs 5.0 µs 5.0 µs 5.0 µs * T = CKI clock cycle for 1X input clock option or T = CKI clock cycle divided by M/(2N) for PLL clock option (see Section 4.12). † tL = PLL lock time (see Table 64). The power dissipation listed is for internal power dissipation only. Total power dissipation can be calculated on the basis of the application by adding C x VDD/2 x f for each output, where C is the additional load capacitance and f is the output frequency. 74 Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 9 Electrical Characteristics and Requirements (continued) Power dissipation due to the input buffers is highly dependent on the input voltage level. At full CMOS levels, essentially no dc current is drawn. However, for levels between the power supply rails, especially at or near the threshold of VDD/2, high currents can flow. Although input and I/O buffers may be left untied (since the input voltage levels of the input and I/O buffers are designed to remain at full CMOS levels when not driven by the DSP), it is still recommended that unused input and I/O pins be tied to VSS or VDD through a 10 kΩ resistor to avoid application ambiguities. Further, if I/O pins are tied high or low, they should be pulled fully to V SS or VDD. WARNING: The device needs to be clocked for at least six CKI cycles during reset after powerup. Otherwise, high currents may flow. 10 Timing Characteristics for 5 V Operation The following timing characteristics and requirements are preliminary information and are subject to change. Timing characteristics refer to the behavior of the device under specified conditions. Timing requirements refer to conditions imposed on the user for proper operation of the device. All timing data is valid for the following conditions: TA = –40 °C to +85 °C (See Section 8.3.) VDD = 5 V ± 5%, VSS = 0 V (See Section 8.3.) Capacitance load on outputs (CL) = 50 pF, except for CKO, where CL = 20 pF. Output characteristics can be derated as a function of load capacitance (CL). All outputs: 0.03 ns/pF ≤ dt/dCL ≤ 0.06 ns/pF for 10 ≤ CL ≤ 100 pF at VIH for rising edge and at VIL for falling edge. For example, if the actual load capacitance is 30 pF instead of 50 pF, the derating for a rising edge is (30 – 50) pF x 0.06 ns/pF = 1.2 ns less than the specified rise time or delay that includes a rise time. Test conditions for inputs: ■ Rise and fall times of 4 ns or less ■ Timing reference levels for delays = VIH, VIL Test conditions for outputs (unless noted otherwise): ■ CLOAD = 50 pF; except for CKO, where CLOAD = 20 pF ■ Timing reference levels for delays = VIH, VIL ■ 3-state delays measured to the high-impedance state of the output driver For the timing diagrams, see Table 62 for input clock requirements. Unless otherwise noted, CKO in the timing diagrams is the free-running CKO. Lucent Technologies Inc. 75 Data Sheet March 2000 DSP1627 Digital Signal Processor 10 Timing Characteristics for 5 V Operation (continued) 10.1 DSP Clock Generation t1 t3 t2 1X CKI* t4 t5 CKO† t6, t6a CKO‡ EXTERNAL MEMORY CYCLE W = 1§ 5-4009 (F).a * See Table 62 for input clock electrical requirements. † Free-running clock. ‡ Wait-stated clock (see Table 38). § W = number of wait-states. Figure 11. I/O Clock Timing Diagram Table 66. Timing Requirements for Input Clock Abbreviated Reference t1 t2 t3 14 ns and 11 ns* Min Max Unit 20 —† ns 10 — ns 10 — ns Parameter Clock In Period (high to high) Clock In Low Time (low to high) Clock In High Time (high to low) * Device speeds greater than 50 MIPS do not support 1X operation. Use the PLL. † Device is fully static, t1 is tested at 100 ns for 1X input clock option, and memory hold time is tested at 0.1 s. Table 67. Timing Characteristics for Input Clock and Output Clock Abbreviated Reference t4 t5 t6 t6a Parameter Clock Out High Delay Clock Out Low Delay (high to low) Clock Out Period (low to low) Clock Out Period with SLOWCKI Bit Set in powerc Register (low to low) 14 ns Min Max — 10 — 10 T* — 0.74 1.6 11 ns Min Max — 8 — 8 T* — 0.74 1.6 Unit ns ns ns µs * T = internal clock period, set by CKI or by CKI and the PLL parameters. 76 Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 10 Timing Characteristics for 5 V Operation (continued) 10.2 Reset Circuit The DSP1627 has a powerup reset circuit that automatically clears the JTAG controller upon powerup. If the supply voltage falls below VDD MIN* and a reset is required, the JTAG controller must be reset—even if the JTAG port isn’t being used—by applying six low-to-high clock edges on TCK with TMS held high, followed by the usual RSTB and CKI reset sequence. Figure 12 shows two separate events: an initial powerup and a powerup following a drop in the power supply voltage. * See Table 60, Recommended Operating Condiitons. VDD RAMP VDD MIN 0.4 V V DD MIN 0.4 V t146 t9 t8 t151 t152 t8 t9 CKI TCK TMS VIH t153 t153 VIH RSTB VIL t10 t10 t11 t11 PINS VOH VOL 5-2253 (F).a Notes: See Table 62 for CKI electrical requirements and Table 71 for TCK timing requirements. When both INT0 and RSTB are asserted, all output and bidirectional pins (except TDO, which 3-states by JTAG control) are put in a 3-state condition. With RSTB asserted and INT0 not asserted, EROM, ERAMHI, ERAMLO, IO, DSEL, and RWN outputs remain high, and CKO remains a free-running clock. TMS and TDI signals have internal pull-up devices. Figure 12. Powerup Reset and Chip Reset Timing Diagram Table 68. Timing Requirements for Powerup Reset and Chip Reset Abbreviated Reference Parameter t8 Reset Pulse (low to high) t9 VDD Ramp t146 VDD MIN to RSTB Low CMOS Crystal* Small-signal t151 t152 TMS High JTAG Reset to RSTB Low CMOS Crystal* Small-Signal t153 RSTB Rise (low to high) Min 6T — 2T 20 20 Max — 10 — — — Unit ns ms ns ms µs 6 * TTCK† 2T 20 ms – 6 * TTCK if 6 * TTCK < 20 ms 0 if 6 * TTCK ≥ 20 ms 20 µs – 6 * TTCK if 6 * TTCK < 20 µs 0 if 6 * TTCK ≥ 20 µs — — — — — — — 95 ns ns ns * With external components as specified in Table 62. † TTCK = t12 = TCK period. See Table 71 for TCK timing requirements. Lucent Technologies Inc. 77 Data Sheet March 2000 DSP1627 Digital Signal Processor 10 Timing Characteristics for 5 V Operation (continued) Table 69. Timing Characteristics for Powerup Reset and Chip Reset Abbreviated Reference t10 t11 Parameter RSTB Disable Time (low to 3-state) RSTB Enable Time (high to valid) Min — — Max 100 100 Unit ns ns Note: The device needs to be clocked for at least six CKI cycles during reset after powerup. Otherwise, high currents may flow. 10.3 Reset Synchronization t5 + 2 x t6 CKI* VIH VIL t126 RSTB VIH VIL CKO VIH VIL CKO VIH VIL 5-4011 (F).a * See Table 62 for input clock electrical requirements. Note: CKO1 and CKO2 are two possible CKO states before reset. CKO is free-running. Figure 13. Reset Synchronization Timing Table 70. Timing Requirements for Reset Synchronization Timing Abbreviated Reference t126 78 Parameter Reset Setup (high to high) Min 1.5 Max T/2 – 5 Unit ns Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 10 Timing Characteristics for 5 V Operation (continued) 10.4 JTAG I/O Specifications t12 t155 t13 TCK t14 VIH VIL t15 t156 t16 TMS VIH VIL t17 t18 VIH TDI VIL t19 t20 TDO VOH VOL 5-4017 (F) Figure 14. JTAG Timing Diagram Table 71. Timing Requirements for JTAG Input/Output Abbreviated Reference t12 t13 t14 t15 t16 t17 t18 Parameter TCK Period (high to high) TCK High Time (high to low) TCK Low Time (low to high) TMS Setup Time (valid to high) TMS Hold Time (high to invalid) TDI Setup Time (valid to high) TDI Hold Time (high to invalid) Min 50 22.5 22.5 7.5 2 7.5 2 Max — — — — — — — Unit ns ns ns ns ns ns ns Table 72. Timing Characteristics for JTAG Input/Output Abbreviated Reference t19 t20 Lucent Technologies Inc. Parameter TDO Delay (low to valid) TDO Hold (low to invalid) Min — 0 Max 19 — Unit ns ns 79 Data Sheet March 2000 DSP1627 Digital Signal Processor 10 Timing Characteristics for 5 V Operation (continued) 10.5 Interrupt CKO* V OH V OL t21 INT[1:0] V IH V IL t22 t23 IACK† t25 V OH V OL t24 t26 VEC[3:0] V OH V OL 5-4018 (F) * CKO is free-running. † IACK assertion is guaranteed to be enclosed by VEC[3:0] assertion. Figure 15. Interrupt Timing Diagram Table 73. Timing Requirements for Interrupt Abbreviated Reference t21 t22 Parameter Interrupt Setup (high to low) INT Assertion Time (high to low) Min 15 2T Max — — Unit ns ns Min — — — — Max T/2 + 7.5 9.5 7.5 9.5 Unit ns ns ns ns Note: Interrupt is asserted during an interruptible instruction and no other pending interrupts. Table 74. Timing Characteristics for Interrupt Abbreviated Reference t23 t24 t25 t26 Parameter IACK Assertion Time (low to high) VEC Assertion Time (low to high) IACK Invalid Time (low to low) VEC Invalid Time (low to low) Note: Interrupt is asserted during an interruptible instruction and no other pending interrupts. 80 Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 10 Timing Characteristics for 5 V Operation (continued) 10.6 Bit Input/Output (BIO) t144 CKO V OH V OL t29 V OH IOBIT (OUTPUT) V OL VALID OUTPUT t28 t27 IOBIT (INPUT) V IH DATA INPUT V IL 5-4019 (F).a Figure 16. Write Outputs Followed by Read Inputs (cbit = Immediate; a1 = sbit) Table 75. Timing Requirements for BIO Input Read Abbreviated Reference t27 t28 Parameter IOBIT Input Setup Time (valid to high) IOBIT Input Hold Time (high to invalid) Min 12 0 Max — — Unit ns ns Min — 1 Max 7.5 — Unit ns ns Table 76. Timing Characteristics for BIO Output Abbreviated Reference t29 t144 Parameter IOBIT Output Valid Time (low to valid) IOBIT Output Hold Time (low to invalid) t144 CKO VOH – V O L– t29 IOBIT VOH – (OUTPUT) V O L– VALID OUTPUT t141 IOBIT (INPUT) VIH – V I L– t142 TEST INPUT 5-4019 (F).b Figure 17. Write Outputs and Test Inputs (cbit = Immediate) Table 77. Timing Requirements for BIO Input Test Abbreviated Reference t141 t142 Lucent Technologies Inc. Parameter IOBIT Input Setup Time (valid to low) IOBIT Input Hold Time (low to invalid) Min 12 0 Max — — Unit ns ns 81 Data Sheet March 2000 DSP1627 Digital Signal Processor 10 Timing Characteristics for 5 V Operation (continued) 10.7 External Memory Interface The following timing diagrams, characteristics, and requirements do not apply to interactions with delayed external memory enables unless so stated. See the DSP1611/17/18/27 Digital Signal Processor Information Manual for a detailed description of the external memory interface including other functional diagrams. CKO VOH VOL t33 t34 VOH ENABLE VOL W* = 0 5-4020 (F).b * W = number of wait-states. Figure 18. Enable Transition Timing Table 78. Timing Characteristics for External Memory Enables (EROM, ERAMHI, IO, ERAMLO) Abbreviated Reference t33 t34 Parameter CKO to ENABLE Active (low to low) CKO to ENABLE Inactive (low to high) Min 0 –1 Max 7 6 Unit ns ns Table 79. Timing Characteristics for Delayed External Memory Enables (ioc = 0x000F) Abbreviated Reference t33 82 Parameter CKO to Delayed ENABLE Active (low to low) Min T/2 – 2 Max T/2 + 7 Unit ns Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 10 Timing Characteristics for 5 V Operation (continued) (MWAIT = 0 x 2222) W* = 2 CKO V OH V OL t127 ENABLE V OH V OL t129 t130 DB V IH READ DATA V IL t150 t128 AB V OH V OL READ ADDRESS 5-4021 (F).a * W = number of wait-states. Figure 19. External Memory Data Read Timing Diagram Table 80. Timing Characteristics for External Memory Access Abbreviated Reference t127 t128 Parameter Enable Width (low to high) Address Valid (enable low to valid) Min T(1 + W) – 4 — Max — 2 Unit ns ns Table 81. Timing Requirements for External Memory Read (EROM, ERAMHI, IO, ERAMLO) Abbreviated Reference t129 t130 t150 14 ns 11 ns Unit Min Max Min Max Read Data Setup (valid to enable high) 12 — 11 — ns Read Data Hold (enable high to hold) 0 — 0 — ns External Memory Access Time (valid to valid) — T(1 + W) – 13 — T(1 + W) – 12 ns Lucent Technologies Inc. Parameter 83 Data Sheet March 2000 DSP1627 Digital Signal Processor 10 Timing Characteristics for 5 V Operation (continued) (MWAIT = 0x1002) W* = 2 CKO ERAMLO DB W* = 1 VOH VOL VOH VOL VOH VOL WRITE DATA READ t131 EROM VOH VOL t132 t133 RWN t134 VOH VOL t135 t136 AB VOH WRITE ADDRESS VOL READ ADDRESS 5-4022 (F).a * W = number of wait-states. Figure 20. External Memory Data Write Timing Diagram Table 82. Timing Characteristics for External Memory Data Write (All Enables) Abbreviated Reference t131 t132 t133 t134 t135 t136 84 Parameter 14 ns 11 ns Min Max Min Max Write Overlap (enable low to 3-state) — 0 — 0 RWN Advance (RWN high to enable high) 0 — 0 — RWN Delay (enable low to RWN low) 0 — 0 — Write Data Setup (data valid to RWN high) T(1 + W)/2 – 3 — T(1 + W)/2 – 2 — RWN Width (low to high) T(1 + W) – 5.5 — T(1 + W) – 5.5 — Write Address Setup (address valid to RWN 0 — 0 — low) Unit ns ns ns ns ns ns Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 10 Timing Characteristics for 5 V Operation (continued) (MWAIT = 0x1002) W* = 2 CKO ERAMLO EROM W* = 1 VOH VOL VOH VOL VOH VOL t131 DB VOH VOL WRITE READ t137 t138 RWN VOH VOL t139 AB VOH VOL WRITE ADDRESS READ ADDRESS 5-4023 (F).a * W = number of wait-states. Figure 21. Write Cycle Followed by Read Cycle Table 83. Timing Characteristics for Write Cycle Followed by Read Cycle Abbreviated Reference t131 t137 t138 t139 Lucent Technologies Inc. Parameter Write Overlap (enable low to 3-state) Write Data 3-state (RWN high to 3-state) Write Data Hold (RWN high to data hold) Write Address Hold (RWN high to address hold) Min — — 0 0 Max 0 2 — — Unit ns ns ns ns 85 Data Sheet March 2000 DSP1627 Digital Signal Processor 10 Timing Characteristics for 5 V Operation (continued) 10.8 PHIF Specifications For the PHIF, "READ" means read by the external user (output by the DSP); "WRITE" is similarly defined. The 8-bit reads/ writes are identical to one-half of a 16-bit access. 