Rev: 6, 12/04 Ultra-High-Speed Flash Microcontroller User’s Guide 33 16kB FLASH MEMORY 1kB SRAM WITH AUTOSELECT INCREMENT/ DECREMENT HIGH-SPEED ONE CLOCK-CYCLE 8051 MICROPROCESSOR FOUR 8-BIT PARALLEL PORTS MIPS 25 DUAL DATA POINTERS 5 1 DUAL SERIAL PORTS 0 ORIGINAL 8051 ULTRA-HIGH-S PEED FLASH MICROCONTRO LLER The Ultra-High-Speed Flash Microcontroller User’s Guide should be used in conjunction with the data sheets(s) for all ultra-high-speed flash microcontrollers. TABLE OF CONTENTS SECTION 1: INTRODUCTION SECTION 2: ORDERING INFORMATION SECTION 3: ARCHITECTURE ALU Special-Function Registers Accumulator B Register Program Status Word Data Pointer(s) Stack Pointer I/O Ports Timer/Counters UARTs Scratchpad Registers (RAM) Stack Working Registers Program Counter Address/Data Bus Watchdog Timer Power Monitor Interrupts Timing Control Flash Memory 5 5 5 5 6 6 6 6 6 6 6 6 6 6 6 7 7 7 7 7 7 7 7 ________________________________________________________________________ Maxim Integrated Products 1 Ultra-High-Speed Flash Microcontroller User’s Guide SECTION 4: PROGRAMMING MODEL Memory Organization Memory Map Register Map Bit Addressable Locations Working Registers Stack Address Modes Program Status Flags Special-Function Register Locations Special-Function Registers 8 8 8 10 10 10 10 12 13 16 SECTION 5: CPU TIMING Oscillator XTAL1 XTAL2 Oscillator Characteristics Crystal Selection System Clock Divide Control Instruction Timing Nonpage Mode External Timing Page Mode 1 External Timing—Pages Page Mode 1 External Timing—Pages Page Mode 1 External Timing—Pages Page Mode 2 External Timing—Pages Comparison to the 8051 8 42 1:0 1:0 1:0 1:0 = = = = 10b (Four Cycles) 01b (Two Cycles) 00b (One Cycle) 11b SECTION 6: MEMORY ACCESS Internal Flash Memory ROMSize Feature Flash Security Block/Lock Bits Option Control Register Byte Internal SRAM Memory Program Memory Interface—Nonpage Mode Program Memory Interface—Page Modes Data Memory Interface Nonpage Mode Data Memory Timing Page Mode 1 Data Memory Timing—Pages 1:0 = 10b (Four Cycles) Page Mode 1 Data Memory Timing—Pages 1:0 = 01b (Two Cycles) Page Mode 1 Data Memory Timing—Pages 1:0 = 00b (One Cycle) Page Mode 2 Data Memory Timing—Pages 1:0 = 11b (Four Cycles) 42 42 42 42 42 43 44 45 47 50 51 53 53 58 58 58 59 59 60 61 62 65 70 71 72 74 76 ______________________________________________________________________________________________ 2 Ultra-High-Speed Flash Microcontroller User’s Guide SECTION 7: POWER MANAGEMENT 79 Power Management Summary Precision Voltage Monitor Early-Warning Power-Fail Interrupt Power-Fail Reset Power-On Reset Bandgap Select Watchdog Wake-Up from Idle Power Saving Clock Divide Control Idle Mode Stop Mode Ring Oscillator Wake-Up from Stop Speed Reduction Power Management Modes Switchback Clock Source Selection Ring Oscillator Resume from Stop 80 81 81 81 82 82 82 82 82 83 84 84 85 85 87 88 89 SECTION 8: RESET CONDITIONS 89 Reset Sources Power-On/Power-Fail Reset Watchdog Timer Reset Oscillator Fail-Detect Reset Determining the Cause of a Reset SECTION 9: INTERRUPTS Interrupt Interrupt Interrupt Interrupt Interrupt Overview Sources Priorities Acknowledge Cycle Latency SECTION 10: I/O PORTS Parallel I/O Port 0 Port 2 Port 1 Port 3 Alternate Functions of Ports 1 and 3 Read-Modify-Write Output Functions Input Functions 3 ______________________________________________________________________________________________ 89 89 90 90 90 91 91 92 93 94 95 95 96 96 97 97 97 97 99 99 100 Ultra-High-Speed Flash Microcontroller User’s Guide SECTION 11: PROGRAMMABLE TIMERS 16-Bit Timers Timer 0, Timer 1 Modes Mode 0 Mode 1 Mode 2 Mode 3 Timer 2 Modes Watchdog Timer 100 101 101 102 102 103 103 109 SECTION 12: SERIAL I/O Serial Mode Summary Mode 0 Mode 1 Mode 2 Mode 3 Serial Port Initialization Baud Rates Serial I/O Description Framing Error Detection Multiprocessor Communication SECTION 13: TIMED-ACCESS PROTECTION Protected Bits Protection Scheme Valid Timed-Access Procedure Invalid Timed-Access Procedure Timed-Access Protects Watchdog SECTION 14: INSTRUCTION SET DETAILS SECTION 15: PROGRAM LOADING ROM Loader Mode AutoBaud-Rate Detection Command Line Interface Command Line Syntax Command Summaries Error Messages Parallel Programming Mode User Code In-Application Programming Mode REVISION HISTORY 100 111 111 111 112 112 112 112 113 116 122 122 123 123 123 124 124 124 125 131 131 133 133 135 136 137 138 139 139 ______________________________________________________________________________________________ 4 Ultra-High-Speed Flash Microcontroller User’s Guide SECTION 1: INTRODUCTION Dallas Semiconductor’s ultra-high-speed flash microcontroller is an 8051-compatible microcontroller that provides improved performance and power consumption when compared to the original 8051 version. It retains instruction set and object code compatibility with the 8051, yet performs the same operations in fewer clock cycles. Consequently, greater throughput is possible for the same crystal speed. As an alternative, the device can be run at a reduced frequency to save power. The more efficient design allows a much slower crystal speed to get the same results as an original 8051, using much less power. an on-chip watchdog timer, five levels of interrupt priority, and a crystal multiplier. The device provides 256 bytes of RAM for variables and stack; 128 bytes can be reached using direct or indirect addressing, or using indirect addressing only. The fundamental innovation of the ultra-high-speed flash microcontroller is the use of only one clock per instruction cycle compared with 12 for the original 8051. This results in up to 12 times improvement in performance over the original 8051 architecture and up to four times improvement over other Dallas Semiconductor high-speed microcontrollers. The device provides several peripherals and features in addition to all of the standard features of an 80C32. These include 16kB/32kB/64kB of on-chip flash memory, 1kB of on-chip RAM, four 8-bit I/O ports, three 16-bit timer/counters, two on-chip UARTs, dual data pointers, The device incorporates a power-management mode that allows the device to dynamically vary the internal clock speed from 1 clock per cycle (default) to 1024 clocks per cycle. Because power consumption is directly proportional to clock speed, the device can reduce its operating frequency during periods of little switchback. This greatly reduces power consumption. The switchback feature allows the device to quickly return to highest speed operation upon receipt of an interrupt or serial port activity, allowing the device to respond to external events while in power-management mode. In addition to improved efficiency, it can operate at a maximum clock rate of 33MHz. Combined with the 12 times performance, this allows for a maximum performance of 33 million instructions per second (MIPs). This level of computing power is comparable to many 16-bit processors, but without the added expense and complexity if implementing a 16-bit interface. SECTION 2: ORDERING INFORMATION The ultra-high-speed flash microcontroller family follows the part numbering convention shown below. Note that not all combinations of devices may be currently available. Contact a Maxim/Dallas Semiconductor sales office for up-to-date details. DS89C420-QCL SPEED: L 33MHz TEMPERATURE: C 0°C to +70°C N -40°C to +85°C M PDIP Q PLCC E Thin Quad Flat Pack (TQFP) OPERATING VOLTAGE: 0 +5V MEMORY TYPE: 9 Flash PACKAGE: SECTION 3: ARCHITECTURE The architecture is based on the industry-standard 87C52 and executes the standard 8051 instruction set. The core is an accumulator-based architecture using internal registers for data storage and peripheral control. This section provides a brief description of each architecture feature. Details concerning the programming model, instruction set, and register description are provided in Section 4. ALU The ALU is responsible for math functions, comparisons, and general decision making. The ALU is not used explicitly by software. Instruction decoding prepares the ALU automatically and passes it the appropriate data. The ALU primarily uses two special-function registers (SFRs) as the source and destination for all operations. These are the accumulator and B register. The ALU also provides status information in the program status register. The SFRs are described in the following pages. 5 ______________________________________________________________________________________________ Ultra-High-Speed Flash Microcontroller User’s Guide Special-Function Registers All peripherals and operations that are not explicitly controlled by instructions are controlled through SFRs. All SFRs are described in Section 4. The most commonly used registers that are basic to the architecture are also described in the following pages. Accumulator The accumulator is a source and destination for many operations involving math, data movement, and decisions. Although it can be bypassed, most high-speed instructions require the use of the accumulator (A or ACC) as one argument. B Register The B register is used as the second 8-bit argument in multiply and divide operations. When not used for these purposes, the B register can be used as a general-purpose register. Program Status Word The program status word holds a selection of bit flags that include the carry flag, auxiliary carry flag, general-purpose flag, register bank select, overflow flag, and parity flag. Data Pointer(s) The data pointers (DPTR and DPTR1) are used to assign a memory address for the MOVX instructions. This address can point to a data memory location, either on- or off-chip, or a memory-mapped peripheral. When moving data from one memory area to another or from memory to a memory-mapped peripheral, a pointer is needed for both the source and destination. The user can select the active pointer through a dedicated SFR bit (Sel = DPS.0), or can activate an automatic toggling feature for altering the pointer selection (TSL = DPS.5). An additional feature, if selected, provides automatic incrementing or decrementing of the current DPTR. Stack Pointer The stack pointer denotes the register location at the top of the stack, which is the last used value. The user can place the stack anywhere in the scratchpad RAM by setting the stack pointer to the desired location, although the lower bytes are normally used for working registers. I/O Ports Four 8-bit I/O ports are available. Each I/O port is represented by an SFR location, and can be written or read. The I/O port has a latch that contains the value written by software. In general, software reads the state of external pins during a read operation. Timer/Counters Three 16-bit timer/counters are available. Each timer is contained in two SFR locations that can be written or read by software. The timers are controlled by other SFRs described in Section 4. UARTs The two UARTs are controlled and accessed by SFRs. Each UART has an address that is used to read and write the UART. The same address is used for both read and write operations, which are distinguished by the instruction. Each UART is controlled by its own SFR control register. Scratchpad Registers (RAM) The high-speed core provides 256 bytes of scratchpad RAM for general-purpose data and variable storage. The first 128 bytes are directly available to software. The second 128 are available through indirect addressing. Selected portions of this RAM have other optional functions. Stack The stack is a RAM area that stores return address information during calls and interrupts. The user can also place variables on the stack when necessary. The stack pointer designates the RAM location that is the top of the stack. Thus, depending on the value of the stack pointer, the stack can be located anywhere in the 256 bytes of RAM. A common location would be in the upper 128 bytes of RAM, as these locations are accessible through indirect addressing only. ______________________________________________________________________________________________ 6 Ultra-High-Speed Flash Microcontroller User’s Guide Working Registers The first 32 bytes of the scratchpad RAM can be used as four banks of eight working registers for high-speed data movement. Using four banks, software can quickly change context by changing to a different bank. In addition to the accumulator, the working registers are commonly used as data source or destination. Some of the working registers can also be used as pointers to other RAM locations (indirect addressing). Program Counter The program counter (PC) is a 16-bit value that designates the next program address to be fetched. On-chip hardware automatically increments the PC value to move to the next program memory location. Address/Data Bus The device addresses a 64kB program and 64kB data memory area that resides in a combination of internal and external memory. When external memory is accessed, ports 0 and 2 are used as a multiplexed address and data bus. Three external memory bus structures are supported. The nonpage mode (traditional 8051) bus structure provides the address MSB on port 2 and multiplexes port 0 between address LSB and data. The page mode 1 bus structure uses port 0 exclusively for data and multiplexes port 2 between address MSB and address LSB. The page mode 2 bus structure uses port 0 exclusively for address LSB and multiplexes port 2 between address MSB and data. These addressing modes are detailed later. Watchdog Timer The watchdog timer provides a supervisory function for applications that cannot afford to run out of control. The watchdog timer is a programmable, free-running timer. If allowed to reach the termination of its count, if enabled, the watchdog resets the CPU software must prevent this by clearing or resetting the watchdog prior to its timeout. Power Monitor A bandgap reference and analog circuitry are incorporated to monitor the power-supply conditions. When VCC begins to drop out of tolerance, the power monitor issues an optional early warning power-fail interrupt. If power continues to fall, the power monitor invokes a reset condition. This remains until power returns to normal operating voltage. The power monitor also functions on power-up, holding the microcontroller in a reset state until power is stable. Interrupts The device is capable of evaluating 13 interrupt sources simultaneously. Each interrupt has an associated interrupt vector, flag, priority, and enable. These interrupts can be globally enabled or disabled. Timing Control The microcontroller provides an on-chip oscillator for use with an external crystal. This can be bypassed by injecting a clock source into the XTAL1 pin. The clock source is used to create machine cycle timing (four clocks), ALE, PSEN, watchdog, timer, and serial baudrate timing. In addition, an on-chip ring oscillator can be used to provide an approximately 10MHz clock source. A frequency multiplier feature is included, which can be selected by SFR control to multiply the input clock source by either two or four. This allows lower frequency (and cost) crystals to be used while still allowing internal operation up to the full 33MHz limit. Flash Memory On-chip program memory is implemented in flash memory. This can be programmed in-system with the standard 5V VCC supply through a serial port (in-system) using a built-in program memory loader, or by a standard flash or EPROM programmer. Full programming details are given in Section 15. A memory management unit (MMU) and other hardware supports any of the three programming methods. The MMU controls program and data memory access, and provides sequencing and timing controls for programming of the on-chip program memory. There is also a separate security flash block that is used to support a standard three-level lock, a 64-byte encryption array, and other flash options. The full on-chip program memory range can be fetched by the processor automatically. Reset routines and all interrupt vectors are located in the lower 128 bytes of the on-chip program memory area. 7 ______________________________________________________________________________________________ Ultra-High-Speed Flash Microcontroller User’s Guide SECTION 4: PROGRAMMING MODEL This section provides a programmer’s overview of the ultra-high-speed microcontroller core. It includes information on the memory map, on-chip RAM, SFRs, and instruction set. The programming model of the ultra-high-speed microcontroller is very similar to that of the industry-standard 80C52. The memory map is identical. It uses the same instruction set, with improved instruction timing. Several new SFRs have been added. MEMORY ORGANIZATION The ultra-high-speed flash microcontroller, like the 8052, uses several distinct memory areas. These areas include registers, program memory, and data memory. Registers serve to control on-chip peripherals and as RAM. Note that registers (on-chip RAM) are separate from data memory. Registers are divided into three categories including directly addressed on-chip RAM, indirectly addressed onchip RAM, and SFRs. The program and data memory areas are discussed in the Memory Map section. The registers are discussed in the Register Map section. MEMORY MAP The ultra-high-speed microcontroller uses a memory-addressing scheme that separates program memory from data memory. Each area is 64kB beginning at address 0000h and ending at FFFFh, as shown in Figure 4-1. The program and data segments can overlap since they are accessed in different ways. Program memory is fetched by the microcontroller automatically. These addresses are never written by software. In fact, there are no instructions that allow the program area to be written. There is one instruction (MOVC) that is used to explicitly read the program area. This is commonly used to read lookup tables. The data memory area is accessed explicitly using the MOVX instruction. This instruction provides multiple ways of specifying the target address. It is used to access the 64kB of data memory. The address and data range of devices with on-chip program and data memory overlap the 64k memory space. When on-chip memory is enabled, accessing memory in the on-chip range causes the device to access internal memory. Memory accesses beyond the internal range are addressed externally through ports 0 and 2. The ROMSIZE feature allows software to dynamically configure the maximum address of on-chip program memory. This allows the device to act as a bootstrap loader for an external flash or nonvolatile SRAM. Secondly, this method can also be used to increase the amount of available program memory from 64kB to 80kB without bank switching. For more information on this feature, see Section 6. Program and data memory can also be increased beyond the 64kB limit using bank-switching techniques. This is described in Application Note 81: Memory Expansion with the High-Speed Microcontroller Family. REGISTER MAP The register map is illustrated in Figure 4-2. It is entirely separate from the program and data memory areas mentioned above. A separate class of instructions is used to access the registers. There are 256 potential register location values. In practice, the ultra-highspeed microcontroller has 256 bytes of scratchpad RAM and up to 128 SFRs. This is possible since the upper 128 scratchpad RAM locations can only be accessed indirectly. That is, the contents of a working register (R0 or R1) or the stack pointer designates the RAM location. A direct reference to one of the lower 128 addresses (0h-7Fh) accesses the scratchpad RAM. A direct reference to one of the upper 128 addresses (80h-FFh) must be an SFR access. In contrast, indirect references can access the entire scratchpad RAM range (0h-FFh). Scratchpad RAM is available for general-purpose data storage. It is commonly used in place of off-chip RAM when the total data contents are small. When off-chip RAM is needed, the scratchpad area still provides the fastest general-purpose access. Within the 256 bytes of RAM, there are several special purpose areas. These are described as follows: ______________________________________________________________________________________________ 8 Ultra-High-Speed Flash Microcontroller User’s Guide FFFFh FFh 64k PROGRAM MEMORY DATA MEMORY 255 INDIRECT RAM 7Fh 128 DIRECT RAM 0000h 0000h Figure 4-2. Register Map Figure 4-1. Memory Map FFh INDIRECT RAM 7Fh DIRECT RAM 2Fh 2Eh 2Dh 2Ch 2Bh 2Ah 29h 28h 27h 26h 25h 24h 23h 22h 21h 20h 1Fh 7F 77 6F 67 5F 57 4F 47 3F 37 2F 27 1F 17 0F 07 7E 76 6E 66 5E 56 4E 46 3E 36 2E 26 1E 16 0E 06 7D 75 6D 65 5D 55 4D 45 3D 35 2D 25 1D 15 0D 05 7C 74 6C 64 5C 54 4C 44 3C 34 2C 24 1C 14 0C 04 7B 73 6B 63 5B 53 4B 43 3B 333 2B 23 1B 13 0B 03 7A 72 6A 62 5A 52 4A 42 3A 32 2A 22 1A 12 0A 02 79 71 69 61 59 51 49 41 39 31 29 21 19 11 09 01 78 70 68 60 58 50 48 40 38 30 28 20 18 10 08 00 BANK 3 18h 17h BANK 2 10h 0Fh BANK 1 08h 07h BANK 0 00h MSB LSB Figure 4-3. Scratchpad Register Addressing 9 ______________________________________________________________________________________________ DIRECT SPECIAL FUNCTION REGISTERS FFh 7Fh Ultra-High-Speed Flash Microcontroller User’s Guide BIT ADDRESSABLE LOCATIONS In addition to direct register access, some individual bits are also accessible. These are individually addressable bits in both the RAM and SFR area. In the scratchpad RAM area, registers 20h to 2Fh are bit addressable. This provides 128 (16 x 8) individual bits available to software. A bit access is distinguished from a full register access by the type of instruction. Addressing modes are discussed later in this section. In the SFR area, any register location ending in a 0 or 8 is bit addressable. Figure 4-3 shows details of the on-chip RAM addressing, including the locations of individual RAM bits. WORKING REGISTERS As part of the lower 128 bytes of RAM, there are four banks of working registers (8 bytes each). The working registers are general-purpose RAM locations that can be addressed in a special way. They are designated R0 through R7. Since there are four banks, the currently selected bank is used by any instruction using R0–R7. This allows software to change context by simply switching banks. This is controlled through the program status word register in the next SFR area. The working registers also allow their contents to be used for indirect addressing of the upper 128 bytes of RAM. Thus, an instruction can designate the value stored in R0 (for example) to address the upper RAM. This value might be the result of another calculation. STACK Another use of the scratchpad area is for the programmer’s stack. This area is selected using the stack pointer (SP;81h) SFR. Whenever a call or interrupt is invoked, the return address is placed on the stack. It also is available to the programmer for variables, etc. Since the stack can be moved, there is no fixed location within the RAM designated as stack. The stack pointer defaults to 07h upon reset. The user can then move it as needed. A convenient location would be the upper RAM area (>7Fh), since this is only available indirectly. The SP points to the last used value. Therefore, the next value placed on the stack is put at SP + 1. Each PUSH or CALL increments the SP by the appropriate value. Each POP or RET decrements as well. ADDRESS MODES The DS89C420 uses the standard 8051 instruction set that is supported by a wide range of third-party assemblers and compilers. Like the 8051, the DS89C420 uses three memory areas. These are program memory, data memory, and registers. The program and data areas are 64kB each. They extend from 0000h to FFFFh. The register areas are located between 00h and FFh, but do not overlap with the program and data segments. This is because the ultra-high-speed flash microcontroller uses different modes of addressing to reach each memory segment. These modes are described below. Program memory is the area from which all instructions are fetched. It is inherently read only. This is because the 8051 instruction set provides no instructions that write to this area. Read/write access is for data memory and registers only. No special action is required to fetch from program memory. Each instruction fetch is performed automatically by the on-chip CPU. In versions that contain on-chip memory, the hardware decides whether the fetch is on-chip or off-chip based on the address. Explicit addressing modes are needed for the data memory and register areas. These modes determine which register area is accessed or if off-chip data memory is used. The ultra-high-speed microcontroller supports eight addressing modes: • Register addressing • Direct addressing • Register indirect addressing • Immediate addressing • Register indirect addressing with displacement • Relative addressing • Page addressing • Extended addressing Five of the eight addressing modes are used to address operands. The remainder are used for program control and branching. When writing assembly language instructions that use arguments, the convention is “destination, source.” Each mode of addressing is summarized on the following pages. Note that many instructions (such as ADD) have multiple-addressing modes available. _____________________________________________________________________________________________ 10 Ultra-High-Speed Flash Microcontroller User’s Guide Register Addressing Register addressing is used for operands that are located in one of the eight working registers (R7–R0). The eight working registers can be located in one of four working register banks found in the lower 32 bytes of scratchpad RAM, as determined by the current register bank-select bits. A register bank is selected using two bits in the program status word (PSW;D0h). This addressing mode is powerful, since it uses the active bank without knowing which bank is selected. Thus, one instruction can have multiple uses by simply switching banks. Register addressing is also a high-speed instruction, requiring only one machine cycle. Two examples of register addressing are provided below: ADD A, R4 ;Add register R4 to Accumulator INC R2 ;Increment the value in register R2 In the first case, the value in R4 is the source of the operation. In the latter, R2 is the destination. These instructions do not consider the absolute address of the register. They act on whichever bank has been selected. Any working register can also be accessed by direct addressing. In order to do this, the absolute address must be specified. Direct Addressing Direct addressing is the mode used to access the entire lower 128 bytes of scratchpad RAM and the SFR area. It is commonly used to move the value in one register to another. Two examples are shown below: MOV 72h, 74h ;Move the value in register 74 to ;register 72. MOV 90h, 20h ;Move the value in register 20 to ;the SFR at 90h (Port 1) Note that there is no instruction difference between a RAM access and an SFR access. The SFRs are register locations above 7Fh. Direct addressing also extends to bit addressing. There is a group of instructions that explicitly use bits. The address information provided to such an instruction is the bit location, rather than the register address. Registers between 20h and 2Fh contain bits that are individually addressable. SFRs that end in 0 or 8 are bit addressable. An example of direct bit addressing is as follows: SETB 00h MOV C, 0B7h ;Set bit 00 in the RAM. This is the ;LSb of the register at address 20h ;as shown earlier in this section. ;Move the contents of bit B7 to the ;Carry flag. Bit B7 is the MSb of ;register B0 (Port 3). Register Indirect Addressing This mode is used to access the scratchpad RAM locations above 7Fh. It can also be used to reach the lower RAM (0h–7Fh), if needed. The address is supplied by the contents of the working register specified in the instruction. Thus, one instruction can be used to reach many values by altering the contents of the designated working register. Note that, in general, only R0 and R1 can be used as pointers. An example of register indirect addressing follows: ANL A, @R0 ;Logical AND the Accumulator ;with the contents of the register ;pointed to by the value stored in R0. This mode is also used for stack manipulation. This is because all stack references are directed by the value in the stack pointer register. The push and pop instructions use this method of addressing. An example is as follows: ;Saves the contents of the ;accumulator on the stack. Register indirect addressing is used for all off-chip data memory accesses. These involve the MOVX instruction. The pointer registers can be R0, R1, DPTR0 and DPTR1. Both R0 and R1 reside in the working register area of the scratchpad RAM. They can be used to reference a 256-byte area of off-chip data memory. When using this type of addressing, the upper address byte is supplied by the value in the port 2 latch. This value must be selected by software prior to the MOVX instruction. An example is as follows: 11 PUSH A MOVX @R0, A ;Write the value in the accumulator ;to the address pointed to by R0 in ;the page pointed to by P2. _____________________________________________________________________________________________ Ultra-High-Speed Flash Microcontroller User’s Guide The 16-bit data pointers (DPTRs) can be used as an absolute off-chip reference. This gives access to the entire 64kB data memory map. An example is as follows: MOVX @DPTR, A ;Write the value in the accumulator ;to the address referenced by the ;selected data pointer. Immediate Addressing Immediate addressing is used when one of the operands is predetermined and coded into the software. This mode is commonly used to initialize SFRs and to mask particular bits without affecting others. An example is as follows: ORL A, #40h ;Logical OR the Accumulator with 40h. Register Indirect with Displacement Register indirect addressing with displacement is used to access data in lookup tables in program memory space. The location is created using a base address with an index. The base address can be either the PC or the DPTR. The index is the accumulator. The result is stored in the accumulator. An example is as follows: MOVC A, @A +DPTR ;Load the accumulator with the contents of program memory ;pointed to by the contents of the DPTR plus the value in ;the accumulator. Relative Addressing Relative addressing is used to determine a destination address for the conditional branch. Each of these instructions includes an 8-bit value that contains a two’s complement address offset (-127 to +128), which is added to the PC to determine the destination address. This destination is branched to when the tested condition is true. The PC points to the program memory location immediately following the branch instruction when the offset is added. If the tested condition is not true, the next instruction is performed. An example is as follows: JZ $–20 ;Branch to the location (PC+2)–20 ;if the contents of the accumulator = 0. Page Addressing Page addressing is used by the branching instructions to specify a destination address within the same 2kB block as the next contiguous instruction. The full 16-bit address is calculated by taking the five highest-order bits for the next instruction (PC + 2) and concatenating them with the lowest order 11-bit field contained in the current instruction. An example is as follows: 0870h ACALL 100h ;Call to the subroutine at address 100h plus the ;current page address. In this example, the current page address is 800h, so the destination address is 900h. Extended Addressing Extended addressing is used by the branching instructions to specify a 16-bit destination address within the 64kB address space. The destination address is fixed in the software as an absolute value. An example is as follows: LJMP 0F732h ;Jump to address 0F732h. Program Status Flags All program status flags are contained in the program status word at SFR location D0h. It contains flags that reflect the status of the CPU and the result of selected operations. The flags are summarized below. The following table shows the instructions that affect each flag. Bit Description*: PSW.7 C Carry Set when the previous operation resulted in a carry (during addition) or a borrow (during subtraction). Otherwise cleared. _____________________________________________________________________________________________ 12 Ultra-High-Speed Flash Microcontroller User’s Guide PSW.6 AC Auxiliary Carry Set when the previous operation resulted in a carry (during addition) or a borrow (during subtraction) from the high-order nibble. Otherwise cleared. PSW.2 OV Overflow For addition, OV is set when a carry is generated into a high order bit (bit 6 or bit 7), but not a carry out of the same high-order bit. For subtraction, OV is set if a borrow is needed into a high order bit (bit 6 or bit 7), but not into the other high-order bit. For multiplication, OV is set when the product exceeds FFh. For division, OV is always cleared. PSW.0 P Parity Set to logic 1 to indicate an odd number of ones in the accumulator (odd parity). Cleared for an even number of ones. This produces even parity. *All of these bits are cleared to a logic 0 for all resets. Table 4-1. Instructions that Affect Flag Settings INSTRUCTION ADD ADDC SUBB MUL DIV DA RRC RLC SETB C C X X X 0 0 X X X 1 FLAGS OV X X X X X INSTRUCTION AC X X X CLR C CPL C ANL C, bit ANL C, bit ORL C, bit ORL C, bit MOV C, bit CJNE C 0 X X X X X X X — — FLAGS OV AC Note: X indicates the modification is according to the result of the instruction. SPECIAL-FUNCTION REGISTER LOCATIONS The ultra-high-speed flash microcontroller, like the 8051, uses SFRs to control peripherals and modes. In many cases, an SFR controls individual functions or report status on individual functions. The SFRs reside in register locations 80h–FFh and are reached using direct addressing. SFRs that end in 0 or 8 are bit addressable. All standard SFR locations from the original 8051 are duplicated, with several additions. Tables are provided to illustrate the locations of the SFRs and the default reset conditions of all SFR bits. Detailed descriptions of each SFR follow. 13 _____________________________________________________________________________________________ Ultra-High-Speed Flash Microcontroller User’s Guide Special-Function Register Locations REGISTER P0 SP DPL DPH DPL1 DPH1 DPS PCON TCON TMOD TL0 TL1 TH0 TH1 CKCON P1 EXIF CKMOD SCON0 SBUF0 ACON P2 IE SADDR0 SADDR1 P3 IP1 IP0 SADEN0 SADEN1 SCON1 SBUF1 ROMSIZE PMR STATUS TA T2CON T2MOD RCAP2L RCAP2H TL2 TH2 PSW FCNTL FDATA WDCON ACC EIE B EIP1 EIP0 ADDRESS 80h 81h 82h 83h 84h 85h 86h 87h 88h 89h 8Ah 8Bh 8Ch 8Dh 8Eh 90h 91h 96h 98h 99h 9Dh A0h A8h A9h AAh B0h B1h B8h B9h BAh C0h C1h C2h C4h C5h C7h C8h C9h CAh CBh CCh CDh D0h D5h D6h D8h E0h E8h F0h F1h F8h BIT 7 P0.7 BIT 6 P0.6 BIT 5 P0.5 BIT 4 P0.4 BIT 3 P0.3 BIT 2 P0.2 BIT 1 P0.1 BIT 0 P0.0 ID1 SMOD_0 TF1 GATE ID0 SMOD0 TR1 C/T TSL OFDF TF0 M1 AID OFDE TR0 M0 — GF1 IE1 GATE — GF0 IT1 C/T — STOP IE0 M1 SEL IDLE IT0 M0 WD1 P1.7 IE5 WD0 P1.6 IE4 T0M P1.3 CKRY T0MH TB8_0 MD1 P1.1 RGSL MD0 P1.0 BGS SM1_0 T1M P1.4 IE2 T1MH REN_0 MD2 P1.2 RGMD SM0/FE_0 T2M P1.5 IE3 T2MH SM2_0 RB8_0 TI_0 RI_0 PAGEE P2.7 EA PAGES1 P2.6 ES1 PAGES0 P2.5 ET2 P2.4 ES0 P2.3 ET1 P2.2 EX1 P2.1 ET0 P2.0 EX0 P3.7 — — P3.6 MPS1 LPS1 P3.5 MPT2 LPT2 P3.4 MPS0 LPS0 P3.3 MPT1 LPT1 P3.2 MPX1 LPX1 P3.1 MPT0 LPT0 P3.0 MPX0 LPX0 SM0/FE_1 SM1_1 SM2_1 REN_1 TB8_1 RB8_1 TI_1 RI_1 CD1 PIS2 CD0 PIS1 SWB PIS0 CTM — PRAME 4X/2X SPTA1 RMS2 ALEON SPRA1 RMS1 DME1 SPTA0 RMS0 DME0 SPRA0 TF2 — EXF2 — RCLK — TCLK — EXEN2 — TR2 — C/T2 T2OE CP/RL2 DCEN CY FBUSY AC FERR F0 RS1 RS0 FC3 OV FC2 F1 FC1 P FC0 SMOD_1 POR EPFI PFI WDIF WTRF EWT RWT — — — EWDI EX5 EX4 EX3 EX2 — — — MPWDI LPWDI MPX5 LPX5 MPX4 LPX4 MPX3 LPX3 MPX2 LPX2 Note: Shaded bits are timed-access protected. _____________________________________________________________________________________________ 14 Ultra-High-Speed Flash Microcontroller User’s Guide Special-Function Register Reset Values REGISTER P0 SP DPL DPH DPL1 DPH1 DPS PCON TCON TMOD TL0 TL1 TH0 TH1 CKCON P1 EXIF CKMOD SCON0 SBUF0 ACON P2 IE SADDR0 SADDR1 P3 IP1 IP0 SADEN0 SADEN1 SCON1 SBUF1 ROMSIZE PMR STATUS TA T2CON T2MOD RCAP2L RCAP2H TL2 TH2 PSW FCNTL FDATA WDCON ACC EIE B EIP1 EIP0 15 ADDRESS 80h 81h 82h 83h 84h 85h 86h 87h 88h 89h 8Ah 8Bh 8Ch 8Dh 8Eh 90h 91h 96h 98h 99h 9Dh A0h A8h A9h AAh B0h B1h B8h B9h BAh C0h C1h C2h C4h C5h C7h C8h C9h CAh CBh CCh CDh D0h D5h D6h D8h E0h E8h F0h F1h F8h BIT 7 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 0 0 0 1 0 0 0 1 1 1 0 0 0 0 1 1 0 1 0 1 0 0 0 0 0 1 0 0 0 1 0 1 1 BIT 6 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 0 0 0 1 0 0 0 1 0 0 0 0 0 0 1 0 0 1 0 1 0 0 0 0 0 0 0 Special 0 1 0 1 1 BIT 5 1 0 0 0 0 0 0 Special 0 0 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 1 0 0 0 0 0 0 1 0 0 1 0 1 0 0 0 0 0 1 0 0 0 1 0 1 1 BIT 4 1 0 0 0 0 0 0 Special 0 0 0 0 0 0 0 1 0 0 0 0 1 1 0 0 0 1 0 0 0 0 0 0 1 0 1 1 0 1 0 0 0 0 0 1 0 Special 0 0 0 0 0 BIT 3 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 Special 0 0 0 1 1 0 0 0 1 0 0 0 0 0 0 0 0 0 1 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 BIT 2 1 1 0 0 0 0 1 0 0 0 0 0 0 0 0 1 Special 1 0 0 1 1 0 0 0 1 0 0 0 0 0 0 1 0 0 1 0 1 0 0 0 0 0 0 0 Special 0 0 0 0 0 _____________________________________________________________________________________________ BIT 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 1 Special 1 0 0 1 1 0 0 0 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 Special 0 0 0 0 0 BIT 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 1 0 0 1 1 0 0 0 1 0 0 0 0 0 0 1 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Ultra-High-Speed Flash Microcontroller User’s Guide SPECIAL-FUNCTION REGISTERS Most of the unique features of the ultra-high-speed microcontroller family are controlled by bits in SFRs located in unused locations in the 8051 SFR map. This allows for increased functionality while maintaining complete instruction set compatibility. The description for each bit indicates its read and write access, as well as its state after a power-on reset. Port 0 (P0) SFR 80h 7 P0.7 RW-1 6 P0.6 RW-1 5 P0.5 RW-1 4 P0.4 RW-1 3 P0.3 RW-1 2 P0.2 RW-1 1 P0.1 RW-1 0 P0.0 RW-1 R = Unrestricted read, W = Unrestricted write, -n = Value after reset P0.7–0 Port 0. This port functions according to the table below where PAGEE = ACON.7 and PAGES = ACON.6-5. Port 0 Function PAGEE 0 0 1 1 PAGES XX XX 00, 01, 10 11 PORT 0 FUNCTION General-Purpose I/O (code execution < ROMSIZE.2-0) Multiplexed Address LSB / Data (code execution > ROMSIZE.2-0) Data Address LSB When serving as general-purpose I/O (GPIO), the port is open-drain and requires pullups. Writing a 1 to one of the bits of this register configures the associated port 0 pin as an input. All read operations, with the exception of read-modify-write instructions, leave the port latch unchanged. During external memory addressing and data memory write cycles, the port has high-and-low drive capability. During external memory data read cycles, the port is held in a high-impedance state. Stack Pointer (SP) SFR 81h 7 SP.7 RW-0 6 SP.6 RW-0 5 SP.5 RW-0 4 SP.4 RW-0 3 SP.3 RW-0 2 SP.2 RW-1 1 SP.1 RW-1 0 SP.0 RW-1 R = Unrestricted read, W = Unrestricted write, -n = Value after reset Stack Pointer. This stack pointer is written by software to identify the location where the stack begins. The stack pointer is incremented before every PUSH operation and is decremented following every POP operation. This register defaults to 07h after reset. SP.7–0 Bits 7–0 Data Pointer Low 0 (DPL) SFR 82h 7 DPL.7 RW-0 6 DPL.6 RW-0 5 DPL.5 RW-0 4 DPL.4 RW-0 3 DPL.3 RW-0 2 DPL.2 RW-0 1 DPL.1 RW-0 0 DPL.0 RW-0 R = Unrestricted read, W = Unrestricted write, -n = Value after reset Data Pointer LOW 0. This register is the low byte of the standard 80C32 16-bit data pointer. DPL and DPH are used to point to nonscratchpad data RAM. DPL.7–0 Bits 7–0 Data Pointer High 0 (DPH) SFR 83h 7 DPH.7 RW-0 6 DPH.6 RW-0 5 DPH.5 RW-0 4 DPH.4 RW-0 3 DPH.3 RW-0 2 DPH.2 RW-0 1 DPH.1 RW-0 0 DPH.0 RW-0 R = Unrestricted read, W = Unrestricted write, -n = Value after reset DPH.7–0 Bits 7–0 Data Pointer High 0. This register is the high byte of the standard 80C32 16-bit data pointer. DPL and DPH are used to point to nonscratchpad data RAM. _____________________________________________________________________________________________ 16 Ultra-High-Speed Flash Microcontroller User’s Guide Data Pointer Low 1 (DPL1) SFR 84h 7 DPL1.7 RW-0 6 DPL1.6 RW-0 5 DPL1.5 RW-0 4 DPL1.4 RW-0 3 DPL1.3 RW-0 2 DPL1.2 RW-0 1 DPL1.1 RW-0 0 DL1H.0 RW-0 R = Unrestricted read, W = Unrestricted write, -n = Value after reset DPL1.7–0 Bits 7–0 Data Pointer Low 1. This register is the low byte of the auxiliary 16-bit data pointer. When the SEL bit (DPS.0) is set, DPL1 and DPH1 are used in place of DPL and DPH during DPTR operations. Data Pointer High 1 (DPH1) SFR 85h 7 DPH1.7 RW-0 6 DPH1.6 RW-0 5 DPH1.5 RW-0 4 DPH1.4 RW-0 3 DPH1.3 RW-0 2 DPH1.2 RW-0 1 DPH1.1 RW-0 0 DPH1.0 RW-0 R = Unrestricted read, W = Unrestricted write, -n = Value after reset DPH1.7–0 Bits 7–0 Data Pointer Low 1. This register is the high byte of the auxiliary 16-bit data pointer. When the SEL bit (DPS.0) is set, DPL1 and DPH1 are used in place of DPL and DPH during DPTR operations. Data Pointer Select (DPS) SFR 86h 7 ID1 RW-0 6 ID0 RW-0 5 TSL RW-0 4 AID R-0 3 — R-0 2 — R-1 1 — R-0 0 SEL RW-0 R = Unrestricted read, W = Unrestricted write, -n = Value after reset ID1 Bit 7 Increment/Decrement Select for DPTR1. This bit determines the effect of the INC DPTR instruction on DPTR1 when selected (SEL = 1) as the active data pointer. 0 = INC DPTR increments DPTR1 (default) 1 = INC DPTR decrements DPTR1 ID0 Bit 6 Increment/Decrement Select for DPTR. This bit determines the effect of the INC DPTR instruction on DPTR when selected (SEL = 0) as the active data pointer. 0 = INC DPTR increments DPTR (default) 1 = INC DPTR decrements DPTR TSL Bit 5 Toggle Select. When clear (= 0), DPTR-related instructions do not affect the SEL bit. When set (= 1), the SEL bit is toggled following execution of any of the below DPTR-related instructions: INC DPTR MOV DPTR, #data16 MOVC A, @A+DPTR MOVX A, @DPTR MOVX @DPTR, A AID Bit 4 Autoincrement/Decrement Enable. When set, the active data pointer is automatically incremented or decremented (as determined by ID1, ID0 bit settings) following execution of any of the below DPTR-related instructions: MOVC A, @A+DPTR MOVX A, @DPTR MOVX @DPTR, A Bits 3, 2, 1 17 Reserved. These bits read 010b. _____________________________________________________________________________________________ Ultra-High-Speed Flash Microcontroller User’s Guide SEL Bit 0 Data Pointer Select. This bit selects the active data pointer. 0 = Instructions that use the DPTR use DPL and DPH. 1 = Instructions that use the DPTR use DPL1 and DPH1. Power Control (PCON) SFR 87h 7 SMOD_0 RW-0 6 SMOD0 RW-0 5 OFDF RW-0* 4 OFDE RW-0* 3 GF1 RW-0 2 GF0 RW-0 1 STOP RW-0 0 IDLE RW-0 R = Unrestricted read, W = Unrestricted write, -n = Value after reset, * = See description SMOD_0 Bit 7 Serial Port 0 Baud Rate Doubler Enable. This bit enables/disables the serial baud rate doubling function for Serial Port 0. 0 = Serial Port 0 baud rate is that defined by baud rate generation equation. 1 = Serial Port 0 baud rate is double that defined by baud rate generation equation. SMOD0 Bit 6 Framing Error Detection Enable. When clear (= 0), SCON1.7 and SCON0.7 serve as mode select bit SM0 for the respective serial ports. When set (= 1), SCON1.7 and SCON0.7 report whether a Framing Error has been detected. OFDF Bit 5 Oscillator Fail Detect Flag. When OFDE = 1, this flag will be set if a reset condition is generated due to oscillator failure. This bit is cleared on a power-on reset and is unchanged by other reset sources. This bit must be cleared by software. OFDE Bit 4 Oscillator Fail Detect Enable. When set (= 1), the oscillator fail detect circuitry and flag generation are enabled. An oscillator fail detection occurs if the crystal oscillator falls below ~20kHz. An oscillator fail detection does not occur if the oscillator is halted through software setting of the STOP bit (PCON.1) or when running from the internal ring oscillator source. When clear (= 0), the oscillator fail detect circuitry is disabled. General-Purpose User Flag 1. This is a general-purpose flag for software control. GF1 Bit 3 GF0 Bit 2 General-Purpose User Flag 0. This is a general-purpose flag for software control. STOP Bit 1 Stop Mode Select. Setting this bit stops program execution, halts the CPU oscillator and internal timers, and places the CPU in a low-power mode. This bit always be reads as a 0. Setting this bit causes the CTM bit (PMR.4) to be cleared. Setting both the STOP bit and the IDLE bit causes the device to enter stop mode; however, doing this is not advised. IDLE Bit 0 Idle Mode Select. Setting this bit stops program execution but leaves the CPU oscillator, timers, serial ports, and interrupts active. This bit is always read as a 0. Setting both the STOP bit and the IDLE bit causes the device to enter stop mode; however, doing this is not advised. _____________________________________________________________________________________________ 18 Ultra-High-Speed Flash Microcontroller User’s Guide Timer/Counter Control (TCON) SFR 88h 7 TF1 RW-0 6 TR1 RW-0 5 TF0 RW-0 4 TR0 RW-0 3 IE1 RW-0 2 IT1 RW-0 1 IE0 RW-0 0 IT0 RW-0 R = Unrestricted read, W = Unrestricted write, -n = Value after reset TF1 Bit 7 Timer 1 Overflow Flag. This bit indicates when Timer 1 overflows its maximum count as defined by the current mode. This bit can be cleared by software and is automatically cleared when the CPU vectors to the Timer 1 interrupt service routine. 0 = No Timer 1 overflow has been detected. 1 = Timer 1 has overflowed its maximum count. TR1 Bit 6 Timer 1 Run Control. This bit enables/disables the operation of Timer 1. 0 = Timer 1 is halted. 1 = Timer 1 is enabled. TF0 Bit 5 Timer 0 Overflow Flag. This bit indicates when Timer 0 overflows its maximum count as defined by the current mode. This bit can be cleared by software and is automatically cleared when the CPU vectors to the Timer 0 interrupt service routine or by software. 0 = No Timer 0 overflow has been detected. 1 = Timer 0 has overflowed its maximum count. TR0 Bit 4 Timer 0 Run Control. This bit enables/disables the operation of Timer 0. 0 = Timer 0 is halted. IE1 Bit 3 Interrupt 1 Edge Detect. This bit is set when an edge/level of the type defined by IT1 is detected. If IT1 = 1, this bit remains set until cleared in software or until the start of the External Interrupt 1 service routine. If IT1 = 0, this bit inversely reflects the state of the INT1 pin. Interrupt 1 Type Select. This bit selects whether the INT1 pin detects edge- or level-triggered interrupts. 1 = Timer 0 is enabled. IT1 Bit 2 0 = INT1 is level triggered. 1 = INT1 is edge triggered. IE0 Bit 1 IT0 Bit 0 Interrupt 0 Edge Detect. This bit is set when an edge/level of the type defined by IT0 is detected. If IT0 = 1, this bit remains set until cleared in software or until the start of the External Interrupt 0 service routine. If IT0 = 0, this bit inversely reflects the state of the INT0 pin. Interrupt 0 Type Select. This bit selects whether the INT0 pin detects edge- or level-triggered interrupts. 0 = INT0 is level triggered. 1 = INT0 is edge triggered. 19 _____________________________________________________________________________________________ Ultra-High-Speed Flash Microcontroller User’s Guide Timer Mode Control (TMOD) SFR 89h 7 GATE RW-0 6 C/T RW-0 5 M1 RW-0 4 M0 RW-0 3 GATE RW-0 2 C/T RW-0 1 M1 RW-0 0 M0 RW-0 R = Unrestricted read, W = Unrestricted write, -n = Value after reset GATE Bit 7 Timer 1 Gate Control. This bit enables/disables the ability of Timer 1 to increment. 0 = Timer 1 clocks when TR1 = 1, regardless of the state of INT. 1 = Timer 1 clocks only when TR1 = 1 and INT1 = 1. C/T Timer 1 Counter/Timer Select. Bit 6 0 = Timer 1 is incremented by internal clocks. 1 = Timer 1 is incremented by pulses on T1 when TR1 (TCON.6) is 1. M1, M0 Bits 5, 4 Timer 1 Mode Select. These bits select the operating mode of Timer 1. Timer 1 Mode Selection M1 0 0 1 1 GATE Bit 3 M0 0 1 0 1 MODE Mode 0: 8 bits with 5-bit prescale Mode 1: 16 bits Mode 2: 8 bits with autoreload Mode 3: Timer 1 is halted, but holds its count Timer 0 Gate Control. This bit enables/disables the ability of Timer 0 to increment. 0 = Timer 0 clocks when TR0 = 1, regardless of the state of INT0. 1 = Timer 0 clocks only when TR0 = 1 and INT0 = 1. C/T Bit 2 Timer 0 Counter/Timer Select. 0 = Timer incremented by internal clocks. 1 = Timer 1 is incremented by pulses on T0 when TR0 (TCON.4) is 1. M1, M0 Bits 1, 0 Timer 0 Mode Select. These bits select the operating mode of Timer 0. When Timer 0 is in mode 3, TL0 is started/stopped by TR0 and TH0 is started/stopped by TR1. Run control from Timer 1 is then provided by the Timer 1 mode selection. Timer 0 Mode Selection M1 0 0 1 1 M0 0 1 0 1 MODE Mode 0: 8 bits with 5-bit prescale Mode 1: 16 bits Mode 2: 8 bits with autoreload Mode 3: Timer 0 is two 8-bit counters _____________________________________________________________________________________________ 20 Ultra-High-Speed Flash Microcontroller User’s Guide Timer 0 LSB (TL0) SFR 8Ah 7 TL0.7 RW-0 6 TL0.6 RW-0 5 TL0.5 RW-0 4 TL0.4 RW-0 3 TL0.3 RW-0 2 TL0.2 RW-0 1 TL0.1 RW-0 0 TL0.0 RW-0 R = Unrestricted read, W = Unrestricted write, -n = Value after reset TL0.7–0 Bits 7–0 Timer 0 LSB. This register contains the least significant byte of Timer 0. Timer 1 LSB (TL1) SFR 8Bh 7 TL1.7 RW-0 6 TL1.6 RW-0 5 TL1.5 RW-0 4 TL1.4 RW-0 3 TL1.3 RW-0 2 TL1.2 RW-0 1 TL1.1 RW-0 0 TL1.0 RW-0 R = Unrestricted read, W = Unrestricted write, -n = Value after reset TL1.7–0 Bits 7–0 Timer 1 LSB. This register contains the least significant byte of Timer 1. Timer 0 MSB (TH0) SFR 8Ch 7 TH0.7 RW-0 6 TH0.6 RW-0 5 TH0.5 RW-0 4 TH0.4 RW-0 3 TH0.3 RW-0 2 TH0.2 RW-0 1 TH0.1 RW-0 0 TH0.0 RW-0 R = Unrestricted read, W = Unrestricted write, -n = Value after reset TH0.7–0 Bits 7–0 Timer 0 MSB. This register contains the most significant byte of Timer 0. Timer 1 MSB (TH1) SFR 8Dh 7 TH1.7 RW-0 6 TH1.6 RW-0 5 TH1.5 RW-0 4 TH1.4 RW-0 3 TH1.3 RW-0 2 TH1.2 RW-0 1 TH1.1 RW-0 0 TH1.0 RW-0 R = Unrestricted read, W = Unrestricted write, -n = Value after reset TH1.7–0 Bits 7–0 Timer 1 MSB. This register contains the most significant byte of Timer 1. Clock Control (CKCON) SFR 8Eh 7 WD1 RW-0 6 WD0 RW-0 5 T2M RW-0 4 T1M RW-0 3 T0M RW-0 2 MD2 RW-0 1 MD1 RW-0 0 MD0 RW-1 R = Unrestricted read, W = Unrestricted write, -n = Value after reset WD1, WD0 Bits 7, 6 21 Watchdog Timer Mode Select 1-0. These bits determine the watchdog timer timeout period for the watchdog timer. The timer divides the crystal (or external oscillator) frequency by a programmable value as shown on the next page. The divider value is expressed in crystal (oscillator) cycles. The settings of the system clock control bits 4X/2X (PMR.3) and CD1:0 (PMR.7-6) affect the clock input to the watchdog timer and therefore its timeout period as shown below. All watchdog timer reset timeouts follow the setting of the interrupt flag by 512 system clocks. _____________________________________________________________________________________________ Ultra-High-Speed Flash Microcontroller User’s Guide Watchdog Interrupt Flag Timeout Periods (in Oscillator Clocks) 4X / 2X 1 0 X X X CD1:0 00 00 01 10 11 WD1:0 = 00 215 216 217 217 227 WD1:0 = 01 218 219 220 220 230 WD1:0 = 10 221 222 223 223 233 WD1:0 = 11 224 225 226 226 236 T2M Bit 5 Timer 2 Clock Select. This bit controls the input clock that drives Timer 2. This bit has no effect when the timer is in baud rate generator or clock output modes. See Timer Operation table. T1M Bit 4 Timer 1 Clock Select. This bit controls the input clock that drives Timer 1. See Timer Operation table. T0M Bit 3 Timer 0 Clock Select. This bit controls the input clock that drives Timer 0. See Timer Operation table. Timer Operation (in Oscillator Clocks) 4X/2X CD1:0 MD2, MD1, MD0 Bits 2, 1, 0 OSCILLATOR CLOCKS PER TIMER (0, 1, 2) CLOCK TxMH, TxM = OSCILLATOR CLOCKS PER TIMER 2 CLOCK (BAUD RATE GEN) T2MH, T2M = 00 01 1X 1 00 12 1 0.25 XX 2 0 00 12 2 0.5 2 X 01 12 4 1 2 X 10 12 4 1 2 X 11 3072 1024 1024 2048 Stretch MOVX Select 2-0. These bits select the time by which external MOVX cycles are to be stretched. This allows slower memory or peripherals to be accessed without using ports or manual software intervention. The RD or WR strobe is stretched by the specified interval, which is transparent to the software except for the increased time to execute to MOVX instruction. All internal MOVX instructions are executed at the two machine cycle rate (0 stretch) independent of these bit settings. MOVX Instruction MD2, MD1, MD0 000 STRETCH VALUE 0 MOVX DURATION 2 machine cycles 001 1 3 machine cycles (default) 010 011 100 101 110 111 2 3 4 5 6 7 4 machine cycles 5 machine cycles 9 machine cycles 10 machine cycles 11 machine cycles 12 machine cycles _____________________________________________________________________________________________ 22 Ultra-High-Speed Flash Microcontroller User’s Guide Port 1 (P1) SFR 90h 7 6 5 4 3 2 1 0 P1.7 INT5 P1.6 INT4 P1.5 INT3 P1.4 INT2 P1.3 TXD1 P1.2 RXD1 P1.1 T2EX P1.0 T2 RW-1 RW-1 RW-1 RW-1 RW-1 RW-1 RW-1 RW-1 R = Unrestricted read, W = Unrestricted write, -n = Value after reset P1.7–0 Bits 7–0 General-Purpose I/O Port 1. This register functions as a general-purpose I/O port. In addition, all the pins have an alternative function listed below. Each of the functions is controlled by several other SFRs. The associated Port 1 latch bit must contain a logic 1 before the pin can be used in its alternate function capacity. INT5 Bit 7 External Interrupt 5. A falling edge on this pin causes an external interrupt 5 if enabled. INT4 Bit 6 External Interrupt 4. A rising edge on this pin causes an external interrupt 4 if enabled. INT3 Bit 5 External Interrupt 3. A falling edge on this pin causes an external interrupt 3 if enabled. INT2 Bit 4 External Interrupt 2. A rising edge on this pin causes an external interrupt 2 if enabled. TXD1 Bit 3 Serial Port 1 Transmit. This pin transmits the serial port 1 data in serial port modes 1, 2, 3 and emits the synchronizing clock in serial port mode 0. RXD1 Bit 2 Serial Port 1 Receive. This pin receives the serial port 1 data in serial port modes 1, 2, 3 and is a bidirectional data transfer pin in serial port mode 0. T2EX Bit 1 Timer 2 Capture/Reload Trigger. A 1-to-0 transition on this pin causes the value in the T2 registers to be transferred into the capture registers if enabled by EXEN2 (T2CON.3). When in autoreload mode, a 1-to-0 transition on this pin reloads the Timer 2 registers with the value in RCAP2L and RCAP2H if enabled by EXEN2 (T2CON.3). T2 Timer 2 External Input. A 1-to-0 transition on this pin causes Timer 2 increment or decrement bit External Interrupt Flag (EXIF) 7 6 5 4 3 2 1 IE5 IE4 IE3 IE2 CKRY RGMD RGSL RW-0 RW-0 RW-0 RW-0 R-* R-* RW-* R = Unrestricted read, W = Unrestricted write, T = Timed-access write only, -n = Value after reset,* = See description. SFR 91h 0 BGS RT-0 Bit 0 depending on the timer configuration. IE5 Bit 7 External Interrupt 5 Flag. This bit is set when a falling edge is detected on INT5. This bit must be cleared manually by software. Setting this bit in software causes an interrupt if enabled. IE4 Bit 6 External Interrupt 4 Flag. This bit is set when a rising edge is detected on INT4. This bit must be cleared manually by software. Setting this bit in software causes an interrupt if enabled. IE3 Bit 5 External Interrupt 3 Flag. This bit is set when a falling edge is detected on INT3. This bit must be cleared manually by software. Setting this bit in software causes an interrupt if enabled. IE2 Bit 4 External Interrupt 2 Flag. This bit is set when a rising edge is detected on INT2. This bit must be cleared manually by software. Setting this bit in software causes an interrupt if enabled. CKRY Bit 3 Clock Ready. This bit indicates the status of the startup period for the crystal oscillator or crystal multiplier warm-up period. This bit is cleared after a reset or when exiting STOP mode. It is also cleared when the clock multiplier is enabled (setting of PMR.4 = 1). Once CKRY is cleared, a 65,536 clock count must take place before CKRY is set and the lockout preventing modification of CD1:CD0 is removed. Once CKRY is set (= 1), the clock multiplier can then be selected as the 23 _____________________________________________________________________________________________ Ultra-High-Speed Flash Microcontroller User’s Guide clock source or switchover from the ring oscillator to the crystal oscillator can occur. RGMD Bit 2 Ring Mode Status. This status bit indicates the current clock source for the device. This bit is cleared to 0 after a power-on reset, and unchanged by all other forms of reset. 0 = Device is operating from the external crystal or oscillator. 1 = Device is operating from the ring oscillator. RGSL Bit 1 Ring Oscillator Select. When set (= 1), this bit enables operation using the on-chip ring oscillator as the clock source until the oscillator warm-up period has completed (CKRY = 1). Using the ring oscillator to resume from stop mode allows almost instantaneous startup. This bit is cleared to 0 after a power-on reset, and unchanged by all other forms of reset. 0 = Device operation is held until completion of the crystal oscillator warm-up delay period. 1 = The device begins operating from the ring oscillator and switch over to the crystal oscillator upon completion of the warm-up delay period. BGS Bit 0 Bandgap Select. This bit enables/disables the bandgap reference during stop mode. Disabling the bandgap reference provides significant power savings in stop mode, but sacrifices the ability to perform a power-fail interrupt or power-fail reset while stopped. This bit can only be modified with a timed access procedure. Timer and Serial Port Clock Mode Register (CKMOD) SFR 96h 7 — RW-1 6 — RW-1 5 T2MH RW-0 4 T1MH RW-0 3 T0MH RW-0 2 — RW-1 1 — RW-1 0 — RW-1 R = Unrestricted read, W = Unrestricted write, -n = Value after reset 0 = The bandgap reference is disabled in stop mode but functions during normal operation. 1 = The bandgap reference operates in stop mode. T2MH Bit 5 Timer 2 Clock Mode High-Speed Select. When set (= 1), the system clock is used as the input clock for Timer 2 and the T2M bit (CKCON.5) setting is ignored. When clear (= 0), the input clock for Timer 2 is selected using the T2M bit. T1MH Bit 4 Timer 1 Clock Mode High-Speed Select. When set (= 1), the system clock is used as the input clock for Timer 2 and the T1M bit (CKCON.4) setting is ignored. When clear (= 0), the input clock for Timer 2 is selected using the T1M bit. Serial Port 0 Control (SCON0) SFR 98h 7 SM0/FE_0 RW-0 6 SM1_0 RW-0 5 SM2_0 RW-0 4 REN_0 RW-0 3 TB8_0 RW-0 2 RB8_0 RW-0 1 TI_0 RW-0 0 RI_0 RW-0 R = Unrestricted read, W = Unrestricted write, -n = Value after reset T0MH Bit 3 Timer 0 Clock Mode High-Speed Select. When set (= 1), the system clock is used as the input clock for Timer 2 and the T0M bit (CKCON.3) setting is ignored. When clear (= 0), the input clock for Timer 2 is selected using the T0M bit. SM0–2 Bits 7, 6, 5 Serial Port Mode. These bits control the mode of serial port 0. In addition the SM0 and SM2_0 bits have secondary functions as shown. _____________________________________________________________________________________________ 24 Ultra-High-Speed Flash Microcontroller User’s Guide Serial Port Mode Functions SM0 0 0 0 1 SM1 0 0 1 0 SM2 0 1 X 0 MODE 0 0 1 2 1 0 1 2 Asynchronous with multiprocessor communication 11 See PMR register 1 1 0 3 Asynchronous 11 Timer 1 or 2 baud rate equation 3 Asynchronous with multiprocessor communication 11 Timer 1 or 2 baud rate equation 1 1 1 FUNCTION Synchronous Synchronous Asynchronous Asynchronous LENGTH (BITS) 8 8 10 11 PERIOD See PMR register See PMR register Timer 1 or 2 baud rate equation See PMR register SM0/FE_0 Bit 7 Framing Error Flag. When SMOD0 (PCON.6) = 0, this bit is used as a mode select bit (SM0) for serial port 0. When SMOD0 (PCON.6) = 1, this bit becomes a framing error (FE) bit, which reports detection of an invalid stop bit. When used as FE, this bit must be cleared in software. Once the SMOD0 bit is set, modifications to this bit do not affect the serial port mode settings. Although accessed from the same register, the data for bits SM0 and FE are stored internally in different physical locations. SM1_0 Bit 6 No alternate function. SM2_0 Bit 5 Multiple CPU Communications. The function of this bit is dependent on the serial port 0 mode. Mode 0: Selects period for synchronous serial port 0 data transfers. Mode 1: When set, reception is ignored (RI_0 is not set) if invalid stop bit received. Modes 2/3: When this bit is set, multiprocessor communications are enabled in modes 2 and 3. This prevents the RI_0 bit from being set, and an interrupt being asserted, if the 9th bit received is not 1. REN_0 Bit 4 Receiver Enable. This bit enable/disables the serial port 0 receiver shift register. 0 = Serial port 0 reception disabled. 1 = Serial port 0 receiver enabled (modes 1, 2, 3). Initiate synchronous reception (mode 0). TB8_0 Bit 3 9th Transmission Bit State. This bit defines the state of the 9th transmission bit in serial port 0 modes 2 and 3. RB8_0 Bit 2 9th Received Bit State. This bit identifies that state of the 9th reception bit of received data in serial port 0 modes 2 and 3. In serial port mode 1, when SM2_0 = 0, RB8_0 is the state of the stop bit. RB8_0 is not used in mode 0. TI_0 Bit 1 Transmitter Interrupt Flag. This bit indicates that data in the serial port 0 buffer has been completely shifted out. In serial port mode 0, TI_0 is set at the end of the 8th data bit. In all other modes, this bit is set at the end of the last data bit. This bit must be manually cleared by software. RI_0 Bit 0 Receiver Interrupt Flag. This bit indicates that a byte of data has been received in the serial port 0 buffer. In serial port mode 0, RI_0 is set at the end of the 8th bit. In serial port mode 1, RI_0 is set after the last sample of the incoming stop bit subject to the state of SM2_0. In modes 2 and 3, RI_0 is set after the last sample of RB8_0. This bit must be manually cleared by software. Serial Data Buffer 0 (SBUF0) SFR 99h 7 SBUF0.7 RW-0 6 SBUF0.6 RW-0 5 SBUF0.5 RW-0 4 SBUF0.4 RW-0 3 SBUF0.3 RW-0 2 SBUF0.2 RW-0 1 SBUF0.1 RW-0 0 SBUF0.0 RW-0 R = Unrestricted read, W = Unrestricted write, -n = Value after reset SBUF0.7–0 Bits 7–0 25 Serial Data Buffer 0. Data for serial port 0 is read from or written to this location. The serial transmit and receive buffers are separate registers, but both are addressed at this location. _____________________________________________________________________________________________ Ultra-High-Speed Flash Microcontroller User’s Guide Address Control (ACON) SFR 9Dh 7 PAGEE RT-0 6 PAGES1 RT-0 5 PAGES0 RT-0 4 — R-1 3 — R-1 2 — R-1 1 — R-1 0 — R-1 R = Unrestricted read, W = Unrestricted write, T = Timed-access write only, -n = Value after reset PAGEE Bits 7 Page Mode Enable. When set (= 1), page mode access is enabled for external bus operations as configured by the page mode select bits PAGES1, PAGES0. When clear (= 0), external bus operations default to the standard 8051 expanded bus configuration. PAGES1, PAGES0 Bits 6, 5 Page Mode Select. If PAGEE = 1, these bits select the page mode configuration that is followed for external bus operations. The four possible configurations are summarized in the table below. Mode 1 results in Port 0 serving as the data bus and Port 2 being the multiplexed address MSB/LSB. Mode 2 results in Port 0 being used strictly for address LSB and Port 2 being multiplexed between address MSB and data. Memory Access Cycle PAGES 1-0 00 01 10 11 Bits 4–0 MODE 1 1 1 2 PAGE-HIT 1 2 4 2 PAGE-MISS 2 4 8 4 Reserved. Read data is 1. Port 2 (P2) SFR A0h 7 P2.7 RW-1 6 P2.6 RW-1 5 P2.5 RW-1 4 P2.4 RW-1 3 P2.3 RW-1 2 P2.2 RW-1 1 P2.1 RW-1 0 P2.0 RW-1 R = Unrestricted read, W = Unrestricted write, -n = Value after reset P2.7–0 Bits 7–0 Port 2. This port functions according to the table below where PAGEE = ACON.7 and PAGES = ACON.6-5. Port 2 Functions PAGEE 0 0 1 1 PAGES XX XX 00, 01, 10 11 PORT2 FUNCTION General-Purpose I/0 (code execution < ROMSIZE.2-0) Address MSB (code execution > ROMSIZE.2-0) Multiplexed Address MSB/LSB Multiplexed Address MSB/Data Writing a 1 to an SFR bit configures the associated port pin as an input. All read operations, with the exception of read-modify-write instructions, leave the port latch unchanged. During external memory addressing and data memory write cycles, the port has high and low drive capability. During external memory data read cycles, the port is held in a high-impedance state. _____________________________________________________________________________________________ 26 Ultra-High-Speed Flash Microcontroller User’s Guide Interrupt Enable (IE) SFR A8h 7 EA RW-0 6 ES1 RW-0 5 ET2 RW-0 4 ES0 RW-0 3 ET1 RW-0 2 EX1 RW-0 1 ET0 RW-0 0 EX0 RW-0 R = Unrestricted read, W = Unrestricted write, -n = Value after reset EA Bit 7 Global Interrupt Enable. This bit controls the global masking of all interrupts except power-fail interrupt, which is enabled by the EPFI bit (WDCON.5). 0 = Disable all interrupt sources. This bit overrides individual interrupt mask settings. 1 = Enable all individual interrupt masks. Individual interrupts occur if enabled. ES1 Bit 6 Enable Serial Port 1 Interrupt. This bit controls the masking of the serial port 1 interrupt. 0 = Disable all serial port 1 interrupts. 1 = Enable interrupt requests generated by the RI_1 (SCON1.0) or TI_1 (SCON1.1) flags. ET2 Bit 5 Enable Timer 2 Interrupt. This bit controls the masking of the Timer 2 interrupt. 0 = Disable all Timer 2 interrupts. 1 = Enable interrupt requests generated by the TF2 flag (T2CON.7). ES0 Bit 4 Enable Serial Port 0 Interrupt. This bit controls the masking of the serial port 0 interrupt. 0 = Disable all serial port 0 interrupts. 1 = Enable interrupt requests generated by the RI_0 (SCON0.0) or TI_0 (SCON0.1) flags. ET1 Bit 3 Enable Timer 1 Interrupt. This bit controls the masking of the Timer 1 interrupt. 0 = Disable all Timer 1 interrupts. 1 = Enable all interrupt requests generated by the TF1 flag (TCON.7). EX1 Bit 2 Enable External Interrupt 1. This bit controls the masking of external interrupt 1. 0 = Disable external interrupt 1. 1 = Enable all interrupt requests generated by the INT0 pin. ET0 Bit 1 Enable Timer 0 Interrupt. This bit controls the masking of the Timer 0 interrupt. 0 = Disable all Timer 0 interrupts. 1 = Enable all interrupt requests generated by the TF0 flag (TCON.5). EX0 Bit 0 Enable External Interrupt 0. This bit controls the masking of external interrupt 0. 0 = Disable external interrupt 0. 1 = Enable all interrupt requests generated by the INT0 pin. SADDR0.7–0 Slave Address Register 0. This register is programmed with the given or broadcast address Slave Address Register 0 (SADDR0) . SFR A9h 7 SADDR0.7 RW-0 6 SADDR0.6 RW-0 5 SADDR0.5 RW-0 4 SADDR0.4 RW-0 3 SADDR0.3 RW-0 2 SADDR0.2 RW-0 R = Unrestricted read, W = Unrestricted write, -n = Value after reset Bits 7–0 27 assigned to serial port 0. _____________________________________________________________________________________________ 1 SADDR0.1 RW-0 0 SADDR0.0 RW-0 Ultra-High-Speed Flash Microcontroller User’s Guide Slave Address Register 1 (SADDR1) SFR AAh 7 SADDR1.7 RW-0 6 SADDR1.6 RW-0 5 SADDR1.5 RW-0 4 SADDR1.4 RW-0 3 SADDR1.3 RW-0 2 SADDR1.2 RW-0 1 SADDR1.1 RW-0 0 SADDR1.0 RW-0 R = Unrestricted read, W = Unrestricted write, -n = Value after reset SADDR1.7–0 Bits 7–0 Slave Address Register 1. This register is programmed with the given or broadcast address assigned to serial port 1. Port 3 (P3) SFR B0h 7 6 5 4 3 2 1 0 P3.7 RD P3.6 WR P3.5 T1 P3.4 T0 P3.3 INT1 P3.2 INT0 P3.1 TXD0 P3.0 RXD0 RW-1 RW-1 RW-1 RW-1 RW-1 RW-1 RW-1 RW-1 R = Unrestricted read, W = Unrestricted write, -n = Value after reset P3.7–0 Bits 7–0 Purpose I/O Port 3. This register functions as a general-purpose I/O port. In addition, all the pins have an alternative function listed below. Each of the functions is controlled by several other SFRs. The associated port 3 latch bit must contain a logic 1 before the pin can be used in its alternate function capacity. RD Bit 7 External Data Memory Read Strobe. This pin provides an active-low read strobe to an external memory device. WR Bit 6 External Data Memory Write Strobe. This pin provides an active-low write strobe to an external memory device. T1 Bit 5 Timer/Counter External Input. A 1-to-0 transition on this pin increments Timer 1. T0 Bit 4 Counter External Input. A 1-to-0 transition on this pin increments Timer 0. INT1 Bit 3 External Interrupt 1. A falling edge/low level on this pin causes an external interrupt 1 if enabled. INT0 Bit 2 External Interrupt 0. A falling edge/low level on this pin causes an external interrupt 0 if enabled. TXD0 Bit 1 Serial Port 0 Transmit. This pin transmits the serial port 0 data in serial port modes 1, 2, 3 and emits the synchronizing clock in serial port mode 0. RXD0 Serial Port 0 Receive. This pin receives the serial port 0 data in serial port modes 1, 2, 3 and is a Interrupt Priority 1 (IP1) SFR B1h 7 — R-1 6 MPS1 RW-0 5 MPT2 RW-0 4 MPS0 RW-0 3 MPT1 RW-0 2 MPX1 RW-0 1 MPT0 RW-0 0 MPX0 RW-0 R = Unrestricted read, W = Unrestricted write, -n = Value after reset Bit 0 bidirectional data transfer pin in serial port mode 0. IP17.0 Bit 7 Reserved. Read data is 1. MPS1 Bit 6 Most Significant Priority Select Bit for Serial Port 1 Interrupt. This is the most significant bit of the bit pair MPS1, LPS1 (IP0.6) that designates priority level for the serial port 1 interrupt. _____________________________________________________________________________________________ 28 Ultra-High-Speed Flash Microcontroller User’s Guide MPT2 Bit 5 Most Significant Priority Select Bit for Timer 2 Interrupt. This is the most significant bit of the bit pair MPT2, LPT2 (IP0.5) that designates priority level for the timer 2 interrupt. MPS0 Bit 4 Most Significant Priority Select Bit for Serial Port 0 Interrupt. This is the most significant bit of the bit pair MPS0, LPS0 (IP0.4) that designates priority level for the serial port 0 interrupt MPT1 Bit 3 Most Significant Priority Select Bit for Timer 1 Interrupt. This is the most significant bit of the bit pair MPT1, LPT1 (IP0.3) that designates priority level for the timer 1 interrupt. MPX1 Bit 2 Most Significant Priority Select Bit for External Interrupt 1. This is the most significant bit of the bit pair MPX1, LPX1 (IP0.2) that designates priority level for external interrupt 1 MPT0 Bit 1 Most Significant Priority Select Bit for Timer 0 Interrupt. This is the most significant bit of the bit pair MPT0, LPT0 (IP0.1) that designates priority level for the timer 0 interrupt MPX0 Bit 0 Most Significant Priority Select Bit for External Interrupt 0. This is the most significant bit of the bit pair MPX0, LPX0 (IP0.0) that designates priority level for external interrupt 0. Interrupt priority level for the above sources is assigned using one bit from register IP1 (B1h) and one bit from IP0 (B8h). The bit from IP1 serves as the most significant bit and the bit from IP0 serves as the least significant bit in forming a 2-bit binary number. This number represents the priority level. Higher priority interrupts, when enabled, take precedence over lower priority sources. The powerfail warning interrupt source is assigned priority level 4. Most Significant Priority Select Bit Levels Interrupt Priority 0 (IP0) SFR B8h 7 6 5 4 3 2 1 0 — LPS1 LPT2 LPS0 LPT1 LPX1 LPT0 LPX0 R-1 RW-0 RW-0 RW-0 RW-0 RW-0 RW-0 RW-0 R = Unrestricted read, W = Unrestricted write, -n = Value after reset MP (IP1.X) 0 0 1 1 LP (IP0.X) 0 1 0 1 PRIORITY LEVEL 0 (natural priority) 1 2 3 (high priority) IP07.0 Bit 7 Reserved. Read data is 1. LPS1 Bit 6 Least Significant Priority Select Bit for Serial Port 1 Interrupt. LPS1 is the least significant bit of the bit pair MPS1 (IP1.6), LPS1 that designates priority level for the serial port 1 interrupt. LPT2 Bit 5 Least Significant Priority Select Bit for Timer 2 Interrupt. LPT2 is the least significant bit of the bit pair MPT2 (IP1.5), LPT2 that designates priority level for the Timer 2 interrupt. LPS0 Bit 4 Least Significant Priority Select Bit for Serial Port 0 Interrupt. MPS0 is the least significant bit of the bit pair MPS0 (IP1.4), LPS0 that designates priority level for the serial port 0 interrupt. LPT1 Bit 3 Least Significant Priority Select Bit for Timer 1 Interrupt. MPT1 is the least significant bit of the bit pair MPT1 (IP1.3), LPT1 that designates priority level for the Timer 1 interrupt. LPX1 Bit 2 Least Significant Priority Select Bit for External Interrupt 1. MPX1 is the least significant bit of the bit pair MPX1 (IP1.2), LPX1 that designates priority level for external interrupt 1. LPT0 Bit 1 Least Significant Priority Select Bit for Timer 0 Interrupt. MPT0 is the least significant bit of the bit pair MPT0 (IP1.1), LPT0 that designates priority level for the Timer 0 interrupt. LPX0 Bit 0 Least Significant Priority Select Bit for External Interrupt 0. MPX0 is the least significant bit of the bit pair MPX0 (IP1.0), LPX0 that designates priority level for external interrupt 0. 