16-bit READ 16-bit WRITE VIH– PCSN PODS VIL– VIH– VIL– t41 t42 t44 VIH– PIDS VIL– t43 PBSEL VIH– VIL– t47 t45 PSTAT t48 t46 VIH– VIL– t49 VIH– PB[7:0] t50 t154 t51 t52 VIL– 5-4036 (F) Figure 22. PHIF Intel Mode Signaling (Read and Write) Timing Diagram Table 84. Timing Requirements for PHIF Intel Mode Signaling Abbreviated Reference t41 t42 t43 t44 t45* t46* t47* t48* t51* t52* * Parameter PODS to PCSN Setup (low to low) PCSN to PODS Hold (high to high) PIDS to PCSN Setup (low to low) PCSN to PIDS Hold (high to high) PSTAT to PCSN Setup (valid to low) PCSN to PSTAT Hold (high to invalid) PBSEL to PCSN Setup (valid to low) PCSN to PBSEL Hold (high to invalid) PB Write to PCSN Setup (valid to high) PCSN to PB Write Hold (high to invalid) Min 0 0 0 0 4.5 0 4.5 0 7.5 4 Max — — — — — — — — — — Unit ns ns ns ns ns ns ns ns ns ns This timing diagram for the PHIF port shows accesses using the PCSN signal to initiate and complete a transaction. The transactions can also be initiated and completed with the PIDS and PODS signals. An output transaction (read) is initiated by PCSN or PODS going low, whichever comes last. For example, the timing requirements referenced to PCSN going low, t45 and t49, should be referenced to PODS going low, if PODS goes low after PCSN. An output transaction is completed by PCSN or PODS going high, whichever comes first. An input transaction is initiated by PCSN or PIDS going low, whichever comes last. An input transaction is completed by PCSN or PIDS going high, whichever comes first. All requirements referenced to PCSN apply to PIDS or PODS, if PIDS or PODS is the controlling signal. Table 85. Timing Characteristics for PHIF Intel Mode Signaling Abbreviated Reference t49* t50* * Parameter PCSN to PB Read (low to valid) PCSN to PB Read Hold (high to invalid) Min — 3 Max 13 — Unit ns ns This timing diagram for the PHIF port shows accesses using the PCSN signal to initiate and complete a transaction. The transactions can also be initiated and completed with the PIDS and PODS signals. An output transaction (read) is initiated by PCSN or PODS going low, whichever comes last. For example, the timing requirements referenced to PCSN going low, t45 and t49, should be referenced to PODS going low, if PODS goes low after PCSN. An output transaction is completed by PCSN or PODS going high, whichever comes first. An input transaction is initiated by PCSN or PIDS going low, whichever comes last. An input transaction is completed by PCSN or PIDS going high, whichever comes first. All requirements referenced to PCSN apply to PIDS or PODS, if PIDS or PODS is the controlling signal. 86 Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 10 Timing Characteristics for 5 V Operation (continued) 16-bit READ 16-bit WRITE 8-bit WRITE 8-bit READ t55 t55 PCSN VIH VIL t56 t55 PODS t56 VIH VIL t56 PIDS t56 t55 VIH VIL t56 PBSEL VOH VOL t53 POBE t53 VOH VOL t54 t54 PIBF VOH VOL 5-4037 (F).a Figure 23. PHIF Intel Mode Signaling (Pulse Period and Flags) Timing Diagram Table 86. Timing Requirements for PHIF Intel Mode Signaling Abbreviated Reference t55 t56 Parameter PCSN/PODS/PIDS Pulse Width (high to low) PCSN/PODS/PIDS Pulse Width (low to high) Min 15 15 Max — — Unit ns ns Min — — Max 15 15 Unit ns ns Table 87. Timing Characteristics for PHIF Intel Mode Signaling Abbreviated Reference t53* t54* Parameter PCSN/PODS to POBE† (high to high) PCSN/PIDS to PIBF† (high to high) * t53 should be referenced to the rising edge of PCSN or PODS, whichever comes first. t54 should be referenced to the rising edge of PCSN or PIDS, whichever comes first. † POBE and PIBF may be programmed to be the opposite logic levels shown in the diagram (positive assertion levels shown). t53 and t54 apply to the inverted levels as well as those shown. Lucent Technologies Inc. 87 Data Sheet March 2000 DSP1627 Digital Signal Processor 10 Timing Characteristics for 5 V Operation (continued) 16-bit READ 16-bit WRITE VI H PCSN VIL t42 VI H PDS VIL t41 t43 t44 t47 t48 VI H PRWN VIL t43 PBSEL t44 VI H VIL t45 PSTAT t46 VI H VIL t49 t50 t154 t51 t52 PB[7:0] 5-4038 (F).a Figure 24. PHIF Motorola Mode Signaling (Read and Write) Timing Diagram Table 88. Timing Requirements for PHIF Motorola Mode Signaling Abbreviated Reference t41 t42 Parameter PDS to PCSN Setup (valid to low) Min 0 Max — Unit ns PCSN to PDS† Hold (high to invalid) 0 — ns t43 PRWN to PCSN Setup (valid to low) 4.5 — ns t44 PCSN to PRWN Hold (high to invalid) 0 — ns t45* PSTAT to PCSN Setup (valid to low) 4.5 — ns t46* PCSN to PSTAT Hold (high to invalid) 0 — ns t47* PBSEL to PCSN Setup (valid to low) 4.5 — ns t48* PCSN to PBSEL Hold (high to invalid) 0 — ns t51* PB Write to PCSN Setup (valid to high) 8 — ns t52* PCSN to PB Write Hold (high to invalid) 4 — ns † * This timing diagram for the PHIF port shows accesses using the PCSN signal to initiate and complete a transaction. The transactions can also be initiated and completed with the PDS signal. An input/output transaction is initiated by PCSN or PDS going low, whichever comes last. For example, the timing requirements referenced to PCSN going low, t45 and t49, should be referenced to PDS going low, if PDS goes low after PCSN. An input/output transaction is completed by PCSN or PDS going high, whichever comes first. All requirements referenced to PCSN should be referenced to PDS, if PDS is the controlling signal. PRWN should never be used to initiate or complete a transaction. † PDS is programmable to be active-high or active-low. It is shown active-low in Figures 24 and 25. POBE and PIBF may be programmed to be the opposite logic levels shown in the diagram. t53 and t54 apply to the inverted levels as well as those shown. Table 89. Timing Characteristics for PHIF Motorola Mode Signaling Abbreviated Reference Parameter Min Max Unit t49* PCSN to PB Read (low to valid) — 13 ns t50* PCSN to PB Read (high to invalid) 3 — ns * This timing diagram for the PHIF port shows accesses using the PCSN signal to initiate and complete a transaction. The transactions can also be initiated and completed with the PDS signal. An input/output transaction is initiated by PCSN or PDS going low, whichever comes last. For example, the timing requirements referenced to PCSN going low, t45 and t49, should be referenced to PDS going low, if PDS goes low after PCSN. An input/output transaction is completed by PCSN or PDS going high, whichever comes first. All requirements referenced to PCSN should be referenced to PDS, if PDS is the controlling signal. PRWN should never be used to initiate or complete a transaction. † PDS is programmable to be active-high or active-low. It is shown active-low in Figures 24 and 25. POBE and PIBF may be programmed to be the opposite logic levels shown in the diagram. t53 and t54 apply to the inverted levels as well as those shown. 88 Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 10 Timing Characteristics for 5 V Operation (continued) 16-bit WRITE 16-bit READ 8-bit READ 8-bit WRITE t55 t55 PCSN VIH– VIL– t56 t55 t56 t56 VIH– PDS VIL– t56 PRWN t55 VIH– VIL– t56 PBSEL VOH– VOL– t53 POBE t53 VOH– VOL– t54 t54 PIBF VOH– VOL– 5-4039 (F).a Figure 25. PHIF Motorola Mode Signaling (Pulse Period and Flags) Timing Diagram Table 90. Timing Characteristics for PHIF Motorola Mode Signaling Abbreviated Reference t53* t54* Parameter PCSN/PDS to POBE† (high to high) PCSN/PDS† to PIBF† (high to high) † Min — — Max 15 15 Unit ns ns * An input/output transaction is initiated by PCSN or PDS going low, whichever comes last. For example, t53 and t54 should be referenced to PDS going low, if PDS goes low after PCSN. An input/output transaction is completed by PCSN or PDS going high, whichever comes first. All requirements referenced to PCSN should be referenced to PDS, if PDS is the controlling signal. PRWN should never be used to initiate or complete a transaction. † PDS is programmable to be active-high or active-low. It is shown active-low in Figures 24 and 25. POBE and PIBF may be programmed to be the opposite logic levels shown in the diagram. t53 and t54 apply to the inverted levels as well as those shown. Table 91. Timing Requirements for PHIF Motorola Mode Signaling Abbreviated Reference Parameter t55 PCSN/PDS/PRWN Pulse Width (high to low) t56 PCSN/PDS/PRWN Pulse Width (low to high) Lucent Technologies Inc. Min 15 15 Max — — Unit ns ns 89 Data Sheet March 2000 DSP1627 Digital Signal Processor 10 Timing Characteristics for 5 V Operation (continued) V IH PCSN PODS (PDS*) PIDS (PRWN*) V IL V IH V IL V IH V IL t47 PBSEL t48 V IH V IL t46 t45 PSTAT V IH V IL t49 PB[7:0] V OH t50 t154 V OL 5-4040 (F).a * Motorola mode signal name. Figure 26. PHIF Intel or Motorola Mode Signaling (Status Register Read) Timing Diagram Table 92. Timing Requirements for Intel and Motorola Mode Signaling (Status Register Read) Abbreviated Reference t45† t46‡ t47† t48‡ Parameter PSTAT to PCSN Setup (valid to low) PCSN to PSTAT Hold (high to invalid) PBSEL to PCSN Setup (valid to low) PCSN to PBSEL Hold (high to invalid) Min 4.5 0 4.5 0 Max — — — — Unit ns ns ns ns † t45, t47, and t49 are referenced to the falling edge of PCSN or PODS(PDS), whichever occurs last. ‡ t46, t48, and t50 are referenced to the rising edge of PCSN or PODS(PDS), whichever occurs first. Table 93. Timing Characteristics for Intel and Motorola Mode Signaling (Status Register Read) Abbreviated Reference t49† t50‡ Parameter PCSN to PB Read (low to valid) PCSN to PB Read Hold (high to invalid) Min — 3 Max 13 — Unit ns ns † t45, t47, and t49 are referenced to the falling edge of PCSN or PODS(PDS), whichever occurs last. ‡ t46, t48, and t50 are referenced to the rising edge of PCSN or PODS(PDS), whichever occurs first. 90 Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 10 Timing Characteristics for 5 V Operation (continued) RSTB VIH– VIL– POBE VOH– VOL– PIBF VOH– VOL– t57 t58 5-4775 (F) Figure 27. PHIF, PIBF, and POBE Reset Timing Diagram Table 94. PHIF Timing Characteristics for PHIF, PIBF, and POBE Reset Abbreviated Reference t57 t58 Parameter RSTB Disable to POBE/PIBF* (high to valid) RSTB Enable to POBE/PIBF* (low to invalid) Min Max Unit — 3 19 19 ns ns * After reset, POBE and PIBF always go to the levels shown, indicating output buffer empty and input buffer empty. The DSP program, however, may later invert the definition of the logic levels for POBE and PIBF. t57 and t58 continue to apply. CKO VIH– VIL– † VOH– VOL– PIBF† VOH– VOL– t59 POBE t59 5-4776 (F) † POBE and PIBF can be programmed to be active-high or active-low. They are shown active-high. The timing characteristic for active-low is the same as for active-high. Figure 28. PHIF, PIBF, and POBE Disable Timing Diagram Table 95. PHIF Timing Characteristics for POBE and PIBF Disable Abbreviated Reference t59 Lucent Technologies Inc. Parameter CKO to POBE/PIBF Disable (high/low to disable) Min — Max 15 Unit ns 91 Data Sheet March 2000 DSP1627 Digital Signal Processor 10 Timing Characteristics for 5 V Operation (continued) 10.9 Serial I/O Specifications t70 ICK VIH– VIL– ILD VIH– VIL– DI VIH– VIL– IBF VOH– VOL– t72 t71 t75 t74 t73 t75 t77 t78 B0 B1 BN – 1* B0 t79 5-4777 (F) * N = 16 or 8 bits. Figure 29. SIO Passive Mode Input Timing Diagram Table 96. Timing Requirements for Serial Inputs Abbreviated Reference t70 t71 t72 t73 t74 t75 t77 t78 Parameter Clock Period (high to high)‡ Clock Low Time (low to high) Clock High Time (high to low) Load High Setup (high to high) Load Low Setup (low to high) Load High Hold (high to invalid) Data Setup (valid to high) Data Hold (high to invalid) Min 40 18 18 6 6 0 5 0 Max —† — — — — — — — Unit ns ns ns ns ns ns ns ns † Device is fully static; t70 is tested at 200 ns. ‡ For Multiprocessor mode, see note in Section 10.10. Table 97. Timing Characteristics for Serial Outputs Abbreviated Reference t79 92 Parameter IBF Delay (high to high) Min — Max 22 Unit ns Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 10 Timing Characteristics for 5 V Operation (continued) ICK VOH– VOL– t101 t76a ILD VOH– VOL– * t77 DI VIH– VIL– IBF VOH– VOL– t78 B0 BN – 1 B1 B0 t79 5-4778 (F) * ILD goes high during bit 6 (of 0:15), N = 8 or 16. Figure 30. SIO Active Mode Input Timing Diagram Table 98. Timing Requirements for Serial Inputs Abbreviated Reference t77 t78 Parameter Data Setup (valid to high) Data Hold (high to invalid) Min 5 0 Max — — Unit ns ns Min — 4 — Max 22 — 22 Unit ns ns ns Table 99. Timing Characteristics for Serial Outputs Abbreviated Reference t76a t101 t79 Lucent Technologies Inc. Parameter ILD Delay (high to low) ILD Hold (high to invalid) IBF Delay (high to high) 93 Data Sheet March 2000 DSP1627 Digital Signal Processor 10 Timing Characteristics for 5 V Operation (continued) t80 OCK t82 t81 t85 t84 VIH– VIL– t83 OLD VIH– VIL– DO* VOH– VOL– t85 t88 B0 t94 SADD t87 t92 VOH– VOL– AD0 t90 B1 t90 B7 t93 AD1 BN – 1 t89 t93 AD7 AS7 t95 DOEN VIH– VIL– OBE VOH– VOL– t96 