29 _____________________________________________________________________________________________ Ultra-High-Speed Flash Microcontroller User’s Guide Least Significant Priority Select Bit Levels MP (IP1.X) 0 0 1 1 LP (IP0.X) 0 1 0 1 PRIORITY LEVEL 0 (natural priority) 1 2 3 (high priority) Slave Address Mask Enable Register 0 (SADEN0) 7 SADEN0.7 RW-0 SFR B9h 6 SADEN0.6 RW-0 5 SADEN0.5 RW-0 4 SADEN0.4 RW-0 3 SADEN0.3 RW-0 2 SADEN0.2 RW-0 1 SADEN0.1 RW-0 0 SADEN0.0 RW-0 R = Unrestricted read, W = Unrestricted write, -n = Value after reset SADEN0.7–0 Bits 7–0 Slave Address Mask Enable Register 0. This register functions as a mask when comparing serial port 0 addresses for automatic address recognition. When a bit in this register is set, the corresponding bit location in the SADDR0 register is exactly compared with the incoming serial port 0 data to determine if a receiver interrupt should be generated. When a bit in this register is cleared, the corresponding bit in the SADDR0 register becomes a “don’t care” and is not compared against the incoming data. All incoming data generates a receiver interrupt when this register is cleared. Slave Address Mask Enable Register 1 (SADEN1) 7 SADEN1.7 RW-0 SFR BAh 6 SADEN1.6 RW-0 5 SADEN1.5 RW-0 4 SADEN1.4 RW-0 3 SADEN1.3 RW-0 2 SADEN1.2 RW-0 1 SADEN1.1 RW-0 0 SADEN1.0 RW-0 R = Unrestricted read, W = Unrestricted write, -n = Value after reset SADEN1.7–0 Bits 7–0 Slave Address Mask Enable Register 1. This register functions as a mask when comparing serial port 1 addresses for automatic address recognition. When a bit in this register is set, the corresponding bit location in the SADDR1 register is exactly compared with the incoming serial port 1 data to determine if a receiver interrupt should be generated. When a bit in this register is cleared, the corresponding bit in the SADDR1 register becomes a “don’t care” and is not compared against the incoming data. All incoming data generates a receiver interrupt when this register is cleared. Serial Port 1 Control (SCON1) 7 SM0/FE_1 RW-0 SFR C0h 6 SM1_1 RW-0 5 SM2_1 RW-0 4 REN_1 RW-0 3 TB8_1 RW-0 2 RB8_1 RW-0 1 TI_1 RW-0 0 RI_1 RW-0 R = Unrestricted read, W = Unrestricted write, -n = Value after reset SM0–2 Bits 7, 6, 5 Serial Port 1 Mode. These bits control the mode of serial port 1 as shown in the following table. In addition, the SM0 and SM2 bits have secondary functions as shown. Serial Port 1 Modes and Functions SM0 0 0 0 1 SM1 0 0 1 0 SM2 0 1 X 0 MODE 0 0 1 2 1 0 1 2 Asynchronous with multiprocessor communication 11 See PMR register 1 1 0 3 Asynchronous 11 Timer 1 or 2 baud rate equation 3 Asynchronous with multiprocessor communication 11 Timer 1 or 2 baud rate equation 1 1 1 FUNCTION Synchronous Synchronous Asynchronous Asynchronous LENGTH (BITS) 8 8 10 11 PERIOD See PMR register See PMR register Timer 1 or 2 baud rate equation See PMR register _____________________________________________________________________________________________ 30 Ultra-High-Speed Flash Microcontroller User’s Guide SM0/FE_1 Bit 7 Framing Error Flag. When SMOD0 (PCON.6) = 0, this bit is used as a mode select bit (SM0) for serial port 1. When SMOD0 (PCON.6) = 1, this bit becomes a framing error (FE) bit, which reports detection of an invalid stop bit. When used as FE, this bit must be cleared in software. Once the SMOD0 bit is set, modifications to this bit does not affect the serial port mode settings. Although accessed from the same register, the data for bits SM0 and FE are stored internally in different physical locations. SM1_1 Bit 6 No alternate function. SM2_1 Bit 5 Multiple CPU Communications. The function of this bit is dependent on the serial port 1 mode. Mode 0: Selects period for synchronous port 1 data transfers. Mode 1: When this bit is set, reception is ignored (RI_1 is not set) if invalid stop bit received. Modes 2/3: When this bit is set, multiprocessor communications are enabled in mode 2 and 3. This prevents RI_1 from being set, and an interrupt being asserted, if the 9th bit received is not 1. REN_1 Bit 4 Receive Enable. This bit enables/disables the serial port 1 receiver shift register. 0 = Serial port 1 reception disabled. 1 = Serial port 1 receiver enabled (modes 1, 2, 3). Initiate synchronous reception (mode 0). TB8_1 Bit 3 9th Transmission Bit State. This bit defines the state of the 9th transmission bit in serial port 1 modes 2 and 3. RB8_1 Bit 2 9th Received Bit State. This bit identifies the state for the 9th reception bit received data in serial port 1 modes 2 and 3. In serial port mode 1, when SM2_1 = 0, RB8_1 is the state of the stop bit. RB8_1 is not used in mode 0. TI_1 Bit 1 Transmitter Interrupt Flag. This bit indicates that data in the serial port 1 buffer has been completely shifted out. In serial port mode 0, TI_1 is set at the end of the 8th data bit. In all other modes, this bit is set at the end of the last data bit. This bit must be manually cleared by software. RI_1 Receiver Interrupt Flag. This bit indicates that a byte of data has been received in the serial Bit 0 port 1 buffer. In serial port mode 1, RI_1 is set at the end of the 8th bit. In serial port mode 1, RI_1 is set after the last sample of the incoming stop bit subject to the state of SM2_1. In modes 2 and 3, RI_1 is set after the last sample of RB8_1. This bit must be manually cleared by software. Serial Data Buffer 1 (SBUF1) SFR C1h 7 SBUF1.7 RW-0 6 SBUF1.6 RW-0 5 SBUF1.5 RW-0 4 SBUF1.4 RW-0 3 SBUF1.3 RW-0 2 SBUF1.2 RW-0 1 SBUF1.1 RW-0 0 SBUF1.0 RW-0 R = Unrestricted read, W = Unrestricted write, -n = Value after reset SBUF1.7–0 Bits 7–0 31 Serial Data Buffer 1. Data for serial port 1 is read from or written to this location. The serial transmit and receive buffers are separate registers, but both are addressed at this location. _____________________________________________________________________________________________ Ultra-High-Speed Flash Microcontroller User’s Guide ROM Size Select (ROMSIZE) SFR C2h 7 — R-1 6 — R-1 5 — R-1 4 — R-1 3 PRAME RT-0 2 RMS2 RT-1 1 RMS1 RT-0 0 RMS0 RT-1 R = Unrestricted read, W = Unrestricted write, T = Timed-access write only, -n = Value after reset Bits 7–4 These bits are reserved. Read data is 1. PRAME Bit 3 Program RAM Enable. When set (= 1), the internal 1k RAM is mapped as internal program space between addresses 0400h–07FFh. All program fetches and MOVC accesses are directed to this 1k RAM. When serving as program memory, the RAM continues to be accessible as MOVX data space (if DME0 = 1). The 1k RAM is not accessible as program space when EA = 0. When clear (= 0), the internal 1k RAM is not accessible as program space. RMS2–0 Bits 2–0 ROM Memory Size Select 2-0. This register is used to select the maximum on-chip decoded address. Care must be taken that the memory location of the current program counter is valid both before and after modification. These bits can only be modified using a timed-access procedure. The EA pin overrides the function of these bits when asserted, forcing the device to access external program memory only. Configuring this register to a setting that exceeds the maximum amount of internal memory can corrupt device operation. These bits default on reset to the maximum amount of internal program memory (i.e., 16k for DS89C420). On-Chip ROM Address . RS2 0 0 0 0 1 1 1 1 RS1 0 0 1 1 0 0 1 1 RS0 0 1 0 1 0 1 0 1 MAXIMUM ON-CHIP ROM ADDRESS 0kB/Disable on-chip ROM 1kB/03FFh 2kB/07FFh 4kB/0FFFh 8kB/1FFFh 16kB/3FFFh (DS89C420/430 default) 32kB/7FFFh (DS89C440 default) 64kB/FFFFh (DS89C450 default) Power Management Register (PMR) SFR C4h 7 CD1 RW*-1 6 CD0 RW*-0 5 SWB RW-0 4 CTM RW*-0 3 4X / 2X RW*-0 2 ALEON RW-0 1 DME1 RW-0 0 DME0 RW-0 R = Unrestricted read, W = Unrestricted write, -n = Value after reset, * = See description CD1, CD0 Bits 7, 6 Clock Divide Control 1-0. These bits select the number of crystal oscillator clocks required to generate one machine cycle. Switching between modes requires a transition through the default divide-by-1 mode (CD1, CD0 = 10b). Attempts to perform an invalid transition are ignored. For example, going from the crystal multiplier 2X mode to the divide-by-1024 mode would require first switching from the 2X crystal multiplier mode to the divide-by-1 mode, followed by the switch from the divide-by-1 to the divide-by-1024 mode. These bits cannot be modified when running from the internal ring oscillator (RGMD = 1). The divide-by-1024 setting (CD1,CD0 = 11b) cannot be selected when switchback is enabled (SWB = 1) and a switchback source (serial port or external interrupt) is active. CD1, CD0 00 01 10 11 CLOCK FUNCTION Crystal multiplier (4X or 2X mode as determined by PMR.3) Reserved (forced into divide-by-1 mode if set) Divide-by-1 (default state) Divide-by-1024 The setting of these bits affects timer and serial port operation. Tables located in the SFR decription for CKCON (8Eh) detail the respective operational dependencies on these bits. _____________________________________________________________________________________________ 32 Ultra-High-Speed Flash Microcontroller User’s Guide Serial Port Operation (in Oscillator Clocks) CLOCK (MODE 0) 4X/2X CD1:0 1 CLOCK (MODE 2) SM2 = 0 SM2 = 1 SMOD = 0 00 3 1 64 0 00 6 2 64 32 X 01 12 4 64 32 X 10 12 4 64 32 X 11 3072 1024 16384 8192 SMOD = 1 32 SWB Bit 5 Switchback Enable. This bit allows an enabled external interrupt or serial port activity to force the clock divide control bits to the divide-by-1 state (01b). Upon acknowledgement of an external interrupt source, the device switches modes in order to service the interrupt. Note that this means that an external interrupt must actually be recognized (i.e., be enabled and not masked by higher priority interrupts) for the switchback to occur. For serial port reception, the switch occurs at the start of the instructions following the falling edge of the start bit. CTM Bit 4 Crystal Multiplier Enable. This bit enables (= 1) or disables (= 0) the crystal multiplier function. When set (= 1), the CKRY bit (EXIF.3) is cleared and the multiplier circuitry begins a stabilization warm-up period to provide the clock multiplication factor specified by the 4X/2X bit (PMR.3). Upon completion of the warm-up delay, the CKRY bit is set and the user can then modify CD1,CD0 (PMR.7, PMR.6) to select the crystal multiplier clock output. When clear (= 0), the crystal multipler circuitry is disabled to conserve power. The CTM bit cannot be changed unless CD1,CD0 = 10b and RGMD (EXIF.2) is cleared to 0. This bit is automatically cleared to 0 when the processor enters stop mode. 4X/2X Bit 3 Clock Multiplier Selection. This bit selects the clock multiplication factor as shown. 4X/2X = 0 The frequency multiplier is set to two times the incoming clock by 4X/2X = 0. 4X/2X = 1 sets the frequency multiplier to 4 times the incoming clock. This bit can only be altered when the crystal multiplier enable bit (CTM) is cleared. Therefore, it must be set for the desired multiplication factor prior to setting the CTM bit. ALE Enable. When set (= 1), this bit enables the ALE signal output during on-chip program and data memory accesses. When clear (= 0), the ALE signal output is disabled during on-chip program and data memory accesses. External memory access automatically enables ALE independent of the state of ALEON. ALEON Bit 2 DME1, DME0 Bits 1, 0 Data Memory Enable 1-0. These bits determine the functional relationship of the first 1024 bytes of data memory. Two memory configurations are supported to allow either external data memory access through the expanded bus of port 0 and port 2, or internal SRAM data memory access. Note these bits are cleared after a reset, so access to the internal SRAM is prohibited until these bits are modified. Data Memory Access DME1 DME0 DATA MEMORY ADDRESS RANGE 0 0 0000h–FFFFh External Data Memory (default) X 1 0000h–03FFh 0400h–FFFFh Internal SRAM Data Memory External Data Memory 1 0 Reserved MEMORY ACCESS Reserved Status Register (STATUS) SFR C5 7 PIS2 R-0 6 PIS1 R-0 5 PIS0 R-0 4 — R-1 3 SPTA1 R-0 2 SPRA1 R-0 1 SPTA0 R-0 0 SPRA0 R-0 R = Unrestricted read, W = Unrestricted write, -n = Value after reset PIS2-0 Bit 7, 6, 5 33 Priority Interrupt Status Bits 2-0. These bits indicate the level of interrupt that is currently being serviced. (Interrupt levels 0-3 are associated with interrupt sources using the MP,LP bits found in the IP1 and IP0 SFRS.) _____________________________________________________________________________________________ Ultra-High-Speed Flash Microcontroller User’s Guide PIS2-0 000 001 010 011 100 101 INTERRUPT PRIORITY LEVEL No interrupt in progress Level 0 interrupt in progress Level 1 interrupt in progress Level 2 interrupt in progress Level 3 interrupt in progress Power-fail warning interrupt in progress Bit 4 This bit is reserved and reads a logic 1. SPTA1 Bit 3 Serial Port 1 Transmit Activity Monitor. When set, this bit indicates that data is currently being transmitted by serial port 1. It is cleared when the internal hardware sets the TI_1 bit. Do not alter the clock divide control bits (PMR.7-6) while this bit is set or serial port data can be lost. SPRA1 Bit 2 Serial Port 1 Receive Activity Monitor. When set, this bit indicates that data is currently being received by serial port 1. It is cleared when the internal hardware sets the RI_1 bit. Do not alter the clock divide control bits (PMR.7-6) while this bit is set or serial port data can be lost. SPTA0 Bit 1 Serial Port 0 Transmit Activity Monitor. When set, this bit indicates that data is currently being transmitted by serial port 0. It is cleared when the internal hardware sets the TI_1 bit. Do not alter the clock divide control bits (PMR.7-6) while this bit is set or serial port data can be lost. SPRA0 Bit 0 Serial Port 0 Receive Activity Monitor. When set, this bit indicates that data is currently being received by serial port 0. It is cleared when the internal hardware sets the RI_1 bit. Do not alter the clock divide control bits (PMR.7-6) while this bit is set or serial port data can be lost. Timed Access Register (TA) SFR C7h 7 TA.7 W-1 6 TA.6 W-1 5 TA.5 W-1 4 TA.4 W-1 3 TA.3 W-1 2 TA.2 W-1 1 TA.1 W-1 0 TA.0 W-1 W = Unrestricted write, -n = Value after reset TA.7–0 Bits 7–0 Timed Access. Correctly accessing this register permits modification of timed access protected bits. Write AAh to this register first, followed within 3 cycles by writing 55h. Timed access protected bits can then be modified for a period of 3 cycles measured from the writing of the 55h. Timer 2 Control (T2CON) SFR C8h 7 TF2 RW-0 6 EXF2 RW-0 5 RCLK RW-0 4 TCLK RW-0 3 EXEN2 RW-0 2 TR2 RW-0 1 C/T2 RW-0 0 CP/RL2 RW-0 R = Unrestricted read, W = Unrestricted write, -n = Value after reset TF2 Bit 7 Timer 2 Overflow Flag. This flag is set when Timer 2 overflows from FFFFh or the count equal to the capture register in down count mode. It must be cleared by software. TF2 is only set if RCLK and TCLK are both cleared to 0. EXF2 Bit 6 Timer 2 External Flag. A negative transition on the T2EX pin (P1.1) or timer 2 underflow/overflow causes this flag to set based on the CP/RL2 (T2CON.0), EXEN2 (T2CON.3), and DCEN (T2MOD.0) bits (see the following table). If set by a negative transition, this flag must be cleared to 0 by software. Setting this bit in software or detection of a negative transition on the T2EX pin forces a timer interrupt if enabled. _____________________________________________________________________________________________ 34 Ultra-High-Speed Flash Microcontroller User’s Guide CP/RL2 EXEN2 RESULT 0 X Negative transitions on P1.1 do not affect this bit. 1 1 X Negative transitions on P1.1 set this bit. 0 0 0 Negative transitions on P1.1 do not affect this bit. 0 1 0 Negative transitions on P1.1 set this bit. 1 Bit toggles whenever Timer 2 underflows/overflows and can be used as a 17th bit of resolution. In this mode, EXF2 does not cause an interrupt. 0 RCLK Bit 5 DCEN 1 X Receive Clock Flag. This bit determines the serial port 0 time base when receiving data in serial modes 1 or 3. Setting this bit forces Timer 2 into baud-rate generation mode. The timer operates from a divide-by-2 of the external clock. 0 = Timer 1 overflow is used to determine receiver baud rate for serial port 0. 1 = Timer 2 overflow is used to determine receiver baud rate for serial port 0. TCLK Bit 4 Transmit Clock Flag. This bit determines the serial port 0 time base when transmitting data in serial modes 1 or 3. Setting this bit forces Timer 2 into baud rate generation mode. The timer operates from a divide-by-2 of the external clock. 0 = Timer 1 overflow is used to determine transmitter baud rate for serial port 0. 1 = Timer 2 overflow is used to determine transmitter baud rate for serial port 0. EXEN2 Bit 3 Timer 2 External Enable. This bit enables the capture/reload function on the T2EX pin if Timer 2 is not generating baud rates for the serial port. 0 = Timer 2 ignores all external events at T2EX. 1 = Timer 2 captures or reload a value if a negative transition is detected on the T2EX pin. TR2 Bit 2 Timer 2 Run Control. This bit enables/disables the operation of Timer 2. Halting this timer preserves the current count in TH2, TL2. 0 = Timer 2 is halted. 1 = Timer 2 is enabled. C/T2 Bit 1 Counter/Timer Select. This bit determines whether Timer 2 functions as a timer or counter. Independent of this bit, Timer 2 runs at 2 clocks per tick when used in either baud-rate generator or clock-output mode. 0 = Timer 2 function as a timer. 1 = Timer 2 counts negative transitions on the T2 pin (P1.0). CP/RL2 Bit 0 Capture/Reload Select. This bit determines whether the capture or reload function is used for Timer 2. When set (= 1), Timer 2 captures occur when a falling edge is detected on T2EX(P1.1) if EXEN2 = 1. When clear (= 0), Timer 2 functions in an autoreload mode. An autoreload occurs following each overflow if RCLK or TCLK is set or if a falling edge is detected on T2EX if EXEN2 = 1. 0 = Autoreloads occur when Timer 2 overflows or a falling edge is detected on T2EX if EXEN2 = 1. 1 = Timer 2 captures occur when a falling edge is detected on T2EX if EXEN2 = 1. 35 _____________________________________________________________________________________________ Ultra-High-Speed Flash Microcontroller User’s Guide Timer 2 Mode (T2MOD) SFR C9h 7 — R-1 6 — R-1 5 — R-1 4 — R-1 3 — R-1 2 — R-1 1 T2OE RW-0 0 DCEN RW-0 R = Unrestricted read, W = Unrestricted write, -n = Value after reset T2MOD 7–2 Bits 7–2 Reserved. Read data is 1. T2OE Bit 1 Timer 2 Output Enable. This bit enables/disables the clock output function of the T2 pin (P1.0). When set (= 1), Timer 2 drives the T2 pin with a clock output if C/(T2CON.1) = 0. For this setting, Timer 2 rollovers do not cause interrupts. When clear (= 0), the T2 pin functions as either a standard port pin or as a counter input for Timer 2. DCEN Bit 0 Down Count Enable. This bit, in conjunction with the T2EX (P1.1) pin, controls the direction that Timer 2 counts in 16-bit autoreload mode. DCEN 1 1 0 T2EX 1 0 X DIRECTION Up Down Up Timer 2 Capture LSB (RCAP2L) SFR CAh 7 RCAP2L.7 RW-0 6 RCAP2L.6 RW-0 5 RCAP2L.5 RW-0 4 RCAP2L.4 RW-0 3 RCAP2L.3 RW-0 2 RCAP2L.2 RW-0 1 RCAP2L.1 RW-0 0 RCAP2L.0 RW-0 R = Unrestricted read, W = Unrestricted write, -n = Value after reset RCAP2L.7–0 Bits 7–0 Timer 2 Capture LSB. This register is used to capture the TL2 value when Timer 2 is configured in capture mode. RCAP2L is also used as the LSB of a 16-bit reload value when Timer 2 is configured in autoreload mode. Timer 2 Capture LSB (RCAP2H) SFR CBh 7 RCAP2H.7 RW-0 6 RCAP2H.6 RW-0 5 RCAP2H.5 RW-0 4 RCAP2H.4 RW-0 3 RCAP2H.3 RW-0 2 RCAP2H.2 RW-0 1 RCAP2H.1 RW-0 0 RCAP2H.0 RW-0 R = Unrestricted read, W = Unrestricted write, -n = Value after reset RCAP2H.7–0 Bits 7–0 Timer 2 Capture MSB. This register is used to capture the TH2 value when Timer 2 is configured in capture mode. RCAP2H is also used as the MSB of a 16-bit reload value when Timer 2 is configured in autoreload mode. Timer 2 LSB (TL2) SFR CCh 7 TL2.7 RW-0 6 TL2.6 RW-0 5 TL2.5 RW-0 4 TL2.4 RW-0 3 TL2.3 RW-0 2 TL2.2 RW-0 1 TL2.1 RW-0 0 TL2.0 RW-0 R = Unrestricted read, W = Unrestricted write, -n = Value after reset TL2.7–0 Bits 7–0 Timer 2 LSB. This register contains the least significant byte of Timer 2. _____________________________________________________________________________________________ 36 Ultra-High-Speed Flash Microcontroller User’s Guide Timer 2 MSB (TH2) SFR CDh 7 TH2.7 RW-0 6 TH2.6 RW-0 5 TH2.5 RW-0 4 TH2.4 RW-0 3 TH2.3 RW-0 2 TH2.2 RW-0 1 TH2.1 RW-0 0 TH2.0 RW-0 R = Unrestricted read, W = Unrestricted write, -n = Value after reset TH2.7–0 Bits 7–0 Timer 2 MSB. This register contains the most significant byte of Timer 2. Program Status Word (PSW) SFR D0h 7 CY RW-0 6 AC RW-0 5 F0 RW-0 4 RS1 RW-0 3 RS0 RW-0 2 OV RW-0 1 F1 RW-0 0 PARITY RW-0 R = Unrestricted read, W = Unrestricted write, -n = Value after reset CY Bit 7 Carry Flag. This bit is set if the last arithmetic operation resulted in a carry (during addition) or a borrow (during subtraction). Otherwise, it is cleared to 0 by all arithmetic operations. AC Bit 6 Auxiliary Carry Flag. This bit is set to 1 if the last arithmetic operation resulted in a carry into (during addition) or a borrow (during subtraction) from the high-order nibble. Otherwise, it is cleared to 0 by all arithmetic operations. F0 User Flag 0. This is a bit-addressable, general-purpose flag for software control. Bit 5 RS1, RS0 Bits 4-3 Register Bank Select 1–0. These bits select which register bank is addressed during register accesses. Register Bank Addresses RS1 0 0 1 1 RS0 0 1 0 1 REGISTER BANK 0 1 2 3 ADDRESS 00h – 07h 08h – 0Fh 10h – 17h 18h – 1Fh OV Bit 2 Overflow Flag. This bit is set to 1 if the last arithmetic operation resulted in a carry (addition), borrow (subtraction), or overflow (multiplication or division). Otherwise it is cleared to 0 by all arithmetic operations. F1 User Flag 1. This is a bit-addressable, general-purpose flag for software control. Bit 1 PARITY Bit 0 37 Parity Flag. This bit is set to 1 if the module-2 sum of the 8 bits of the accumulator is 1 (odd parity), and cleared to 0 on even parity. _____________________________________________________________________________________________ Ultra-High-Speed Flash Microcontroller User’s Guide Watchdog Control (WDCON) SFR D8h 7 SMOD_1 RW-0 6 POR RT-* 5 EPFI RW-0 4 PFI RW-* 3 WDIF RT-0 2 WTRF RW-* 1 EWT RT-* 0 RWT RT-0 R = Unrestricted read, W = Unrestricted write, T = Timed-access write only, -n = Value after reset, * = see description SMOD_1 Serial Modification. This bit controls the doubling of the serial port 1 baud rate in modes 1, 2, and 3. Bit 7 0 = Serial port 1 baud rate operates at normal speed. 1 = Serial port 1 baud rate is doubled. POR Bit 6 Power-On Reset Flag. This bit indicates whether the last reset was a power-on reset. This bit is typically interrogated following a reset to determine if the reset was caused by a power-on reset. It must be cleared by a timed access write before the next reset of any kind or user software may erroneously determine that another power-on reset has occurred. This bit is set following a poweron reset and unaffected by all other resets. This bit automatically cleared when the ROM loader is invoked. 0 = Last reset was from a source other than a power-on reset. 1 = Last reset was a power-on reset. EPFI Bit 5 Enable Power-Fail Interrupt. This bit enables/disables the ability of the internal bandgap reference to generate a power-fail interrupt when VCC falls below approximately 4.5V. While in stop mode, both this bit and the bandgap Select bit, BGS (EXIF.0), must be set to enable the power-fail interrupt. 0 = Power-fail interrupt disabled. 1 = Power-fail interrupt enabled during normal operation. Power-fail interrupt enabled in stop mode if BGS is set. PFI Bit 4 Power-Fail Interrupt Flag. When set, this bit indicates that a power-fail interrupt has occurred. This bit must be cleared in software before exiting the interrupt service routine, or another interrupt is generated. Setting this bit in software generate a power-fail interrupt, if enabled. This bit is automatically cleared when the ROM loader is invoked. WDIF Bit 3 Watchdog Interrupt Flag. This bit indicates if a watchdog timer event has occurred. The timeout period of the watchdog timer is controlled by the Watchdog Timer Mode Select bits (CKCON.7-6). The Watchdog Timer Interrupt Enable bit, EWDI (EIE.4), and Enable Watchdog Timer Reset bit, EWT (WDCON.1), determine what action is taken. This bit must be cleared in software before exiting the interrupt service routine, or another interrupt is generated. Setting this bit in software generates a watchdog interrupt if enabled. This bit can only be modified using a Timed Access Procedure. WTRF Bit 2 Watchdog Timer Reset Flag. When set, this bit indicates that a watchdog timer reset has occurred. It is typically interrogated to determine if a reset was caused by watchdog timer reset. It is cleared by a power-on reset but otherwise must be cleared by software before the next reset of any kind or software can erroneously determine that a watchdog timer reset has occurred. Setting this bit in software does not generate a watchdog timer reset. If the EWT bit is cleared, the watchdog timer has no effect on this bit. This bit is automatically cleared when the ROM loader is invoked. EWT Bit 1 Enable Watchdog Timer Reset. This bit enables/disables the generation of a watchdog timer reset 512 system clocks after the occurrence of a watchdog timeout. This bit can only be modified using a Timed Access Procedure and is unaffected by all other resets. The default power-on reset state of EWT is determined by Option Control Register bit 3 (OCR.3) located in flash memo ry. This bit will automatically be cleared when the ROM loader is invoked. 0 = A watchdog reset is not generated after a watchdog timeout 1 = A watchdog reset is generated 512 system clocks after a watchdog timeout unless RWT is strobed or EWT is cleared. _____________________________________________________________________________________________ 38 Ultra-High-Speed Flash Microcontroller User’s Guide RWT Bit 0 Reset Watchdog Timer. Setting this bit resets the watchdog timer count. This bit must be set using a Timed Access procedure before the watchdog timer expires, or a watchdog timer reset and/or interrupt is generated if enabled. The timeout period is defined by the Watchdog Timer Mode Select bits (CKCON.7-6). This bit is always be 0 when read. Accumulator (A or ACC) SFR E0h 7 ACC.7 RW-0 6 ACC.6 RW-0 5 ACC.5 RW-0 4 ACC.4 RW-0 3 ACC.3 RW-0 2 ACC.2 RW-0 1 ACC.1 RW-0 0 ACC.0 RW-0 R = Unrestricted read, W = Unrestricted write, -n = Value after reset ACC.7–0 Accumulator. This register serves as the accumulator for arithmetic operations. Bits 7–0 Extended Interrupt Enable (EIE) SFR E8h 7 — R-1 6 — R-1 5 — R-1 4 EWDI RW-0 3 EX5 RW-0 2 EX4 RW-0 1 EX3 RW-0 0 EX2 RW-0 R = Unrestricted read, W = Unrestricted write, -n = Value after reset Bits 7–5 Reserved. Read data is 1. EWDI Watchdog Interrupt Enable. This bit enables/disables the watchdog interrupt. Bit 4 0 = Disable the watchdog interrupt. 1 = Enable interrupt requests generated by the watchdog timer. EX5 External Interrupt 5 Enable. This bit enables/disables external interrupt 5. Bit 3 0 = Disable external interrupt 5. 1 = Enable interrupt requests generated by the INT5 pin. EX4 External Interrupt 4 Enable. This bit enables/disables external interrupt 4. Bit 2 0 = Disable external interrupt 4. 1 = Enable interrupt requests generated by the INT4 pin. EX3 External Interrupt 3 Enable. This bit enables/disables external interrupt 3. Bit 1 0 = Disable external interrupt 3. EX2 External Interrupt 2 Enable. This bit enables/disables external interrupt 2. 1 = Enable interrupt requests generated by the INT3 pin. Bit 0 0 = Disable external interrupt 2. 1 = Enable interrupt requests generated by the INT2 pin. B Register (B) SFR F0h 7 B.7 RW-0 6 B.6 RW-0 5 B.5 RW-0 4 B.4 RW-0 3 B.3 RW-0 2 B.2 RW-0 1 B.1 RW-0 0 B.0 RW-0 R = Unrestricted read, W = Unrestricted write, -n = Value after reset B.7–0 B Register. This register serves as a second accumulator for certain arithmetic operations. Bits 7–0 39 _____________________________________________________________________________________________ Ultra-High-Speed Flash Microcontroller User’s Guide Extended Interrupt Priority 1 (EIP1) SFR F1h 7 — R-1 6 — R-1 5 — R-1 4 MPWDI RW-0 3 MPX5 RW-0 2 MPX4 RW-0 1 MPX3 RW-0 0 MPX2 RW-0 R = Unrestricted read, W = Unrestricted write, -n = Value after reset Bits 7–5 Reserved. Read data is 1. MPWDI Bit 4 Most significant priority select bit for watchdog interrupt. Most significant bit of the bit pair MPWDI, LPWDI (EIP0.4) that designates priority level for the watchdog interrupt. MPX5 Bit 3 Most significant priority select bit for external interrupt 5. Most significant bit of the bit pair MPX5, LPX5 (EIP0.3) that designates priority level for external interrupt 5. MPX4 Bit 2 Most significant priority select bit for external interrupt 4. Most significant bit of the bit pair MPX4, LPX4 (EIP0.2) that designates priority level for external interrupt 4. MPX3 Bit 1 Most significant priority select bit for external interrupt 3. Most significant bit of the bit pair MPX3, LPX3 (EIP0.1) that designates priority level for external interrupt 3. MPX2 Bit 0 Most significant priority select bit for external interrupt 2. Most significant bit of the bit pair MPX2, LPX2 (EIP0.0) that designates priority level for external interrupt 2. Interrupt priority level for the above sources is assigned using one bit from register EIP1 (F1h) and one bit from EIP0 (F8h). The bit from EIP1 serves as the most significant bit and the bit from EIP0 serves as the least significant bit, in forming a 2-bit binary number. This number represents the pri ority level. Higher priority interrupts, when enabled, take precedence over lower priority sources. The power-fail warning interrupt source is assigned Priority Level 4. Most Significant Priority Select Bit Levels MP (EIP1.x) 0 0 1 1 LP (EIP0.x) 0 1 0 1 PRIORITY LEVEL 0 (natural priority) 1 2 3 (high priority) _____________________________________________________________________________________________ 40 Ultra-High-Speed Flash Microcontroller User’s Guide Extended Interrupt Priority 0 (EIP0) SFR F8h 7 — R-1 6 — R-1 5 — R-1 4 LPWDI RW-0 3 LPX5 RW-0 2 LPX4 RW-0 1 LPX3 RW-0 0 LPX2 RW-0 R = Unrestricted read, W = Unrestricted write, -n = Value after reset Bits 7–5 Reserved. Read data is 1. LPWDI Bit 4 Least significant priority select bit for watchdog interrupt. This is the least significant bit of the bit pair MPWDI (EIP1.4), LPWDI that designates priority level for the watchdog interrupt. LPX5 Bit 3 Least significant priority select bit for external interrupt 5. This is the least significant bit of the bit pair MPX5 (EIP1.3), LPX5 that designates priority level for external interrupt 5. LPX4 Bit 2 Least significant priority select bit for external interrupt 4. This is the least significant bit of the bit pair MPX4 (EIP1.2), LPX4 that designates priority level for external interrupt 4. LPX3 Bit 1 Least significant priority select bit for external interrupt 3. This is the least significant bit of the bit pair MPX3 (EIP1.1), LPX3 that designates priority level for external interrupt 3. LPX2 Bit 0 Least significant priority select bit for external interrupt 2. This is the least significant bit of the bit pair MPX2 (EIP1.0), LPX2 that designates priority level for external interrupt 2. Least Significant Priority Select Bit Levels MP (IP1.X) 0 0 1 1 41 LP (IP0.X) 0 1 0 1 PRIORITY LEVEL 0 (natural priority) 1 2 3 (high priority) _____________________________________________________________________________________________ Ultra-High-Speed Flash Microcontroller User’s Guide SECTION 5: CPU TIMING The timing of the ultra-high-speed microcontroller is the area with the greatest departure from the original 8051 series. This section explains the timing and compares it to the original 8051. OSCILLATOR The ultra-high-speed microcontroller provides an on-chip oscillator circuit that can be driven by an external crystal or by an off-chip TTL clock source. The oscillator circuit provides the internal clocking signals to the on-chip CPU and I/O circuits. In many designs, a crystal is the preferred clock source. Figure 5-1 shows the required connections for a crystal and typical capacitor values. Some designs may prefer using an off-chip clock oscillator as the primary clock source. This configuration is illustrated in Figure 5-2. When using an off-chip oscillator, the duty cycle becomes important. As near as possible, a 50% duty cycle should be supplied. XTAL1 This pin is the input to an inverting high-gain amplifier. It also serves as the input for an off-chip oscillator. Note that, when using an offchip oscillator, XTAL2 is left unconnected. XTAL2 This pin is the output of the crystal amplifier. It can be used to distribute the clock to other devices on the same board. If using a crystal, the loading on this pin should be kept to a minimum, especially capacitive loading. OSCILLATOR CHARACTERISTICS The ultra-high-speed microcontroller was designed to operate with a parallel resonant AT-cut crystal. The crystal should resonate at the desired frequency in its primary or fundamental mode. The oscillator employs a high-gain amplifier to assure a clean waveform at high frequency. Due to the high-performance nature of the product, both clock edges are used for internal timing. Therefore, the duty cycle of the clock source is of importance. A crystal circuit balances itself automatically. Thus crystal users do not need to take extra precautions concerning duty cycle. CRYSTAL SELECTION The ultra-high-speed microcontroller family was designed to operate with fundamental mode crystals for improved stability. Although most high-speed (i.e., greater than 25MHz) crystals operate from their third overtone, fundamental mode crystals are available from most major crystal suppliers. Designers are cautioned to ensure that high-speed crystals being specified for use in their application do operate at the rated frequency in their fundamental mode. The use of a third overtone crystal will typically result in oscillation rates at one-third the desired speed. ULTRA-HIGH-SPEED MICROCONTROLLER XTAL1 18pF 18pF Figure 5-1. Crystal Connection ULTRA-HIGH-SPEED MICROCONTROLLER TO INTERNAL CIRCUITS CLOCK OSCILLATOR XTAL2 XTAL1 TO INTERNAL CIRCUITS XTAL2 Figure 5-2. Clock Source Input _____________________________________________________________________________________________ 42 Ultra-High-Speed Flash Microcontroller User’s Guide SYSTEM CLOCK DIVIDE CONTROL The ultra-high-speed microcontroller provides the ability to speed up or slow down the system clock that is used internally by the CPU. The system clock divide ratio can be configured to 0.25 (4X multiply mode), 0.5 (2X multiply mode), 1 (default), or 1024 (power management mode) and is controlled by the CD1:0 bits (PMR.7, PMR.6). To use the crystal multiply mode, the multiplier circuit must be prompted to warmup in the desired 4X or 2X configuration. The 4X/2X bit defines the crystal multiplying factor. This bit can be altered only from the divide-by-1 (default) mode, while the crystal multiplier is disabled (CTM = 0). Once the 4X/2X bit has been configured as desired, setting the CTM bit (PMR.4) initiates the crystal multiplier warmup period. The CTM bit can only be altered when the CD1:0 bits are set to divide-by-1 mode and the RGMD bit is cleared to 0. During the multiplier warmup period the CKRY bit remains cleared and the CD1:0 clock control bits cannot be set to 00b. When the crystal multiplier circuit has completed the warmup and is ready for use, the CKRY (EXIF.3) bit set to a logic 1. At this point, the CD1:0 bits can be modified to select the multiplier output for use as the internal system clock. Specifics of hardware restrictions associated with the use of the 4X/2X CTM, CKRY, CD1, and CD0 bits are outlined in the SFR descriptions. The prescribed sequence for selecting the the crystal multiplier is as follows: 1) Ensure that the current clock mode is set to divide-by-1 (CD1:0 = 10b) and that RGMD (EXIF.2) = 0. 2) Clear the CTM bit. 3) Put the 4X/2X bit in the desired state. 4) Set the CTM bit. 5) Poll for the CKRY (EXIF.3) bit to be set (= 1). This takes ~65536 external clock cycles. 6) Set CD1:0 = 00b. The frequency multiplier is engaged on the memory cycle following the writing of these bits. An additional circuit provides a divide-by-1024 clock source that can be selected as the internal system clock. When programmed to the divide-by-1024 mode, the user may wish to set the switchback bit (PMR.5: SWB) to force the clock divide control bits automatically back to the divide-by-1 mode whenever the system detects an externally enabled interrupt or an incoming serial port start bit. This automatic switchback is only enabled during divide-by-1024 mode, and all other clock control settings are unaffected by interrupts and serial port activity. The power management mode is detailed further in Section 7 (Power Management). It is important to remember that changing the system clock frequency affects all aspects of system operation, including timers and serial port baud rates. These effects are detailed in Section 11 (Programmable Timers) and Section 12 (Serial I/O). The following diagram illustrates the system clock control function. 4X / 2X CTM CRYSTAL OSCILLATOR CLOCK MULTIPLIER MUX 00 01, 10 11 DIVIDE 1024 SYSTEM CLOCK RING OSCILLATOR RING ENABLE CD0 CD1 SELECTOR Figure 5-3. System Clock Sources 43 _____________________________________________________________________________________________ Ultra-High-Speed Flash Microcontroller User’s Guide MUX SYSTEM CLOCK 00 01, 10 11 EXECUTION INTERNAL EXTERNAL NON PAGE PAGE MODE 1 (1-CYCLE) PAGE MODE 1 (2-CYCLE) PAGE MODE 1 (3-CYCLE) PAGE MODE 2 MEMORY CYCLE SYSCLK/1 SYSCLK/4 SYSCLK/1 SYSCLK/2 SYSCLK/4 SYSCLK/2 (PROG) SYSCLK/4 (DATA) MEMORY CYCLE Figure 5-4. Instruction Memory Cycle Determination INSTRUCTION TIMING The ultra-high-speed microcontroller executes the industry standard 8051 instruction set. Each instruction requires a minimum of one memory cycle of execution time, and may require as many as ten memory cycles (DIV AB only). The number of memory cycles required to execute any given 8051 instruction is documented at the end of this section and can be found in Section 14 (Instruction Set Details). A memory cycle is the basic timing unit for the ultra-high-speed microcontroller. If internal program code is being executed, a memory cycle always consists of one system clock. If external program code is being executed, a memory cycle is then composed of 1, 2, or 4 system clocks, as defined by the external bus configuration (non page mode, page mode 1, or page mode 2). Calculating the number of external crystal or oscillator clock periods (tCLCL) per memory cycle additionally depends upon how the user has configured the system clock as a function of the external clock. The system clock control function was covered earlier in the section. As an example, if the crystal multiplier is used to generate a system clock frequency four times the frequency of the external clock source, a nonpaged mode external memory cycle would consist of one external clock. All instructions are coded within an 8-bit field called an op code. This single byte must be fetched from program memory. The CPU decodes the op code to determine what action the microcontroller must take or whether additional information is needed from memory. If no other memory is needed, then only 1 byte was required. Thus, the instruction is called a 1-byte instruction. In some cases, more data is needed. These are 2- or 3-byte instructions. Single-Byte Instructions A single-byte instruction can require anywhere between one and ten memory cycles to execute. When the execution cycle count exceeds the byte count, the program counter must stall until instruction execution is completed. All MOVX data memory access instructions have a single-byte op code, but require more memory cycles so that data may be accessed. The MOVX instruction timing is covered in Section 6 (Memory Access). Following are examples of single byte instructions, each requiring a different number of execution cycles: RRC A DA A RET MUL AB DIV AB OPCODE 13h D4h 22h A4h 84h NO. OF CYCLES 1 2 3 9 10 2-Byte Instructions All 2-byte instructions require a minimum of two cycles, since fetching each byte requires a separate memory access. The first byte is the instruction op code that is decoded by the CPU. The second byte is normally an operand, or it can specify the location of the operand. For example, “ADD A, direct” is a 2-byte, two cycle instruction where the second byte specifies the direct address location of the operand. Due to internal access restrictions, certain direct addressing instructions require one extra memory cycle when operating on the PSW, SP, DPS, IE, EIE, IP0, IP1, EIP0, or EIP1 register. Following are examples of these and other 2-byte instructions: ADD A, direct ADD A, #data SJMP rel OPCODE 25h 24h 80h OPERAND/LOCATION <addr7-0> <data7-0> <addr7-0> NO. OF CYCLES 2 2 3 _____________________________________________________________________________________________ 44 Ultra-High-Speed Flash Microcontroller User’s Guide ANL direct, A ORL direct, A DJNZ Rn, direct 52h 42h D8h-DFh <addr7-0> <addr7-0> <addr7-0> 2 or 3 2 or 3 4 3-Byte Instructions Three-byte instructions require a minimum of three cycles since each byte fetch requires one memory cycle. The first byte, the opcode, instructs the CPU on how to handle the next two bytes. Most 3-byte instructions involve comparison or branching, but not all. Just like the 2-byte instructions, certain 3-byte instructions may require 1 extra memory cycle when operating on the PSW, SP, DPS, IE, EIE, IP0, IP1, EIP0, or EIP1 register. Following are examples of 3-byte instructions. LJMP addr16 MOV dptr, #data16 MOV direct, direct JBC bit, rel DJNZ direct, rel OPCODE 02h 90h 85h 10h D5h OPERAND(s)/LOCATION(s) <addr15-8><addr7-0> <data15-8><data7-0> <addr7-0><addr7-0> <addr7-0><addr7-0> <addr7-0><addr7-0> NO. Of CYCLES 3 3 3 or 4 4 or 5 5 NONPAGE MODE EXTERNAL TIMING The ultra-high-speed flash microcontroller defaults to a nonpage mode external memory interface. The nonpage mode bus structure requires four system clock cycles per memory cycle. In the nonpage mode, the ALE signal latches the address LSB on each program fetch. When the cycle count of an instruction exceeds the byte count, “dummy” fetches are performed each cycle until instruction execution is complete. The following diagrams demonstrate the basic timing for nonpage mode instruction execution. The first diagram below shows the execution of the DA A instruction (1 byte, two cycles) followed by execution of the RRC A (1 byte, one cycle) instruction. When a code fetch is made from a different 256-byte page, the new address MSB is presented on port 2. The second diagram below shows the execution of the INC direct instruction (2 bytes) for the cases where an extra memory cycle is not (INC DPL) and is (INC DPS) required. NONPAGE MODE: DA A – RRC A SYSCLK ALE PSEN PORT 2 MSB ADDRESS MSB ADDRESS PORT 0 LSB D4 LSB 13 LSB 13 LSB DA A RRC A NONPAGE MODE: INC DIRECT (2 CYCLES) – INC DIRECT (3 CYCLES) SYSCLK ALE PSEN MSB ADDRESS PORT 2 PORT 0 05 LSB 82 INC DPL 45 LSB 05 LSB 86 LSB E0 INC DPS _____________________________________________________________________________________________ LSB E0 Ultra-High-Speed Flash Microcontroller User’s Guide NONPAGE MODE EXTERNAL TIMING (CONTINUED) The first diagram below illustrates an ACALL instruction (2 bytes, two cycles) with a destination address residing on a different 256byte page. This is indicated only by the MSB address change on port 2. The memory cycle duration remains constant. The second diagram below shows execution of the RET instruction (1 byte, three cycles). Because the cycle count of the RET instruction exceeds the byte count, two stall cycles (“dummy” fetches) are inserted to allow execution to complete. In this example, the return address and the RET instruction are on different 256-byte pages (signified by the MSB address change on port 2). NONPAGE MODE: ACALL – NOP SYSCLK ALE PSEN MSB ADDRESS PORT 2 PORT 0 LSB 71 LSB MSB ADDRESS 33 LSB 00 LSB ACALL NOP NONPAGE MODE: RET – NOP SYSCLK ALE PSEN MSB ADDRESS PORT 2 PORT 0 LSB 22 LSB MSB ADDRESS LSB RET LSB 00 LSB NOP _____________________________________________________________________________________________ 46 Ultra-High-Speed Flash Microcontroller User’s Guide PAGE MODE 1 EXTERNAL TIMING—PAGES 1:0 = 10b (FOUR CYCLES) The page mode 1 external bus structure multiplexes port 2 to provide the address MSB and LSB. Data transactions occur exclusively on port 0. ALE is used to latch the address MSB only when needed, and PSEN serves as the enable for external program memory. Page mode 1 must be initiated by internal code memory. To invoke 4-cycle page mode 1 operation, the PAGES1:0 bits must be set to 10b, followed by the setting of the PAGEE bit. In the four-cycle page mode 1 configuration, a page-hit memory cycle is four system clocks in length, while the page-miss memory cycle requires eight system clocks. The first diagram below shows the fetch of the DA A instruction (1 byte, two cycles) during a page-miss memory cycle as would occur when a page boundary is crossed. Like nonpage mode operation, a “dummy” or stall cycle must then be inserted for the single-byte DA A instruction, since it requires two cycles of execution time. After stalling for one cycle, the real fetch of the RRC A instruction takes place. The second diagram below illustrates the fetch of the DA A instruction as the last byte of a 256-byte page. In this case, the stall cycle needed in executing the DA A instruction coincides with a page-miss memory cycle instead of a page hit (as in the first diagram). FOUR-CYCLE PAGE MODE 1: (PAGE MISS) – DA A – RRC A SYSCLK ALE PSEN MISS PORT 2 MSB ADDRESS STALL HIT LSB ADDRESS PORT 0 HIT LSB ADDRESS D4 13 LSB ADDRESS 13 DA A RRC A NONPAGE MODE: DA A – (PAGE MISS) – RRC A SYSCLK ALE PSEN STALL HIT PORT 2 PORT 0 HIT MISS LSB ADDRESS LSB ADDRESS MSB ADDRESS D4 13 HIT LSB ADDRESS 13 DA A 47 LSB ADDRESS _____________________________________________________________________________________________ RRC A Ultra-High-Speed Flash Microcontroller User’s Guide PAGE MODE 1 EXTERNAL TIMING—PAGES 1:0 = 10b (FOUR CYCLES) (CONTINUED) The first diagram below shows execution of the INC direct instruction (2 byte, two or three cycles) for the cases where an extra memory cycle is not (INC DPL) and is (INC DPS) required. The second diagram illustrates execution of the ACALL instruction whose destination address is on a different 256-byte page. Therefore, the second execution cycle of the ACALL instruction is a page-miss memory cycle that requires an ALE signal toggle to be used in order to latch a new address MSB. FOUR-CYCLE PAGE MODE 1: INC DIRECT (TWO CYCLES) – INC DIRECT (THREE CYCLES) SYSCLK ALE PSEN HIT PORT 2 HIT MSB ADDRESS PORT 0 05 HIT LSB ADDRESS LSB ADDRESS 05 82 HIT STALL LSB ADDRESS 86 INC DPL E0 E0 INC DPS FOUR-CYCLE PAGE MODE: ACALL – (PAGE MISS) SYSCLK ALE PSEN HIT PORT 2 PORT 0 MISS HIT LSB ADDRESS 71 LSB ADDRESS MSB ADDRESS HIT LSB ADDRESS LSB ADDRESS 33 ACALL _____________________________________________________________________________________________ 48 Ultra-High-Speed Flash Microcontroller User’s Guide PAGE MODE 1 EXTERNAL TIMING—PAGES 1:0 = 10b (FOUR CYCLES) (CONTINUED) The two diagrams below demonstrate the execution of the RET (1 byte, three cycles) instruction. In the first diagram, the return address resides on the same 256-byte page as that of the executed RET instruction. Two stall cycles are inserted followed by a page-hit memory cycle. In the second diagram, the return address is on a different 256-byte page from where the RET instruction was executed. In this case, two stall cycles are inserted, followed by a page-miss memory cycle. FOUR-CYCLE PAGE MODE 1: RET SYSCLK ALE PSEN HIT PORT 2 STALL LSB ADDRESS PORT 0 STALL LSB ADDRESS HIT HIT LSB ADDRESS LSB ADDRESS 22 RET FOUR-CYCLE PAGE MODE 1: RET – (PAGE MISS) SYSCLK ALE PSEN HIT PORT 2 PORT 0 STALL LSB ADDRESS MISS STALL LSB ADDRESS MSB ADDRESS 22 ACALL 49 _____________________________________________________________________________________________ LSB ADDRESS Ultra-High-Speed Flash Microcontroller User’s Guide PAGE MODE 1 EXTERNAL TIMING—PAGES 1:0 = 01b (TWO CYCLES) The page mode 1 external bus structure multiplexes port 2 to provide the address MSB and LSB. Data transactions occur exclusively on port 0. ALE is used to latch the Address MSB only when needed, and PSEN serves as the enable for external program memory. To invoke two-cycle page mode 1 operation, the PAGES1:0 bits must be set to 01b, followed by the setting of the PAGEE bit. In the two-cycle page mode 1 configuration, a page-hitmemory cycle is two system clocks in length, while the page-miss memory cycle requires four system clocks. The first diagram below shows the fetch of the CLR C instruction (1 byte, one cycle) during a page-miss memory cycle, followed by the fetch of the RRC A instruction (1 byte, one cycle) during a page-hit memory cycle. Since the next instruction, XCH A, @R0 (1 byte, three cycles), requires three memory cycles to execute, two stall cycles must be inserted for it to complete prior to the next instruction being read. The second diagram below illustrates the LJMP (3 bytes, three cycles) instruction, whose destination address is on a different 256-byte page than the LJMP instruction, thus resulting in a page-miss memory cycle. TWO-CYCLE PAGE MODE 1: (PAGE MISS) – CLR C – RRC A – XCH A, @R0 SYSCLK ALE PSEN MISS PORT 2 MSB HIT HIT LSB PORT 0 LSB C3 STALL LSB 13 CLR C STALL HIT LSB ADDRESS HIT LSB C6 RRC A XCH A, @R0 TWO-CYCLE PAGE MODE 1: (PAGE MISS) – LJMP ADDR16 – (PAGE MISS) SYSCLK ALE PSEN MISS PORT 2 PORT 0 MSB HIT HIT LSB LSB 02 LSB 50 HIT MISS MSB LSB LSB HIT LSB 00 LJMP ADDR16 _____________________________________________________________________________________________ 50 Ultra-High-Speed Flash Microcontroller User’s Guide PAGE MODE 1 EXTERNAL TIMING—PAGES 1:0 = 00b (ONE CYCLE) The page mode 1 external bus structure multiplexes port 2 to provide the address MSB and LSB. Data transactions occur exclusively on port 0. ALE is used to latch the address MSB only when needed, and PSEN serves as the enable for external program memory. Note that the one-cycle configuration differs slightly from the two-cycle and four-cycle configurations of the page mode 1 bus structure in that PSEN does not toggle for consecutive page hits, but stays in the active-low state. To invoke one-cycle page mode 1 operation, the PAGES 1:0 bits must be set to 00b, followed by the setting of the PAGEE bit. In the 1-cycle Page Mode 1 configuration, a page-hit memory cycle is one system clock in length, while the page-miss memory cycle requires two system clocks. In the following first diagram, the CLR C (1 byte, one cycle) instruction fetch occurs during a page-miss memory cycle, followed by the RRC A instruction (1 byte, 1 cycle) instruction fetch during a page-hit memory cycle. The MUL AB (1 byte, nine cycles) instruction, which occurs next, requires that the program counter be stalled for eight additional memory cycles so that execution can complete. In a similar fashion, the DA A (1 byte, two cycles) instruction, which follows the multiply, requires that one stall be inserted. The second diagram illustrates the memory cycle dependence of some direct instructions on the SFR addressed. The ORL direct, A is shown for cases where P1 and IE are being addressed. ONE-CYCLE PAGE MODE 1: (PAGE MISS) – CLR C – MUL AB – DA A – NOP SYSCLK ALE PSEN MISS HIT HIT STALLS HIT STALL HIT HIT LSB ADDRESS PORT 2 C3 PORT 0 13 A4 D4 CLR C RRC A 00 MUL AB DA A NOP ONE-CYCLE PAGE MODE 1: (PAGE MISS) – ORL DIRECT, A (TWO CYCLES) – ORL DIRECT, A (THREE CYCLES) – NOP SYSCLK ALE PSEN MISS HIT HIT HIT HIT STALL HIT PORT 2 PORT 0 45 90 ORL, P1, A 51 HIT HIT HIT HIT HIT LSB 45 A8 MUL AB 00 NOP _____________________________________________________________________________________________ HIT HIT HIT Ultra-High-Speed Flash Microcontroller User’s Guide PAGE MODE 1 EXTERNAL TIMING—PAGES 1:0 = 00b (ONE CYCLE) (CONTINUED) The first diagram below illustrates the JBC bit, rel (3 bytes, four cycles) instruction for the case where the tested bit is clear and the jump is not taken. Note that one stall cycle must be inserted since the cycle count exceeds the byte count by one. The RET (1 byte, three cycles) instruction that follows requires insertion of two stall cycles. In this example, the return address is on a different 256-byte page than the RET instruction, thus resulting in a page-miss memory cycle. The MOV direct, #data (3 bytes, three cycles) executed next provides an example of an instruction not requiring any stall cycles. The second diagram shows the same JBC bit, rel instruction for the case where the tested bit is set and the jump is taken. Since the bit must be cleared and involves one of the special registers (PSW, SP, DPS, IE, EIE, IP0, IP1, EIP0, EIP1), a fifth memory cycle is required. For this example, the jump taken by the JBC instruction crosses a 256-byte page boundary, while the RET instruction stays on the same page. 1-CYCLE PAGE MODE 1: (PAGE MISS) – JBC bit, rel (4 CYCLES) – RET – (PAGE MISS) – MOV direct, data SYSCLK ALE PSEN MISS HIT HIT HIT stall stall LSB PORT2 D1 10 PORT0 1D stall MISS HIT HIT HIT HIT HIT LSB 75 22 RET JBC PSW.1, $+20h HIT 55 90 MOV P1,#55h 1-CYCLE PAGE MODE 1: (PAGE MISS) – JBC bit, rel (5 CYCLES) – (PAGE MISS) -- RET – MOV direct, data SYSCLK ALE PSEN MISS HIT HIT stall MISS stall stall HIT HIT HIT HIT HIT HIT LSB PORT2 PORT0 stall 10 D1 1D JBC PSW.1, $+20h 22 75 RET 90 55 MOV P1,#55h _____________________________________________________________________________________________ 52 Ultra-High-Speed Flash Microcontroller User’s Guide PAGE MODE 2 EXTERNAL TIMING—PAGES 1:0 = 11b The page mode 2 external bus structure multiplexes port 2 between address MSB and data. The address LSB is provided exclusively on port 0. ALE is used to latch the address MSB only when needed, and PSEN serves as the enable for external program memory. To invoke page mode 2 operation, the PAGES 1:0 bits must be set to 11b, followed by the setting of the PAGEE bit. In the page mode 2 configuration, a page-hit program memory cycle is two system clocks in length, while the page-miss program memory cycle requires four system clocks. All data memory cycles are four system clocks in length. The first diagram below shows the fetch of the CLR C instruction (1 byte, 1 cycle) during a page-miss memory cycle, followed by the fetch of the RRC A instruction (1 byte, one cycle) during a page-hit memory cycle. The next instruction, XCH A, @R0 (1 byte, three cycles), requires three memory cycles to execute, so two stall cycles must be inserted for it to complete prior to the next instruction being read. The second diagram below illustrates the LJMP (3 bytes, three cycles) instruction, whose destination address is on a different 256-byte page than the LJMP instruction, thus resulting in a page-miss memory cycle. PAGE MODE 2: (PAGE MISS) – CLR C – RRC A – XCH A, @R0 SYSCLK ALE PSEN MISS PORT2 MSB HIT C3 LSB PORT0 stall HIT 13 LSB stall HIT C6 LSB CLR C HIT LSB LSB Address RRC A XCH A, @R0 PAGE MODE 2: (PAGE MISS) – LJMP addr16 – (PAGE MISS) SYSCLK ALE PSEN MISS PORT2 PORT0 MSB HIT 02 LSB HIT 50 LSB MISS 00 LSB HIT HIT MSB LSB LSB LSB LJMP addr16 COMPARISON TO THE 8051 The original 8051 needed 12 clocks per machine cycle and most instructions executed in either one or two machine cycles. Thus, except for the MUL and DIV instructions, the 8051 used either 12 or 24 clocks for each instruction. Furthermore, each machine cycle in the 8051 used two memory fetches. In many cases the second fetch was a dummy, and the extra clock cycles were wasted. The ultra-high-speed microcontroller uses one clock per memory (or machine) cycle. Where there were primarily one- and two-cycle instructions before, an instruction on the ultra-high-speed microcontroller may take between one and ten cycles. The divide instruction, for example, requires 10 cycles. Note however, that the 10 cycles needed for the DIV AB instruction can be executed at one clock per cycle (10 x 1 = 10 total clock cycles). The instruction is executed 4.8 times faster than the original 8051 architecture which required four cycles at a rate of 12 clocks per cycle (4 x 12 = 48 total clock cycles). Each instruction is at least four times faster, with the highest throughput improvement being 24 times that of the original 8051 architecture. 53 _____________________________________________________________________________________________ Ultra-High-Speed Flash Microcontroller User’s Guide Table 5-1 shows each instruction, the number of clocks used in the ultra-high-speed microcontroller, and the number used in the 8051 for comparison. The factor by which the ultra-high-speed microcontroller improves on the 8051 is shown as the speed advantage. A speed advantage of 12 means that the ultra-high-speed microcontroller performs the same instruction 12 times faster than the original 8051. Table 5-2 provides a summary by instruction type. Note that many of the instructions provide multiple op codes. As an example, the ADD A, Rn instruction can act on one of eight working registers. There are eight op codes for this instruction because it can be used on eight independent locations. Table 5-2 shows totals for both number of instructions and number of op codes. Averages are provided in the tables. However, the real speed improvement seen in any system depends on the instruction mix. Table 5-1. INSTRUCTION TIMING COMPARISON Ultra-high-speed microcontroller is abbreviated as UHSM. UHSM UHSM 8051 8051 UHSM vs. HEX CLOCK TIME CLOCK TIME 8051 SPEED INSTRUCTION CODE CYCLES @ 25MHz CYCLES @ 25MHz ADVANTAGE ADD A, Rn 28..2F 1 40 ns 12 480 ns 12 25 2 80 ns 12 480 ns 6 ADD A, direct ADD A, @Ri ADD A, #data ADDC A, Rn ADDC A, direct ADDC A, @Ri ADDC A, #data SUBB A, Rn SUBB A, direct 26..27 2 80 ns 12 480 ns 6 24 2 80 ns 12 480 ns 6 38..3F 1 40 ns 12 480 ns 12 35 2 80 ns 12 480 ns 6 36..37 2 80 ns 12 480 ns 6 34 2 80 ns 12 480 ns 6 98..9F 1 40 ns 12 480 ns 12 95 2 80 ns 12 480 ns 6 96..97 2 80 ns 12 480 ns 6 94 2 80 ns 12 480 ns 6 INC A 04 1 40 ns 12 480 ns 12 INC Rn 08..0F 1 40 ns 12 480 ns 12 05 2 80 ns 12 480 ns 6 06..07 2 80 ns 12 480 ns 6 SUBB A, @Ri SUBB A, #data INC direct INC @Ri INC DPTR A3 1 40 ns 24 960 ns 24 DEC A 14 1 40 ns 12 480 ns 12 DEC Rn 18..1F 1 40 ns 12 480 ns 12 DEC direct 15 2 80 ns 12 480 ns 6 DEC @Ri 16..17 2 80 ns 12 480 ns 6 MUL AB A4 9 360 ns 48 960 ns 5.33 DIV AB 84 10 400 ns 48 960 ns 4.80 DA A D4 2 80 ns 12 480 ns 6 58..5F 1 40 ns 12 480 ns 12 55 2 80 ns 12 480 ns 6 ANL A, Rn ANL A, direct ANL A, @Ri 56..57 2 80 ns 12 480 ns 6 ANL A, #data 54 2 80 ns 12 480 ns 6 ANL direct, A 52 2 80 ns 12 480 ns 6 ANL direct, #data 53 3 120 ns 24 960 ns 8 _____________________________________________________________________________________________ 54 Ultra-High-Speed Flash Microcontroller User’s Guide UHSM UHSM 8051 8051 UHSM vs. HEX CLOCK TIME CLOCK TIME 8051 SPEED INSTRUCTION CODE CYCLES @ 25MHz CYCLES @ 25MHz ADVANTAGE ORL A, Rn 48..4F 1 40 ns 12 480 ns 12 45 2 80 ns 12 480 ns 6 46..47 2 80 ns 12 480 ns 6 ORL A, #data 44 2 80 ns 12 480 ns 6 ORL direct, A 42 2 80 ns 12 480 ns 6 ORL direct, #data 43 3 120 ns 24 960 ns 8 68..6F 1 40 ns 12 480 ns 12 65 2 80 ns 12 480 ns 6 66..67 2 80 ns 12 480 ns 6 ORL A, direct ORL A, @Ri XRL A, Rn XRL A, direct XRL A, @Ri XRL A, #data 64 2 80 ns 12 480 ns 6 XRL direct, A 62 2 80 ns 12 480 ns 6 XRL direct, #data 63 3 120 ns 24 960 ns 8 CLR A E4 1 40 ns 12 480 ns 12 CPL A F4 1 40 ns 12 480 ns 12 RL A 23 1 40 ns 12 480 ns 12 RLC A 33 1 40 ns 12 480 ns 12 RR A 03 1 40 ns 12 480 ns 12 RRC A 13 1 40 ns 12 480 ns 12 SWAP A C4 1 40 ns 12 480 ns 12 MOV A, Rn MOV A, direct MOV A, @Ri MOV A, #data E8..EF 1 40 ns 12 480 ns 12 E5 2 80 ns 12 480 ns 6 E6..E7 2 80 ns 12 480 ns 6 74 2 80 ns 12 480 ns 6 MOV Rn, A F8..FF 1 40 ns 12 480 ns 12 MOV Rn, direct A8..AF 2 80 ns 24 960 ns 12 MOV Rn, #data 78..7F 2 80 ns 12 480 ns 6 MOV direct, A F5 2 80 ns 12 480 ns 6 MOV direct, Rn 88..8F 2 80 ns 24 960 ns 12 85 3 120 ns 24 960 ns 8 MOV direct, direct MOV direct, @Ri 86..87 2 80 ns 24 960 ns 12 75 3 120 ns 24 960 ns 8 MOV @Ri, A F6..F7 1 40 ns 12 480 ns 12 MOV @Ri, direct A6..A7 2 80 ns 24 960 ns 12 MOV @Ri, #data MOV direct, #data 76..77 2 80 ns 12 480 ns 6 MOV DPTR, #data 16 90 3 120 ns 24 960 ns 8 MOVC A, @A+DPTR 93 3 120 ns 24 960 ns 8 MOVC A, @A+PC MOVX A, @Ri MOVX A, @DPTR MOVX @Ri, A 55 83 3 120 ns 24 960 ns 8 E2..E3 2 80 ns 24 960 ns 12 E0 2 80 ns 24 960 ns 12 F2..F3 2 80 ns 24 960 ns 12 _____________________________________________________________________________________________ Ultra-High-Speed Flash Microcontroller User’s Guide UHSM UHSM 8051 8051 UHSM vs. HEX CLOCK TIME CLOCK TIME 8051 SPEED INSTRUCTION CODE CYCLES @ 25MHz CYCLES @ 25MHz ADVANTAGE MOVX @DPTR, A F0 2 80 ns 24 960 ns 12 PUSH direct C0 2 80 ns 24 960 ns 12 POP direct D0 2 80 ns 24 960 ns 12 XCH A, Rn C8..CF 2 80 ns 12 480 ns 6 C5 3 120 ns 12 480 ns 4 XCH A, @Ri C6..C7 3 120 ns 12 480 ns 4 XCHD A, @Ri XCH A, direct D6..D7 3 120 ns 12 480 ns 4 CLR C C3 1 40 ns 12 480 ns 12 CLR bit C2 2 80 ns 12 480 ns 6 SETB C D3 1 40 ns 12 480 ns 12 SETB bit D2 2 80 ns 12 480 ns 6 CPL C B3 1 40 ns 12 480 ns 12 CPL bit B2 2 80 ns 12 480 ns 6 ANL C, bit 82 2 80 ns 24 960 ns 12 ANL C, B0 2 80 ns 24 960 ns 12 ORL C, bit 72 2 80 ns 24 960 ns 12 ORL C, A0 2 80 ns 24 960 ns 12 MOV C, bit A2 2 80 ns 12 480 ns 6 MOV bit, C 92 2 80 ns 24 960 ns 12 Byte 1 2 80 ns 24 960 ns 12 12 3 120 ns 24 960 ns 8 ACALL addr 11 Hex codes = 11, 31, 51, Hex code 71, 91, B1, D1, or F1 LCALL addr 16 RET 22 3 120 ns 24 960 ns 8 RETI 32 3 120 ns 24 960 ns 8 Byte 1 2 80 ns 24 960 ns 12 02 3 120 ns 24 960 ns 8 AJMP addr 11 Hex code = 01, 21, 41, Hex code 61, 81, A1, C1, or E1 LJMP addr 16 JMP @A+DPTR 73 3 120 ns 24 960 ns 8 SJMP rel 80 3 120 ns 24 960 ns 8 JZ rel 60 3 120 ns 24 960 ns 8 JNZ rel 70 3 120 ns 24 960 ns 8 JC rel 40 3 120 ns 24 960 ns 8 JNC rel 50 3 120 ns 24 960 ns 8 JB bit, rel 20 4 160 ns 24 960 ns 6 JNB bit, rel 30 4 160 ns 24 960 ns 6 JBC bit, rel 10 4 160 ns 24 960 ns 6 CJNE A, direct, rel B5 5 200 ns 24 960 ns 4.8 _____________________________________________________________________________________________ 56 Ultra-High-Speed Flash Microcontroller User’s Guide UHSM UHSM 8051 8051 UHSM vs. HEX CLOCK TIME CLOCK TIME 8051 SPEED INSTRUCTION CODE CYCLES @ 25MHz CYCLES @ 25MHz ADVANTAGE CJNE A, #data, rel B4 4 160 ns 24 960 ns 6 CJNE Rn, #data, rel B8..BF 4 160 ns 24 960 ns 6 CJNE @Ri, #data, rel B6..B7 5 200 ns 24 960 ns 4.8 DJNZ Rn, rel D8..DF 4 160 ns 24 960 ns 6 DJNZ direct, rel D5 5 200 ns 24 960 ns 4.8 NOP 00 1 40 ns 12 480 ns 12 Table 5-2. INSTRUCTION SPEED SUMMARY INSTRUCTION CATEGORY Total instructions: 1 byte Total instructions: 2 byte Total instructions: 3 byte Average across all instructions OPCODE CATEGORY Total opcodes: 1 byte Total opcodes: 2 byte Total opcodes: 3 byte Average across all opcodes 57 SPEED ADVANTAGE QUANTITY 4.0 4.8 5.3 6.0 8.0 12.0 24.0 4.0 6.0 8.0 12.0 4.8 6.0 8.0 8.5 2 1 1 12 5 27 1 1 27 5 13 3 5 8 111 SPEED ADVANTAGE QUANTITY 4.0 4.8 5.3 6.0 8.0 12.0 24.0 4.0 6.0 8.0 12.0 4.8 6.0 8.0 9.4 4 1 1 35 5 93 1 1 42 5 43 4 12 8 255 _____________________________________________________________________________________________ Ultra-High-Speed Flash Microcontroller User’s Guide SECTION 6: MEMORY ACCESS The ultra-high-speed flash microcontroller supports the memory interface convention established for the industry standard 80C51, but also implements two new page mode memory interfaces needed to support ultra-high-speed external operation. These external page mode interfaces are described later in this section. Program and data memory areas can be implemented on-chip, off-chip, or as a combination. When opting not to use the internal memory provided, or when exceeding the maximum address of on-chip program or data memory, the device performs an external memory access using the expanded memory bus on ports 0 and 2. While serving as a memory bus, port 0 and port 2 cannot function as I/O ports. The PSEN signal is driven active low to function as a chip enable or output enable when performing external code memory fetches. The RD and WR signals serve as enables when accessing external SRAM data memory. Program execution always begins at the reset vector, address 0000h. If on-chip program memory is enabled, program execution begins at internal location 0000h; otherwise, external program memory is used. Any reset causes the next program fetch to begin at this location. Subsequent branches and interrupts determine how program memory fetches deviate from sequential addressing. INTERNAL FLASH MEMORY The ultra-high-speed flash microcontroller contains five physically distinct blocks of embedded flash memory. The two largest blocks are each half of the total amount of internal program memory. A 64-byte flash security block has been incorporated to allow encryption during program memory verify operations. To further protect internal code against undesirable access, a three-level lock system has been implemented in a separate flash memory block. This single-byte block contains three lock bits (LB1, LB2, LB3), each of which can individually enable higher lock levels and greater code protection. The fifth flash memory block is the option control register. This byte contains a bit to enable or disable the watchdog timer reset function (EWT = WDCON.1) on a power-on reset. The two program memory blocks form a contiguous address range extending from 0000h through the maximum amount of on-chip program memory. The on-chip decoded address range is controlled in hardware by the EA pin, and in software through the ROMSIZE feature. The EA pin enables or disables the ability to access internal program memory and overrides any software configured bit settings. The logic state of the EA pin should be changed only when the microcontroller is being held in reset. The EA pin is sampled on each exit from the reset state to determine whether program fetching should begin internally or externally. When the EA pin is low, all code fetches are done externally through the expanded bus. When the EA pin is high, code fetches begin from internal program memory. Code fetches exceeding the maximum address of on-chip program memory cause the device to access off-chip program memory. The maximum on-chip decoded address is selectable by software using the ROMSIZE feature. ROMSIZE FEATURE Using the ROMSIZE feature, software can imitate a device with less on-chip memory. The maximum memory size is dynamically variable. Thus, a portion of memory can be removed from the memory map to access off-chip memory, then restored to access on-chip memory. In fact, all of the on-chip memory can be removed from the memory map, allowing the full 64kB of external memory space to be addressed. The ROMSIZE feature has two primary uses. In the first instance, it allows the device to act as a bootstrap loader for a flash memory or nonvolatile SRAM (NVSRAM). The internal program memory can contain a bootstrap loader, which can program the external memory device. Secondly, this method can be used to increase the amount of available program memory from 64kB to 80kB without bank switching. The maximum amount of on-chip memory is selected by configuring the ROM size select register bits RMS2, RMS1, RMS0 (ROMSIZE.2-0). The reset default condition gives access to the maximum on-chip program memory. In this configuration, only code addresses greater than the maximum amount of on-chip program memory result in external program memory accesses. The possible settings for the ROM size select register are shown in the ROMSIZE special-function register. _____________________________________________________________________________________________ 58 Ultra-High-Speed Flash Microcontroller User’s Guide Modification of the ROMSIZE (C2h) special function register requires using the timed access procedure and must be followed by a two machine cycle delay, such as executing two NOP instructions, before jumping to the new address range. Interrupts must be disabled during this operation, because a call to an interrupt vector during the changing of the memory map can cause erratic results. To select a different internal program memory size, software must alter bits RMS2–RMS0. The procedure to reconfigure the size of on-chip memory should be done as follows: 1) Jump to a location in program memory that is unaffected by the change. 2) Disable interrupts by clearing the EA bit (IE.7). 3) Write AAh to the timed access register (TA;C7h). 4) Write 55h to the timed access register (TA;C7h). 5) Modify the ROM size select bits (RMS2-RMS0). 6) Delay 2 machine cycles (2 NOP instructions). 