5-4796 (F) * See sioc register, MSB field to determine if B0 is the MSB or LSB. See sioc register, ILEN field to determine if the DO word length is 8 bits or 16 bits. Figure 31. SIO Passive Mode Output Timing Diagram Table 100. Timing Requirements for Serial Inputs Abbreviated Reference t80 t81 t82 t83 t84 t85 Parameter Clock Period (high to high)‡ Clock Low Time (low to high) Clock High Time (high to low) Load High Setup (high to high) Load Low Setup (low to high) Load Hold (high to invalid) Min 40 18 18 6 6 0 Max —† — — — — — Unit ns ns ns ns ns ns † Device is fully static; t80 is tested at 200 ns. ‡ For multiprocessor mode, see note in Section 10.10. Table 101. Timing Characteristics for Serial Outputs Abbreviated Reference t87 t88 t89 t90 t92 t93 t94 t95 t96 94 Parameter Data Delay (high to valid) Enable Data Delay (low to active) Disable Data Delay (high to 3-state) Data Hold (high to invalid) Address Delay (high to valid) Address Hold (high to invalid) Enable Delay (low to active) Disable Delay (high to 3-state) OBE Delay (high to high) Min — — — 4 — 4 — — — Max 22 22 22 — 22 — 22 22 22 Unit ns ns ns ns ns ns ns ns ns Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 10 Timing Characteristics for 5 V Operation (continued) OCK VOH– VOL– t102 t86a OLD VOH– VOL– DO VOH– VOL– * t88 B0 t94 SADD VOH– VOL– DOEN VIH– VIL– OBE VOH– VOL– t87 t92 AD0 t90 B1 t90 B7 t93 AD1 BN – 1 t89 t93 AD7 AS7 t95 t96 5-4797 (F) * OLD goes high at the end of bit 6 of 0:15. Figure 32. SIO Active Mode Output Timing Diagram Table 102. Timing Characteristics for Serial Outputs Abbreviated Reference t86a t102 t87 t88 t89 t90 t92 t93 t94 t95 t96 Lucent Technologies Inc. Parameter OLD Delay (high to low) OLD Hold (high to invalid) Data Delay (high to valid) Enable Data Delay (low to active) Disable Data Delay (high to 3-state) Data Hold (high to invalid) Address Delay (high to valid) Address Hold (high to invalid) Enable Delay (low to active) Disable Delay (high to 3-state) OBE Delay (high to high) Min — 4 — — — 4 — 4 — — — Max 22 — 22 22 22 — 22 — 22 22 22 Unit ns ns ns ns ns ns ns ns ns ns ns 95 Data Sheet March 2000 DSP1627 Digital Signal Processor 10 Timing Characteristics for 5 V Operation (continued) CKO VOH– VOL– ICK VOH– VOL– OCK VOH– VOL– ICK/OCK* ILD OLD SYNC t97 t98 t99 t100 VOH– t76a t101 t76b t101 t86a t102 t86b t102 t103 t105 t104 t105 VOH– VOL– VOH– VOL– VOH– VOL– 5-4798 (F) * See sioc register, LD field. Figure 33. Serial I/O Active Mode Clock Timing Table 103. Timing Characteristics for Signal Generation Abbreviated Reference t97 t98 t99 t100 t76a t76b t101 t86a t86b t102 t103 t104 t105 96 Parameter ICK Delay (high to high) ICK Delay (high to low) OCK Delay (high to high) OCK Delay (high to low) ILD Delay (high to low) ILD Delay (high to high) ILD Hold (high to invalid) OLD Delay (high to low) OLD Delay (high to high) OLD Hold (high to invalid) SYNC Delay (high to low) SYNC Delay (high to high) SYNC Hold (high to invalid) Min — — — — — — 4 — — 4 — — 4 Max 15 15 15 15 22 22 — 22 22 — 22 22 — Unit ns ns ns ns ns ns ns ns ns ns ns ns ns Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 10 Timing Characteristics for 5 V Operation (continued) 10.10 Multiprocessor Communication TIME SLOT 1 TIME SLOT 2 OCK/ICK t113 t112 t112 SYNC VIH– VIL– DO/D1 VOH– VOL– t113 * t116 B15 B0 B1 t117 B7 B8 B15 B0 t114 t122 t121 AD0 SADD AD1 AD7 AS0 t120 DOEN t115 AS7 AD0 t120 VOH– VOL– 5-4799 (F) * Negative edge initiates time slot 0. Figure 34. SIO Multiprocessor Timing Diagram Note: All serial I/O timing requirements and characteristics still apply, but the minimum clock period in passive multiprocessor mode, assuming 50% duty cycle, is calculated as (t77 + t116) x 2. Table 104. Timing Requirements for SIO Multiprocessor Communication Abbreviated Reference t112 t113 t114 t115 Parameter Sync Setup (high/low to high) Sync Hold (high to high/low) Address Setup (valid to high) Address Hold (high to invalid) Min 22 0 9 0 Max — — — — Unit ns ns ns ns Min — — — — — Max 22 20 16 22 20 Unit ns ns ns ns ns Table 105. Timing Characteristics for SIO Multiprocessor Communication Abbreviated Reference* t116 t117 t120 t121 t122 Parameter Data Delay (bit 0 only) (low to valid) Data Disable Delay (high to 3-state) DOEN Valid Delay (high to valid) Address Delay (bit 0 only) (low to valid) Address Disable Delay (high to 3-state) * With capacitance load on ICK, OCK, DO, SYNC, and SADD = 100 pF, add 4 ns to t116—t122. Lucent Technologies Inc. 97 DSP1627 Digital Signal Processor Data Sheet March 2000 11 Timing Characteristics for 3.0 V Operation The following timing characteristics and requirements are preliminary information and are subject to change. Timing characteristics refer to the behavior of the device under specified conditions. Timing requirements refer to conditions imposed on the user for proper operation of the device. All timing data is valid for the following conditions: TA = –40 °C to +85 °C (See Section 8.3.) VDD = 3.0 V to 3.6 V, VSS = 0 V (See Section 8.3.) Capacitance load on outputs (CL) = 50 pF, except for CKO, where CL = 20 pF. Output characteristics can be derated as a function of load capacitance (CL). All outputs: 0.03 ns/pF ≤ dt/dCL ≤ 0.07 ns/pF for 10 ≤ CL ≤ 100 pF at VIH for rising edge and at VIL for falling edge. For example, if the actual load capacitance is 30 pF instead of 50 pF, the derating for a rising edge is (30 – 50) pF x 0.06 ns/pF = 1.2 ns less than the specified rise time or delay that includes a rise time. Test conditions for inputs: ■ Rise and fall times of 4 ns or less ■ Timing reference levels for delays = VIH, VIL Test conditions for outputs (unless noted otherwise): ■ CLOAD = 50 pF; except for CKO, where CLOAD = 20 pF ■ Timing reference levels for delays = VIH, VIL ■ 3-state delays measured to the high-impedance state of the output driver For the timing diagrams, see Table 62 for input clock requirements. Unless otherwise noted, CKO in the timing diagrams is the free-running CKO. 98 Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 11 Timing Characteristics for 3.0 V Operation (continued) 11.1 DSP Clock Generation t1 t3 t2 1X CKI* t4 t5 CKO† t6, t6a CKO ‡ EXTERNAL MEMORY CYCLE W = 1§ 5-4009 (F).a * † ‡ § See Table 62 for input clock electrical requirements. Free-running clock. Wait-stated clock (see Table 38). W = number of wait-states. Figure 35. I/O Clock Timing Diagram Table 106. Timing Requirements for Input Clock Abbreviated Reference Parameter t1 Clock In Period (high to high) Min 20 t2 t3 Clock In Low Time (low to high) Clock In High Time (high to low) 10 10 10 ns* Max † — — — Unit ns ns ns * Device speeds greater than 50 MIPS do not support 1X operation. Use the PLL. † Device is fully static, t1 is tested at 100 ns for 1X input clock option, and memory hold time is tested at 0.1 s. Table 107. Timing Characteristics for Input Clock and Output Clock Abbreviated Reference Parameter t4 t5 t6 t6a Clock Out High Delay Clock Out Low Delay (high to low) Clock Out Period (low to low) Clock Out Period with SLOWCKI Bit Set in powerc Register (low to low) 10 ns Min — — T* 0.74 Unit Max 10 10 — 3.8 ns ns ns µs * T = internal clock period, set by CKI or by CKI and the PLL parameters. Lucent Technologies Inc. 99 Data Sheet March 2000 DSP1627 Digital Signal Processor 11 Timing Characteristics for 3.0 V Operation (continued) 11.2 Reset Circuit The DSP1627 has a powerup reset circuit that automatically clears the JTAG controller upon powerup. If the supply voltage falls below VDD MIN* and a reset is required, the JTAG controller must be reset—even if the JTAG port isn’t being used—by applying six low-to-high clock edges on TCK with TMS held high, followed by the usual RSTB and CKI reset sequence. Figure 60 shows two separate events: an initial powerup and a powerup following a drop in the power supply voltage. * See Table 60, Recommended Operating Condiitons. VDD RAMP VDD MIN 0.4 V V DD MIN 0.4 V t146 t9 t8 t151 t152 t8 t9 CKI TCK TMS VIH t153 RSTB t153 VIH VIL t10 t11 t10 t11 PINS VOH VOL 5-2253 (F).a Notes: See Table 62 for CKI electrical requirements and Table 151 for TCK timing requirements. When both INT0 and RSTB are asserted, all output and bidirectional pins (except TDO, which 3-states by JTAG control) are put in a 3-state condition. With RSTB asserted and INT0 not asserted, EROM, ERAMHI, ERAMLO, IO, DSEL, and RWN outputs remain high, and CKO remains a free-running clock. TMS and TDI signals have internal pull-up devices. Figure 36. Powerup Reset and Chip Reset Timing Diagram Table 108. Timing Requirements for Powerup Reset and Chip Reset Abbreviated Reference Parameter Min Max Unit t8 Reset Pulse (low to high) 6T — ns t9 VDD Ramp — 10 ms 2T — ns t146 VDD MIN to RSTB Low CMOS Crystal* 20 — ms µs Small-Signal 20 — t151 TMS High — ns 6 * TTCK† ns 2T — t152 JTAG Reset to CMOS 20 ms – 6 * TTCK if 6 * TTCK < 20 ms — RSTB Low Crystal* — 0 if 6 * TTCK ≥ 20 ms Small-Signal 20 µs – 6 * TTCK if 6 * TTCK < 20 µs — — 0 if 6 * TTCK ≥ 20 µs t153 RSTB (low to high) — 54 ns * With external components as specified in Table 62. † TTCK = t12 = TCK period. See Table 151 for TCK timing requirements. 100 Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 11 Timing Characteristics for 3.0 V Operation (continued) Table 109. Timing Characteristics for Powerup Reset and Chip Reset Abbreviated Reference t10 t11 Parameter RSTB Disable Time (low to 3-state) RSTB Enable Time (high to valid) Min — — Max 100 100 Unit ns ns Note: The device needs to be clocked for at least six CKI cycles during reset after powerup. Otherwise, high currents may flow. 11.3 Reset Synchronization t5 + 2 x t6 CKI* VIH VIL t126 RSTB VIH VIL CKO VIH VIL CKO VIH VIL 5-4011 (F).a * See Table 62 for input clock electrical requirements. Note: CKO1 and CKO2 are two possible CKO states before reset. CKO is free-running. Figure 37. Reset Synchronization Timing Table 110. Timing Requirements for Reset Synchronization Timing Abbreviated Reference t126 Lucent Technologies Inc. Parameter Reset Setup (high to high) Min 3 Max T/2 – 5 Unit ns 101 Data Sheet March 2000 DSP1627 Digital Signal Processor 11 Timing Characteristics for 3.0 V Operation (continued) 11.4 JTAG I/O Specifications t12 t155 t13 TCK t14 VIH VIL t156 t15 t16 TMS VIH VIL t17 t18 VIH TDI VIL t19 t20 TDO VOH VOL 5-4017 (F) Figure 38. JTAG Timing Diagram Table 111. Timing Requirements for JTAG Input/Output Abbreviated Reference t12 t13 t14 t15 t16 t17 t18 Parameter TCK Period (high to high) TCK High Time (high to low) TCK Low Time (low to high) TMS Setup Time (valid to high) TMS Hold Time (high to invalid) TDI Setup Time (valid to high) TDI Hold Time (high to invalid) Min 50 22.5 22.5 7.5 2 7.5 2 Max — — — — — — — Unit ns ns ns ns ns ns ns Table 112. Timing Characteristics for JTAG Input/Output Abbreviated Reference t19 t20 102 Parameter TDO Delay (low to valid) TDO Hold (low to invalid) Min — 0 Max 19 — Unit ns ns Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 11 Timing Characteristics for 3.0 V Operation (continued) 11.5 Interrupt CKO* V OH V OL t21 INT[1:0] V IH V IL t22 t23 IACK† t25 V OH V OL t24 t26 VEC[3:0] V OH V OL 5-4018 (F) * CKO is free-running. † IACK assertion is guaranteed to be enclosed by VEC[3:0] assertion. Figure 39. Interrupt Timing Diagram Table 113. Timing Requirements for Interrupt Abbreviated Reference t21 t22 Parameter Interrupt Setup (high to low) INT Assertion Time (high to low) Min 19 2T Max — — Unit ns ns Min — — — — Max T/2 + 10 12.5 10 12.5 Unit ns ns ns ns Note: Interrupt is asserted during an interruptible instruction and no other pending interrupts. Table 114. Timing Characteristics for Interrupt Abbreviated Reference t23 t24 t25 t26 Parameter IACK Assertion Time (low to high) VEC Assertion Time (low to high) IACK Invalid Time (low to low) VEC Invalid Time (low to low) Note: Interrupt is asserted during an interruptible instruction and no other pending interrupts. Lucent Technologies Inc. 103 Data Sheet March 2000 DSP1627 Digital Signal Processor 11 Timing Characteristics for 3.0 V Operation (continued) 11.6 Bit Input/Output (BIO) t144 CKO V OH V OL t29 V OH IOBIT (OUTPUT) V OL VALID OUTPUT t28 t27 IOBIT (INPUT) V IH DATA INPUT V IL 5-4019 (F).a Figure 40. Write Outputs Followed by Read Inputs (cbit = Immediate; a1 = sbit) Table 115. Timing Requirements for BIO Input Read Abbreviated Reference t27 t28 Parameter IOBIT Input Setup Time (valid to high) IOBIT Input Hold Time (high to invalid) Min 15 0 Max — — Unit ns ns Min — 1 Max 9 — Unit ns ns Table 116. Timing Characteristics for BIO Output Abbreviated Reference t29 t144 Parameter IOBIT Output Valid Time (low to valid) IOBIT Output Hold Time (low to invalid) t144 CKO VOH – V O L– t29 IOBIT VOH – (OUTPUT) V O L– VALID OUTPUT t141 IOBIT (INPUT) VIH – V I L– t142 TEST INPUT 5-4019 (F).b Figure 41. Write Outputs and Test Inputs (cbit = Immediate) Table 117. Timing Requirements for BIO Input Test Abbreviated Reference t141 t142 104 Parameter IOBIT Input Setup Time (valid to low) IOBIT Input Hold Time (low to invalid) Min 15 0 Max — — Unit ns ns Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 11 Timing Characteristics for 3.0 V Operation (continued) 11.7 External Memory Interface The following timing diagrams, characteristics, and requirements do not apply to interactions with delayed external memory enables