7) Enable interrupts by setting the EA bit (IE.7). As noted in the first step above, ensure that changes to the ROMSIZE register do not corrupt program execution. For example, assume that a 16kB DS89C430 is executing instructions from internal program memory near the 12kB boundary (~3000h) and the ROMSIZE register is still configured to the default internal program space. If software reconfigures the ROMSIZE register for a maximum of 4kB (0000h–0FFFh) internal program space (RMS2–0 = 011b), the device immediately accesses external program memory since current program execution no longer resides within the new on-chip decoded range. This could result in code misalignment and execution of an invalid instruction. The recommended method is to modify the ROMSIZE register from a location in memory that is internal (or external) both before and after the operation. In the above example, the instruction which modifies the ROMSIZE register should be located below the 4kB (1000h) boundary or above the maximum boundary, so that it is unaffected by the memory modification. The same rule applies when executing from external program memory and increasing the on-chip decoded address range. If the 0kB of internal program memory setting is selected, take extra precautions. In this case, it is necessary to duplicate the interrupt vector table in external program memory. This is because the interrupt vector table is located in the lower 1kB of memory, and the device automatically redirects any fetches from the interrupt vector table to external memory. Be careful when assembling or compiling the program so that all the modules are located at the correct starting address, including the interrupt vector table. FLASH SECURITY BLOCK/LOCK BITS The device incorporates a 64-byte encryption array, allowing the user to verify program codes while viewing the data in encrypted form. The encryption array, often referred to as the security block, has the same electrical and timing characteristics as the on-chip program memory. Once the encryption array is programmed to non-FFh, the data presented in the verify mode is encrypted. Each byte of data is XNORed with a byte in the encryption array during verification. If the security block is used, program unused portions of the internal flash program memory range with random data so that the encryption vector cannot be easily extracted. The single byte, which contains the 3 lock bits, logically resides at byte address 40h of the security block. The 3 lock bits (LB3, LB2, and LB1) can be accessed in bit positions 5, 4, and 3, respectively. By programming the 3 lock bits, the user may select a level of security as specified in table below. Once a security level is selected and programmed, the setting of the lock bits remains. Only a mass erase erases these bits and allows reprogramming the security level to a less restricted protection. Table 6-1. Flash Memory Lock Bits LEVEL 1 LB1 1 LB2 1 LB3 1 PROTECTION No program lock. Encrypted verify if encryption array is programmed. 2 0 1 1 Prevent MOVC in external memory from reading program code in internal memory. EA is sampled and latched on reset. Allow no further parallel or program memory Loader programming. 3 X 0 1 Level 2 plus no verify operation. Also prevent MOVX in external memory from reading internal SRAM. 4 X X 0 Level 3 plus no external execution. The lock bits affect the read/write accessibility in program memory loader and parallel programming modes. 59 _____________________________________________________________________________________________ Ultra-High-Speed Flash Microcontroller User’s Guide OPTION CONTROL REGISTER BYTE User-selectable options are present that must be set before beginning software execution. The option control register uses flash bits, rather than SFRs, and is individually erasable and programmable as a byte-wide register. Bit 3 of this register is defined as the watchdog POR default. Setting this bit to 1 disables the watchdog reset function on power-up, and clearing this bit to 0 enables the watchdog reset function automatically. Other bits of this register are undefined and are at logic 1 when read. The value of this register can be read at address FCh in parallel programming mode or by executing the verify option control register instruction in ROM Loader mode. FFFF FFFF Note: The hatched areas shown on the internal and external memory are disabled on power-up (Default) INTERNAL MEMORY 03FF INTERNAL REGISTERS SCRATCH PAD 128 Bytes SFR 0000 1kB x 8 SRAM External Program Memory Data OR prog mem addr from 400–7FF External Data Memory 4000 3FFF FF 8kB x 8 Flash Memory (Program) 128 Bytes Indirect Addressing 2000 80 7F 1FFF 2F 20 1F 8kB x 8 Flash Memory (Program) Bit Addressable Bank 3 Bank 2 Bank 1 00 Bank 0 0000 03FF 0000 0000 Non-usable if Internal SRAM is activated Figure 6-1. Memory Map for the DS89C420/430 INTERNAL SRAM MEMORY The ultra-high-speed microcontroller incorporates an internal 1kB SRAM that is usable as data, program, or merged program/data memory. Upon a power-on reset, the internal 1kB memory is disabled and transparent to both program and data memory maps. When used for data, the memory is addressed through MOVX commands, and is in addition to the 256 bytes of scratchpad memory. To enable the 1kB SRAM as internal data memory, software must set the DME0 bit (PMR.0). After setting this bit, all MOVX accesses within the first 1kB (0000h–03FFh) is directed to the internal SRAM. Any data memory accesses outside of this range are still directed to the expanded bus. One advantage of using the internal data memory is that MOVX operations automatically default to the fastest access possible. Note that the DME0 bit is cleared after any reset, so access to the internal data memory is prohibited until this bit is modified. The contents of the internal data memory are not affected by the changing of the data memory enable (DME0) bit. Table 6-2 shows how the DME1, DME0 bits affect the data memory map. _____________________________________________________________________________________________ 60 Ultra-High-Speed Flash Microcontroller User’s Guide Table 6-2. Data Memory Access Control DME1 0 X 1 DME0 0 1 0 DATA MEMORY ADDRESS RANGE 0000h–FFFFh 0000h–03FFh Reserved DATA MEMORY LOCATION External Data Memory (default) Internal Data Memory E lD M Reserved When configured as program memory, code fetches and MOVC read operations can be directed to this 1kB internal SRAM. To enable the 1kB SRAM as internal program memory, software must set the PRAME bit (ROMSIZE.3). After setting this bit, code accesses to the address range 0400h–07FFh are made to the internal 1kB SRAM in place of the program memory previously mapped to that address range. For applications using only external program memory (EA = 0), the internal 1kB SRAM cannot be enabled as program space. The internal 1kB SRAM can serve as merged program/data memory if both the DME0 and PRAME bits have been set. This feature can be effective for changing small pieces of frequently executed code, but be cautious when employing self-modifying code techniques. PROGRAM MEMORY INTERFACE—NONPAGE MODE The ultra-high-speed flash microcontroller defaults to a nonpage mode, external program memory interface. This memory interconnect scheme is the same as is used for the high-speed microcontroller family, and is shown in Figure 6-2. This example uses the DS89C420 and one 64k x 8 memory device. The program store enable (PSEN) signal is used to provide an output enable to the memory. It can also be used to provide a chip enable, but this generally results in less-favorable timing. The address LSB and data are multiplexed on port 0, and the address MSB is provided on port 2. An external latch, shown in the diagram as a 74F373, is used to latch the lower byte of the address to the memory device. The address latch enable (ALE) signal controls the timing of the latch so that the operation is performed in the proper sequence. The signals and relative timing for a program access are shown in Figure 6-3. When implementing a high-speed memory interface, the F series (or faster) logic should be used. HC logic has worst-case propagation delays that are too long. Specifications for all devices should be checked. More information on the nonpage mode memory interface timing can be found in Application Note 57 (DS80C320 Memory Interface Timing) and Application Note 85 (High-Speed Microcontroller Interface Timing). The DS89C420 provides an extremely high-speed interface to external memory. This allows for use of the slowest, and least expensive, memory device for a given crystal speed. The ultra-high-speed flash microcontroller provides very fast slew rates to allow the maximum possible time for memory access. Refer to the electrical specifications for exact timing. Figure 6-3 shows the timing relationship for internal and external nonpage mode code fetches when CD1:0 = 10b. Note that an external program fetch takes four system clocks, and an internal program fetch requires only one system clock. PSEN ALE PORT 0 OE CK 74F373 LATCH LSB ADDRESS (8) 64kB X 8 MEMORY DATA BUS (8) CE DS89C4x0 MSB ADDRESS (7) PORT 2 Figure 6-2. Program Memory Interconnect (Nonpage Mode) 61 _____________________________________________________________________________________________ Ultra-High-Speed Flash Microcontroller User’s Guide Internal Memory Cycles Ext Memory Cycle C1 C2 C3 C4 Ext Memory Cycle C1 C2 C3 C4 XTAL1 ALE PSEN LSB Add Data LSB Add Data Port 0 MSB Add Port 2 MSB Add Figure 6-3. External Program Memory Access (Nonpage Mode and CD1:0 = 10b) As illustrated in that same figure, ALE is deasserted when executing an internal memory fetch. The microcontroller provides a programmable user option (ALEON bit = PMR.2) to turn on the ALE signal during internal program memory operation. The ALE signal is automatically enabled for external code fetches, independent of the setting of this bit. PSEN is asserted only for external code fetches, and is inactive during internal execution. PROGRAM MEMORY INTERFACE—PAGE MODES Page mode retains the basic external circuitry requirements as the original 8051 external memory interface, but modifies the address/data roles of P0 and P2 in order to achieve the most efficient single-cycle external operation possible. The functions of ALE and PSEN are also altered to support page mode operation. Page mode is enabled by setting the PAGEE (ACON.7) bit to a logic 1. Clearing the PAGEE bit disables the page mode and returns to the traditional external bus structure of the 8051 (nonpage mode). Page mode is supported in two external bus structures. The page mode select bits (PAGES1:0) contained in the ACON register determine the external bus structure and the number of system clocks per basic memory cycle. The following table summarizes the four options available through the PAGES bits. The first three selections all represent the page mode 1 external bus structure, but with different memory cycle timings. The last configuration (PAGES = 11b) selects the page mode 2 bus structure. Table 6-3. Page Mode Select EXTERNAL ADDRESSING MODE Page mode 1 (1 cycle) Page mode 1 (2 cycle) Page mode 1 (4 cycle) Page mode 2 PAGES1:PAGES0 00 01 10 11 CLOCKS PER MEMORY CYCLE PAGE HIT PAGE MISS 1 2 2 4 4 8 2* 4 EXTERNAL BUS STRUCTURE PAGE MODE 1 PAGE MODE 1 PAGE MODE 1 PAGE MODE 2 Note: External data memory accesses always require four clock cycles, regardless of page hit or page miss. _____________________________________________________________________________________________ 62 Ultra-High-Speed Flash Microcontroller User’s Guide PAGE MODE 1: P0: Primary data bus. P2: Primary address bus, multiplexing the upper byte and lower byte of address. PAGE MODE 2: P0: Lower address byte. P2: Upper address byte is multiplexed with the data byte. In addition to being accessible to the user application code, the page mode enable and select bits can also be modified while in ROM loader mode. This allows in-system MOVX read/write access to external memory already connected according to the page mode 1 or page mode 2 bus structure. Since all resets, including the one generated when exiting ROM loader mode, return to the nonpage mode external bus structure, user application code must always configure the ACON register appropriately before addressing page mode external memory. Write access to the ACON register requires using the timed access procedure. PAGE MODE 1 BUS STRUCTURE The page mode 1 external bus structure uses P2 as the primary address bus (multiplexing both the most significant byte and least significant byte of the address for each external memory cycle), and P0 is used as the primary data bus. This program memory interconnect scheme is depicted in Figure 6-4. PSEN ALE PORT 2 OE CK 74F373 LATCH MSB ADDRESS (8) 64kB X 8 MEMORY LSB ADDRESS (8) CE DS89C4x0 DATA BUS (8) PORT 0 Figure 6-4 Program Memory Interconnect (Page Mode 1) During external code fetches, P0 is held in a high-impedance state by the processor. Opcodes are driven by the external memory onto P0 and latched on the rising edge of PSEN at the end of the external fetch cycle. • A page miss occurs when the most significant byte of the subsequent address is different from the last address. The exter nal memory machine cycle can be 2, 4, or 8 system clocks in length for a page miss. • A page hit occurs when the most significant byte of the subsequent address does not change from the last address. The external memory machine cycle can be 1, 2, or 4 system clocks in length for a page hit. During a page hit, P2 drives Addr [7:0] of the 16-bit address while the most significant address byte is held in the external address latches. PSEN, RD, and RD strobe accordingly for the appropriate operation on the P0 data bus. There is no ALE assertion for page hits. 63 _____________________________________________________________________________________________ Ultra-High-Speed Flash Microcontroller User’s Guide During a page miss, P2 drives the Addr [15:8] of the 16-bit address and holds it for the duration of the first half of the memory cycle to allow the external address latches to latch the new most significant address byte. ALE is asserted to strobe the external address latches. During this operation, PSEN, RD, and WR are all held in inactive states and P0 is in a high-impedance state. The following halfmemory cycle is executed as a page-hit cycle and the appropriate operation takes place. A page-miss can occur at set intervals or during external operations that require a memory access into a page of memory that has not been accessed during the last external cycle. Generally, the first external memory access causes a page miss. The new page address is stored internally and is used to detect a page miss for the current external memory cycle. Note that there are a few exceptions for this mode of operation when PAGES1 and PAGES2 are set to 00b: • PSEN is asserted for both page hit and page miss for a full clock cycle. • The execution of external MOVX instruction causes a page miss. • A page miss occurs when fetching the next external instruction following the execution of an external MOVX instruction. The figure below shows external memory cycles for the page mode 1 bus structure. The first case illustrates a back-to-back MOVX execution sequence for one-cycle page mode (PAGES 1:0 = 00b). PSEN remains active during page-hit cycles, and page misses are forced during and after MOVX executions, independent of the most significant byte of the subsequent addresses. The second case illustrates a MOVX execution sequence for two-cycle page mode (PAGES 1:0 = 01b). PSEN is active for a full clock cycle in code fetches. Note that the page misses in this sequence are caused by changing of the most significant byte of the data address. The third case illustrates a MOVX execution sequence for four-cycle page mode (PAGES 1:0 = 10b). There is no page-miss in this execution cycle, as the most significant byte of the data address is assumed to match the last program address. Internal Memory Cycles External Memory Cycles XTAL1 ALE PSEN RD / WR PAGES=00 Port 0 Port 2 Inst MSB LSB Inst LSB Page Miss MOVX LSB MOVX LSB Page Hit Inst Data MSB LSB MSB Data Access LSB MSB Page Miss MOVX executed Data LSB MSB Data Access MOVX executed ALE PSEN RD / WR PAGES=01 Port 0 Port 2 MOVX MSBAdd LSB Add Page Miss Inst Data LSB Add MSBAdd Page Hit LSB Add Data Access MOVX executed MSBAdd Page Miss next instruction ALE PSEN RD / WR PAGES=10 Port 0 Port 2 Inst MSBAdd LSB Add Page Miss Data LSB Add Data Access Figure 6-5. Page Mode 1 External Memory Cycle (CD1:0 = 10b) _____________________________________________________________________________________________ 64 Ultra-High-Speed Flash Microcontroller User’s Guide PAGE MODE 2 BUS STRUCTURE The page mode 2 external bus structure multiplexes the most significant address byte with data on P2 and uses P0 for the least significant address byte. An illustration of this memory interface is provided in the Figure 6-6. This bus structure speeds up external code fetches only. Aside from the different functions of P0 and P2 when operating in page mode 2, the external memory accesses are equal in duration and timing to those made in the nonpage mode. Figure 6-7 illustrates memory cycles for the page mode 2 bus structure. DATA MEMORY INTERFACE As described in Section 4, the ultra-high-speed microcontroller provides a small amount of RAM mapped as registers for on-chip direct access. This is not considered data memory and does not fall into the memory map. Systems that require more RAM or memorymapped peripherals must use the data memory area. This segment is a 64kB space located between 0000h and FFFFh. It is reached PSEN ALE PORT 2 OE CK 74F373 LATCH MSB ADDRESS (8) 64kB X 8 MEMORY DATA BUS (8) DS89C4x0 CE LSB ADDRESS (8) PORT 0 Figure 6-6. Program Memory Interconnect (Page Mode 2) Ext Code Fetches Internal Memory Cycles Page Miss C1 C2 C3 Page Hit C4 C1 C2 Page Hit C1 C2 XTAL1 ALE PSEN Port 0 Port 2 LSB Add MSB Add LSB Add Data Figure 6-7. Page Mode 2 External Code Fetch Cycle (CD1:0 = 10B) 65 _____________________________________________________________________________________________ Data LSB Add Data Ultra-High-Speed Flash Microcontroller User’s Guide using the MOVX instruction. Any use of this instruction automatically accesses the data area. Although the original 8051 convention placed all data memory off-chip, the device incorporates 1kB of on-chip data memory. The means for enabling and accessing this 1kB SRAM was covered earlier in this section. From a software standpoint, the physical location of the data area is not relevant because the same instructions are used. Like the program segment, if software accesses a data address that is above the on-chip data area, this access is automatically routed to the expanded bus. Thus, data or peripherals that are off-chip can be used in conjunction with on-chip memory by selecting addresses that do not overlap. For example, since the microcontroller has 1kB of on-chip data memory, an MOVX instruction at location 0400h is directed off-chip through the expanded bus. The external data memory interface follows the same bus structure as defined for program memory. The page mode enable (PAGEE) and page mode select (PAGES 1:0) bits control whether the external bus structure follows the nonpage mode, page mode 1, or page mode 2 scheme. During external data read/write operations, P0 or P2 (depending upon external memory mode) serves as the bidirectional data bus. This port is held in a high-impedance state for external reads from data memory, and driven with data during external writes to data memory. The read and write strobes used to access external data memory are provided on P3.7 and P3.6, respectively. EXTERNAL DATA MEMORY INTERFACE—NONPAGE MODE Data memory is accessed through use of the MOVX instruction. This instruction requires two basic memory cycles: a program-fetch memory access, and then a read or write memory access. Just like the program memory cycle, a basic internal data memory cycle contains one system clock, and a basic external data memory cycle contains four system clocks for nonpage mode operation. The program-fetch memory cycle for an MOVX instruction is no different from any other instruction. The unique timing occurs for the second memory cycle when data is accessed. The ultra-high-speed flash microcontroller allows software to adjust the speed of external data memory access by stretching the memory bus cycle. The MD2:0 bits contained in the CKCON (8Eh) SFR provide the means to modify the stretch value. This stretch feature allows the application to dynamically select the minimum (fastest) access time to each data memory peripheral device. The table below shows the data memory cycle stretch values and their effect on the read and write control signals associated with the external MOVX memory bus cycle. A stretch machine cycle always contains four system clocks. As illustrated in Table 6-4, the stretch feature supports eight external data memory access cycles, which can be categorized into three timing groups. When the stretch value is cleared to 000b, there is no stretch on external data memory access and a MOVX instruction is completed in two basic memory cycles. When the stretch value is set to 001b, 010b, or 011b, the external data memory access is extended by 1, 2, or 3 stretch machine cycles, respectively. Note that the 001b stretch value does not add four system clocks to the RD or WR control signals but instead uses one system clock to create additional address setup and data bus float time and one system clock to create additional address and data hold time. When using very slow RAM and peripherals, a larger stretch value (4–7) can be selected. In this stretch category, one stretch machine cycle (four system clocks) is used to stretch the ALE pulse width, one stretch machine cycle is used to create additional setup, and one stretch machine cycle is used to create additional hold time. Table 6-4. Nonpage Mode Data Memory Stretch Values MD2: MD0 (STRETCH VALUE) STRETCH CYCLES 4X/2X, CD1, CD0 = 100 4X/2X, CD1, CD0 = 000 4X/2X, CD1, CD0 = X10 4X/2X, CD1, CD0 = X11 000 0 0.5 1 2 2048 001 1 1 2 4 4096 010 2 2 4 8 8192 011 3 3 6 12 12288 100 7 4 8 16 16384 101 8 5 10 20 20480 110 9 6 12 24 24576 111 10 7 14 28 28672 _____________________________________________________________________________________________ 66 Ultra-High-Speed Flash Microcontroller User’s Guide CK 74F373 LATCH ALE PORT 0 LSB ADDRESS (8) 64kB X 8 SRAM DATA BUS (8) CE DS89C4x0 MSB ADDRESS (8) PORT 2 RD (P3.7) OE WR (P3.6) WE Figure 6-8. Data Memory Interconnect (Nonpage Mode) EXTERNAL DATA MEMORY INTERFACE—PAGE MODES The ultra-high-speed flash microcontroller allows software to adjust the speed of external data memory access by stretching the memory bus cycle in page mode operation just like nonpage mode operation. The tables below summarize the stretch values for page mode 1 and page mode 2. The number of stretch cycles added to the external MOVX operation and the control signal pulse width (in terms of the number of oscillator clocks) are provided. A stretch machine cycle always contains four system clocks, independent of the logic value of the page mode select bits. Just like nonpage mode operation, the stretch feature supports eight stretched external data memory access cycles that can be categorized into three timing groups. When the stretch value is cleared to 000b, there is no stretch on external data memory access and a MOVX instruction is completed in two basic memory cycles. When the stretch value is set to 001b, 010b, or 011b, the external data memory access is extended by 1, 2, or 3 stretch machine cycles, respectively. The 001b stretch value does not add four system clocks to the RD or WR control signals, but instead uses one system clock to create additional address setup and data bus float time and one system clock to create additional address and data hold time. When using very slow RAM and peripherals, a larger stretch value (4–7) can be selected. In this stretch category, one stretch machine cycle (four system clocks) is used to stretch the ALE pulse width, one stretch machine cycle is used to create additional setup and one stretch machine cycle is used to create additional hold time. Table 6-5. Page Mode 1—Data Memory Stretch Values 1 Cycle (PAGES 1:0 = 00b) RD/WR PULSE WIDTH (IN NUMBER OF OSCILLATOR CLOCKS) MD2:MD0 (STRETCH VALUE) STRETCH CYCLES 4X/2X, CD1, CD0 = 100 4X/2X, CD1, CD0 = 000 4X/2X, CD1, CD0 = X10 4X/2X, CD1, CD0 = X11 000 001 010 011 100 101 110 111 0 1 2 3 7 8 9 10 0.25 0.75 1.75 2.75 3.75 4.75 5.75 6.75 0.5 1.5 3.5 5.5 7.5 9.5 11.5 13.5 1 3 7 11 15 19 23 27 1024 3072 7168 11264 15360 19456 23552 27648 67 _____________________________________________________________________________________________ Ultra-High-Speed Flash Microcontroller User’s Guide Table 6-6. Page Mode 1—Data Memory Stretch Values Two Cycles (PAGES 1:0 = 01b) RD/WR PULSE WIDTH (IN NUMBER OF OSCILLATOR CLOCKS) MD2:MD0 (STRETCH VALUE) STRETCH CYCLES 4X/2X, CD1, CD0 = 100 4X/2X, CD1, CD0 = 000 4X/2X, CD1, CD0 = X10 4X/2X, CD1, CD0 = X11 000 001 010 011 100 101 110 111 0 1 2 3 7 8 9 10 0.25 0.75 1.75 2.75 3.75 4.75 5.75 6.75 0.5 1.5 3.5 5.5 7.5 9.5 11.5 13.5 1 3 7 11 15 19 23 27 1024 3072 7168 11264 15360 19456 23552 27648 Table 6-7. Page Mode 1—Data Memory Stretch Values Four Cycles (PAGES 1:0 = 10b) RD/WR PULSE WIDTH (IN NUMBER OF OSCILLATOR CLOCKS) MD2:MD0 (STRETCH VALUE) STRETCH CYCLES 4X/2X, CD1, CD0 = 100 4X/2X, CD1, CD0 = 000 4X/2X, CD1, CD0 = X10 4X/2X, CD1, CD0 = X11 000 001 010 011 100 101 110 111 0 1 2 3 7 8 9 10 0.5 1 2 3 4 5 6 7 1 2 4 6 8 10 12 14 2 4 8 12 16 20 24 28 2048 4096 8192 12288 16384 20480 24576 28672 Table 6-8. Page Mode 2—Data Memory Stretch Values (PAGES 1:0 = 11b) RD/WR PULSE WIDTH (IN NUMBER OF OSCILLATOR CLOCKS) MD2:MD0 (STRETCH VALUE) STRETCH CYCLES 4X/2X, CD1, CD0 = 100 4X/2X, CD1, CD0 = 000 4X/2X, CD1, CD0 = X10 4X/2X, CD1, CD0 = X11 000 001 010 011 100 101 110 111 0 1 2 3 7 8 9 10 0.5 1 2 3 4 5 6 7 1 2 4 6 8 10 12 14 2 4 8 12 16 20 24 28 2048 4096 8192 12288 16384 20480 24576 28672 _____________________________________________________________________________________________ 68 Ultra-High-Speed Flash Microcontroller User’s Guide Figures 6-9 and 6-10 below show data memory interconnect examples for page mode 1 and page mode 2. ALE PORT 2 CK 74F373 LATCH MSB ADDRESS (8) 64kB X 8 SRAM LSB ADDRESS (8) DATA BUS (8) DS89C4x0 CE PORT 0 RD (P3.7) OE WR (P3.6) WE Figure 6-9. Data Memory Interconnect (Page Mode 1) ALE PORT 2 CK 74F373 LATCH MSB ADDRESS (8) 64kB X 8 SRAM DATA BUS (8) CE DS89C4x0 LSB ADDRESS (8) PORT 0 RD (P3.7) OE WR (P3.6) WE Figure 6-10. Data Memory Interface (Page Mode 2) The following pages provide timing diagrams to illustrate the external data memory timing for the nonpage and page mode external bus structures. 69 _____________________________________________________________________________________________ Ultra-High-Speed Flash Microcontroller User’s Guide NONPAGE MODE DATA MEMORY TIMING The first diagram below shows execution of the MOVX instruction from internal program memory with stretch value = 0 assigned (MD2:0 = 000b). Note that the internal memory cycles consist of one system clock while the external memory cycles always consist of four system clocks. The second diagram illustrates the same MOVX instruction with a default stretch value (MD2:0 = 001b). The stretch cycle (four system clocks) is distributed as follows: one system clock added for address setup, two system clocks being added to the RD or WR pulse duration, and one system clock added for address/data hold. For subsequent stretch values of 2 or 3, the full stretch cycle is added to the duration of the RD or WR pulse. NONPAGE MODE: MOVX (TWO CYCLES) SYSCLK ALE PSEN WR / RD MOVX MSB PORT2 LSB PORT0 MOVX INST DATA MOVX Data Access INSTRUCTIONS NONPAGE MODE: MOVX (THREE CYCLES) SYSCLK ALE PSEN WR / RD MOVX MSB PORT2 MOVX LSB PORT0 MOVX INST MOVX DATA MOVX Data Access (1 Stretch C y cle) INSTRUCTIONS _____________________________________________________________________________________________ 70 Ultra-High-Speed Flash Microcontroller User’s Guide PAGE MODE 1 DATA MEMORY TIMING–PAGES 1:0 = 10b (FOUR CYCLES) The first diagram below shows execution of the MOVX instruction from internal program memory with stretch value = 0 assigned (MD2:0 = 000b). Note that the internal memory cycles consist of one system clock while the external memory cycles consist of four system clocks (page hit) or eight system clocks (page miss). The second diagram illustrates the same MOVX instruction with a default stretch value (MD2:0 = 001b). The stretch cycle (four system clocks) is distributed as follows: one system clock added for address setup, two system clocks being added to the RD or WR pulse duration, and one system clock added for address/data hold. For subsequent stretch values of 2 or 3, the full stretch cycle is added to the duration of the RD or WR pulse. FOUR-CYCLE PAGE MODE 1: MOVX (TWO CYCLES) SYSCLK ALE PSEN WR / RD MOVX MSB PORT2 MOVX LSB DATA PORT0 MOVX INST MOVX Data Access (Page Miss) INSTRUCTIONS FOUR-CYCLE PAGE MODE 1: MOVX (THREE CYCLES) SYSCLK ALE PSEN WR / RD MOVX MSB PORT2 MOVX LSB MOVX DATA PORT0 MOVX INST MOVX Data Access (Page Miss + 1 Stretch Cycle) = STRETCH CYCLE 71 _____________________________________________________________________________________________ INST INST Ultra-High-Speed Flash Microcontroller User’s Guide PAGE MODE 1 DATA MEMORY TIMING–PAGES 1:0 = 01b (TWO CYCLES) The first diagram below shows execution of back-to-back MOVX instructions from internal flash memory. A stretch value = 0 (MD2:0 = 000b) has been assigned. Note that the internal memory cycles consist of one system clock while the external memory cycles consist of two system clocks (page hit) or four system clocks (page miss). The second diagram below illustrates the timing of the MOVX operation with stretch value = 1 (MD2:0 = 001b). The stretch cycle (four system clocks) is distributed as follows: one system clock added for address setup, two system clocks being added to the RD or WR pulse duration, and one system clock added for address/data hold. For subsequent stretch values of 2 or 3, the full stretch cycle is added to the duration of the RD or WR pulse. 2-CYCLE PAGE MODE 1: MOVX (2 CYCLE) — MOVX (2 CYCLES) SYSCLK ALE PSEN WR / RD MOVX #1 MSB PORT2 MOVX #2 MSB MOVX #1 LSB MOVX #2 LSB MOVX DATA MOVX DATA PORT0 MOVX MOVX #1 #2 MOVX #1 Data Access (Page Miss) MOVX #2 Data Access (Page Miss) INST INSTRUCTIONS 2-CYCLE PAGE MODE 1: MOVX (3 CYCLES) SYSCLK ALE PSEN WR / RD MOVX MSB PORT2 MOVX LSB MOVX DATA PORT0 MOVX INST MOVX Data Access (Page Miss + 1 Stretch Cycle) INSTRUCTIONS = STRETCH CYCLE _____________________________________________________________________________________________ 72 Ultra-High-Speed Flash Microcontroller User’s Guide PAGE MODE 1 DATA MEMORY TIMING–PAGES 1:0 = 01b (TWO CYCLES) (CONTINUED) The first diagram below shows execution of a MOVX instruction with default stretch value = 1 (MD2:0 = 001b) from external program memory. The most probable case, where a page-miss is needed for the MOVX instruction, is given here. However, if the MOVX address happened to coincide with the current code execution page, a page hit would occur. The second diagram illustrates the MOVX timing that would occur if the address MSB for the MOVX data were to coincide with the code execution pages before and after the data access. Since a different MSB would not need to be latched, neither of the page-miss cycles seen in the third diagram would occur. TWO-CYCLE PAGE MODE 1: MOVX (THREE CYCLE) EXTERNAL CODE EXECUTION WITH PAGE MISSES SYSCLK ALE PSEN WR / RD PORT2 LSB LSB MOVX MSB MOVX LSB MSB LSB MOVX DATA PORT0 MOVX INST (Page Hit) MOVX Data Access (Page Miss + 1 Stretch Cycle) INST (Page Miss) TWO-CYCLE PAGE MODE 1: MOVX (THREE CYCLE) EXTERNAL CODE EXECUTION - NO PAGE MISSES SYSCLK ALE PSEN WR / RD PORT2 LSB LSB MOVX LSB LSB LSB LSB INST (Page Hit) INST (Page Hit) INST (Page Hit) MOVX DATA PORT0 MOVX #1 INST (Page Hit) MOVX Data Access (Page Hit + 1 Stretch Cycle) = STRETCH CYCLE 73 _____________________________________________________________________________________________ Ultra-High-Speed Flash Microcontroller User’s Guide PAGE MODE 1 DATA MEMORY TIMING–PAGES 1:0 = 00b (ONE CYCLE) The first diagram below illustrates execution of back-to-back MOVX instructions from internal flash memory. The default MOVX stretch setting (MD2:0 = 001b) has been assumed. The total duration of each MOVX instruction is seven system clocks = one system clock (page-hit memory cycle) + 2 system clocks (page-miss memory cycle) + four system clocks (one stretch cycle). Note that all external MOVX operations in one-cycle page mode 1 result in page-misses. The second diagram illustrates execution of the same back-to-back MOVX instructions with a stretch value of 0 (MD2:0 = 000b). ONE-CYCLE PAGE MODE 1: MOVX (THREE CYCLES) — MOVX (THREE CYCLES) SYSCLK ALE PSEN WR / RD MOVX MSB PORT2 MOVX MSB MOVX#1 LSB Address MOVX#2 LSB Address MOVX data PORT0 MOVX MOVX #1 #2 MOVX #1 Data Access (Page Miss + 1 Stretch Cycle) MOVX data INST MOVX #2 Data Access (Page Miss + 1 Stretch Cycle) INST ONE-CYCLE PAGE MODE 1: MOVX (TWO CYCLES) — MOVX (2 CYCLES) SYSCLK ALE PSEN WR / RD MOVX MSB PORT2 PORT0 MOVX MOVX #1 #2 MOVX LSB MOVX MOVX MSB LSB MOVX DATA MOVX DATA MOVX #1 Data Access (Page Miss) INST MOVX #2 Data Access (Page Miss) INSTRUCTIONS = STRETCH CYCLE _____________________________________________________________________________________________ 74 Ultra-High-Speed Flash Microcontroller User’s Guide PAGE MODE 1 DATA MEMORY TIMING–PAGES 1:0 = 00b (ONE CYCLE) (CONTINUED) The next diagram, still using a MOVX stretch value = 0, shows the back-to-back MOVX instructions being executed from external program memory. The last diagram shows external code memory execution of an external MOVX instruction with stretch value = 4 (MD2:0 = 100b). It has been assumed, for this example, that a page-miss is required for the MOVX data access. A stretch value = 4 results in the addition of 4 stretch cycles beyond the stretch value = 3. The four stretch cycles are distributed as folllows: two stretch cycles added for address setup, one stretch cycle added to RD or WR pulse duration, and 1 stretch cycle added for address/data hold. For subsequent stretch values of 5, 6, or 7, the added stretch cycle increases the RD or WR pulse duration. ONE-CYCLE PAGE MODE 1: MOVX (TWO CYCLES) — MOVX (TWO CYCLES) EXTERNAL CODE EXECUTION SYSCLK ALE PSEN WR / RD PORT2 LSB LSB MOVX MSB MOVX LSB MSB LSB MOVX DATA PORT0 MOVX MOVX #1 #2 MOVX #1 Data Access (Page Miss) MOVX MOVX MSB MSB LSB LSB LSB LSB LSB LSB LSB LSB MOVX DATA INST (Page Miss) MOVX #2 Data Access (Page Miss) INST (Page Miss) INSTRUCTIONS ONE-CYCLE PAGE MODE 1: MOVX (NINE CYCLES) EXTERNAL CODE EXECUTION SYSCLK ALE PSEN Stretch #1 Stretch #2 Stretches #3-6 Stretch #7 WR / RD PORT2 LSB LSB MOVX MSB MOVX LSB LSB MOVX DATA PORT0 MOVX INST 75 MSB MOVX Data Access (Page Miss + 7 Stretch Cycles) _____________________________________________________________________________________________ INST (Page Miss) Ultra-High-Speed Flash Microcontroller User’s Guide PAGE MODE 2 DATA MEMORY TIMING–PAGES 1:0 = 11b (FOUR CYCLES) All external data memory accesses made using the page mode 2 external bus configuration require four system clocks. The MOVX timing looks identical to the nonpage mode MOVX timing except that port 2 multiplexes the MSB and the data, while port 0 serves as the LSB. DATA MEMORY ACCESS As mentioned earlier in this section, the ultra-high-speed microcontroller uses the MOVX instruction for data memory access. This includes off-chip RAM and memory-mapped peripherals needing read/write access. Several aspects of the MOVX operation have been enhanced as compared to the original 8051. The principal improvements are in the areas of the MOVX timing and the data pointer. The MOVX instruction is used to generate read/write access to off-chip address locations. It has several addressing modes. The first uses the MOVX @Ri command to reach a 256-byte block. This instruction uses the value in the designated working register to address one of 256 locations. The upper byte of the address is supplied by the value in the port 2 latch. A second way to access data is the data pointer (DPTR). This 16-bit register provides an absolute address for data memory access. 16-bits cover the entire 64kB area. Thus the DPTR serves as a pointer to memory. Using the DPTR, the relevant instruction is MOVX @DPTR. The original 8051 contained one DPTR. While this provides access to the entire memory area, it is difficult to move data from one address to another. The ultra-high-speed microcontroller provides two data pointers. Thus software can load both a source and a destination address. The MOVX instruction uses the active pointer to direct the off-chip address. The dual data pointers are DPTR0 and DPTR1. DPTR0 is at SFR addresses 82h and 83h. These are the locations used by the original 8051. No modification of standard code is needed to use DPTR0. The new DPTR is located at SFR 84h and 85h. The data pointer select bit (SEL) chooses the active pointer and is located in bit position 0 of the DPS (86h) SFR. When DPS is set to 0, the DPTR0 is active. When set to 1, DPTR1 is used. All DPTR-related instructions use the currently selected DPTR for any activity. Each data pointer (DPTR0, DPTR1) has an associated control bit (ID0, ID1) that determines whether the INC DPTR operation results in an increment or decrement of the pointer. When the active data pointer ID (increment/decrement) control bit is clear, the INC DPTR instruction will increment the pointer, whereas a decrement occurs if the active pointer’s ID bit is set when the INC DPTR instruction is performed. ID0 = DPS.6 ID1 = DPS.7 Using the dual data pointers for large block copy operations results in substantial code savings versus using a single data pointer, since one data pointer can be used for the source address and the second pointer can be used as the destination address. The user switches between data pointers by toggling the SEL bit. One way of accomplishing this is by executing the INC DPS instruction. For these large-block copy operations, the user must execute this instruction frequently to toggle between DPTR0 and DPTR1. To improve the speed and efficiency of moving data with dual data pointers, the ultra-high-speed microcontroller contains a toggle select (TSL) bit. When this TSL bit (DPS.5) is set, execution of certain MOVX instructions automatically toggle the SEL bit in hardware, allowing removal of the INC DPS instruction and increasing execution speed. Copying large blocks of data also requires that the source and destination pointers index byte-by-byte through their respective data ranges. The traditional method for incrementing each pointer is through the use of the INC DPTR instruction. The ultra-high-speed flash microcontroller provides yet another means of accelerating data transfers with the implementation of an auto increment/decrement bit (AID). When this AID bit (DPS.4) is set, execution of certain MOVX instructions automatically increments or decrements the active data pointer. AUTO-TOGGLE (if TSL = 1) AUTO-INC/DEC (if AID = 1) MOVC A, @A+DPTR MOVC A, @A+DPTR MOVX A, @DPTR MOVX A, @DPTR MOVX @DPTR, A MOVX @DPTR, A INC DPTR MOV DPTR, #data16 The following table summarizes the tremendous speed improvements gained through using the dual DPTRs along with autoincrement and autotoggle features. To properly quantify the speed improvement gained with enhanced data pointer operation versus improvement attributed to the single-cycle core architecture, execution time for the DS80C320 high-speed microcontroller (four-cycle core) has been included where applicable. For external page mode 1 (PAGES1:0 = 00b) code execution has been assumed. It is unreasonable to expect that the address MSBs for MOVX read/write operations are the same as the address MSB for code execution. Therefore, one clock cycle has been added to each MOVX instruction (for data access) and to the instruction that follows the MOVX (for code fetch) _____________________________________________________________________________________________ 76 Ultra-High-Speed Flash Microcontroller User’s Guide to account for potential page misses. The sample code listings have been marked accordingly with ‘+D’ to indicate a data access page-miss and ‘+C’ to indicate a code-fetch page-miss. Thus, in the case of back-to-back MOVX operations, the second MOVX operation has two clock cycles added (‘+CD’), one associated with the code fetch and one associated with the data access. Table 6-9. Enhanced Data Pointer Speed Improvement DS80C320 HIGH SPEED DS89C420 ULTRA-HIGH SPEED DATA POINTER OPERATION CLOCK CYCLES (4CLKS/MCLK) EXECUTION TIME (AT 33MHZ) CLOCK CYCLES EXECUTION TIME (AT 33MHZ) Single Data Pointer 1869 x 4 227µs 1933 59µs Dual Data Pointer 1098 x 4 133µs 1291 39µs Dual Data Pointer w/AID — — 1169 35µs Dual Data Pointer w/TSL — — 910 28µs Dual Data Pointer w/AID,TSL — — 782 24µs The sample code listings for these programs appear on the following pages. Program 1 listed below is original code written for an 8051 and utilizes a single data pointer. Program 2 uses the dual data pointer feature. Program 3 uses the dual data pointer with autoincrement enhancement. Program 4 uses the dual data pointer with autotoggle enhancement. Program 5 uses the dual data pointer with autoincrement and autotoggle enhancements. The relevant register and bit locations are summarized as follows: DPL 82h Low-byte original DPTR DPH 83h High-byte original DPTR DPL1 84h Low-byte new DPTR DPH1 85h High-byte new DPTR DPS 86h SEL bit = DPS.0 AID bit = DPS.4 TSL bit = DPS.5 PROGRAM 1: 64-BYTE BLOCK MOVE (WITHOUT DUAL DATA POINTER) ; ; ; ; ; SH and SL are high and low byte source address. DH and DL are high and low byte of destination address. For cycle counts: HSM = High-Speed Microcontroller UHSM = ultra-high-speed microcontroller MOV R5, #64 ; NUMBER OF BYTES TO MOVE MOV DPTR, #SHSL ; LOAD SOURCE ADDRESS MOV R1, #SL ; SAVE LOW BYTE OF SOURCE MOV R2, #SH ; SAVE HIGH BYTE OF SOURCE MOV R3, #DL ; SAVE LOW BYTE OF DESTINATION MOV R4, #DH ; SAVE HIGH BYTE OF DESTINATION MOVE: ; THIS LOOP IS PERFORMED R5 TIMES, IN THIS EXAMPLE 64 MOVX A, @DPTR ; READ SOURCE DATA BYTE MOV R1, DPL ; SAVE NEW SOURCE POINTER MOV R2, DPH ; MOV DPL, R3 ; LOAD NEW DESTINATION MOV DPH, R4 ; 77 _____________________________________________________________________________________________ # HSM/UHSM CYCLES 2/2 3/3 2/2 2/2 2/2 2/2 2/3 +D 2/3 +C 2/2 2/2 2/2 Ultra-High-Speed Flash Microcontroller User’s Guide MOVX INC MOV MOV MOV MOV INC DJNZ @DPTR, A DPTR R3, DPL R4, DPH DPL, R1 DPH, R2 DPTR R5, MOVE ; ; ; ; ; ; ; ; WRITE DATA TO DESTINATION NEXT DESTINATION ADDRESS SAVE NEW DESTINATION POINTER GET NEW SOURCE POINTER NEXT SOURCE ADDRESS FINISHED WITH TABLE? # HSM/UHSM CYCLES 2/3 +D 3/2 +C 2/2 2/2 2/2 2/2 3/1 3/4 PROGRAM 2: 64-BYTE BLOCK MOVE (DUAL DATA POINTER) ; ; ; ; ; ; SH and SL are high and low byte source address. DH and DL are high and low byte of destination address. DPS is the data pointer select. Reset condition DPTR0. For cycle counts: HSM = High-Speed Microcontroller UHSM = ultra-high-speed microcontroller # HSM/UHSM CYCLES DPS EQU 86h ; TELL ASSEMBLER ABOUT DPS MOV R5, #64 ; NUMBER OF BYTES TO MOVE MOV DPTR, #DHDL ; LOAD DESTINATION ADDRESS INC DPS ; CHANGE ACTIVE DPTR MOV DPTR, #SHSL ; LOAD SOURCE ADDRESS MOVE: ; THIS LOOP IS PERFORMED R5 TIMES, IN THIS EXAMPLE 64 MOVX A, @DPTR ; READ SOURCE DATA BYTE INC DPS ; CHANGE DPTR TO DESTINATION MOVX @DPTR, A ; WRITE DATA TO DESTINATION INC DPTR ; NEXT DESTINATION ADDRESS INC DPS ; CHANGE DATA POINTER TO SOURCE INC DPTR ; NEXT SOURCE ADDRESS DJNZ R5, MOVE ; FINISHED WITH TABLE? 2/2 3/3 2/3 3/3 2/3 2/4 2/3 3/2 2/3 3/1 3/4 +D +C +D +C PROGRAM 3: 64-BYTE BLOCK MOVE (DUAL DATA POINTER, AID) ; SH and SL are high and low byte source address. ; DH and DL are high and low byte of destination address. ; DPS is the data pointer select. Reset condition DPTR0. # UHSM CYCLES DPS EQU 86h ; TELL ASSEMBLER ABOUT DPS MOV R5, #64 ; NUMBER OF BYTES TO MOVE ORL DPS, #10h ; SET AUTO-INC/DEC (AID) MOV DPTR, #DHDL ; LOAD DESTINATION ADDRESS INC DPS ; CHANGE ACTIVE DPTR MOV DPTR, #SHSL ; LOAD SOURCE ADDRESS MOVE: ; THIS LOOP IS PERFORMED R5 TIMES, IN THIS EXAMPLE 64 MOVX A, @DPTR ; READ SOURCE DATA BYTE INC DPS ; CHANGE DPTR TO DESTINATION MOVX @DPTR, A ; WRITE DATA TO DESTINATION INC DPS ; CHANGE DATA POINTER TO SOURCE DJNZ R5, MOVE ; FINISHED WITH TABLE? ANL DPS, #0EFH ; CLEAR AUTO-INC/DEC 2 3 3 3 3 3 4 3 4 4 3 +D +C +D +C _____________________________________________________________________________________________ 78 Ultra-High-Speed Flash Microcontroller User’s Guide PROGRAM 4: 64-BYTE BLOCK MOVE (DUAL DATA POINTER, TSL) ; SH and SL are high and low byte source address. ; DH and DL are high and low byte of destination address. ; DPS is the data pointer select. Reset condition DPTR0. # UHSM CYCLES DPS EQU 86h ; TELL ASSEMBLER ABOUT DPS MOV R5, #64 ; NUMBER OF BYTES TO MOVE ORL DPS, #20h ; SET TOGGLE SELECT (TSL) MOV DPTR, #SHSL ; LOAD SOURCE ADDRESS MOV DPTR, #DHDL ; LOAD DESTINATION ADDRESS MOVE: ; THIS LOOP IS PERFORMED R5 TIMES, IN THIS EXAMPLE 64 MOVX A, @DPTR ; READ SOURCE DATA BYTE MOVX @DPTR, A ; WRITE DATA TO DESTINATION INC DPTR ; NEXT SOURCE ADDRESS INC DPTR ; NEXT DESTINATION ADDRESS DJNZ R5, MOVE ; FINISHED WITH TABLE? ANL DPS, #0DFh ; CLEAR TOGGLE SELECT 2 3 3 3 3 +D 4 +CD 2 +C 1 4 3 PROGRAM 5: 64-BYTE BLOCK MOVE (DUAL DATA POINTER, AID, TSL) ; SH and SL are high and low byte source address. ; DH and DL are high and low byte of destination address. ; DPS is the data pointer select. Reset condition DPTR0. # UHSM CYCLES DPS EQU 86h ; TELL ASSEMBLER ABOUT DPS MOV R5, #64 ; NUMBER OF BYTES TO MOVE 2 ORL DPS, #30h ; SET TOGGLE SELECT, AUTO-INC/DEC 3 MOV DPTR, #SHSL ; LOAD SOURCE ADDRESS 3 MOV DPTR, #DHDL ; LOAD DESTINATION ADDRESS 3 MOVE: ; THIS LOOP IS PERFORMED R5 TIMES, IN THIS EXAMPLE 64 MOVX A, @DPTR ; READ SOURCE DATA BYTE 3 +D MOVX @DPTR, A ; WRITE DATA TO DESTINATION 4 +CD DJNZ R5, MOVE ; FINISHED WITH TABLE? 5 +C ANL DPS, #0CFh ; CLEAR TSL, AID 3 Note that since each pass through the loop saves additional clock cycles when compared to the single DPTR approach, efficiency improvement when moving larger blocks is even greater using these features. Further speed improvement can be gained when executing from internal flash program memory, since no code-fetch page misses (+C) would occur. For example, running Program 5 from internal memory at 33MHz would require only 19.8µs (= 1/33MHz x (14 + 64 x 10)). SECTION 7: POWER MANAGEMENT The ultra-high-speed flash microcontroller has several features that relate to power consumption and management. They provide a combination of controlled operation in unreliable power applications and reduced power consumption in portable or battery-powered applications. The range of features is shown below with details to follow. POWER MANAGEMENT Precision Voltage Monitor Early-Warning Power-Fail Interrupt Power-Fail/Power-On Reset Bandgap Select Watchdog Wake-Up From Idle POWER SAVING Clock Divide Control Idle Mode Stop Mode Ring Wake-Up From Stop Power Management Mode 79 _____________________________________________________________________________________________ Ultra-High-Speed Flash Microcontroller User’s Guide POWER MANAGEMENT SUMMARY The following is a summary of the power management bits and those that are useful or related. They are contained in the register locations WDCON;D8h, EIE;E8h, EXIF;91h, and PCON; 87h. WDCON.6 POR: Power-on reset. Hardware sets this bit on a power-up condition. Software can read it, but must clear it manually. This bit assists software in determining the cause of a reset. WDCON.5 EPFI: enable power-fail interrupt. Setting this bit to 1 enables the power-fail interrupt. This occurs when VCC drops to approximately 4.375V, and the processor vectors to location 33h. Setting this bit to a 0 turns off the power-fail interrupt. WDCON.4 PFI: Power-fail interrupt flag. Hardware sets this bit to a 1 when a power-fail condition occurs. Software must clear the bit manually. Writing a 1 to this bit forces an interrupt, if enabled. WDCON.3 WDIF: Watchdog interrupt flag. If the watchdog interrupt is enabled (EIE.4), hardware sets this bit to indicate that the watchdog interrupt has occurred. If the interrupt is not enabled, this bit indicates that the timeout has passed. If the watchdog reset is enabled (WDCON.1), the user has 512 system clocks to strobe the watchdog prior to a reset. Software or any reset can clear this flag. WDCON.2 WTRF: Watchdog timer reset flag. Hardware sets this bit when the watchdog timer causes a reset. Software can read it, but must clear it manually. A power-fail reset also clears the bit. This bit assists software in determining the cause of a reset. If EWT = 0, the watchdog timer has no affect on this bit. WDCON.1 EWT: Enable watchdog timer reset. Setting this bit turns on the watchdog timer reset function. The interrupt does not occur unless the EWDI bit in the EIE register is set. A reset occurs according to the WD1 and WD0 bits in the CKCON register. Setting this bit to a 0 disables the resets but leaves the timer running. WDCON.0 RWT: Reset watchdog timer. This bit serves as the strobe for the watchdog function. During the timeout period, software must set the RWT bit if the watchdog is enabled. Failing to set the RWT causes a reset when the timeout has elapsed. There is no need to set the RWT bit to a 0 because it is self-clearing. EIE.4 EWDI: Enable watchdog interrupt. Setting this bit in software enables the watchdog interrupt. EXIF.0 BGS: Bandgap select. Setting this bit to a 1 allows the use of the bandgap voltage reference while in stop mode. Since this function uses as much as 75µA, the bandgap is optional in stop mode. Setting this bit to a 0 turns off the bandgap while in stop mode. When BGS = 0, no power-fail interrupt or power-fail reset is available in stop mode. PCON.1 STOP. When this bit is set, the program stops execution, clocks are stopped, and the CPU enters power-down mode. PCON.0 IDLE. Program execution halts, leaving timers, serial ports, and clocks running. EXIF.2 RGMD: Ring oscillator mode. Hardware sets this status bit to a 1 when the clock source is the ring oscillator. Hardware sets this status bit to a 0 when the crystal is the clock source. Refer to RGSL for operation of the ring oscillator. EXIF.1 RGSL: Ring oscillator select. When set to a 1 by software, the microcontroller uses a ring oscillator to come out of stop mode without waiting for crystal startup. This allows an instantaneous startup when coming out of stop mode. It is useful if software needs to perform a short task, and then return to stop. It is also useful if software must respond quickly to an external event. After the crystal has performed 65,536 cycles, hardware switches to the crystal as its clocksource. The RGMD status bit reports on this changeover. When RGSL is set to a 0, the microcontroller delays software execution until after the 65,536 clock crystal startup time. RGSL is only cleared by a power-on reset and is not altered by other forms of reset. _____________________________________________________________________________________________ 80 Ultra-High-Speed Flash Microcontroller User’s Guide VCC VPFW VRST VSS INTERRUPT SERVICE ROUTINE tCSU XTAL1 tPOR INTERNAL RESET Figure 7-1. Power Cycle Operaton Precision Voltage Monitor A precision bandgap reference and other analog circuits monitor the state of the power supply during power-up and power-down transitions. This obviates the need for external circuits to perform these functions that other microcontroller systems would require. The bandgap reference provides a precise voltage to compare with VCC. When VCC begins to drop, the power monitor compares it to its reference. This enables the analog circuits to detect when VCC passes through predetermined thresholds, VPFW and VRST. These are specified in the product data sheet. Early-Warning Power-Fail Interrupt The precision voltage reference has the ability to generate a power-fail interrupt and/or reset in response to a low-supply voltage. When VCC reaches the VPFW threshold, the microcontroller can generate a power-fail interrupt. This early warning of supply voltage failure allows the system time to save critical parameters in nonvolatile memory and put external functions in a safe state. The power-fail interrupt is optional and is enabled using the enable power-fail warning interrupt (EPFI) bit at WDCON.5. If enabled, VCC dropping below VPFW causes the device to vector to address 33h. The power-fail interrupt status bit, PFI (WDCON.4), is set anytime VCC transitions below VPFW. This flag is not cleared when VCC is above VPFW, and software should clear it immediately after reading it. As long as the condition exists, PFI is immediately set again by hardware. A typical application of the PFI is to place the device into a “safe mode” when a power loss appears imminent. When the interrupt occurs, the code vectors to location 33h. At this time, software can disable the interrupt, save any critical data, clear PFI, and then continually poll the status of the power supply through the PFI flag. As long as PFI is set, power is still below VPFW. If power returns to the proper level, PFI is not set once cleared by software. This indicates a safe operating condition. If power continues to fall, a power-fail reset is invoked automatically. Power-Fail Reset The power-fail reset automatically invokes a reset when VCC drops below VRST. This halts device operation and places all outputs in their reset state. This state continues to be held until VCC drops below the voltage necessary to power the port pins. Because VRST. is lower than VPFW, the microcontroller has the option to use the power-fail interrupt to place the device into a “safe” state before the device halts operation with a power-fail reset. The power-fail reset function cannot be disabled. 81 _____________________________________________________________________________________________ Ultra-High-Speed Flash Microcontroller User’s Guide Power-On Reset When VCC is applied to a system, the device holds itself in reset until power is within tolerance and stable. The internal bandgap reference provides a highly accurate and stable means of detecting power-supply levels. It requires no external circuits to accomplish this. As power rises, the processor stays in a reset state until VCC > VRST. As VCC rises above VRST, internal analog circuits detect this and activate the on-chip crystal oscillator. On-chip hardware then counts 65536 oscillator clocks. During this count, VCC must remain above VRST. or the process restarts. If an off-chip clock source is used, clock counting still begins once VCC > VRST. This count period is used to make certain that power is within tolerance and that the oscillator has time to stabilize. This provides a very controlled and predictable startup condition. Once the 65536 count period has elapsed, the reset condition is removed automatically, and software execution begins at the reset vector location of 0000h. Software is able to detect the power-on reset condition using the power-on reset (POR) flag. POR is located at WDCON.6. This bit is high to indicate that a power-on reset has occurred. It should then be cleared by software. The complete power cycle operation is shown in Figure 7-1. Note that the interrupt threshold is fixed, but the interrupt itself is optional. Reset thresholds are also fixed and the reset operation is transparent. It requires no external components and no action by software to control reset operation. Bandgap Select The bandgap is normally disabled automatically upon entering stop mode to provide the lowest power state. Since the bandgap is inactive, there can be no power-fail interrupt and no power-fail reset, similar to a traditional 8051. If the use of the power-fail features is desired in stop mode, the BGS bit (EXIF, 91h) can be used. When set to a logic 1 by software, the bandgap reference and associated power monitor circuits remain active in stop mode. The price of this feature is higher powersupply current requirements. In stop mode with the bandgap reference disabled (default), the processor draws approximately 10µA. With the bandgap enabled, it draws approximately 75µA. BGS allows the user to decide whether the control circuitry and its associated power consumption are needed. If the application is such that power does not fail while in stop, or if it does not matter that power fails, the BGS should be set to 0 (default). If power can fail at any time and cause problems, the BGS should be set to 1. Watchdog Wake-Up From Idle The watchdog wake-up is more of an application than a feature. It allows a system to enter the idle mode for power savings, then to wake up periodically to sample the external world. Idle mode is a low-power state described below. Any of the programmable timers can perform this function, but the watchdog allows a much longer period to be selected. At 33MHz, the maximum watchdog timeout is over 2s. This contrasts with 23.8ms using the 16-bit timers. Software that uses the watchdog as a wake-up alarm should enable only the watchdog interrupt and not the reset. Note that the watchdog cannot be used to wake the system while in stop mode since no clocks are running. Stop mode is described in the Power Management Summary section below. POWER SAVING The ultra-high-speed flash microcontroller is implemented using full CMOS circuitry for low-power operation. It is fully static, so the clock speed can be run down to DC. Like other CMOS, the power consumption is also a function of operating frequency. Although the microcontroller is designed for maximum performance, it also provides improved power versus work relationships compared with standard 8051 devices. These topics are discussed in detail in the following pages. Clock Divide Control The programmable clock divide control bits CD1 and CD0 (PMR, C4h) provide the processor with the ability to adapt to different crystals and also to slow the system clocks, providing lower power operation when required. An on-chip crystal multiplier allows the ultrahigh-speed flash microcontroller to operate at two or four times the crystal frequency by setting the 4X/2X bit and is enabled by setting the CTM bit to a logic 1. An additional circuit provides a clock source at divide-by-1024. When used with a 10MHz crystal, for example, the processor executes machine cycle in times ranging from 25ns (divide-by-0.25) to 102.4µs (divide-by-1024) and maintains a highly accurate, serial port baud rate while allowing the use of more cost-effective, lower-frequency crystals. Although the clock divide control bits can be written at any time, certain hardware features have been provided to enhance the use of these clock controls to guarantee proper serial port operation, and also to allow for a high-speed response to an external interrupt. The 01b setting of CD1 and CD0 is reserved and has the same effect as the setting of 10b, which forces the system clock into a divide-by-1 mode. The ultra-high-speed flash microcontroller defaults to divide-by-1 clock mode on all forms of reset. When programmed to the divide-by-1024 mode, and the switchback bit (PMR.5: SWB) is also set, the system forces the clock divide control bits to reset automatically to the divide-by-1 mode whenever the system has detected externally enabled interrupts. _____________________________________________________________________________________________ 82 Ultra-High-Speed Flash Microcontroller User’s Guide 4X/2X CTM Clock Multiplier Crystal Oscillator MUX Divide-by-1024 System Clock Ring Oscillator Ring Enable CD0 Selector CD1 Figure 7-2. System Clock Sources The oscillator divide ratios of 0.25, 0.5 and 1 are also used to provide standard baud rate generation for the serial ports through a forced divide-by-12 input clocks (TxMH,TxM = 00b, x = 1, 2, or 3) to the timers. When in divide-by-1024 mode, in order to allow a quick response to incoming data on a serial port, the system utilizes the switchback mode to automatically revert to divide-by-1 mode whenever a start bit is detected. This automatic switchback is only enabled during divide-by-1024 mode and all other clock modes are unaffected by interrupts and serial port activity. See power management mode for more details. Use of the divide-by-0.25 or 0.5 option through the clock divide control bits requires that the crystal multiplier be enabled and the specific system clock multiply value be established by the 4X/2X bit in the PMR register. The multiplier is enabled by the CTM (PMR.4) bit but cannot be automatically selected until a startup delay has been established through the CKRY bit in the status register. The 4X/2X bit can only be altered when the CTM bit is cleared to a logic 0. This prevents the system from changing the multiplier until the system has moved back to the divide-by-1 mode and the multiplier has been disabled through the CTM bit. The CTM bit can only be altered when the CD1 and CD0 bits are set to divide-by-1 mode and the RGMD bit is cleared to 0. Setting the CTM to a logic 1 from a previous logic 0 automatically clears the CKRY bit in the status register and starts the multiplier startup timeout in the multiplier startup counter. During the multiplier startup period, the CKRY bit remains cleared and the CD1 and CD0 clock controls cannot be set to 00b. The CTM bit is cleared to a logic 0 on all resets. Figure 7-2 (System Clock Sources) gives a simplified description of the generation of the system clocks. Specifics of hardware restrictions associated with the use of the 4X/2X, CTM, CKRY, CD1, and CD0 bits are outlined in the SFR description. The microcontroller provides two modes (other than operating) that allow power conservation. They are similar, but have different merits and drawbacks. These modes are idle and stop. In the original 8051, the stop mode is called power-down. These modes are invoked in the same manner as the original 8051 series. Idle Mode Idle mode suspends all CPU processing by holding the program counter in a static state. No program values are fetched and no processing occurs. This saves considerable power versus full operation. The virtue of idle mode is that it uses half the power of the operating state, yet reacts instantly to any interrupt conditions. All clocks remain active so the timers, watchdog, serial port, and power monitor functions are all working. Since all clocks are running, the CPU can exit the idle state using any of the interrupt sources. 83 _____________________________________________________________________________________________ Ultra-High-Speed Flash Microcontroller User’s Guide Software can invoke the idle mode by setting the IDLE bit in the PCON register at location 87h. The bit is located at PCON.0. The instruction that executes this step is the last instruction prior to freezing the program counter. Once in idle, all resources are preserved. There are two ways to exit the idle mode. First, any interrupt (that is enabled) will cause an exit. This results in a jump to the appropriate interrupt vector. The IDLE bit in the PCON register is cleared automatically. Upon returning from this vector using the RETI instruction, the next address is the one immediately after the instruction that invoked the idle state. The idle mode can also be removed using a reset. Any of the three reset sources can do this. On receiving the reset stimulus, the CPU is placed in a reset state and the idle condition cleared. When the reset stimulus is removed, software begins execution as for any reset. Since all clocks are active, there is no delay after the reset stimulus is removed. Note that if enabled, the watchdog timer continues to run during idle and must be supported. Stop Mode Stop mode is the lowest power state available. This is achieved by stopping all on-chip clocks, resulting in a fully static condition. No processing is possible, timers are stopped, and no serial communication is possible. Software can invoke stop mode by setting the STOP bit in the PCON register at location 87h. The bit is located at PCON.1. Processor operation halts on the instruction that sets the STOP bit. The internal amplifier that excites the external crystal is disabled, halting crystal oscillation in stop mode. Stop mode takes precedence if application code attempts to set both the STOP and IDLE bits. However, doing this is not suggested. Table 7-1 shows the state of the processor pins in idle and stop modes. Stop mode can be exited in two ways. First, like the 8052 microcontrollers, a nonclocked interrupt such as the external interrupts or the power-fail interrupt can be used. Clocked interrupts, such as the watchdog timer, internal timers, and serial ports do not operate in stop mode. Note that the bandgap reference must be enabled in order to use the power-fail interrupt to exit stop mode, which increases stop mode current. Processor operation resumes with the fetching of the interrupt vector associated with the interrupt that caused the exit from stop mode. When the interrupt service routine is complete, an RETI returns the program to the instruction immediately following the one that invoked the stop mode. A second method of exiting stop mode is with a reset. The watchdog timer reset is not available as a reset source because no timers are running in stop mode. An external reset by the RST pin unconditionally exits the device from stop mode. If the BGS bit is set, the device provides a reset while in stop mode if VCC should drop below the VRST level. If the BGS bit is 0, then a dip in power below VRST does not cause a reset. For example, if VCC drops to a level of VRST -0.5V, then returns to the full level, no reset is generated. For this reason, use of the bandgap reference is recommended if a brownout condition is possible in stop mode. If power fails completely (VCC = 0V), then a power-on reset is still performed when VCC is reapplied, regardless of the state of the BGS bit. Processor operation resumes execution from address 0000h like any other reset. Table 7-1. Pin States in Power-Saving Modes DEVICE EXECUTION Internal External nonpage External page mode 1 External page mode 2 External (any) 1Port 2Port 3Port 4Port 5Port MODE Idle or stop Idle Idle Idle Stop ALE 1 1 1 1 1 PSEN 1 1 1 1 1 P0 Port data2 Latched1 Latched1 Latched5 Port data 2 P1 Port data2 Port data2 Port data2 Port data2 Port data2 P2 Port data2 Latched3 Latched5 Latched1 Port data4 P3 Port data2 Port data2 Port data2 Port data2 Port data2 exhibits opcode following instruction that sets the idle bit. reflects data stored in corresponding port SFR. Port 0 functions as an open-drain output in this mode. exhibits address MSB of opcode following instruction that sets the idle bit. reflects data stored in corresponding port SFR. In this mode, the port uses weak pullups. exhibits address LSB of opcode following instruction that sets the idle bit. Ring Oscillator Wake-Up From Stop A typical low-power application is to keep the processor in stop mode most of the time. Periodically, the system wakes up (using an external interrupt), takes a reading of some condition, and then returns to sleep. The duration of full-power operation is as short as possible. One disadvantage to this method is that the clock must be restarted prior to performing a meaningful operation. This startup period is a waste of time and power since no work can be performed. The ultra-high-speed flash microcontroller provides an alternative. _____________________________________________________________________________________________ 84 Ultra-High-Speed Flash Microcontroller User’s Guide If the ring select (RGSL) is enabled, the microcontroller can exit stop mode running from an internal ring oscillator. Upon receipt of an interrupt, this oscillator can start instantaneously, allowing software execution to begin immediately while the oscillator is stabilizing. Ring oscillator execution cannot be used to support accurate baud-rate generation or precise timer/counter operations. Once 65,536 clock cycles have been detected, the CPU automatically switches to the normal oscillator as its clock source. However, if the required interrupt response is very short, the software can reenter stop mode before the crystal is even stable. In this case, stop mode can be invoked and both oscillators are stopped. Speed Reduction The ultra-high-speed flash microcontroller is a fully CMOS 8051-compatible device. It can use significantly less power than other 8051 versions, because it is more efficient. As an average, software runs 10 times faster on the ultra-high-speed flash microcontroller than on other 8051 derivatives. Thus, the same job can be accomplished by slowing down the crystal by a factor of 10. For example, an existing 8051 design that runs at 12MHz can run at approximately 1.2MHz on the ultra-high-speed flash microcontroller. At this reduced speed, the ultra-high-speed microcontroller has lower power consumption than an 8051, yet performs the same job. Using the 10X factor, Table 7-2 shows the approximate speed at which the ultra-high-speed flash microcontroller can accomplish the same work as an 8051. The exact improvement varies depending on the actual instruction mix. Available crystal speeds must also be considered. Refer to Section 14 for information on instruction timing. Table 7-2. Crystal vs. MIPS Comparison ORIGINAL 8051 CRYSTAL SPEED (MHz) MIPS ULTRA-HIGH-SPEED FLASH MICROCONTROLLER CRYSTAL SPEED (MHz) 16 20 24 33 40 1.3 1.6 2.0 2.7 3.3 1.6 2.0 2.4 3.3 4.0 Power Management Modes Power consumption in CMOS microcontrollers is a function of operating frequency. The power management mode (PMM) feature allows software to dynamically match operating frequency and current consumption with the need for processing power. Instead of the default one clock per machine cycle, PMM utilizes 1024 clocks per cycle to conserve power. Several special features have been added to enhance the function of the PMM. The switchback feature allows the device to almost instantaneously return to divide-by-1 mode upon detection of an enabled external interrupt or the receipt of a falling edge on a serial port receiver pin. The advantages of this become apparent when one calculates the increased interrupt service time of a device operating in PMM. In addition, a device operating in PMM would normally be unable to sample an incoming serial transmission to properly receive it. The switchback feature, explained below, allows a device to return to divide-by-4 operation in time to receive incoming serial port data and process interrupts with no loss in performance. A status register (STATUS;C5h) prevents the device from accidentally reducing the clock rate during the servicing of an external interrupt or serial port activity. This register can be interrogated to determine whether an interrupt is in progress, or if serial port activity is occurring. Based on this information the software can delay or reject a planned change in the clock divider rate. In addition, the ultra-high-speed flash microcontroller has the capability to operate from the internal ring oscillator during normal operation, not only during the crystal warmup period. Table 7-3 summarizes the new control bits associated with the power management features. 