unless so stated. See the DSP1611/17/18/27 Digital Signal Processor Information Manual for a detailed description of the external memory interface including other functional diagrams. CKO VOH VOL t33 t34 VOH ENABLE VOL W* = 0 5-4020 (F).b * W = number of wait-states. Figure 42. Enable Transition Timing Table 118. Timing Characteristics for External Memory Enables (EROM, ERAMHI, IO, ERAMLO) Abbreviated Reference t33 t34 Parameter CKO to ENABLE Active (low to low) CKO to ENABLE Inactive (low to high) Min 0 –1 Max 5 4.5 Unit ns ns Table 119. Timing Characteristics for Delayed External Memory Enables (ioc = 0x000F) Abbreviated Reference t33 Lucent Technologies Inc. Parameter CKO to Delayed ENABLE Active (low to low) Min T/2 – 2 Max T/2 + 7 Unit ns 105 Data Sheet March 2000 DSP1627 Digital Signal Processor 11 Timing Characteristics for 3.0 V Operation (continued) (MWAIT = 0 x 2222) W* = 2 CKO V OH V OL t127 ENABLE V OH V OL t129 t130 DB V IH READ DATA V IL t150 t128 AB V OH V OL READ ADDRESS 5-4021 (F).a * W = number of wait-states. Figure 43. External Memory Data Read Timing Diagram Table 120. Timing Characteristics for External Memory Access Abbreviated Reference Parameter t127 Enable Width (low to high) t128 Address Valid (enable low to valid) Min T(1 + W) – 1.5 — Max — 2 Unit ns ns Table 121. Timing Requirements for External Memory Read (EROM, ERAMHI, IO, ERAMLO) Abbreviated Reference t129 t130 t150 106 Parameter Read Data Setup (valid to enable high) Read Data Hold (enable high to hold) External Memory Access Time (valid to valid) 10 ns Min 13 0 — Max — — T(1 + W) – 13 Unit ns ns ns Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 11 Timing Characteristics for 3.0 V Operation (continued) (MWAIT = 0x1002) W* = 2 CKO ERAMLO DB W* = 1 VOH VOL VOH VOL VOH VOL WRITE DATA READ t131 EROM VOH VOL t132 t133 RWN t134 VOH VOL t135 t136 AB VOH WRITE ADDRESS VOL READ ADDRESS 5-4022 (F).a * W = number of wait-states. Figure 44. External Memory Data Write Timing Diagram Table 122. Timing Characteristics for External Memory Data Write (All Enables) Abbreviated Reference Parameter t131 t132 t133 t134 Write Overlap (enable low to 3-state) RWN Advance (RWN high to enable high) RWN Delay (enable low to RWN low) Write Data Setup (data valid to RWN high) t135 t136 RWN Width (low to high) Write Address Setup (address valid to RWN low) Lucent Technologies Inc. 10 ns Min — 0 0 T(1 + W)/2 – 3 T(1 + W) – 4 0 Unit Max 0 — — — ns ns ns ns — — ns ns 107 Data Sheet March 2000 DSP1627 Digital Signal Processor 11 Timing Characteristics for 3.0 V Operation (continued) (MWAIT = 0x1002) W* = 2 CKO ERAMLO EROM W* = 1 VOH VOL VOH VOL VOH VOL t131 DB VOH VOL WRITE READ t137 t138 RWN VOH VOL t139 AB VOH VOL WRITE ADDRESS READ ADDRESS 5-4023 (F).a * W = number of wait-states. Figure 45. Write Cycle Followed by Read Cycle Table 123. Timing Characteristics for Write Cycle Followed by Read Cycle Abbreviated Reference t131 t137 t138 t139 108 Parameter Write Overlap (enable low to 3-state) Write Data 3-state (RWN high to 3-state) Write Data Hold (RWN high to data hold) Write Address Hold (RWN high to address hold) Min — — 0 0 Max 0 2 — — Unit ns ns ns ns Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 11 Timing Characteristics for 3.0 V Operation (continued) 11.8 PHIF Specifications For the PHIF, "READ" means read by the external user (output by the DSP); "WRITE" is similarly defined. The 8-bit reads/ writes are identical to one-half of a 16-bit access. 16-bit READ 16-bit WRITE VIH– PCSN PODS VIL– VIH– VIL– t41 t42 t44 VIH– PIDS VIL– t43 PBSEL VIH– VIL– t47 t45 PSTAT t48 t46 VIH– VIL– t49 VIH– PB[7:0] t50 t154 t51 t52 VIL– 5-4036 (F) Figure 46. PHIF Intel Mode Signaling (Read and Write) Timing Diagram Table 124. Timing Requirements for PHIF Intel Mode Signaling Abbreviated Reference t41 t42 t43 t44 t45* t46* t47* t48* t51* t52* Parameter PODS to PCSN Setup (low to low) PCSN to PODS Hold (high to high) PIDS to PCSN Setup (low to low) PCSN to PIDS Hold (high to high) PSTAT to PCSN Setup (valid to low) PCSN to PSTAT Hold (high to invalid) PBSEL to PCSN Setup (valid to low) PCSN to PBSEL Hold (high to invalid) PB Write to PCSN Setup (valid to high) PCSN to PB Write Hold (high to invalid) Min 0 0 0 0 6 0 6 0 10 5 Max — — — — — — — — — — Unit ns ns ns ns ns ns ns ns ns ns * This timing diagram for the PHIF port shows accesses using the PCSN signal to initiate and complete a transaction. The transactions can also be initiated and completed with the PIDS and PODS signals. An output transaction (read) is initiated by PCSN or PODS going low, whichever comes last. For example, the timing requirements referenced to PCSN going low, t45 and t49, should be referenced to PODS going low, if PODS goes low after PCSN. An output transaction is completed by PCSN or PODS going high, whichever comes first. An input transaction is initiated by PCSN or PIDS going low, whichever comes last. An input transaction is completed by PCSN or PIDS going high, whichever comes first. All requirements referenced to PCSN apply to PIDS or PODS, if PIDS or PODS is the controlling signal. Table 125. Timing Characteristics for PHIF Intel Mode Signaling Abbreviated Reference t49* t50* Parameter PCSN to PB Read (low to valid) PCSN to PB Read Hold (high to invalid) Min — 3 Max 17 — Unit ns ns * This timing diagram for the PHIF port shows accesses using the PCSN signal to initiate and complete a transaction. The transactions can also be initiated and completed with the PIDS and PODS signals. An output transaction (read) is initiated by PCSN or PODS going low, whichever comes last. For example, the timing requirements referenced to PCSN going low, t45 and t49, should be referenced to PODS going low, if PODS goes low after PCSN. An output transaction is completed by PCSN or PODS going high, whichever comes first. An input transaction is initiated by PCSN or PIDS going low, whichever comes last. An input transaction is completed by PCSN or PIDS going high, whichever comes first. All requirements referenced to PCSN apply to PIDS or PODS, if PIDS or PODS is the controlling signal. Lucent Technologies Inc. 109 Data Sheet March 2000 DSP1627 Digital Signal Processor 11 Timing Characteristics for 3.0 V Operation (continued) 16-bit WRITE 16-bit READ 8-bit READ 8-bit WRITE t55 t55 PCSN VIH VIL t56 t55 PODS t56 VIH VIL t56 PIDS t56 t55 VIH VIL t56 PBSEL VOH VOL t53 POBE t53 VOH VOL t54 t54 PIBF VOH VOL 5-4037 (F).a Figure 47. PHIF Intel Mode Signaling (Pulse Period and Flags) Timing Diagram Table 126. Timing Requirements for PHIF Intel Mode Signaling Abbreviated Reference t55 t56 Parameter PCSN/PODS/PIDS Pulse Width (high to low) PCSN/PODS/PIDS Pulse Width (low to high) Min 20.5 20.5 Max — — Unit ns ns Min — — Max 20 20 Unit ns ns Table 127. Timing Characteristics for PHIF Intel Mode Signaling Abbreviated Reference t53* t54* Parameter PCSN/PODS to POBE† (high to high) PCSN/PIDS to PIBF† (high to high) * t53 should be referenced to the rising edge of PCSN or PODS, whichever comes first. t54 should be referenced to the rising edge of PCSN or PIDS, whichever comes first. † POBE and PIBF may be programmed to be the opposite logic levels shown in the diagram (positive assertion levels shown). t53 and t54 apply to the inverted levels as well as those shown. 110 Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 11 Timing Characteristics for 3.0 V Operation (continued) 16-bit READ 16-bit WRITE VI H PCSN V IL t42 VI H PDS VIL t41 PRWN t47 t48 t44 VI H VIL t45 PSTAT t44 VIL t43 PBSEL t43 VI H t46 VI H VIL t49 t50 t154 t51 t52 PB[7:0] 5-4038 (F).a Figure 48. PHIF Motorola Mode Signaling (Read and Write) Timing Diagram Table 128. Timing Requirements for PHIF Motorola Mode Signaling Abbreviated Reference t41 t42 t43 t44 t45* t46* t47* t48* t51* t52* Parameter † PDS to PCSN Setup (valid to low) PCSN to PDS† Hold (high to invalid) PRWN to PCSN Setup (valid to low) PCSN to PRWN Hold (high to invalid) PSTAT to PCSN Setup (valid to low) PCSN to PSTAT Hold (high to invalid) PBSEL to PCSN Setup (valid to low) PCSN to PBSEL Hold (high to invalid) PB Write to PCSN Setup (valid to high) PCSN to PB Write Hold (high to invalid) Min 0 0 6 0 6 0 6 0 10 5 Max — — — — — — — — — — Unit ns ns ns ns ns ns ns ns ns ns * This timing diagram for the PHIF port shows accesses using the PCSN signal to initiate and complete a transaction. The transactions can also be initiated and completed with the PDS signal. An input/output transaction is initiated by PCSN or PDS going low, whichever comes last. For example, the timing requirements referenced to PCSN going low, t45 and t49, should be referenced to PDS going low, if PDS goes low after PCSN. An input/output transaction is completed by PCSN or PDS going high, whichever comes first. All requirements referenced to PCSN should be referenced to PDS, if PDS is the controlling signal. PRWN should never be used to initiate or complete a transaction. † PDS is programmable to be active-high or active-low. It is shown active-low in Figures 72 and 73. POBE and PIBF may be programmed to be the opposite logic levels shown in the diagram. t53 and t54 apply to the inverted levels as well as those shown. Table 129. Timing Characteristics for PHIF Motorola Mode Signaling Abbreviated Reference t49* t50* Parameter PCSN to PB Read (low to valid) PCSN to PB Read (high to invalid) Min — 3 Max 17 — Unit ns ns * This timing diagram for the PHIF port shows accesses using the PCSN signal to initiate and complete a transaction. The transactions can also be initiated and completed with the PDS signal. An input/output transaction is initiated by PCSN or PDS going low, whichever comes last. For example, the timing requirements referenced to PCSN going low, t45 and t49, should be referenced to PDS going low, if PDS goes low after PCSN. An input/output transaction is completed by PCSN or PDS going high, whichever comes first. All requirements referenced to PCSN should be referenced to PDS, if PDS is the controlling signal. PRWN should never be used to initiate or complete a transaction. Lucent Technologies Inc. 111 Data Sheet March 2000 DSP1627 Digital Signal Processor 11 Timing Characteristics for 3.0 V Operation (continued) 16-bit READ 16-bit WRITE 8-bit READ 8-bit WRITE t55 t55 PCSN VIH– VIL– t56 t55 PDS t56 VIL– t56 PRWN t56 VIH– t55 VIH– VIL– t56 PBSEL VOH– VOL– t53 POBE t53 VOH– VOL– t54 t54 PIBF VOH– VOL– 5-4039 (F).a Figure 49. PHIF Motorola Mode Signaling (Pulse Period and Flags) Timing Diagram Table 130. Timing Characteristics for PHIF Motorola Mode Signaling Abbreviated Reference t53* t54* Parameter † † PCSN/PDS to POBE (high to high) PCSN/PDS† to PIBF† (high to high) Min — Max 20 Unit ns — 20 ns * An input/output transaction is initiated by PCSN or PDS going low, whichever comes last. For example, t53 and t54 should be referenced to PDS going low, if PDS goes low after PCSN. An input/output transaction is completed by PCSN or PDS going high, whichever comes first. All requirements referenced to PCSN should be referenced to PDS, if PDS is the controlling signal. PRWN should never be used to initiate or complete a transaction. † PDS is programmable to be active-high or active-low. It is shown active-low in Figures 48 and 49. POBE and PIBF may be programmed to be the opposite logic levels shown in the diagram. t53 and t54 apply to the inverted levels as well as those shown. Table 131. Timing Requirements for PHIF Motorola Mode Signaling Abbreviated Reference Parameter t55 PCSN/PDS/PRWN Pulse Width (high to low) t56 PCSN/PDS/PRWN Pulse Width (low to high) 112 Min 20 20 Max — — Unit ns ns Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 11 Timing Characteristics for 3.0 V Operation (continued) V IH PCSN PODS (PDS*) PIDS (PRWN*) V IL V IH V IL V IH V IL t47 PBSEL t48 V IH V IL t46 t45 PSTAT V IH V IL t49 PB[7:0] V OH t50 t154 V OL 5-4040 (F).a * Motorola mode signal name. Figure 50. PHIF Intel or Motorola Mode Signaling (Status Register Read) Timing Diagram Table 132. Timing Requirements for Intel and Motorola Mode Signaling (Status Register Read) Abbreviated Reference t45† t46‡ t47† t48‡ Parameter PSTAT to PCSN Setup (valid to low) PCSN to PSTAT Hold (high to invalid) PBSEL to PCSN Setup (valid to low) PCSN to PBSEL Hold (high to invalid) Min 6 0 6 0 Max — — — — Unit ns ns ns ns † t45, t47, and t49 are referenced to the falling edge of PCSN or PODS(PDS), whichever occurs last. ‡ t46, t48, and t50 are referenced to the rising edge of PCSN or PODS(PDS), whichever occurs first. Table 133. Timing Characteristics for Intel and Motorola Mode Signaling (Status Register Read) Abbreviated Reference t49† t50‡ Parameter PCSN to PB Read (low to valid) PCSN to PB Read Hold (high to invalid) Min — 3 Max 17 — Unit ns ns † t45, t47, and t49 are referenced to the falling edge of PCSN or PODS(PDS), whichever occurs last. ‡ t46, t48, and t50 are referenced to the rising edge of PCSN or PODS(PDS), whichever occurs first. Lucent Technologies Inc. 113 Data Sheet March 2000 DSP1627 Digital Signal Processor 11 Timing Characteristics for 3.0 V Operation (continued) RSTB VIH– VIL– t57 POBE VOH– VOL– PIBF VOH– VOL– t58 5-4775 (F) Figure 51. PHIF, PIBF, and POBE Reset Timing Diagram Table 134. PHIF Timing Characteristics for PHIF, PIBF, and POBE Reset Abbreviated Reference t57 t58 Parameter RSTB Disable to POBE/PIBF* (high to valid) RSTB Enable to POBE/PIBF* (low to invalid) Min — Max 25 Unit ns 3 25 ns * After reset, POBE and PIBF always go to the levels shown, indicating output buffer empty and input buffer empty. The DSP program, however, may later invert the definition of the logic levels for POBE and PIBF. t57 and t58 continue to apply. CKO VIH– VIL– † VOH– VOL– t59 POBE t59 PIBF† VOH– VOL– 5-4776 (F) † POBE and PIBF can be programed to be active-high or active-low. They are shown active-high. The timing characteristic for active-low is the same as for active-high. Figure 52. PHIF, PIBF, and POBE Disable Timing Diagram Table 135. PHIF Timing Characteristics for POBE and PIBF Disable Abbreviated Reference t59 Parameter CKO to POBE/PIBF* Disable (high/low to disable) Min — Max 20 Unit ns * POBE and PIBF can be programmed to be active-high or active-low. They are shown active-high. The timing characteristic for active-low is the same as for active-high. 114 Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 11 Timing Characteristics for 3.0 V Operation (continued) 11.9 Serial I/O Specifications t70 ICK ILD VIH– VIL– VIH– VIL– t72 t71 t75 t74 t73 t75 t77 DI VIH– VIL– IBF VOH– VOL– t78 B0 B1 BN – 1* B0 t79 5-4777 (F) * N = 16 or 8 bits. Figure 53. SIO Passive Mode Input Timing Diagram Table 136. Timing Requirements for Serial Inputs Abbreviated Reference t70 t71 t72 t73 t74 t75 t77 t78 Parameter Clock Period (high to high)‡ Min 40 Max —† Unit ns Clock Low Time (low to high) Clock High Time (high to low) Load High Setup (high to high) Load Low Setup (low to high) Load High Hold (high to invalid) Data Setup (valid to high) Data Hold (high to invalid) 18 18 8 8 0 7 0 — — — — — — — ns ns ns ns ns ns ns † Device is fully static; t70 is tested at 200 ns. ‡ For multiprocessor mode, see note in Section 12.10. Table 137. Timing Characteristics for Serial Outputs Abbreviated Reference t79 Lucent Technologies Inc. Parameter IBF Delay (high to high) Min — Max 35 Unit ns 115 Data Sheet March 2000 DSP1627 Digital Signal Processor 11 Timing Characteristics for 3.0 V Operation (continued) ICK VOH– VOL– t101 t76a ILD VOH– VOL– DI VIH– VIL– IBF VOH– VOL– * t77 t78 B0 B1 BN – 1 B0 t79 5-4778 (F) * ILD goes high during bit 6 (of 0:15), N = 8 or 16. Figure 54. SIO Active Mode Input Timing Diagram Table 138. Timing Requirements for Serial Inputs Abbreviated Reference t77 t78 Parameter Data Setup (valid to high) Data Hold (high to invalid) Min 7 0 Max — — Unit ns ns Min — 5 — Max 35 — 35 Unit ns ns ns Table 139. Timing Characteristics for Serial Outputs Abbreviated Reference t76a t101 t79 116 Parameter ILD Delay (high to low) ILD Hold (high to invalid) IBF Delay (high to high) Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 11 Timing Characteristics for 3.0 V Operation (continued) t80 OCK t82 t81 t85 t84 VIH– VIL– t83 OLD VIH– VIL– DO* VOH– VOL– t85 t88 B0 t94 SADD VOH– VOL– DOEN VIH– VIL– OBE VOH– VOL– t87 t92 AD0 t90 B1 t90 B7 t93 AD1 BN – 1 t89 t93 AD7 AS7 t95 t96 5-4796 (F) * See sioc register, MSB field, to determine if B0 is the MSB or LSB. See sioc register, ILEN field, to determine if the DO word length is 8 bits or 16 bits. Figure 55. SIO Passive Mode Output Timing Diagram Table 140. Timing Requirements for Serial Inputs Abbreviated Reference t80 t81 t82 t83 t84 t85 Parameter Clock Period (high to high)‡ Clock Low Time (low to high) Clock High Time (high to low) Load High Setup (high to high) Load Low Setup (low to high) Load Hold (high to invalid) Min 40 18 18 8 8 0 Max — — — — — — Unit ns ns ns ns ns ns Min — — — 5 — 5 — — — Max 35 35 35 — 35 — 35 35 35 Unit ns ns ns ns ns ns ns ns ns † † Device is fully static; t80 is tested at 200 ns. ‡ For multiprocessor mode, see note in Section 12.10. Table 141. Timing Characteristics for Serial Outputs Abbreviated Reference t87 t88 t89 t90 t92 t93 t94 t95 t96 Lucent Technologies Inc. Parameter Data Delay (high to valid) Enable Data Delay (low to active) Disable Data Delay (high to 3-state) Data Hold (high to invalid) Address Delay (high to valid) Address Hold (high to invalid) Enable Delay (low to active) Disable Delay (high to 3-state) OBE Delay (high to high) 117 Data Sheet March 2000 DSP1627 Digital Signal Processor 11 Timing Characteristics for 3.0 V Operation (continued) OCK VOH– VOL– t102 t86a OLD VOH– VOL– DO VOH– VOL– * t88 B0 t94 SADD VOH– VOL– DOEN VIH– VIL– OBE VOH– VOL– t87 t92 AD0 t90 B1 t90 B7 t93 AD1 BN – 1 t89 t93 AD7 AS7 t95 t96 5-4797 (F) * OLD goes high at the end of bit 6 of 0:15. Figure 56. SIO Active Mode Output Timing Diagram Table 142. Timing Characteristics for Serial Output Abbreviated Reference t86a t102 t87 t88 t89 t90 t92 t93 t94 t95 t96 118 Parameter OLD Delay (high to low) OLD Hold (high to invalid) Data Delay (high to valid) Enable Data Delay (low to active) Disable Data Delay (high to 3-state) Data Hold (high to invalid) Address Delay (high to valid) Address Hold (high to invalid) Enable Delay (low to active) Disable Delay (high to 3-state) OBE Delay (high to high) Min — 5 — — — 5 — 5 — — — Max 35 — 35 35 35 — 35 — 35 35 35 Unit ns ns ns ns ns ns ns ns ns ns ns Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 11 Timing Characteristics for 3.0 V Operation (continued) CKO VOH– VOL– ICK VOH– VOL– OCK VOH– VOL– ICK/OCK* ILD OLD SYNC t97 t98 t99 t100 VOH– t76a t101 t76b t101 t86a t102 t86b t102 t103 t105 t104 t105 VOH– VOL– VOH– VOL– VOH– VOL– 5-4798 (F) * See sioc register, LD field. Figure 57. Serial I/O Active Mode Clock Timing Table 143. Timing Characteristics for Signal Generation Abbreviated Reference t97 t98 t99 t100 t76a t76b t101 t86a t86b t102 t103 t104 t105 Lucent Technologies Inc. Parameter ICK Delay (high to high) ICK Delay (high to low) OCK Delay (high to high) OCK Delay (high to low) ILD Delay (high to low) ILD Delay (high to high) ILD Hold (high to invalid) OLD Delay (high to low) OLD Delay (high to high) OLD Hold (high to invalid) SYNC Delay (high to low) SYNC Delay (high to high) SYNC Hold (high to invalid) Min — — — — — — 5 — — 5 — — 5 Max 18 18 18 18 35 35 — 35 35 — 35 35 — Unit ns ns ns ns ns ns ns ns ns ns ns ns ns 119 Data Sheet March 2000 DSP1627 Digital Signal Processor 11 Timing Characteristics for 3.0 V Operation (continued) 11.10 Multiprocessor Communication TIME SLOT 1 TIME SLOT 2 OCK/ICK t113 t112 t112 SYNC VIH– VIL– DO/D1 VOH– VOL– t113 * t116 B15 B0 B1 t117 B7 B8 B15 B0 t114 t122 t121 AD0 SADD AD1 AD7 AS0 t120 DOEN t115 AS7 AD0 t120 VOH– VOL– 5-4799 (F) * Negative edge initiates time slot 0. Figure 58. SIO Multiprocessor Timing Diagram Note: All serial I/O timing requirements and characteristics still apply, but the minimum clock period in passive multiprocessor mode, assuming 50% duty cycle, is calculated as (t77 + t116) x 2. Table 144. Timing Requirements for SIO Multiprocessor Communication Abbreviated Reference t112 t113 t114 t115 Parameter Sync Setup (high/low to high) Sync Hold (high to high/low) Address Setup (valid to high) Address Hold (high to invalid) Min 35 0 12 0 Max — — — — Unit ns ns ns ns Min — — — — — Max 35 30 25 35 30 Unit ns ns ns ns ns Table 145. Timing Characteristics for SIO Multiprocessor Communication Abbreviated Reference* t116 t117 t120 t121 t122 Parameter Data Delay (bit 0 only) (low to valid) Data Disable Delay (high to 3-state) DOEN Valid Delay (high to valid) Address Delay (bit 0 only) (low to valid) Address Disable Delay (high to 3-state) * With capacitance load on ICK, OCK, DO, SYNC, and SADD = 100 pF, add 4 ns to t116—t122. 120 Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 12 Timing Characteristics for 2.7 V Operation The following timing characteristics and requirements are preliminary information and are subject to change. Timing characteristics refer to the behavior of the device under specified conditions. Timing requirements refer to conditions imposed on the user for proper operation of the device. All timing data is valid for the following conditions: TA = –40 °C to +85 °C (See Section 8.3.) VDD = 3 V ± 10%, VSS = 0 V (See Section 8.3.) Capacitance load on outputs (CL) = 50 pF, except for CKO, where CL = 20 pF. Output characteristics can be derated as a function of load capacitance (CL). All outputs: 0.03 ns/pF ≤ dt/dCL ≤ 0.07 ns/pF for 10 ≤ CL ≤ 100 pF at VIH for rising edge and at VIL for falling edge. For example, if the actual load capacitance is 30 pF instead of 50 pF, the derating for a rising edge is (30 – 50) pF x 0.06 ns/pF = 1.2 ns less than the specified rise time or delay that includes a rise time. Test conditions for inputs: ■ Rise and fall times of 4 ns or less ■ Timing reference levels for delays = VIH, VIL Test conditions for outputs (unless noted otherwise): ■ CLOAD = 50 pF; except for CKO, where CLOAD = 20 pF ■ Timing reference levels for delays = VIH, VIL ■ 3-state delays measured to the high-impedance state of the output driver For the timing diagrams, see Table 62 for input clock requirements. Unless otherwise noted, CKO in the timing diagrams is the free-running CKO. Lucent Technologies Inc. 121 Data Sheet March 2000 DSP1627 Digital Signal Processor 12 Timing Characteristics for 2.7 V Operation (continued) 12.1 DSP Clock Generation t1 t3 t2 1X CKI* t4 t5 CKO† t6, t6a CKO‡ EXTERNAL MEMORY CYCLE W = 1§ 5-4009 (F).a * † ‡ § See Table 62 for input clock electrical requirements. Free-running clock. Wait-stated clock (see Table 38). W = number of wait-states. Figure 59. I/O Clock Timing Diagram Table 146. Timing Requirements for Input Clock t1 Clock In Period (high to high) 20 ns and 12.5 ns* Min Max Unit ns 20 —† t2 t3 Clock In Low Time (low to high) Clock In High Time (high to low) 10 10 Abbreviated Reference Parameter — — ns ns * Device speeds greater than 50 MIPS do not support 1 X operation. Use the PLL. † Device is fully static, t1 is tested at 100 ns for 1X input clock option, and memory hold time is tested at 0.1 s. Table 147. Timing Characteristics for Input Clock and Output Clock Abbreviated Reference Parameter t4 t5 t6 t6a Clock Out High Delay Clock Out Low Delay (high to low) Clock Out Period (low to low) Clock Out Period with SLOWCKI Bit Set in powerc Register (low to low) 20 ns Min Max — 14 — 14 T* — 0.74 3.8 12.5 ns Min Max — 10 — 10 T* — 0.74 3.8 Unit ns ns ns µs * T = internal clock period, set by CKI or by CKI and the PLL parameters. 122 Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 12 Timing Characteristics for 2.7 V Operation (continued) 12.2 Reset Circuit The DSP1627 has a powerup reset circuit that automatically clears the JTAG controller upon powerup. If the supply voltage falls below VDD MIN* and a reset is required, the JTAG controller must be reset—even if the JTAG port isn’t being used—by applying six low-to-high clock edges on TCK with TMS held high, followed by the usual RSTB and CKI reset sequence. Figure 60 shows two separate events: an initial powerup and a powerup following a drop in the power supply voltage. * See Table 60, Recommended Operating Condiitons. VDD RAMP VDD MIN 0.4 V V DD MIN 0.4 V t146 t9 t8 t151 t152 t8 t9 CKI TCK TMS VIH t153 t153 VIH RSTB VIL t10 t11 t10 t11 PINS VOH VOL 5-2253 (F).a Notes: See Table 62 for CKI electrical requirements and Table 151 for TCK timing requirements. When both INT0 and RSTB are asserted, all output and bidirectional pins (except TDO, which 3-states by JTAG control) are put in a 3-state condition. With RSTB asserted and INT0 not asserted, EROM, ERAMHI, ERAMLO, IO, DSEL, and RWN outputs remain high, and CKO remains a free-running clock. TMS and TDI signals have internal pull-up devices. Figure 60. Powerup Reset and Chip Reset Timing Diagram Table 148. Timing Requirements for Powerup Reset and Chip Reset Abbreviated Reference Parameter t8 Reset Pulse (low to high) t9 VDD Ramp t146 VDD MIN to RSTB Low CMOS Crystal* Small-signal t151 t152 TMS High JTAG Reset to RSTB Low t153 RSTB (low to high) Min 6T — 2T 20 20 6 * TTCK† 2T CMOS Crystal* 20 ms – 6 * TTCK if 6 * TTCK < 20 ms 0 if 6 * TTCK ≥ 20 ms Small-Signal 20 µs – 6 * TTCK if 6 * TTCK < 20 µs 0 if 6 * TTCK ≥ 20 µs — Max — 10 — — — Unit ns ms ns ms µs — — — — — — ns ns 54 ns * With external components as specified in Table 62. † TTCK = t12 = TCK period. See Table 151 for TCK timing requirements. Lucent Technologies Inc. 123 Data Sheet March 2000 DSP1627 Digital Signal Processor 12 Timing Characteristics for 2.7 V Operation (continued) Table 149. Timing Characteristics for Powerup Reset and Chip Reset Abbreviated Reference t10 t11 Parameter RSTB Disable Time (low to 3-state) RSTB Enable Time (high to valid) Min — — Max 100 100 Unit ns ns Note: The device needs to be clocked for at least six CKI cycles during reset after powerup. Otherwise, high currents may flow. 12.3 Reset Synchronization t5 + 2 x t6 CKI* VIH VIL t126 RSTB VIH VIL CKO VIH VIL CKO VIH VIL 5-4011 (F).a * See Table 62 for input clock electrical requirements. Note: CKO1 and CKO2 are two possible CKO states before reset. CKO is free-running. Figure 61. Reset Synchronization Timing Table 150. Timing Requirements for Reset Synchronization Timing Abbreviated Reference t126 124 Parameter Reset Setup (high to high) Min 3 Max T/2 – 5 Unit ns Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 12 Timing Characteristics for 2.7 V Operation (continued) 12.4 JTAG I/O Specifications t12 t155 t13 TCK t14 VIH VIL t15 t156 t16 TMS VIH VIL t17 t18 VIH TDI VIL t19 t20 TDO VOH VOL 5-4017 (F) Figure 62. JTAG Timing Diagram Table 151. Timing Requirements for JTAG Input/Output Abbreviated Reference t12 t13 t14 t15 t16 t17 t18 Parameter TCK Period (high to high) TCK High Time (high to low) TCK Low Time (low to high) TMS Setup Time (valid to high) TMS Hold Time (high to invalid) TDI Setup Time (valid to high) TDI Hold Time (high to invalid) Min 50 22.5 22.5 7.5 2 7.5 2 Max — — — — — — — Unit ns ns ns ns ns ns ns Table 152. Timing Characteristics for JTAG Input/Output Abbreviated Reference t19 t20 Lucent Technologies Inc. Parameter TDO Delay (low to valid) TDO Hold (low to invalid) Min — 0 Max 19 — Unit ns ns 125 Data Sheet March 2000 DSP1627 Digital Signal Processor 12 Timing Characteristics for 2.7 V Operation (continued) 12.5 Interrupt CKO* V OH V OL t21 INT[1:0] V IH V IL t22 t23 IACK † t25 V OH V OL t24 t26 VEC[3:0] V OH V OL 5-4018 (F) * CKO is free-running. † IACK assertion is guaranteed to be enclosed by VEC[3:0] assertion. Figure 63. Interrupt Timing Diagram Table 153. Timing Requirements for Interrupt Abbreviated Reference t21 t22 Parameter Interrupt Setup (high to low) INT Assertion Time (high to low) Min 19 2T Max — — Unit ns ns Max T/2 + 10 12.5 10 12.5 Unit ns ns ns ns Note: Interrupt is asserted during an interruptible instruction and no other pending interrupts. Table 154. Timing Characteristics for Interrupt Abbreviated Reference t23 t24 t25 t26 Parameter IACK Assertion Time (low to high) VEC Assertion Time (low to high) IACK Invalid Time (low to low) VEC Invalid Time (low to low) Min — — — — Note: Interrupt is asserted during an interruptible instruction and no other pending interrupts. 