85 _____________________________________________________________________________________________ Ultra-High-Speed Flash Microcontroller User’s Guide Table 7-3. Power Management and Status Bit Summary BIT NAME CD1, CD0 SWB RESET STATE READ/WRITE ACCESS 10 Write: 10 anytime; 00, 01, and 11 only when previously in 10 state. Unrestricted read. Switchback enable 0 = Interrupts and serial port activity will not affect clock divider control bits 1 = Enabled interrupts and serial port activity will cause a switchback 0 Unrestricted Priority Interrupt Status 101 = Level 4 interrupt (power fail) in progress 100 = Level 3 interrupt in progress 011 = Level 2 interrupt in progress 010 = Level 1 interrupt in progress 001 = Level 0 interrupt in progress 000 = No interrupt in progress 0 Read only LOCATION PMR.7–6 PMR.5 FUNCTION Clock divider control CD1 CD0 osc cycles per system clock cycle 0 0 Crystal multiplier 0 1 Reserved 1 0 1 (reset default) 1 1 1024 (PMM) PIS2:PIS0 STATUS.7:5 SPTA1 STATUS.3 Serial port 1 transmitter activity status 0 = Serial port 1 transmitter inactive 1 = Serial port 1 transmitter active 0 Read only SPRA1 STATUS.2 Serial port 1 receiver activity status 0 = Serial port 1 receiver inactive 1 = Serial port 1 receiver active 0 Read only SPTA0 STATUS.1 Serial port 0 transmitter activity status 0 = Serial port 0 transmitter inactive 1 = Serial port 0 transmitter active 0 Read only SPRA0 STATUS.0 Serial port 0 receiver activity status 0 = Serial port 0 receiver inactive 1 = Serial port 0 receiver active 0 Read only Power Management Mode Timing The power management mode reduces power consumption by internally dividing the clock signal to the device, causing it to operate at a reduced speed. When PMM is invoked, the external crystal continues to operate at full speed. The difference is that the device uses 1024 external clocks to generate each system clock cycle as opposed to one clock per internal system clock cycle in the default state. Relative timing relationships of all signals when the device is operating in PMM remains the same as the one cycle timing. Note that all internal functions, on-board timers (including serial port baud-rate generation), watchdog timer, and software timing loops also runs at the reduced speed. Most applications do not find it necessary to attend to this much detail, but the information is provided for calculating critical timings. Figure 7-2 demonstrates the internal timing relationships during PMM. PMM is entered and exited by setting the clock rate divider bits (PMR.7-6). In addition, it is possible use the switchback feature to effect a return to the divide-by-1 mode from the power management mode. This allows both hardware and software to cause an exit from PMM. Entry to or exit from PMM must be done through the divide-by-1 mode (CD1:0 = 10b). This means that to switch from divide-by1024 to the crystal multiplier 4X mode or vice versa, one must first switch back to divide-by-1 mode. Attempts to execute an illegal speed change are ignored, and the bits remain unchanged. It is the responsibility of the software to test for serial port activity before attempting to change speed, as a modification of the clock divider bits during a serial port operation corrupts the data. _____________________________________________________________________________________________ 86 Ultra-High-Speed Flash Microcontroller User’s Guide MINIMUM INSTRUCTION CYCLE INTERNAL SYSTEM CLOCK (PMM) EXTERNAL CLOCK 1024 CLOCKS Figure 7-3. Internal Timing Relationships in PMM PMM and Peripheral Functions Timers 0, 1, and 2 default on reset to a 12 clock per cycle operation to remain compatible with the original 8051 timing. The timers can be individually configured to run at the fastest instruction cycle timing (divide-by-1) or to a system clock divide-by-4 input by setting the relevant bits in the clock control register (CKCON;8Eh). Because the timers derive their time base from the internal clock, timers 0, 1, and 2 operate at reduced clock rates during PMM. This also affects the operation of the serial ports in PMM. In general, it is not possible to generate standard baud rates while in PMM, and the user is advised to avoid PMM, or use the switchback feature, if serial port operation is desired. Table 7-4 shows the effect of the PMM clock divider option on timer and serial port operation. Table 7-4. Effect of PMM Clock Mode on Timer, Serial Operation CD1:0 11 OSC CYCLES PER TIMER 0/1/2 CLOCK OSC CYCLES PER MACHINE CYCLE 1024 00 3072 TxMH, TxM = 01 1x 1024 1024 OSC CYCLES PER TIMER 2 CLOCK, BAUDRATE GEN. TxMH, TxM = xx 2048 OSC CYCLES PER SERIAL PORT CLOCK MODE 0 SM2 = 0 1 3072 1024 OSC CYCLES PER SERIAL PORT CLOCK MODE 2 SMOD = 0 1 16,348 8192 SWITCHBACK The switchback feature solves one of the most vexing problems faced by power-conscious systems. Many applications are unable to use the stop and idle modes because they require constant computation. Traditionally, system designers could not reduce the operating speed below that required to process the fastest event. This meant that system architects would be forced to operate their systems at the highest rate of speed, even when it was not required. The switchback feature allows a system to operate at a relatively slow speed and burst to a faster mode when required by an external event. When this feature is enabled by setting the switchback enable bit (SWB), (PMR.5), a qualified interrupt, serial port reception, or transmission causes the device to return to the default divide-by-1 mode. A qualified interrupt is defined as an interrupt that has occurred and been acknowledged. This means that an interrupt must be enabled and also not blocked by a higher priority interrupt. After the event is complete, software can manually return the device to PMM. The following sources can trigger a switchback: 87 • External interrupt 0/1/2/3/4/5 • Serial start bit detected, serial port 0/1 • Transmit buffer loaded, serial port 0/1 _____________________________________________________________________________________________ Ultra-High-Speed Flash Microcontroller User’s Guide • Watchdog timer reset • Power-on reset • External reset In the case of a serial port-initiated switchback, the switchback is not generated by the associated interrupt. This is because a device operating in PMM is not able to correctly receive a byte of data to generate an interrupt. Instead, a switchback is generated by a serial port reception on the falling edge associated with the start bit, if the associated receiver enable bit (SCON0.4 or SCON1.4) is set. For serial port transmissions, a switchback is generated when the serial port buffer (SBUF0;99h or SBUF1;C1h) is loaded. This ensures the device is operating in divide-by-1 mode when the data is transmitted, and eliminates the need for a write to the CD1, CD0 bits to exit PMM before transmitting. The switchback feature is unaffected by the state of the serial port interrupt flags (RI_0, TI_0, RI_1, TI_1). The timing of the switchback is dependent on the source. Interrupt-initiated switchbacks occur at the start of the first clock cycle following the event initiating the switchback. In PMM, each internal clock cycle is 1024 external clock cycles. If the current instruction in progress is a write to the IE, IP, EIE, or EIP registers, interrupt processing is delayed until the completion of the following instruction. Serial transmit-initiated switchbacks occur at the start of the instruction following the MOV that loads SBUF0 or SBUF1. Serial reception-initiated switchbacks occur during the cycle in which the falling edge was detected. A few points must be considered when using a serial port reception to generate a switchback. Under normal circumstances, noise on the line or an aborted transmission causes the serial port to time out and the data to be ignored. This presents a problem if the switchback is used, however, because a switchback would occur without indication to the system. If PMM and serial port switchback functions are used in a noisy environment, the user is advised to periodically check if the device has accidentally exited PMM. A similar problem can occur if multiprocessor communication protocols are used in conjunction with PMM. The ultra-high-speed flash microcontroller family supports both the use of the SM2 flag (SCON0.5 or SCON1.5), and the slave address-recognition registers (SADDR0;A9h, SADDR1;AAh, SADEN0;B9h, SADEN1;BAh) for multiprocessor communications. The problem is that an invalid address, which should be ignored by a particular processor, still generates a switchback. As a result, it is not recommended to use a multiprocessor communication scheme in conjunction with PMM. If the system power considerations allow for an occasional erroneous switchback, a polling scheme can be used to place the device back into PMM. CLOCK SOURCE SELECTION The ultra-high-speed flash microcontroller family supports three clock sources for operation. As with most microcontrollers, the device can be clocked from an external crystal using the on-board crystal amplifier, or a clock source can be supplied by an external oscillator. In addition, some members of the family incorporate an on-board ring oscillator to provide a quick resumption from stop mode. The ring oscillator is a low power digital oscillator internal to the microcontroller. When enabled, it provides an approximately 10MHz clock source for device operation without external components. The ring oscillator is not as stable as an external crystal, and software should refrain from performing timing-dependent operations, including serial port activity, while operating from the ring oscillator. The ring oscillator provides many advantages to the designers of microcontroller-based systems. One is that it allows Dallas Semiconductor microcontrollers to perform a fast resume from stop mode, eliminating the crystal warmup delay when restarting the device. The microcontroller must begin operation following a power-on reset from an external clock source, either an external crystal or oscillator. The control and status bits which support the new and/or enhanced features are shown in Table 7-5. Table 7-5. Clock Control and Status Bit Summary BIT NAME RGMD RGSL LOCATION RESET WRITE ACCESS EXIF.2 Ring Oscillator Mode Status. 1 = Ring oscillator is current clock source, 0 = Crystal or external clock is current clock source. FUNCTION 0 None EXIF.1 Ring Oscillator Select, Stop Mode. 1 = Ring oscillator will be the clock source when resuming from stop mode, 0 = Crystal or external clock will be the clock source when resuming from stop mode Note: Upon completion of crystal warm up period, the device will switch to the crystal. — Unrestricted _____________________________________________________________________________________________ 88 Ultra-High-Speed Flash Microcontroller User’s Guide RING OSCILLATOR RESUME FROM STOP To achieve the minimum power consumption during periods of processor inactivity, software can place the device into stop mode. Such systems typically resume operation using an external interrupt, perform some activity, and then return to stop mode. Traditional designs that rely upon an external crystal as the clock source must incur the startup delay of the crystal when resuming from stop mode. This is a waste of time and power, as no work can be performed until the crystal has stabilized. Although the ring oscillator provides an approximately 10MHz clock source for device operation, it is not as stable as an external crystal. As a result, high-accuracy timing operations should be avoided while running from the ring oscillator. This includes using the timers for pulse measurement, and the use of the serial ports in asynchronous modes (1, 2, 3). Serial ports operating in mode 0 are unaffected by the stability of the clock source as a separate synchronizing clock is generated. If the ring oscillator select bit RGSL, (EXIF.1) is set, the device resumes operation immediately using the internal ring oscillator as the clock source. The device continues to run from the ring oscillator until the crystal warmup period of 65,536 clock cycles (measured from the external source) has completed. At this time, the device switches to clock source active before it enters stop mode and continues operation. This allows software execution to begin immediately upon resuming from stop mode. The current clock source is indicated by the ring oscillator mode bit, RGMD (EXIF.2). In stop mode, enabled interrupts become true edge triggered interrupts, compared with the sampled edge detection used during normal operation. This means that external interrupts are more sensitive to noise in stop mode than during normal operation. Applications should be carefully designed to ensure that noise will not cause an erroneous exit from stop mode. SECTION 8: RESET CONDITIONS The condition that causes the microcontroller to vector to address 0000h is a reset. This can happen internally or external to the microcontroller. The reset condition puts the microcontroller in a known state following a course of events not anticipated by the designer. The circuit could be subjected to numerous conditions, such as power brownout, noise due to lightning strike, or corrupted code. RESET SOURCES The microcontroller can enter a reset condition if invoked in one of five ways: • Power-on/power-fail reset • Watchdog timer reset • Oscillator-fail detect reset • Internal system reset • External reset The reset state is the same, regardless of the source of the reset. When in reset, the oscillator is running, but no program execution is allowed. When the reset source is external, the user must remove the reset stimulus to continue operation. When power is applied to the device, the power-on delay removes the stimulus automatically. Power-On/Power-Fail Reset The ultra-high-speed flash microcontroller incorporates an internal voltage reference, which holds the device in power-on reset while VCC is out of tolerance. Once VCC has risen above the threshold, the device restarts the external crystal oscillator and counts 65,536 clock cycles before program execution begins at location 0000h. The power monitor invokes a reset state when VCC drops below the threshold condition. The condition remains in effect while power is below the minimum voltage level. When power returns above the reset threshold, a full power-on reset is performed. This mechanism provides a controlled and predictable startup condition. The processor exits the reset condition automatically once VCC meets the minimum voltage requirement. This helps the system maintain reliable operation by only permitting processor operation when voltage is in a known good state. Software can determine that a power-on reset has occurred by checking the power-on reset flag (POR) in the WDCON register. Software should clear the POR bit after it is read. 89 _____________________________________________________________________________________________ Ultra-High-Speed Flash Microcontroller User’s Guide Watchdog Timer Reset The ultra-high-speed flash microcontroller incorporates a safety feature to prevent corrupted software from controlling the CPU. This feature is called the watchdog timer. It is a free-running timer with a programmable interval. The watchdog supervises the processor operation by requiring software to clear the timer before an overflow occurs. If the timer is enabled and software fails to clear it before this interval expires, the ultra-high-speed flash microcontroller is placed into a reset state. The reset state maintains for 13 clock cycles. Once the reset is removed, the processor resumes execution at address 0000h. Software can determine if a reset is caused by a watchdog timeout by checking the watchdog timer reset flag (WTRF) in the WDCON register. This flag is cleared by software only. Oscillator Fail-Detect Reset Oscillator fail-detect circuitry monitors the on-chip oscillator activity. When enabled, this circuit causes a reset if the oscillator frequency falls below ~20kHz, and holds the chip in reset until the oscillator frequency rises back above ~20kHz. The circuitry is enabled by setting the OFDE (PCON.4) bit to a logic 1. The OFDE bit can be cleared by software or by the occurrence of a power-fail reset. A reset caused by an oscillator failure sets the OFDF (PCON.5) flag bit to a logic 1. This flag can be cleared by software or by a power-on reset. The oscillator fail-detect circuitry utilizes the internal ring oscillator to clock the chip into the reset state and maintain the reset state while the oscillator is below the minimum frequency. Note, however, that the circuitry does not force a reset when the oscillator is purposely stopped when software invokes stop mode. External Reset If the RST input is asserted to logic 1, the device is forced into a reset state. An external reset is accomplished by holding the RST pin high at least four clock cycles while the oscillator is running. Once the reset state is invoked, it is maintained as long as RST is asserted at logic 1. When the RST is removed, the processor exits the reset state within four clock cycles and begins execution at address 0000h. If an RST is applied while the processor is in the stop mode, the RST causes the oscillator to begin running and forces the program counter to 0000h. The reset delay is 65,536 clock cycles to allow the oscillator to stabilize. The RST pin is a bidirectional I/O. If a reset is caused by a power fail reset, a watchdog timer reset, an oscillator fail detect reset, or an internal system reset, a positive output level is also generated at the RST pin. This reset level is asserted as long as an internal reset is asserted. The drive capability of this I/O port may be insufficient if the RST pin is connected to a RC reset circuit. Connecting the RST pin to a capacitor would not affect the internal reset condition. Determining The Cause Of a Reset During the debugging process, it might be necessary to isolate the cause of a device reset. Because resets are initiated by a limited number of sources, it is relatively easy to determine their source by interrogating the flag bits associated with the reset sources. The table below lists the reset sources and flag bits. Although no flag bits are associated with the internal system reset generated by issuing a system reset or complement bank-select flash command, it is unlikely that these would occur unintentionally, given that the flash command bits (FCNTL.3-0) require a timed-access write. Reset Source Flag Associations RESET SOURCE Power-on reset Watchdog reset Oscillator fail-detect reset Internal system reset External reset FLAG BIT POR – WDCON.6 WTRF – WDCON.2 OFDF – PCON.5 None None _____________________________________________________________________________________________ 90 Ultra-High-Speed Flash Microcontroller User’s Guide SECTION 9: INTERRUPTS The ultra-high-speed microcontroller family improves upon the traditional 8051 architecture by utilizing a five-priority interrupt system. The five priority levels, from highest priority to lowest, are 4, 3, 2, 1, and 0. The power-fail interrupt, when enabled, always receives the highest priority (level 4), while other interrupt sources can be configured to level 3, 2, 1, or 0. Each source has independent priority bits, flag(s), interrupt vector, and enable. In addition, interrupts can be globally enabled (or disabled). The interrupt system is compatible with the original 8051 family, having all of the original interrupts available. A summary of all interrupt sources is provided in the table below. Table 9-1. Interrupt Summary PRIORITY CONTROL INTERRUPT INTERRUPT VECTOR NATURAL ORDER FLAG ENABLE Power-fail 33h 0 (highest) PFI (WDCON.4) EPFI (WDCON.5) N/A External interrupt 0 03h 1 IE0 (TCON.1)** EX0 (IE.0) MPX0 (IP1.0) LPX0 (IP0.0) Timer 0 overflow 0Bh 2 TF0 (TCON.5)* ET0 (IE.1) MPT0 (IP1.1) LPT0 (IP0.1) External interrupt 1 13h 3 IE1 (TCON.3)** EX1 (IE.2) MPX1 (IP1.2) LPX1 (IP0.2) Timer 1 overflow 1Bh 4 TF1 (TCON.7)* ET1 (IE.3) MPT1 (IP1.3) LPT1 (IP0.3) Serial port 0 23h 5 RI_0 (SCON0.0), TI_0 (SCON0.1) ES0 (IE.4) MPS0 (IP1.4) LPS0 (IP0.4) Timer 2 overflow 2Bh 6 TF2 (T2CON.7) EXF2(T2CON.6) ET2 (IE.5) MPT2 (IP1.5) LPT2 (IP0.5) Serial port 1 3Bh 7 RI_1 (SCON1.0), TI_1 (SCON1.1) ES1 (IE.6) MPS1 (IP1.6) LPS1 (IP0.6) External interrupt 2 43h 8 IE2 (EXIF.4) EX2 (EIE.0) MPX2 (EIP1.0) LPX2 (EIP0.0) External interrupt 3 4Bh 9 IE3 (EXIF.5) EX3 (EIE.1) MPX3 (EIP1.1) LPX3 (EIP0.1) External interrupt 4 53h 10 IE4 (EXIF.6) EX4 (EIE.2) MPX4 (EIP1.2) LPX4 (EIP0.2) External interrupt 5 5Bh 11 IE5 (EXIF.7) EX5 (EIE.3) MPX5 (EIP1.3) LPX5 (EIP0.3) Watchdog interrupt 63h 12 WDIF (WDCON.3) EWDI (EIE.4) MPWDI (EIP1.4) LPWDI (EIP0.4) Unless marked, these flags must be cleared manually by software. *Cleared automatically by hardware when the service routine is vectored to. **If edge-triggered, cleared automatically by hardware when the service routine is vectored to. If level-triggered, flag follows the state of the pin. INTERRUPT OVERVIEW An interrupt allows the software to react to unscheduled or asynchronous events. When an interrupt occurs, the CPU is expected to “service” the interrupt. This service takes the form of an interrupt service routine (ISR). The ISR resides at a predetermined address, as shown in Table 9-1. When the interrupt occurs, the CPU vectors to this address and runs code created to service the interrupt. The CPU stays in an interrupt service state until the return from interrupt instruction (RETI) is executed at completion of the ISR. When an RETI is performed, the processorl returns to the instruction that would have been next when the interrupt occurred. Once an ISR has begun, it can be interrupted only by a higher priority interrupt. 91 _____________________________________________________________________________________________ Ultra-High-Speed Flash Microcontroller User’s Guide When an interrupt condition occurs, the processor indicates this by setting a flag bit. This flag bit cannot alone cause an interrupt, and is set regardless of whether the interrupt is enabled. Most flags must be cleared manually by software. However, IE0 and IE1 are cleared automatically by hardware upon vectoring to the service routine if the interrupt was edge-triggered. In level-triggered mode, the IE0 or IE1 flags will follow the state of the pin. Flags TF0 and TF1 are always cleared automatically when the service routine is vectored to. Refer to the individual bit descriptions for more details. Each source must be individually enabled in order to generate CPU interrupts. Each interrupt source has an independent enable, as shown in Table 9-1. In order for the processor to acknowledge the interrupt and vector to the ISR, the enable all bit (IE.7: EA) must be set to globally enable interrupt sources. Clearing the EA bit to a logic 0 disables all interrupts, regardless of the individual interrupt enables. The power-fail warning interrupt source is the only exception, requiring only its individual enable bit be set (WDCON.5: EPFI) to be recognized by the CPU. The EA bit has no effect on the power-fail interrupt. INTERRUPT SOURCES The interrupt sources present on the ultra-high-speed microcontroller can broken into several categories: external, timer-based, serial communication, and power-fail. Each type is described in the following pages. Interrupt sources are evaluated during the final memory cycle of each instruction to determine whether and which interrupt is serviced. If the interrupt source goes active after this evaluation, it is not considered until the final memory cycle of the next instruction. External Interrupts The ultra-high-speed microcontroller has six external interrupt sources. These include the standard two interrupts of the 8051 architecture and four new sources. The original interrupts are INT0 and INT1. These are active-low and can be programmed to be edge- or level-sensitive. The detection mode for each source is controlled through TCON register bits IT0 and IT1, respectively. When ITx = 0, the interrupt is triggered by a logic 0 on the appropriate interrupt pin. The interrupt condition remains in effect as long as the pin is low. When ITx = 1, the interrupt is pseudo-edge-triggered. This means that the interrupt is activated, if on successive samples the pin is found to be in a low state, indicating that a falling edge occurred. Since the external interrupts are sampled, the pin driver of an edgetriggered interrupt should hold both the high and then the low condition for at least two system clock cycles (each) to ensure detection. This means maximum sampling frequency on any interrupt pin is one-fourth of the system clock frequency. It is important to note that level-sensitive interrupts are not latched. This is most important if using other interrupts of equal or higher priority, because the level-sensitive interrupt request may not receive immediate service by the processor. A level-sensitive interrupt request is missed unless the condition is held until it can be serviced. The remaining four external interrupts are similar in nature, with two differences. First, the four new interrupts are edge-detect only. They do not have level-detect modes. Second, INT2 and INT4 are positive-edge sensitive instead of negative-edge sensitive. All associated bits and flags operate the same and have the same polarity as the original two. A logic 1 indicates the presence of a condition, not the logic state of the pin. If the power management mode is utilized, the designer must remember that detection of edge-triggered interrupts is defined in relation to the system clock (= 1024 oscillator clock cycles). This means that it requires 2048 external clock cycles before detecting that an edge has just occurred. As a result, the latency for these interrupts is much longer in power management mode. Timer Interrupts The ultra-high-speed flash microcontroller incorporates three 16-bit programmable timers, each of which can generate an interrupt, and a programmable watchdog timer. The three 16-bit programmable timers operate in the same manner as the 80C52. Each timer has an independent interrupt enable, flag, vector, and priority. The watchdog timer also has its own interrupt enable, flag, and priority. Timers 0 and 1 set their respective interrupt flags when the timer overflows from a full condition, depending on its mode. This flag is set regardless of the interrupt enable state. If the interrupt is enabled, this event generates a call to the corresponding interrupt vector. For timers 0 and 1, the flags are cleared when the processor jumps to the interrupt vector. Thus, these flags are not available for use by the interrupt service routine (ISR), but are available outside of the ISR and in applications that do not acknowledge the interrupt (i.e., jump to the vector). If the interrupt is not acknowledged, then software must manually clear the flag bit. In addition to having an interrupt flag for an overflow condition (as is the case for timers 0 and 1), timer 2 has a second interrupt flag (EXF2) that is associated with detection of a falling edge on the T2EX (P1.1) pin. When timer 2 has been configured in capture mode (CP/RL2 = 1, EXEN2 = 1) or autoreload mode (CP/RL2 = 0, EXEN2 = 1, DCEN = 0), a negative transition on T2EX causes the EXF2 interrupt flag to be set. For timer 2 interrupts, jumping to the interrupt vector does not clear either of the flags. Instead, software must ascertain which flag caused the interrupt and clear it manually. Timer 0 and 1 flag bits reside in the TCON register. Timer 2 flag bits reside in the T2CON register. The interrupt enables and priorities for timers 0, 1, and 2 reside in the IE and IP0, IP1 registers. _____________________________________________________________________________________________ 92 Ultra-High-Speed Flash Microcontroller User’s Guide The watchdog interrupt usually has a different connotation than the timer interrupts. Unless the watchdog is being used as a very long timer, the interrupt means that the software has failed to reset the timer and may be lost. The watchdog ISR can attempt to determine the system state or allow the CPU to be reset if the watchdog reset function has been enabled (EWT = 1). Like other sources, the watchdog timer has a flag bit, enable bit, priority bits, and its own vector. Serial Communication Interrupts Each UART is capable of generating an interrupt. Each UART has its own interrupt enable, vector, and priority. Each UART interrupt has two flags (RI, TI) that are used by the ISR to determine whether the interrupt comes from a received word or a transmitted one. Unlike the timers, the UART flags are not altered when the interrupt is serviced. Software must change them manually. When a UART finishes the transmission of a word, the TI bit is set and an interrupt is generated (if enabled). Likewise, the UART sets the RI bit and generates an interrupt when a word is completely received. The CPU is not notified until the word is completely received or transmitted. Power-Fail Interrupt The ultra-high-speed microcontroller has the ability to generate an interrupt when VCC drops below a predetermined level. By comparing a fixed ratio of VCC versus an internal reference, the microcontroller can assess when VCC drops below the VPFW level and cause an interrupt (if enabled). The level of VPFW is provided in the data sheet on DC electrical specifications. The power-fail interrupt is a level-sensitive condition, and remains in effect as long as VCC remains below VPFW. The power-fail interrupt has the highest priority level, which cannot be altered by the user. The EPFI bit solely controls the enabling or disabling of the power-fail interrupt source, and is not subject to the global interrupt enable (EA). The EPFI bit should always be cleared to a logic 0 state if the power-fail interrupt is not needed. Simulated Interrupts Software can simulate any interrupt source by setting the corresponding flag bit. This forces an interrupt condition that is acknowledged (if enabled) and is otherwise indistinguishable from the real thing. Thus, an interrupt flag bit should never be set to a logic 1 by software inadvertently. Once an interrupt has been acknowledged, software cannot prevent or end the interrupt by clearing its flag. If, however, software clears an interrupt flag before the interrupt is acknowledged, the interrupt does occur. INTERRUPT PRIORITIES The ultra-high-speed microcontroller has five interrupt priority levels. The five priority levels, from highest priority to lowest, are 4, 3, 2, 1, and 0. The power-fail interrupt, when enabled, always receives the highest priority (level 4), while the remaining interrupt sources can individually be programmed to level 3, 2, 1, or 0. The lowest priority (level 0) is the default condition for the other sources. An interrupt being serviced can only be interrupted by a higher priority interrupt. The power-fail interrupt source, assigned priority level 4, therefore, has the ability to interrupt the service routine of any other source. No interrupt source with equal or lesser priority to one currently being serviced can interrupt the service routine. If two interrupt sources of equal priority levels are requested simultaneously, natural priority is used to arbitrate. The natural priority is given in Table 9-1. Note that natural priority is only used to resolve simultaneous requests. Once an interrupt of a given priority is invoked, only a source that is programmed with a higher priority can intercede. Interrupt Register Conflicts During normal operation, there is a small but finite probability that application software might try to read or modify a register associated with interrupt functions at the same time that the interrupt hardware is modifying the register. In general, these hardware/software interrupt conflicts are resolved according to the “hardware wins” philosophy: In the event of a conflict, the hardware modification of a register takes precedence over the software action to ensure that the interrupt event is not missed. To assist in prevention of hardware/software conflicts, the interrupt selection process that normally occurs in the final memory cycle of each instruction is aborted for any instructions that write to the IP0, IP1, EIP0, EIP1, IE, or EIE registers. When the evaluation takes place in the subsequent instruction, the interrupt source incorporates the new priority and enable values from the previous instruction. If this situation occurs, it lengthens the interrupt latency by the length of the instruction that modified the register. 