126 Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 12 Timing Characteristics for 2.7 V Operation (continued) 12.6 Bit Input/Output (BIO) t144 CKO V OH V OL t29 V OH IOBIT (OUTPUT) V OL VALID OUTPUT t28 t27 IOBIT (INPUT) V IH DATA INPUT V IL 5-4019 (F).a Figure 64. Write Outputs Followed by Read Inputs (cbit = Immediate; a1 = sbit) Table 155. Timing Requirements for BIO Input Read Abbreviated Reference t27 t28 Parameter IOBIT Input Setup Time (valid to high) IOBIT Input Hold Time (high to invalid) Min 15 0 Max — — Unit ns ns Min — 1 Max 9 — Unit ns ns Table 156. Timing Characteristics for BIO Output Abbreviated Reference t29 t144 Parameter IOBIT Output Valid Time (low to valid) IOBIT Output Hold Time (low to invalid) t144 CKO VOH – V O L– t29 IOBIT VOH – (OUTPUT) V O L– VALID OUTPUT t141 IOBIT (INPUT) VIH – V I L– t142 TEST INPUT 5-4019 (F).b Figure 65. Write Outputs and Test Inputs (cbit = Immediate) Table 157. Timing Requirements for BIO Input Test Abbreviated Reference t141 t142 Lucent Technologies Inc. Parameter IOBIT Input Setup Time (valid to low) IOBIT Input Hold Time (low to invalid) Min 15 0 Max — — Unit ns ns 127 Data Sheet March 2000 DSP1627 Digital Signal Processor 12 Timing Characteristics for 2.7 V Operation (continued) 12.7 External Memory Interface The following timing diagrams, characteristics, and requirements do not apply to interactions with delayed external memory enables unless so stated. See the DSP1611/17/18/27 Digital Signal Processor Information Manual for a detailed description of the external memory interface including other functional diagrams. CKO VOH VOL t33 t34 VOH ENABLE VOL W* = 0 5-4020 (F).b * W = number of wait-states. Figure 66. Enable Transition Timing Table 158. Timing Characteristics for External Memory Enables (EROM, ERAMHI, IO, ERAMLO) Abbreviated Reference t33 t34 Parameter CKO to ENABLE Active (low to low) CKO to ENABLE Inactive (low to high) Min 0 –1 Max 5 4.5 Unit ns ns Table 159. Timing Characteristics for Delayed External Memory Enables (ioc = 0x000F) Abbreviated Reference t33 128 Parameter CKO to Delayed ENABLE Active (low to low) Min T/2 – 2 Max T/2 + 7 Unit ns Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 12 Timing Characteristics for 2.7 V Operation (continued) (MWAIT = 0 x 2222) W* = 2 CKO V OH V OL t127 ENABLE V OH V OL t129 t130 V IH DB READ DATA V IL t150 t128 AB V OH V OL READ ADDRESS 5-4021 (F).a * W = number of wait-states. Figure 67. External Memory Data Read Timing Diagram Table 160. Timing Characteristics for External Memory Access Abbreviated Reference t127 t128 Parameter Enable Width (low to high) Address Valid (enable low to valid) Min T(1 + W) – 1.5 — Max — 2 Unit ns ns Table 161. Timing Requirements for External Memory Read (EROM, ERAMHI, IO, ERAMLO) Abbreviated Reference Parameter t129 t130 t150 Read Data Setup (valid to enable high) Read Data Hold (enable high to hold) External Memory Access Time (valid to valid) Lucent Technologies Inc. 20 ns Min Max 15 — 0 — — T(1 + W) – 15 12.5 ns Unit Min Max 13 — ns 0 — ns — T(1 + W) – 14 ns 129 Data Sheet March 2000 DSP1627 Digital Signal Processor 12 Timing Characteristics for 2.7 V Operation (continued) (MWAIT = 0x1002) W* = 2 CKO ERAMLO DB W* = 1 VOH VOL VOH VOL VOH VOL WRITE DATA READ t131 EROM VOH VOL t132 t133 RWN t134 VOH VOL t135 t136 AB VOH WRITE ADDRESS VOL READ ADDRESS 5-4022 (F).a * W = number of wait-states. Figure 68. External Memory Data Write Timing Diagram Table 162. Timing Characteristics for External Memory Data Write (All Enables) Abbreviated Reference t131 t132 t133 t134 t135 t136 130 Parameter 20 ns 12.5 ns Min Max Min Max Write Overlap (enable low to 3-state) — 0 — 0 RWN Advance (RWN high to enable high) 0 — 0 — RWN Delay (enable low to RWN low) 0 — 0 — Write Data Setup (data valid to RWN high) T(1 + W)/2 – 4 — T(1 + W)/2 – 3 — RWN Width (low to high) T(1 + W) – 5 — T(1 + W) – 4 — Write Address Setup (address valid to RWN 0 — 0 — low) Unit ns ns ns ns ns ns Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 12 Timing Characteristics for 2.7 V Operation (continued) (MWAIT = 0x1002) W* = 2 CKO ERAMLO EROM W* = 1 VOH VOL VOH VOL VOH VOL t131 DB VOH VOL WRITE READ t137 t138 RWN VOH VOL t139 AB VOH VOL WRITE ADDRESS READ ADDRESS 5-4023 (F).a * W = number of wait-states. Figure 69. Write Cycle Followed by Read Cycle Table 163. Timing Characteristics for Write Cycle Followed by Read Cycle Abbreviated Reference t131 t137 t138 t139 Lucent Technologies Inc. Parameter Write Overlap (enable low to 3-state) Write Data 3-state (RWN high to 3-state) Write Data Hold (RWN high to data hold) Write Address Hold (RWN high to address hold) Min — — 0 0 Max 0 2 — — Unit ns ns ns ns 131 Data Sheet March 2000 DSP1627 Digital Signal Processor 12 Timing Characteristics for 2.7 V Operation (continued) 12.8 PHIF Specifications For the PHIF, "READ" means read by the external user (output by the DSP); "WRITE" is similarly defined. The 8-bit reads/ writes are identical to one-half of a 16-bit access. 16-bit READ 16-bit WRITE VIH– PCSN PODS VIL– VIH– VIL– t41 t42 t44 VIH– PIDS VIL– t43 PBSEL VIH– VIL– t47 t45 PSTAT t48 t46 VIH– VIL– t49 VIH– PB[7:0] t50 t154 t51 t52 VIL– 5-4036 (F) Figure 70. PHIF Intel Mode Signaling (Read and Write) Timing Diagram Table 164. Timing Requirements for PHIF Intel Mode Signaling Abbreviated Reference t41 t42 t43 t44 t45* t46* t47* t48* t51* t52* Parameter PODS to PCSN Setup (low to low) PCSN to PODS Hold (high to high) PIDS to PCSN Setup (low to low) PCSN to PIDS Hold (high to high) PSTAT to PCSN Setup (valid to low) PCSN to PSTAT Hold (high to invalid) PBSEL to PCSN Setup (valid to low) PCSN to PBSEL Hold (high to invalid) PB Write to PCSN Setup (valid to high) PCSN to PB Write Hold (high to invalid) Min 0 0 0 0 6 0 6 0 10 5 Max — — — — — — — — — — Unit ns ns ns ns ns ns ns ns ns ns * This timing diagram for the PHIF port shows accesses using the PCSN signal to initiate and complete a transaction. The transactions can also be initiated and completed with the PIDS and PODS signals. An output transaction (read) is initiated by PCSN or PODS going low, whichever comes last. For example, the timing requirements referenced to PCSN going low, t45 and t49, should be referenced to PODS going low, if PODS goes low after PCSN. An output transaction is completed by PCSN or PODS going high, whichever comes first. An input transaction is initiated by PCSN or PIDS going low, whichever comes last. An input transaction is completed by PCSN or PIDS going high, whichever comes first. All requirements referenced to PCSN apply to PIDS or PODS, if PIDS or PODS is the controlling signal. Table 165. Timing Characteristics for PHIF Intel Mode Signaling Abbreviated Reference t49* t50* Parameter PCSN to PB Read (low to valid) PCSN to PB Read Hold (high to invalid) Min — 3 Max 17 — Unit ns ns * This timing diagram for the PHIF port shows accesses using the PCSN signal to initiate and complete a transaction. The transactions can also be initiated and completed with the PIDS and PODS signals. An output transaction (read) is initiated by PCSN or PODS going low, whichever comes last. For example, the timing requirements referenced to PCSN going low, t45 and t49, should be referenced to PODS going low, if PODS goes low after PCSN. An output transaction is completed by PCSN or PODS going high, whichever comes first. An input transaction is initiated by PCSN or PIDS going low, whichever comes last. An input transaction is completed by PCSN or PIDS going high, whichever comes first. All requirements referenced to PCSN apply to PIDS or PODS, if PIDS or PODS is the controlling signal. 132 Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 12 Timing Characteristics for 2.7 V Operation (continued) 16-bit READ 16-bit WRITE 8-bit WRITE 8-bit READ t55 t55 PCSN VIH VIL t56 t55 PODS t56 VIH VIL t56 PIDS t56 t55 VIH VIL t56 PBSEL VOH VOL t53 POBE t53 VOH VOL t54 t54 PIBF VOH VOL 5-4037 (F).a Figure 71. PHIF Intel Mode Signaling (Pulse Period and Flags) Timing Diagram Table 166. Timing Requirements for PHIF Intel Mode Signaling Abbreviated Reference t55 t56 Parameter PCSN/PODS/PIDS Pulse Width (high to low) PCSN/PODS/PIDS Pulse Width (low to high) Min 20.5 20.5 Max — — Unit ns ns Table 167. Timing Characteristics for PHIF Intel Mode Signaling Abbreviated Reference t53* t54* Parameter † PCSN/PODS to POBE (high to high) PCSN/PIDS to PIBF† (high to high) Min — Max 20 Unit ns — 20 ns * t53 should be referenced to the rising edge of PCSN or PODS, whichever comes first. t54 should be referenced to the rising edge of PCSN or PIDS, whichever comes first. † POBE and PIBF may be programmed to be the opposite logic levels shown in the diagram (positive assertion levels shown). t53 and t54 apply to the inverted levels as well as those shown. Lucent Technologies Inc. 133 Data Sheet March 2000 DSP1627 Digital Signal Processor 12 Timing Characteristics for 2.7 V Operation (continued) 16-bit READ 16-bit WRITE VI H PCSN VIL t42 VI H PDS VIL t41 PRWN t47 t48 t44 VI H VIL t45 PSTAT t44 VIL t43 PBSEL t43 VI H t46 VI H VIL t49 t50 t154 t51 t52 PB[7:0] 5-4038 (F).a Figure 72. PHIF Motorola Mode Signaling (Read and Write) Timing Diagram Table 168. Timing Requirements for PHIF Motorola Mode Signaling Abbreviated Reference t41 t42 t43 t44 t45* t46* t47* t48* t51* t52* Parameter PDS to PCSN Setup (valid to low) PCSN to PDS† Hold (high to invalid) PRWN to PCSN Setup (valid to low) PCSN to PRWN Hold (high to invalid) PSTAT to PCSN Setup (valid to low) PCSN to PSTAT Hold (high to invalid) PBSEL to PCSN Setup (valid to low) PCSN to PBSEL Hold (high to invalid) PB Write to PCSN Setup (valid to high) PCSN to PB Write Hold (high to invalid) † Min 0 0 6 0 6 0 6 0 10 5 Max — — — — — — — — — — Unit ns ns ns ns ns ns ns ns ns ns * This timing diagram for the PHIF port shows accesses using the PCSN signal to initiate and complete a transaction. The transactions can also be initiated and completed with the PDS signal. An input/output transaction is initiated by PCSN or PDS going low, whichever comes last. For example, the timing requirements referenced to PCSN going low, t45 and t49, should be referenced to PDS going low, if PDS goes low after PCSN. An input/output transaction is completed by PCSN or PDS going high, whichever comes first. All requirements referenced to PCSN should be referenced to PDS, if PDS is the controlling signal. PRWN should never be used to initiate or complete a transaction. † PDS is programmable to be active-high or active-low. It is shown active-low in Figures 72 and 73. POBE and PIBF may be programmed to be the opposite logic levels shown in the diagram. t53 and t54 apply to the inverted levels as well as those shown. Table 169. Timing Characteristics for PHIF Motorola Mode Signaling Abbreviated Reference t49* t50* Parameter PCSN to PB Read (low to valid) PCSN to PB Read (high to invalid) Min — 3 Max 17 — Unit ns ns * This timing diagram for the PHIF port shows accesses using the PCSN signal to initiate and complete a transaction. The transactions can also be initiated and completed with the PDS signal. An input/output transaction is initiated by PCSN or PDS going low, whichever comes last. For example, the timing requirements referenced to PCSN going low, t45 and t49, should be referenced to PDS going low, if PDS goes low after PCSN. An input/output transaction is completed by PCSN or PDS going high, whichever comes first. All requirements referenced to PCSN should be referenced to PDS, if PDS is the controlling signal. PRWN should never be used to initiate or complete a transaction. 134 Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 12 Timing Characteristics for 2.7 V Operation (continued) 16-bit READ 16-bit WRITE 8-bit WRITE 8-bit READ t55 t55 PCSN VIH– VIL– t56 t55 PDS t56 VIL– t56 PRWN t56 VIH– t55 VIH– VIL– t56 PBSEL VOH– VOL– t53 POBE t53 VOH– VOL– t54 t54 PIBF VOH– VOL– 5-4039 (F).a Figure 73. PHIF Motorola Mode Signaling (Pulse Period and Flags) Timing Diagram Table 170. Timing Characteristics for PHIF Motorola Mode Signaling Abbreviated Reference t53* t54* Parameter PCSN/PDS to POBE† (high to high) PCSN/PDS† to PIBF† (high to high) † Min — — Max 20 20 Unit ns ns * An input/output transaction is initiated by PCSN or PDS going low, whichever comes last. For example, t53 and t54 should be referenced to PDS going low, if PDS goes low after PCSN. An input/output transaction is completed by PCSN or PDS going high, whichever comes first. All requirements referenced to PCSN should be referenced to PDS, if PDS is the controlling signal. PRWN should never be used to initiate or complete a transaction. † PDS is programmable to be active-high or active-low. It is shown active-low in Figures 72 and 73. POBE and PIBF may be programmed to be the opposite logic levels shown in the diagram. t53 and t54 apply to the inverted levels as well as those shown. Table 171. Timing Requirements for PHIF Motorola Mode Signaling Abbreviated Reference Parameter t55 PCSN/PDS/PRWN Pulse Width (high to low) t56 PCSN/PDS/PRWN Pulse Width (low to high) Lucent Technologies Inc. Min 20 20 Max — — Unit ns ns 135 Data Sheet March 2000 DSP1627 Digital Signal Processor 12 Timing Characteristics for 2.7 V Operation (continued) V IH PCSN V IL V IH PODS (PDS*) V IL V IH PIDS (PRWN*) V IL t47 t48 V IH PBSEL V IL t46 t45 V IH PSTAT V IL t49 PB[7:0] V OH t50 t154 V OL 5-4040 (F).a * Motorola mode signal name. Figure 74. PHIF Intel or Motorola Mode Signaling (Status Register Read) Timing Diagram Table 172. Timing Requirements for Intel and Motorola Mode Signaling (Status Register Read) Abbreviated Reference t45† Parameter PSTAT to PCSN Setup (valid to low) Min 6 Max — Unit ns t46‡ PCSN to PSTAT Hold (high to invalid) 0 — ns t475† PBSEL to PCSN Setup (valid to low) 6 — ns PCSN to PBSEL Hold (high to invalid) 0 — ns t48 ‡ † t45, t47, and t49 are referenced to the falling edge of PCSN or PODS(PDS), whichever occurs last. ‡ t46, t48, and t50 are referenced to the rising edge of PCSN or PODS(PDS), whichever occurs first. Table 173. Timing Characteristics for Intel and Motorola Mode Signaling (Status Register Read) Abbreviated Reference t49† t50‡ Parameter PCSN to PB Read (low to valid) PCSN to PB Read Hold (high to invalid) Min — 3 Max 17 — Unit ns ns † t45, t47, and t49 are referenced to the falling edge of PCSN or PODS(PDS), whichever occurs last. ‡ t46, t48, and t50 are referenced to the rising edge of PCSN or PODS(PDS), whichever occurs first. 136 Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 12 Timing Characteristics for 2.7 V Operation (continued) VIH– VIL– RSTB t57 POBE VOH– VOL– PIBF VOH– VOL– t58 5-4775 (F) Figure 75. PHIF, PIBF, and POBE Reset Timing Diagram Table 174. PHIF Timing Characteristics for PHIF, PIBF, and POBE Reset Abbreviated Reference Parameter Min Max Unit t57 POBE/PIBF* — 25 ns 3 25 ns RSTB Disable to (high to valid) RSTB Enable to POBE/PIBF* (low to invalid) t58 * After reset, POBE and PIBF always go to the levels shown, indicating output buffer empty and input buffer empty. The DSP program, however, may later invert the definition of the logic levels for POBE and PIBF. t57 and t58 continue to apply. CKO VIH– VIL– † VOH– VOL– t59 POBE t59 PIBF† VOH– VOL– 5-4776 (F) † POBE and PIBF can be programmed to be active-high or active-low. They are shown active-high. The timing characteristic for active-low is the same as for active-high. Figure 76. PHIF, PIBF, and POBE Disable Timing Diagram Table 175. PHIF Timing Characteristics for POBE and PIBF Disable Abbreviated Reference t59 Parameter CKO to POBE/PIBF Disable (high/low to disable) Min — Max 20 Unit ns * POBE and PIBF can be programmed to be active-high or active-low. They are shown active-high. The timing characteristic for active-low is the same as for active-high. Lucent Technologies Inc. 137 Data Sheet March 2000 DSP1627 Digital Signal Processor 12 Timing Characteristics for 2.7 V Operation (continued) 12.9 Serial I/O Specifications t70 ICK VIH– VIL– ILD VIH– VIL– DI VIH– VIL– t72 t71 t75 t74 t73 t75 t77 t78 B0 BN – 1* B1 B0 t79 IBF VOH– VOL– 5-4777 (F) * N = 16 or 8 bits. Figure 77. SIO Passive Mode Input Timing Diagram Table 176. Timing Requirements for Serial Inputs Abbreviated Reference t70 t71 t72 t73 t74 t75 t77 t78 Parameter Clock Period (high to high)‡ Clock Low Time (low to high) Clock High Time (high to low) Load High Setup (high to high) Load Low Setup (low to high) Load High Hold (high to invalid) Data Setup (valid to high) Data Hold (high to invalid) Min 40 18 18 8 8 0 7 0 Max —† — — — — — — — Unit ns ns ns ns ns ns ns ns † Device is fully static; t70 is tested at 200 ns. ‡ For multiprocessor mode, see note in Section 12.10. Table 177. Timing Characteristics for Serial Outputs Abbreviated Reference t79 138 Parameter IBF Delay (high to high) Min — Max 35 Unit ns Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 12 Timing Characteristics for 2.7 V Operation (continued) ICK VOH– VOL– t101 t76a ILD VOH– VOL– DI VIH– VIL– IBF VOH– VOL– * t77 t78 B0 B1 BN – 1 B0 t79 5-4778 (F) * ILD goes high during bit 6 (of 0:15), N = 8 or 16. Figure 78. SIO Active Mode Input Timing Diagram Table 178. Timing Requirements for Serial Inputs Abbreviated Reference t77 t78 Parameter Data Setup (valid to high) Data Hold (high to invalid) Min 7 0 Max — — Unit ns ns Table 179. Timing Characteristics for Serial Outputs Abbreviated Reference t76a t101 t79 Lucent Technologies Inc. Parameter ILD Delay (high to low) ILD Hold (high to invalid) IBF Delay (high to high) Min — 5 — Max 35 — 35 Unit ns ns ns 139 Data Sheet March 2000 DSP1627 Digital Signal Processor 12 Timing Characteristics for 2.7 V Operation (continued) t80 OCK t82 t81 t85 t84 VIH– VIL– t83 OLD VIH– VIL– DO* VOH– VOL– t85 t88 t90 B0 t94 SADD t87 B1 t92 VOH– VOL– t90 B7 t93 AD0 AD1 BN – 1 t89 t93 AS7 AD7 t95 DOEN VIH– VIL– OBE VOH– VOL– t96 5-4796 (F) * See sioc register, MSB field, to determine if B0 is the MSB or LSB. See sioc register, ILEN field, to determine if the DO word length is 8 bits or 16 bits. Figure 79. SIO Passive Mode Output Timing Diagram Table 180. Timing Requirements for Serial Inputs Abbreviated Reference t80 t81 t82 t83 t84 t85 Parameter Clock Period (high to high)‡ Clock Low Time (low to high) Clock High Time (high to low) Load High Setup (high to high) Load Low Setup (low to high) Load Hold (high to invalid) Min 40 18 18 8 8 0 Max —† — — — — — Unit ns ns ns ns ns ns † Device is fully static; t80 is tested at 200 ns. ‡ For multiprocessor mode, see note in Section 12.10. Table 181. Timing Characteristics for Serial Outputs Abbreviated Reference t87 t88 t89 t90 t92 t93 t94 t95 t96 140 Parameter Data Delay (high to valid) Enable Data Delay (low to active) Disable Data Delay (high to 3-state) Data Hold (high to invalid) Address Delay (high to valid) Address Hold (high to invalid) Enable Delay (low to active) Disable Delay (high to 3-state) OBE Delay (high to high) Min — — — 5 — 5 — — — Max 35 35 35 — 35 — 35 35 35 Unit ns ns ns ns ns ns ns ns ns Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 12 Timing Characteristics for 2.7 V Operation (continued) OCK VOH– VOL– t102 t86a OLD VOH– VOL– DO VOH– VOL– * t88 t90 B0 t94 SADD t87 t92 VOH– VOL– AD0 B1 t90 B7 t93 AD1 BN – 1 t89 t93 AS7 AD7 t95 DOEN VIH– VIL– OBE VOH– VOL– t96 5-4797 (F) * OLD goes high at the end of bit 6 of 0:15. Figure 80. SIO Active Mode Output Timing Diagram Table 182. Timing Characteristics for Serial Output Abbreviated Reference t86a t102 t87 t88 t89 t90 t92 t93 t94 t95 t96 Lucent Technologies Inc. Parameter OLD Delay (high to low) OLD Hold (high to invalid) Data Delay (high to valid) Enable Data Delay (low to active) Disable Data Delay (high to 3-state) Data Hold (high to invalid) Address Delay (high to valid) Address Hold (high to invalid) Enable Delay (low to active) Disable Delay (high to 3-state) OBE Delay (high to high) Min — 5 — — — 5 — 5 — — — Max 35 — 35 35 35 — 35 — 35 35 35 Unit ns ns ns ns ns ns ns ns ns ns ns 141 Data Sheet March 2000 DSP1627 Digital Signal Processor 12 Timing Characteristics for 2.7 V Operation (continued) CKO VOH– VOL– ICK VOH– VOL– OCK VOH– VOL– ICK/OCK* ILD OLD SYNC t97 t98 t99 t100 VOH– t76a t101 t76b t101 t86a t102 t86b t102 t103 t105 t104 t105 VOH– VOL– VOH– VOL– VOH– VOL– 5-4798 (F) * See sioc register, LD field. Figure 81. Serial I/O Active Mode Clock Timing Table 183. Timing Characteristics for Signal Generation Abbreviated Reference t97 t98 t99 t100 t76a t76b t101 t86a t86b t102 t103 t104 t105 142 Parameter ICK Delay (high to high) ICK Delay (high to low) OCK Delay (high to high) OCK Delay (high to low) ILD Delay (high to low) ILD Delay (high to high) ILD Hold (high to invalid) OLD Delay (high to low) OLD Delay (high to high) OLD Hold (high to invalid) SYNC Delay (high to low) SYNC Delay (high to high) SYNC Hold (high to invalid) Min — — — — — — 5 — — 5 — — 5 Max 18 18 18 18 35 35 — 35 35 — 35 35 — Unit ns ns ns ns ns ns ns ns ns ns ns ns ns Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 12 Timing Characteristics for 2.7 V Operation (continued) 12.10 Multiprocessor Communication TIME SLOT 1 TIME SLOT 2 OCK/ICK t113 t112 t112 SYNC VIH– VIL– DO/D1 VOH– VOL– t113 * t116 B15 B0 B1 t117 B7 B8 B15 B0 t114 t122 t121 AD0 SADD AD1 AD7 AS0 t120 DOEN t115 AS7 AD0 t120 VOH– VOL– 5-4799 (F) * Negative edge initiates time slot 0. Figure 82. SIO Multiprocessor Timing Diagram Note: All serial I/O timing requirements and characteristics still apply, but the minimum clock period in passive multiprocessor mode, assuming 50% duty cycle, is calculated as (t77 + t116) x 2. Table 184. Timing Requirements for SIO Multiprocessor Communication Abbreviated Reference t112 t113 t114 t115 Parameter Sync Setup (high/low to high) Sync Hold (high to high/low) Address Setup (valid to high) Address Hold (high to invalid) Min 35 0 12 0 Max — — — — Unit ns ns ns ns Min — — — — — Max 35 30 25 35 30 Unit ns ns ns ns ns Table 185. Timing Characteristics for SIO Multiprocessor Communication Abbreviated Reference* t116 t117 t120 t121 t122 Parameter Data Delay (bit 0 only) (low to valid) Data Disable Delay (high to 3-state) DOEN Valid Delay (high to valid) Address Delay (bit 0 only) (low to valid) Address Disable Delay (high to 3-state) * With capacitance load on ICK, OCK, DO, SYNC, and SADD = 100 pF, add 4 ns to t116—t122. Lucent Technologies Inc. 143 Data Sheet March 2000 DSP1627 Digital Signal Processor 13 Crystal Electrical Characteristics and Requirements If the option for using the external crystal is chosen, the following electrical characteristics and requirements apply. 13.1 External Components for the Crystal Oscillator The crystal oscillator is enabled by connecting a crystal across CKI and CKI2, along with one external capacitor from each of these pins to ground (see Figure 83). For most applications, 10 pF external capacitors are recommended; however, larger values allow for better frequency precision (see Section 13.4, Frequency Accuracy Considerations). The crystal should be either fundamental or overtone mode, parallel resonant, with a rated power (drive level) of at least 1 mW, and specified at a load capacitance equal to the total capacitance seen by the crystal (including external capacitors and strays). The series resistance of the crystal should be specified to be less than half the absolute value of the negative resistance shown in Figure 84 or Figure 85 for the crystal frequency. The frequency of the signal at the CKI input pin is equal to the crystal frequency. CKI CKI2 XTAL C1 C2 5-4041 (F).a Figure 83. Fundamental Crystal Configuration The following guidelines should be followed when designing the printed-circuit board layout for a crystal-based application: 1. Keep crystal and external capacitors as close to CKI and CKI2 pins as possible to minimize board stray capacitance. 2. Keep high-frequency digital signals such as CKO away from CKI and CKI2 traces to avoid coupling. 13.2 Power Dissipation Figures 86 and 87 indicate the typical power dissipation of the on-chip crystal oscillator circuit versus frequency. Note that these curves are intended to show the relative effects of load capacitance on supply current and that the actual supply current measured depends on crystal resistance. For typical crystals, measured supply current at the VDDA pin should be less than that shown in the figures. 144 Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 13 Crystal Electrical Characteristics and Requirements (continued) CKI CKI2 C1 = C2 = CEXT C0 = PARASITIC CAPACITANCE OF CRYSTAL (7 pF MAXIMUM) C0 C1 C2 Z(ω) 0 –40 0 C1, C2 = 50 pF –40 –80 C1, C2 = 50 pF –80 –120 C1, C2 = 10 pF –120 –160 –160 –200 –200 C1, C2 = 20 pF –240 C1, C2 = 10 pF –240 –280 –280 C1, C2 = 20 pF –320 –320 Re {Z} (Ω) Re [Z] (Ω) –360 –360 –400 –400 –440 –440 –480 –480 –520 –520 –560 –560 –600 –600 –640 –640 –680 –680 –720 –720 –760 –760 –800 0 5 10 15 20 25 30 35 40 FREQUENCY (MHz) –800 0 5 10 15 20 25 30 35 40 FREQUENCY (MHz) 5-3529 (F).b 5-3527 (F).b Figure 84. Negative Resistance of Crystal Oscillator Circuit, VDD = 4.75 V Figure 85. Negative Resistance of Crystal Oscillator Circuit, VDD = 2.7 V Lucent Technologies Inc. 145 Data Sheet March 2000 DSP1627 Digital Signal Processor AVERAGE OSCILLATOR CURRENT (mA) 13 Crystal Electrical Characteristics and Requirements (continued) 7.0 6.5 6.0 5.5 5.0 C1 = C2 = 10 pF 4.5 C1 = C2 = 50 pF 4.0 3.5 3.0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 FREQUENCY (MHz) 5-5188 (F) AVERAGE OSCILLATOR CURRENT (mA) Figure 86. Typical Supply Current of Crystal Oscillator Circuit, VDD = 5.0 V, 25 °C 2.0 1.5 1.0 C1 = C2 = 10 pF 0.5 C1 = C2 = 50 pF 0.0 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 FREQUENCY (MHz) 5-5189 (F) Figure 87. Typical Supply Current of Crystal Oscillator Circuit, VDD = 2.7 V, 25 °C 146 Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 13 Crystal Electrical Characteristics and Requirements (continued) 13.3 LC Network Design for Third Overtone Crystal Circuits For certain crystal applications, it is cheaper to use a third overtone crystal instead of a fundamental mode crystal. When using third overtone crystals, it is necessary, however, to filter out the fundamental frequency so that the circuit will oscillate only at the third overtone. There are several techniques that will accomplish this; one of these is described below. Figure 88 shows the basic setup for third overtone operation. CKI CKI2 XTAL C1 L1 C2 C3 5-4043 (F).a Figure 88. Third Overtone Crystal Configuration The parallel combination of L1 and C1 forms a resonant circuit with a resonant frequency between the first and third harmonic of the crystal such that the LC network appears inductive at the fundamental frequency and capacitive at the third harmonic. This ensures that a 360° phase shift around the oscillator loop will occur at the third overtone frequency but not at the fundamental. The blocking capacitor, C3, provides dc isolation for the trap circuit and should be chosen to be large compared to C1. For example, suppose it is desired to operate with a 40 MHz, third overtone, crystal: Let: f3 = operating frequency of third overtone crystal (40 MHz in this example) f1 = fundamental frequency of third overtone crystal, or f3/3 (13.3 MHz in this example) fT = 1 resonant frequency of trap = -------------------------2π L 1 C 1 C2 = external load capacitor (10 pF in this example) C3 = dc blocking capacitor (0.1 µF in this example) Arbitrarily, set trap resonance to geometric mean of f1 and f3. Since f1 = f3/3, the geometric mean would be: f3 MHz f T = ------- = 40 -------------------- = 23 MHz 3 Lucent Technologies Inc. 3 147 Data Sheet March 2000 DSP1627 Digital Signal Processor 13 Crystal Electrical Characteristics and Requirements (continued) At the third overtone frequency, f3, it is desirable to have the net impedance of the trap circuit (XT) equal to the impedance of C2 (XC2), i.e., X T = X C2 = X C1 || ( X C3 + X L1 ) Selecting C3 so that XC3 << XL1 yields, X T = X C2 = X C1 || X L1 For a capacitor, –j X C = -------ωC where ω = 2π f For an inductor, X L = jωL 2 Solving for C1, and realizing that L1C1 = 3/ω3 yields, 3 C 1 = --- C 2 2 Hence, for C2 = 10 pF, C1 = 15 pF. Since the impedance of the trap circuit in this example would be equal to the impedance of a 10 pF capacitor, the negative resistance and supply current curves for C1 = C2 = 10 pF at 40 MHz would apply to this example. Finally, solving for the inductor value, 1 L 1 = --------------------------2 2 4π f T C 1 For the above example, L1 is 3.2 µH. 148 Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 13 Crystal Electrical Characteristics and Requirements (continued) 13.4 Frequency Accuracy Considerations For frequency accuracy implications of using the PLL, see Section 4.12, Clock Synthesis. For most applications, clock frequency errors in the hundreds of parts per million can be tolerated with no adverse effects. However, for applications where precise average frequency tolerance on the order of 100 ppm is required, care must be taken in the choice of external components (crystal and capacitors) as well as in the layout of the printed-circuit board. Several factors determine the frequency accuracy of a crystal-based oscillator circuit. Some of these factors are determined by the properties of the crystal itself. Generally, a low-cost, standard crystal will not be sufficient for a high-accuracy application, and a custom crystal must be specified. Most crystal manufacturers provide extensive information concerning the accuracy of their crystals, and an applications engineer from the crystal vendor should be consulted prior to specifying a crystal for a given application. In addition to absolute, temperature, and aging tolerances of a crystal, the operating frequency of a crystal is also determined by the total load capacitance seen by the crystal. When ordering a crystal from a vendor, it is necessary to specify a load capacitance at which the operating frequency of the crystal will be measured. Variations in this load capacitance due to temperature and manufacturing variations will cause variations in the operating frequency of the oscillator. Figure 89 illustrates some of the sources of this variation. CKI CEXT XTAL CD CB CO CKI2 CEXT CB CD CL 5-4045 (F).a Notes: Cext = External load capacitor (one each required for CKI and CKI2). CD = Parasitic capacitance of the DSP1627 itself. CB = Parasitic capacitance of the printed-wiring board. CO = Parasitic capacitance of crystal (not part of CL but still a source of frequency variation). , Figure 89. Components of Load Capacitance for Crystal Oscillator The load capacitance, CL, must be specified to the crystal vendor. The crystal manufacturer will cut the crystal so that the frequency of oscillation will be correct when the crystal sees this load capacitance. Note that CL refers to a capacitance seen across the crystal leads, meaning that for the circuit shown in Figure 89, CL is the series combination of the two external capacitors (Cext/2) plus the equivalent board and device strays (CB/2 + CD/2). For example, if 10 pF external capacitors were used and parasitic capacitance is neglected, then the crystal should be specified for a load capacitance of 5 pF. If the load capacitance deviates from this value due to the tolerance on the external capacitors or the presence of strays, then the frequency will also deviate. This change in frequency as function of load capacitance is known as pullability and is expressed in units of ppm/pF. For small deviations of a few pF, pullability can be determined by the equation below: 6 ( C 1 ) ( 10 ) pullability (ppm/pF) = -------------------------------2 2(C0 + CL) where C0 = parasitic capacitance of crystal in pF. C1 = motional capacitance of crystal in pF (usually between 1 fF to 25 fF, value available from crystal vendor). CL = total load capacitance in pF seen by crystal. Lucent Technologies Inc. 149 Data Sheet March 2000 DSP1627 Digital Signal Processor 13 Crystal Electrical Characteristics and Requirements (continued) Note that for a given crystal, the pullability can be reduced, and, hence, the frequency stability improved, by making CL as large as possible while still maintaining sufficient negative resistance to ensure start-up per the curves shown in Figures 86 and 87. Since it is not possible to know the exact values of the parasitic capacitance in a crystal-based oscillator system, the external capacitors are usually selected empirically to null out the frequency offset on a typical prototype board. Thus, if a crystal is specified to operate with a load capacitance of 10 pF, the external capacitors would have to be made slightly less than 20 pF each in order to account for strays. Suppose, for instance, that a crystal for which CL = 10 pF is specified is plugged into the system and it is determined empirical that the best frequency accuracy occurs with Cext = 18 pF. This would mean that the equivalent board and device strays from each leac to ground would be 2 pF. As an example, suppose it is desired to design a 23 MHz, 3.3 V system with ±100 ppm frequency accuracy. The parameters for a typical high-accuracy, custom, 23 MHz fundamental mode crystal are as follows: Initial Tolerance Temperature Tolerance Aging Tolerance Series Resistance Motional Capacitance (C1) Parasitic Capacitance (C0) 10 ppm 25 ppm 6 ppm 20 Ω max. 15 fF max. 7 pF max. In order to ensure oscillator start-up, the negative resistance of the oscillator with load and parasitic capacitance must be at least twice the series resistance of the crystal, or 40 Ω. Interpolating from Figure 89, external capacitors plus strays can be made as large as 30 pF while still achieving 40 Ω of negative resistance. Assume for this example that external capacitors are chosen so that the total load capacitance including strays is 30 pF per lead, or 15 pF total. Thus, a load capacitance, CL = 15 pF would be specified to the crystal manufacturer. From the above equation, the pullability would be calculated as follows: 6 6 ( C 1 ) ( 10 ) ( 0.015 ) ( 10 ) pullability = -------------------------------- = ---------------------------------- = 15.5 ppm/pF 2 2 2 ( 7 + 15 ) 2( C0 + CL) If 2% external capacitors are used, the frequency deviation due to capacitor tolerance is equal to: (0.02)(15 pF)(15.5 ppm/pF) = 4.7 ppm Note: To simplify analysis, Cext is considered to be 30 pF. In practice, it would be slightly less than this value to account for strays. Also, temperature and aging tolerances on the capacitors have been neglected. Typical capacitance variation of the oscillator circuit in the DSP1627 itself across process, temperature, and supply voltage is ±1 pF. Thus, the expected frequency variation due to the DSP1627 is: (1 pF)(15.5 ppm/pF) = 15.5 ppm Approximate variation in parasitic capacitance of crystal = ±0.5 pF. Frequency shift due to variation in C0 = (0.5 pF)(15.5 ppm/pF) = 7.75 ppm Approximate variation in parasitic capacitance of printed-circuit board = ±1.5 pF. Frequency shift due to variation in board capacitance = (1.5 pF)(15.5 ppm/pF) = 23.25 ppm 150 Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 13 Crystal Electrical Characteristics and Requirements (continued) Thus, the contributions to frequency variation add up as follows: Initial Tolerance of Crystal Temperature Tolerance of Crystal Aging Tolerance of Crystal Load Capacitor Variation DSP1627 Circuit Variation1 C0 Variation Board Variation 10.0 ppm 25.0 6.0 4.7 5.5 7.8 23.3 Total 92.3 ppm This type of detailed analysis should be performed for any crystal-based application where frequency accuracy is critical. Lucent Technologies Inc. 151 Data Sheet March 2000 DSP1627 Digital Signal Processor 14 Outline Diagrams 14.1 100-Pin BQFP (Bumpered Quad Flat Pack) All dimensions are in millimeters. 22.860 ± 0.305 22.350 ± 0.255 19.050 ± 0.405 13 1 89 14 88 PIN #1 IDENTIFIER ZONE EDGE CHAMFER 22.350 ± 0.255 19.050 ± 0.405 38 22.860 ± 0.305 64 39 63 DETAIL A DETAIL B 4.570 MAX 3.555 ± 0.255 SEATING PLANE 0.10 0.760 ± 0.255 0.635 TYP 0.255 0.175 ± 0.025 GAGE PLANE SEATING PLANE 0.280 ± 0.075 0.91/1.17 DETAIL A 0.150 M DETAIL B 5-1970 (F)r.10 152 Lucent Technologies Inc. Data Sheet March 2000 DSP1627 Digital Signal Processor 14 Outline Diagrams (continued) 14.2 100-Pin TQFP (Thin Quad Flat Pack) All dimensions are in millimeters. 16.00 ± 0.20 14.00 ± 0.20 PIN #1 IDENTIFIER ZONE 100 76 1 75 14.00 ± 0.20 16.00 ± 0.20 25 51 26 50 DETAIL A DETAIL B 1.40 ± 0.05 1.60 MAX SEATING PLANE 0.08 0.05/0.15 0.50 TYP 1.00 REF 0.106/0.200 0.25 GAGE PLANE 0.19/0.27 SEATING PLANE 0.45/0.75 DETAIL A Lucent Technologies Inc. 0.08 DETAIL B M 5-2146 (F)r.14 153 For additional information, contact your Microelectronics Group Account Manager or the following: http://www.lucent.com/micro INTERNET: [email protected] E-MAIL: N. AMERICA: Microelectronics Group, Lucent Technologies Inc., 555 Union Boulevard, Room 30L-15P-BA, Allentown, PA 18103 1-800-372-2447, FAX 610-712-4106 (In CANADA: 1-800-553-2448, FAX 610-712-4106) ASIA PACIFIC: Microelectronics Group, Lucent Technologies Singapore Pte. Ltd., 77 Science Park Drive, #03-18 Cintech III, Singapore 118256 Tel. (65) 778 8833, FAX (65) 777 7495 CHINA: Microelectronics Group, Lucent Technologies (China) Co., Ltd., A-F2, 23/F, Zao Fong Universe Building, 1800 Zhong Shan Xi Road, Shanghai 200233 P. R. China Tel. (86) 21 6440 0468, ext. 316, FAX (86) 21 6440 0652 JAPAN: Microelectronics Group, Lucent Technologies Japan Ltd., 7-18, Higashi-Gotanda 2-chome, Shinagawa-ku, Tokyo 141, Japan Tel. (81) 3 5421 1600, FAX (81) 3 5421 1700 EUROPE: Data Requests: MICROELECTRONICS GROUP DATALINE: Tel. (44) 7000 582 368, FAX (44) 1189 328 148 Technical Inquiries:GERMANY: (49) 89 95086 0 (Munich), UNITED KINGDOM: (44) 1344 865 900 (Ascot), FRANCE: (33) 1 40 83 68 00 (Paris), SWEDEN: (46) 8 594 607 00 (Stockholm), FINLAND: (358) 9 4354 2800 (Helsinki), ITALY: (39) 02 6608131 (Milan), SPAIN: (34) 1 807 1441 (Madrid) Lucent Technologies Inc. reserves the right to make changes to the product(s) or information contained herein without notice. No liability is assumed as a result of their use or application. No rights under any patent accompany the sale of any such product(s) or information. Copyright © 2000 Lucent Technologies Inc. All Rights Reserved March 2000 DS00-205WTEC (Replaces DS00-061WTEC)