93 _____________________________________________________________________________________________ Ultra-High-Speed Flash Microcontroller User’s Guide INTERRUPT ACKNOWLEDGE CYCLE The process of acknowledging an interrupt begins with the setting of the associated flag. For edge-triggered external interrupts and internal interrupt sources, the interrupt flags are set automatically by hardware. For level-sensitive external interrupts, the flags are actually under control of the external signal, and the flag rises and falls with the pin level. All interrupt flags are evaluated on the final execution cycle of each instruction. A priority decoding process is performed among pending and new interrupt sources in order to select the appropriate interrupt vector address. This decoding process is accomplished in a single memory cycle using combinatorial logic. Hardware then forces an LCALL to the selected vector address in the following memory cycle, unless blocked by one of the following conditions: • An interrupt of equal or greater priority has already been invoked and the RETI instruction has not been issued to terminate it, • The current cycle is not the final cycle in the execution of the current instruction, or • The instruction in progress is an RETI or a write to IP0, IP1, EIP0, EIP1, IE, or EIE. FLAG BITS PFI INTERRUPT ENABLE BITS INTERRUPT SELECTION INTERRUPT PRIORITY BITS HARDWARE HIGHEST PRIORITY INT0 TF0 INT1 IT1 TF1 RI 0 TI_0 INTERRUPT VECTOR TF2 EXF2 RI_1 TI 1 INT2 INT3 INT4 INT5 WATCHDOG INDIVIDUAL GLOBAL ENABLES ENABLE Figure 9-1. Interrupt Functional Description _____________________________________________________________________________________________ 94 Ultra-High-Speed Flash Microcontroller User’s Guide INTERRUPT LATENCY Interrupt response time is normally between 4 and 18 memory cycles, depending on the state of the microcontroller when the interrupt occurs. If the microcontroller is performing an ISR with equal or greater priority, interrupt latency increases because the new interrupt is not invoked. In other cases, the response time depends on the current instruction. The fastest possible response to an interrupt is four memory cycles. The four memory cycle response time includes one cycle for detecting the interrupt and three cycles to perform the LCALL that is inherent in the interrupt request. The maximum response time occurs if the microcontroller is performing a JBC instruction that clears a bit in IE, IP0, EIE, or EIP0, and then executes a DIV as the next instruction. From the time an interrupt source is activated (not detected), the longest reaction time is 18 memory cycles. This includes one cycle to detect the interrupt, four cycles to finish the JBC, ten cycles to perform the DIV, then three cycles for the LCALL to the ISR. This maximum response time of eighteen memory cycles assumes that there are no other pending interrupts of higher priority to be serviced and that the JBC instruction is not preceded and does not jump to any instruction that aborts the priority decoding process (RETI or writes to IP0, IP1, EIP0, EIP1, IE, or EIE). SECTION 10: I/O PORTS The ultra-high-speed flash microcontroller provides 8-bit I/O ports. Each port appears as a special function register that can be addressed as a byte or 8 individual bit locations. In general, the register and the port pin have identical values, and reading or writing a port is the same as reading or writing the SFR for the port. The basic I/O driver function and its electrical characteristics are similar to the drivers used in the DS87C520, with respect to individual port and pin conditions. Port 0 and port 2 can serve either as general-purpose parallel I/O ports or as the expanded memory bus. Ports 1 and 3 can be used as general-purpose parallel I/O ports with optional special functions associated with each pin. Enabling the special function for a pin automatically converts the I/O pin to that function. An optional function of a pin can be turned on and off dynamically to suit the application. Using one or more I/O pins of a port as special functions does not affect the remaining port pins. It should be noted that port 0 drivers are open-drain and require external pullups when used as general-purpose I/O ports. EXTERNAL ADDRESS CONTROL ADDRESS\ DATA VCC PORT 0.n INTERNAL DATA BUS D Q Q WRITE ENABLE READ ENABLE READ LATCH/PIN Figure 10-1. Port 0 Functional Circuitry 95 _____________________________________________________________________________________________ Ultra-High-Speed Flash Microcontroller User’s Guide PARALLEL I/O Each I/O port can be used as a general-purpose, bidirectional parallel I/O port. Data written to the port latch serves to set both the level and the direction of the data on the pin. The output of the port pin is established by writing to the associated port pin latch. When logic 0 is written to the port for output, the port is pulled to ground. When logic 1 is written to ports 1, 2, or 3, a strong driver momentarily drives the pin from 0 to 1, and then a weak pullup maintains the 1. A logic 1 written to port 0 causes those pins to go tri-state, functioning as open-drain outputs. A logic 1 in the port latch also configures the port pin to the input state. Since the pin is either weakly pulled up or in three-state, the pin is the same as the driven logic state. The logic state of the pin itself does not alter the logic value of the port latch. PORT 0 This is an open drain, 8-bit, bidirectional, general-purpose I/O port. A reset condition or logic 1, written to the latches of this port, threestate, the port pins. This condition also serves as an input mode. When used as an I/O port, external pullups are required. As an alternate function, this port can be used as part of the multiplexed address/data bus to access external memory. Both nonpage and page mode are supported. During the original 8051 expanded bus configuration (nonpage mode), when ALE is high, the LSB of the address is presented to P0. When ALE is low, the port transitions to a bidirectional data bus. When used in page mode 1, P0 is used as the primary data bus only. When used in page mode 2, P0 is used for the LSB of the address only. The use of port 0 as general-purpose I/O is not recommended if the device is used to access external memory. In this case, the state of the pins are disturbed during the memory access. In addition, the pullups required to maintain a high state during the use as general-purpose I/O interfere with the complementary drivers employed when the device operates as an expanded memory bus. When port 0 is used as an address bus, the AD0-7 pins provide true drive capability for logic levels 1 and 0. No external pullups are required. In fact, external pullups degrade the memory interface timing. A two-state system is used on AD0-7. This allows the pin to be driven hard for a period of time, allowing the greatest possible setup or access time. The pin states are then held in a weak latch until forced to the next state or overwritten by an external device. This assures a smooth transition between logic states and also allows a longer hold time. ADDRESS A8-A15 ADDRESS CONTROL VCC DELAY = 2Tclk VCC PORT 2.n INTERNAL DATA BUS D Q Q VCC WRITE ENABLE READ ENABLE READ LATCH/PIN Figure 10-2. Port 2 Functional Circuitry _____________________________________________________________________________________________ 96 Ultra-High-Speed Flash Microcontroller User’s Guide Internal Memory Cycles Ext Memory Cycle C1 C2 C3 C4 Ext Memory Cycle C1 C2 C3 C4 XTAL1 ALE PSEN LSB Add Data LSB Add Data Port 0 Port 2 MSB Add MSB Add Figure 10-3. External Program Memory Access (Nonpage Model) PORT 2 Port 2 is an 8-bit bidirectional I/O port. The reset condition sets the port pins to logic 1. In this state, a weak pullup holds the port pin high. This condition also serves as an input mode, since any external circuit that writes to the port overcomes the weak pullup. Writing a logic 0 to the port pin activates a strong pulldown that remains on until a 1 is written or a reset occurs. Writing a logic 1 after the port has been at 0 causes a strong transition driver to turn on, followed by a weaker sustaining pullup. Once the momentary strong driver turns off, the pin assumes both the output high and input state. As an alternate function, port 2 can function as the MSB of the address bus for external memory access during nonpage mode. When used in page mode 1, P2 is used for both LSB and MSB of external address. When used in page mode 2, P2 is used for the MSB of external address and data. PORT 1 Port 1 functions as both an 8-bit bidirectional I/O port and an alternate functional interface for timer 2 I/O, external interrupts 2, 3, 4, 5, and serial port 1. Reset conditions set these port pins to logic 1 and are held high with weak pullups. This condition also serves as an input mode, since any external circuit that writes to the port overcomes the weak pullup. When a logic 0 is written to any port pin, the port activates a strong pulldown that remains on until a 1 is written or a reset occurs. Writing a logic 1 after the port has been a 0 causes a strong transition driver to turn on, followed by a weaker sustaining pullup. Once the momentary strong driver turns off, the pin assumes both the output high and input state. PORT 3 Port 3 functions as both an 8-bit bidirectional I/O port and an alternate functional interface for external interrupts 0 and 1, serial port 0, timers 0 and 1 inputs, and external data memory strobes. The reset condition sets all bits to logic 1. In this state, a weak pullup holds the port high. This condition also serves as an input mode, since any external circuit that writes to the port overcomes the weak pullup. Writing a logic 0 to any port pin activates a strong pulldown that remains on until a 1 is written or a reset occurs. Writing a logic 1 after the port has been a 0 causes a strong transition driver to turn on, followed by a weaker sustaining pullup. Once the momentary strong driver turns off, the pin assumes both the output high and input state. Alternate Functions of Ports 1 and 3 When any of the pins of ports 1 and 3 is enabled as a special function, the port latch should be programmed to logic 1 to avoid potential bus contention and ensure proper operation. The drive characteristics of these pins do not change when the pins are used for general I/O or as a special function associated with the pin. port 0 and 2 pins, as well as PSEN, ALE, P3.6 and P3.7, incorporate special circuitry to limit the current consumed by the device. 97 _____________________________________________________________________________________________ Ultra-High-Speed Flash Microcontroller User’s Guide MOVX Instruction (Page miss) 1st Cycle 2nd Cycle 3rd Cycle 4th Cycle 9th Cycle XTAL1 ALE PSEN RD / WR Inst Inst Inst Inst Inst Port 0 Port 2 Inst Data LSB LSB LSB LSB LSB MSB MOVX Instruction Fetch LSB LSB Memory Access (Stretch = 4) MOVX Instruction (Page hit) 1st 2nd Cycle Cycle ALE 3rd Cycle 4th Cycle 5th Cycle 9th Cycle PSEN RD / WR Inst Inst Inst Inst Inst Port 0 Port 2 Inst Inst Data LSB LSB LSB LSB LSB MOVX Instruction Fetch LSB LSB LSB Memory Access (Stretch = 4) Figure 10-4. External Program Memory Access (Page Mode 1) INTERNAL MEMORY CYCLES C1 PAGE MISS C2 C3 EXT CODE FETCHES PAGE HIT C4 C1 C2 PAGE HIT C1 C2 XTAL1 ALE PSEN PORT 0 LSB ADD PORT 2 MSB ADD DATA LSB ADD LSB ADD DATA DATA Figure 10-5. External Program Memory Access (Page Mode 2) _____________________________________________________________________________________________ 98 Ultra-High-Speed Flash Microcontroller User’s Guide The special functions and the associated port pins are listed below: P1.0 T2 Timer 2 output P1.1 T2EX Timer 2 capture/reload input P1.2 RXD1 Serial receive UART1 P1.3 TXD1 Serial transmit UART1 P1.4 INT2 External interrupt 2, rising-edge active P1.5 INT3 External interrupt 3, falling-edge active P1.6 INT4 External interrupt 4, rising-edge active P1.7 INT5 External interrupt 5, falling-edge active P3.0 RXD0 Serial receive UART0 P3.1 TXD0 Serial transmit UART0 P3.2 INT0 External interrupt 0, falling-edge active P3.3 INT1 External interrupt 1, falling-edge active P3.4 T0 Timer 0 input P3.5 T1 Timer 1 input P3.6 WR External data memory-write strobe P3.7 RD External data memory-read strobe Read-Modify-Write The normal read instructions associated with the I/O ports read the pin state without regard to the port latches. However, the read-modify-write instructions read the state of the port latches instead of the port pins. This type of instruction reads the contents of an SFR, then modifies it in the ALU, and returns it to the original source. The instructions of this type are listed below: ANL Logical AND ORL Logical OR XRL Logical exclusive OR JBC Branch if bit set and clear bit CPL Complement bit INC Increment DEC Decrement DJNZ Decrement and branch if not zero MOV Px.n, C Move carry bit to bit n of port x CLR Px.n Clear bit n in port x SETB Px.n Set bit n in port x Output Functions The I/O ports appear to be true I/O, but their output characteristics are dependent on the individual port and pin conditions. When software writes logic 0 to the port for output, the port is pulled to ground. When software writes logic 1 to the port for output, ports 1, 2, or 3 drive weak pullups (after the strong transition from 0 to 1). Thus, as long as the port is not heavily loaded, true logic values are output. Port 0, having open-drain outputs, three-states when written to logic 1 and hence requires external pullups be present when used as an output. DC drive capability is provided in the electrical specifications. Note that the DC current available from an I/O port pin is a function of the permissible voltage drop. Transition current is available to help move the port pin from logic 0 to logic 1. Since the logic 0 driver is strong, no additional drive current is needed in the 1 to 0 direction. The transition current is applied when the port latch is changed from logic 0 to logic 1. Writing logic 1 where a 1 was already in place does not change the strength of the pullup. This transition current is applied for two oscillator cycles. The absolute current is not guaranteed, but is approximately 2 mA at 5V. 99 _____________________________________________________________________________________________ Ultra-High-Speed Flash Microcontroller User’s Guide When serving as an I/O port, the drive varies as follows: for logic 0, the port invokes a strong pulldown; for logic 1, the port invokes a strong pullup for two oscillator cycles to assist with the logic transition. Then the port reverts to a weak pullup. This weak pullup is maintained until the port transitions from logic 1 to logic 0. External circuits can overdrive the weak pullup. This allows the logic 1 output state to serve as the input state as well. Substantial DC current is available in both the high and low levels. However, the power dissipation limitations make it inadvisable to heavily load multiple pins. In general, sink and source currents should not exceed 10mA total per port (8 bits) and 25mA total per package. Input Functions The input state of the I/O ports is the same as that of the output logic 1. That is, the pin is pulled weakly to logic 1. This logic 1 state is easily overcome by external components. Thus, after software writes a 1 to the port pin, the port is configured for input. When the port is read by software, the state of the pin is read. The only exception is the read-modify-write instructions, discussed earlier. If the external circuit is driving logic 1, then the pin is logic 1. If the external circuit is driving logic 0, then it overcomes the internal pullup. Thus, the pin is the same as the driven logic state. Note that the port latch is not altered by a read operation. Therefore, if logic 0 is driven onto a port pin from an external source, then removed, the pin reverts to the weak pullup, as determined by the internal latch. SECTION 11: PROGRAMMABLE TIMERS The ultra-high-speed microcontroller incorporates three 16-bit programmable timers and has a watchdog timer with a programmable interval. Because the watchdog timer is significantly different from the other timers, it is described separately. The 16-bit timers are referred to as timers. The three timers offer the same controls and I/O functions that were available in the 80C32. As mentioned, the actual timing of these functions is user selectable to be compatible with the instruction cycle of the older generation of 8051 family (12 clocks per tick) or the new generation (1 clock per tick). The timing for each of the three timers can be selected independently and can be changed dynamically. In most modes, the timers can be used as either counters of external events or timers. When functioning as a counter, 1 to 0 transitions on a port pin are monitored and counted. When functioning as timers, they effectively count oscillator or system clock cycles. The time base for the timer function is detailed later in this section. Because an input clock pulse must be sampled high for two system clock cycles and low for two system clock cycles in order to be recognized, this sets the maximum sampling frequency on any timer input at one-fourth of the main system clock frequency. Since the ultra-high-speed microcontroller timers have a variety of features, the following lists summarize the capabilities: Timer 0 Timer 1 Timer 2 13-bit timer/counter 13-bit timer/counter 16-bit timer/counter 16-bit timer/counter 16-bit timer/counter 16-bit timer with capture 8-bit timer w/ autoreload 8-bit timer w/ autoreload 1 6-bit autoreload timer/counter Two 8-bit timer/counters External control pulse timer/counter 16-bit up/down autoreload External control pulse timer/counter Baud-rate generator Timer/counter Baud-rate generator Timer output clock generator 16-BIT TIMERS Timers 0 and 1 are nearly identical. Timer 2 has several additional features such as up/down counting, capture values, and an optional output pin that make it unique. The following table summarizes the SFR bits that control operation of timers 0, 1, and 2. Detailed bit descriptions can be found in Section 4. After the table, timers 0 and 1 are described first, followed by a separate description for timer 2. As mentioned above, the time base for each timer can be varied, which is discussed in more detail in the following pages. ____________________________________________________________________________________________ 100 Ultra-High-Speed Flash Microcontroller User’s Guide BIT NAMES GATE C/T M1, M0 TF0 TR0 T0M T0MH DESCRIPTION Gate control enable for INT0 pin Counter/timer select Timer mode select bits Timer overflow flag Timer run control Input clock select (/4) Input clock high-speed select (/1) Timer LSB Timer MSB REGISTER LOCATION TMOD – 89h TMOD – 89h TMOD – 89h TCON – 88h TCON – 88h CKCON – 8Eh CKMOD – 96h TL0 – 8Ah TH0 – 8Ch BIT POSITIONS TMOD.3 TMOD.2 TMOD.1,0 TCON.5 TCON.4 CKCON.3 CKMOD.3 GATE C/T M1, M0 TF1 TR1 T1M T1MH Gate control enable for INT1 pin Counter/timer select Timer mode select bits Timer overflow flag Timer run control Input clock select (/4) Input clock high-speed select (/1) Timer LSB Timer MSB TMOD – 89h TMOD – 89h TMOD – 89h TCON – 88h TCON – 88h CKCON – 8Eh CKMOD – 96h TL1 – 8Bh TH1 – 8Dh TMOD.7 TMOD.6 TMOD.5,4 TCON.7 TCON.6 CKCON.4 CKMOD.4 TF2 EXF2 RCLK TCLK EXEN2 TR2 C/T2 CP/RL2 T2OE DCEN T2M T2MH Timer overflow flag Timer external flag Timer 2 receive serial clock enable Timer 2 transmit serial clock enable External enable for T2EX pin Timer run control Counter/timer select Capture/reload select Output enable for T2 pin Down count enable Input clock select (/4) Input clock high-speed select (/1) Timer LSB Timer MSB Timer capture LSB Timer capture MSB T2CON – C8h T2CON – C8h T2CON – C8h T2CON – C8h T2CON – C8h T2CON – C8h T2CON – C8h T2CON – C8h T2MOD – C9h T2MOD – C9h CKCON – 8Eh CKMOD – 96h TL2 – CCh TH2 – CDh RCAP2L – CAh RCAP2H – CBh T2CON.7 T2CON.6 T2CON.5 T2CON.4 T2CON.3 T2CON.2 T2CON.1 T2CON.0 T2MOD.1 T2MOD.0 CKCON.5 CKMOD.5 TIMER 2 TIMER 1 TIMER 0 Table 11-1. Programmable Timers TIMER 0, TIMER 1 MODES Timers 0 and 1 both have three common operating modes. They are 13-bit timer/counter, 16-bit timer/counter, and 8-bit timer/counter with autoreload. Timer 0 can additionally be configured to operate as two 8-bit timers. These four modes, controlled by the TMOD register, are detailed in the following pages. MODE 0 Mode 0 configures either timer 0 or timer 1 for operation as a 13-bit timer/counter. As shown in Figure 11-1, setting TMOD register bits M1, M0 = 00b selects this operating mode for either timer 0 or timer 1. When using timer 0, TL0 uses only bits 0 through 4. These bits serve as the 5 LSbs of the 13-bit timer. TH0 provides the 8 MSbs of the 13-bit timer. Bit 4 of TL0 is used as a ripple out to TH0 bit 0, thereby completely bypassing bits 5 through 7 of TL0. Once the timer is started using the TR0 (TCON.4) timer enable, the timer counts as long as GATE (TMOD.3) is 0 or GATE is 1 and pin INT0 is 1. It counts oscillator or system clock cycles if C/T (TMOD.2) is set to a logic 0 and 1 to 0 transitions on T0 (P3.4) if C/T is set to a 1. When the 13bit count reaches 1FFFh (all ones), the next count causes it to roll over to 0000h. The TF0 (TCON.5) flag is set and an interrupt occurs if enabled. The upper 3 bits of TL0 are indeterminate. 101 ____________________________________________________________________________________________ Ultra-High-Speed Flash Microcontroller User’s Guide Note that when used as a timer, the input clock selection can be affected by the clock divide bits (PMR.7-6), the TxM bit (in the CKCON register), and the TxMH bit (in the CKMOD register). The time base selection is described in more detail later in this section. Mode 0 operates identically when timer 1 is used. The same information applies to TL1 and TH1, which form the 13-bit register. TR1 (TCON.6), INT1 (P3.3), T1 (P3.5) and the relevant C/T (TMOD.6), and GATE (TMOD.7) bits have the same functions. MODE 1 Mode 1 configures the timer for 16-bit operation as either a timer or counter. Figure 11-1 shows that setting the TMOD select bits M1, M0 = 01b invoke this operating mode. For timer n, all of the TLn and THn registers are used. For example, if timer 1 is configured in mode 1, then TL1 holds the LSB and TH1 holds the MSB. Rollover occurs when the timer reaches FFFFh. An interrupt also occurs if enabled and the relevant TFn flag is set. Time-base selection, counter/timer selection, and the gate function operate as described in mode 0. MODE 2 This mode configures the timer as an 8-bit timer/counter with automatic reload of the start value. This configuration is shown in Figure 11-2, and is selected when bits M1and M0 of the TCON register are set to 1 and 0, respectively. When configured in mode 2, the timer uses TLn to count and THn to store the reload value. Software must initialize both TLn and THn with the same starting value for the first count to be correct. Once the TLn reaches FFh, it is automatically loaded with the value in THn. The THn value remains unchanged unless modified by software. Mode 2 is commonly used to generate baud rates since it runs without continued software intervention. As in modes 0 and 1, mode 2 allows counting of either clock cycles or pulses on pin Tn (C/T = 1) when counting is enabled by TRn and the proper setting of GATE and INTn pins. T0M = CKCON.3 (T1M = CKCON.4) DIVIDEBY-12 EXTERNAL OSCILLATOR DIVIDE- 1 BY-4 INPUT TO TIMER CLK MODE SYSCLK DIVIDE-BY-1 OSC / 1 OSC / 0.5 2X 4X OSC / 0.25 T0MH = CKMOD.3 (T1MH = CKMOD.4) 0 C / T = TMOD.2 0 1 T0 = P3.4 (T1 = P3.5) (C / T = TMOD.6) 0 CLK TL0 (TL1) 0 1 00 01 TR0 = TCON.4 (TR1 = TCON.6) GATE = TMOD.3 (GATE = TMOD.7) 7 4 0 TH0 (TH1) MODE 0 M1, M0 = TMOD.1, TMOD.0 (M1, M0 = TMOD.5, TMOD.4) MODE 1 7 INT0 = P3.2 ( INT1 = P3.3) TF0 = TCON.5 (TF1 = TCON.7) INTERRUPT TIMER 1 FUNCTIONS SHOWN IN PARENTHESES () NOTE: FOR POWER-MANAGEMENT MODE (DIVIDE-BY-1024) OPERATION, THE TIMER INPUT CLOCK TO THE TIMER IS OSC / 1024 IF EITHER TXM = 1 OR TXMH = 1. OTHERWISE, THE TIMER INPUT IS OSC / 3072. Figure 11-1. Timers/Counters 0 and 1, Modes 0 and 1 ____________________________________________________________________________________________ 102 Ultra-High-Speed Flash Microcontroller User’s Guide T0M = CKCON.3 (T1M = CKCON.4) EXTERNAL OSCILLATOR INPUT TO TIMER CLK MODE SYSCLK DIVIDE-BY-1 OSC / 1 OSC / 0.5 2X 4X OSC / 0.25 T0 = P3.4 (T1 = P3.5) DIVIDEBY-12 T0MH = CKMOD.3 (T1MH = CKMOD.4) 0 DIVIDE- 1 BY-4 C / T = TMOD.2 0 1 (C / T = TMOD.6) 0 CLK TL0 (TL1) 0 7 RELOAD 1 TR0 = TCON.4 (TR1 = TCON.6) GATE = TMOD.3 (GATE = TMOD.7) 0 TH0 (TH1) 7 INT0 = P3.2 ( INT1 = P3.3) TF0 = TCON.5 (TF1 = TCON.7) INTERRUPT TIMER 1 FUNCTIONS SHOWN IN PARENTHESES () NOTE: FOR POWER-MANAGEMENT MODE (DIVIDE-BY-1024) OPERATION, THE TIMER INPUT CLOCK TO THE TIMER IS OSC / 1024 IF EITHER TXM = 1 OR TXMH = 1. OTHERWISE, THE TIMER INPUT IS OSC / 3072. Figure 11-2. Timers/Counters 0 and 1, Mode 2 MODE 3 This mode provides an 8-bit timer/counter and a second 8-bit timer as indicated in Figure 11-3. In mode 3, TL0 is an 8-bit timer/counter controlled by the normal timer 0 bits (TR0 = TCON.4 and TF0 = TCON.5). TL0 can be used to count clock cycles or 1 to 0 transitions on pin T0, as determined by C/T(TMOD.2). As in the other modes, the GATE function can use INT0 to give external run control of the timer to an outside signal. TH0 becomes an independent 8-bit timer in mode 3; however, it can only count clock cycles as shown in Figure 11-2. In this mode, some of timer 1’s control signals are used to manipulate TH0. That is, TR1 (TCON.6) and TF1 (TCON.7) become the relevant control and flag bits associated with TH0. In mode 3, timer 1 stops counting and holds its value. Thus, timer 1 has no practical application while in mode 3. As mentioned, when timer 0 is in mode 3, it uses some of timer 1’s resources (i.e., TR1 and TF1). Timer 1 can still be used in modes 0, 1, and 2 in this situation, but its flexibility becomes somewhat limited. While it maintains its basic functionality, its inputs and outputs are no longer available. Therefore, when timer 0 is in mode 3, timer 1 can only count clock cycles, and it does not have an interrupt or flag. With these limitations, baud-rate generation is its most practical application, but other time-base functions may be achieved by reading the registers. TIMER 2 MODES Like timers 0 and 1, timer 2 is a full-function timer/counter; however, it has several additional capabilities that make it more useful. Timer 2 has independent control registers in T2CON and T2MOD, and is based on count registers TL2 and TH2. It does not offer the 13-bit or dual 8-bit mode, but instead runs in the 16-bit mode at all times. Also note that while timers 0 and 1 have an 8-bit autoreload mode, timer 2 provides a 16-bit autoreload mode. This mode uses the timer capture registers to hold the reload values. The modes available on timer 2 are described in the following pages. 103 ____________________________________________________________________________________________ Ultra-High-Speed Flash Microcontroller User’s Guide T0M = CKCON.3 EXTERNAL OSCILLATOR INPUT TO TIMER CLK MODE SYSCLK DIVIDE-BY-1 OSC / 1 OSC / 0.5 2X 4X OSC / 0.25 DIVIDEBY-12 T0MH = CKMOD.3 0 DIVIDE- 1 BY-4 C / T = TMOD.2 0 0 CLK 1 0 TL0 7 1 T0 = P3.4 TF0 = TCON.5 INTERRUPT TF1 = TCON.7 INTERRUPT TR0 = TCON.4 GATE = TMOD.3 INT0 = P3.2 TR1 = TCON.6 0 TH0 7 NOTE: FOR POWER-MANAGEMENT MODE (DIVIDE-BY-1024) OPERATION, THE TIMER INPUT CLOCK TO THE TIMER IS OSC / 1024 IF EITHER TXM = 1 OR TXMH = 1. OTHERWISE, THE TIMER INPUT IS OSC / 3072. Figure 11-3. Timer/Counter 0, Mode 3 16-Bit Timer/Counter with Optional Capture In this mode, timer 2 performs a simple timer or counter function where it behaves similarly to mode 1 of timers 0 and 1, but uses 16 instead of 8 bits. This mode, along with the optional capture mode described later, is illustrated in Figure 11-4. The 16-bit count values are found in TL2 and TH2 special function registers (addresses 0CCh and 0CDh, respectively). The selection of whether a timer or counter function is performed is made using bit C/T2 (T2CON.1). When C/T2 is set to a logic 1, timer 2 behaves as a counter where it counts 1 to 0 transitions at the T2 (P1.0) pin. When C/T2 is set to a logic 0, timer 2 functions as a timer where it counts clock cycles. Timing or counting is enabled by setting bit TR2 (T2CON.2) to 1, and disabled by setting it to zero. When the counter rolls over from FFFFh to 0000h, the TF2 flag (T2CON.7) is set and causes an interrupt if timer 2’s interrupt is enabled. A diagram of timer 2’s capture mode is shown in Figure 11-4. In this mode, the timer performs basically the same 16-bit timer/counter function described above. However, a 1 to 0 transition on T2EX (pin P1.1) causes the value in timer 2 to be transferred into the capture registers if enabled by EXEN2 (T2CON.3). The capture registers, RCAP2L and RCAP2H, correspond to TL2 and TH2, respectively. The capture function is enabled by the CP/RL2 (T2CON.0) bit. When this bit is set to logic 1, the timer is in the capture mode just described. When set to logic 0, the timer is in autoreload mode described later. 16-Bit Autoreload Timer/Counter This mode is illustrated in Figure 11-5. When timer 2 reaches an overflow state, i.e., rolls over from FFFFh to 0000, it sets the TF2 flag. This flag can generate an interrupt if enabled. In addition, the timer restores its starting value and begins timing (or counting) again. The starting value is preloaded by software into capture registers RCAP2L and RCAP2H. These registers cannot be used for capture functions while also performing autoreload, so these modes are mutually exclusive. Autoreload is invoked by the CP/RL2 (T2CON.0) bit. When set to a logic 0, the timer is in autoreload mode. When CP/RL2 is set to a logic 1, the timer is in capture mode described above. If the C/T2 bit (T2CON.1) is a logic 0, the timer’s input clock is selectable as a function of the external oscillator or the system clock. Otherwise, pulses on pin T2 (P1.0) are counted when C/T2 = 1. As in other modes, counting or timing is enabled or disabled with TR2 (T2CON.2). When in autoreload mode, timer 2 can also be forced to reload with the T2EX (P1.1) pin. A 1 to 0 transition forces a reload if enabled by the EXEN2 (T2CON.3) bit. If EXEN2 is set to a logic 1, then a 1 to 0 transition on T2EX causes a reload. Otherwise, the T2EX pin is ignored. ____________________________________________________________________________________________ 104 Ultra-High-Speed Flash Microcontroller User’s Guide T2M = CKCON.5 EXTERNAL OSCILLATOR INPUT TO TIMER SYSCLK CLK MODE DIVIDE-BY-1 OSC / 1 2X OSC / 0.5 4X OSC / 0.25 DIVIDEBY-12 T2MH = CKMOD.5 0 DIVIDE- 1 BY-4 C / T2 = T2CON.1 0 0 CLK 1 0 78 15 TL2 TH2 TF2 = T2CON.7 1 T2 = P1.0 TR2 = T2CON.2 RCAP2L 0 T2EX = P1.1 RCAP2H 78 CAPTURE 15 EXF2 = T2CON.6 EXEN2 = T2CON.3 TIMER 2 INTERRUPT NOTE: FOR POWER-MANAGEMENT MODE (DIVIDE-BY-1024) OPERATION, THE TIMER INPUT CLOCK TO THE TIMER IS OSC / 1024 IF EITHER TXM = 1 OR TXMH = 1. OTHERWISE, THE TIMER INPUT IS OSC / 3072. Figure 11-4. Timer/Counter 2 with Optional Capture (CP/RL2 = 1) T2M = CKCON.5 EXTERNAL OSCILLATOR INPUT TO TIMER SYSCLK CLK MODE DIVIDE-BY-1 OSC / 1 OSC / 0.5 2X 4X OSC / 0.25 DIVIDEBY-12 T2MH = CKMOD.5 0 DIVIDE- 1 BY-4 C / T2 = T2CON.1 0 0 CLK 1 0 TL2 78 TH2 15 TF2 = T2CON.7 1 T2 = P1.0 TR2 = T2CON.2 T2EX = P1.1 RCAP2L 0 RCAP2H 78 15 TIMER 2 INTERRUPT EXEN2 = T2CON.3 EXF2 = T2CON.6 NOTE: FOR POWER-MANAGEMENT MODE (DIVIDE-BY-1024) OPERATION, THE TIMER INPUT CLOCK TO THE TIMER IS OSC / 1024 IF EITHER TXM = 1 OR TXMH = 1. OTHERWISE, THE TIMER INPUT IS OSC / 3072. Figure 11-5. Timer/Counter 2 Autoreload Mode (CP/RL2 = 0, DCEN = 0) 105 ____________________________________________________________________________________________ Ultra-High-Speed Flash Microcontroller User’s Guide Up/Down-Count Autoreload Timer/Counter The up/down autoreload counter option is selected by the DCEN (T2MOD.0) bit, and is illustrated in Figure 11-6. When DCEN (T2MOD.0) is set to a logic 1, timer 2 counts up or down as controlled by the state of pin T2EX (P1.1). T2EX causes upward counting when a logic 1 is applied and down counting when a logic 0 is applied. When DCEN = 0, timer 2 only counts up. When an upward counting overflow occurs, the value in RCAP2L and RCAP2H loads into TL2 and TH2. In the down-count direction, an underflow occurs when TL2 and TH2 match the values in RCAP2L and RCAP2H, respectively. When an underflow occurs, FFFFh is loaded into TL2 and TH2 and counting continues. Note that, in this mode, the overflow/underflow output of the timer is provided to an edge-detection circuit as well as to the TF2 bit (T2CON.7). This edge-detection circuit toggles the EXF2 bit (T2CON.6) on every overflow or underflow. Therefore, the EXF2 bit behaves as a 17th bit of the counter, and may be used as such. Baud-Rate Generator Timer 2 can be used to generate baud rates for serial port 0 in serial mode 1 or 3. Baud-rate generator mode is invoked by setting either the RCLK or TCLK bit in the T2CON register to a logic 1, as illustrated in Figure 11-7. In this mode, the timer continues to function in autoreload mode, but instead of setting the interrupt flag TF2 (T2CON.7) and potentially causing an interrupt, the overflow generates the shift clock for the serial port function. As in normal autoreload mode, an overflow causes RCAP2L and RCAP2H to be transferred into T2L and T2H, respectively. Note that, when RCLK or TCLK is set to 1, timer 2 is forced into 16-bit autoreload mode, regardless of the CP/RL2 bit. As explained above, the timer itself cannot set the TF2 interrupt flag and, therefore, cannot generate an interrupt. However, if EXEN2 (T2CON.3) is set to 1, a 1 to 0 transition on the T2EX (P1.1) pin causes the EXF2 (T2CON.6) interrupt flag to be set. If enabled, this causes a timer 2 interrupt to occur. Therefore, in this mode, the T2EX pin may be used as an additional external interrupt, if desired. T2M = CKCON.5 EXTERNAL OSCILLATOR INPUT TO TIMER SYSCLK CLK MODE DIVIDE-BY-1 OSC / 1 OSC / 0.5 2X 4X OSC / 0.25 DIVIDEBY-12 T2MH = CKMOD.5 0 DIVIDE- 1 BY-4 (DOWN-COUNTING RELOAD VALUE) 0FFH 0FFH C / T2 = T2CON.1 0 0 CLK 1 0 TL2 78 TH2 15 1 T2 = P1.0 0 TR2 = T2CON.2 COUNT DIRECTION (1 = UP, 0 = DOWN) T2EX = P1.1 78 RCAP2L 15 R AP2H (UP-COUNTING RELOAD VALUE) TF2 = T2CON.7 TIMER 2 INTERRUPT EXF2 = T2CON.6 NOTE: FOR POWER-MANAGEMENT MODE (DIVIDE-BY-1024) OPERATION, THE TIMER INPUT CLOCK TO THE TIMER IS OSC / 1024 IF EITHER TXM = 1 OR TXMH = 1. OTHERWISE, THE TIMER INPUT IS OSC / 3072. Figure 11-6. Timer/Counter 2 Autoreload Mode (DCEN = 1) ____________________________________________________________________________________________ 106 Ultra-High-Speed Flash Microcontroller User’s Guide TIMER 1 OVERFLOW OSC INPUT TO TIMER C / T2 = T2CON.1 CLK MODE TIMER INPUT DIVIDE BY 1 OSC / 2 0 2X OSC / 2 4X OSC / 2 PMM ( / 1024) OSC / 2048 1 T2 = P1.0 TR2 = T2CON.2 DIVIDEBY-2 0 CLK 0 SMOD_0 = PCON.7 TL2 78 15 TH2 1 RCLK = T2CON.5 1 0 TCLK = T2CON.4 T2EX = P1.1 EXEN2 = T2CON.3 RCAP2L 0 RCAP2H 78 EXF2 = T2CON.6 15 1 TIMER 2 INTERRUPT 0 DIVIDEBY-16 Rx CLOCK DIVIDEBY-1 6 Tx CLOCK Figure 11-7. Timer/Counter 2 Baud-Rate Generator Mode Another feature of the baud-rate generator mode is that the crystal-derived time base for the timer is the crystal frequency divided by 2. No other crystal-divider selection is possible unless operating in power-management mode. If a different time base is desired, bit C/T2 (T2CON.1) may be set to a 1, sourcing the time base from an external clock source supplied by the user on pin T2 (P1.0). The RCAP registers may be read, but not modified, while TR2 = 1. Stop the timer (TR2 = 0) to modify these registers. Timer Output Clock Generator Timer 2 can also be configured to drive a clock output on port pin P1.0 (T2), as shown in Figure 11-8. To configure timer 2 for this mode, it must first be set to 16-bit autoreload timer mode (CP/RL2 = 0, C/T2 = 0). Next, the T2OE (T2MOD.1) bit must be set to a logic 1. TR2 (T2CON.2) must also be set to a logic 1 to enable the timer. This mode produces a 50% duty cycle square-wave output. The frequency of the square wave is given by the formula in Figure 11-7. Each timer overflow causes an edge transition on the pin, i.e., the state of the pin toggles. Note that this mode has two somewhat unique features in common with the baud-rate generation mode. First, the time base is the crystal frequency divided by 2, and other than power-management mode operation, no other divider selection is possible. Second, the timer itself does not generate an interrupt, but, if needed, an additional external interrupt may be caused using T2EX as described above. Because of the two modes’ similarities, the timer can be used to generate both an external clock and a baud-rate clock simultaneously. Once the clock-out mode is established, either TCLK or RCLK is set to 1, and the RCAP2 registers are loaded, the timer provides a clock to both functions. Time-Base Selection The ultra-high-speed microcontroller allows the user to select the time base for each timer independently. In the standard 8051, the timers count the oscillator divided by 12, which is the standard 8051 machine cycle timing. Following a reset, the timers default to an oscillator divided-by-12 input clock to remain drop-in compatible with the original 8051. The ultra-high-speed microcontroller timers can additionally be configured to use the system clock or the system clock divided by 4 for the input clock. These selections, while not affecting the CPU timing, allow for higher precision timing and faster baud rates. As an example, a user might select both the baudrate generator timer and another timer to run at 12 oscillator clocks per timer tick with the third timer running at four system clocks per tick. This allows one timer to measure higher speed events or to gain better resolution. 107 ____________________________________________________________________________________________ Ultra-High-Speed Flash Microcontroller User’s Guide OSC INPUT TO TIMER CLK MODE TIMER INPUT DIVIDE-BY-1 OSC / 2 2X OSC / 2 4X OSC / 2 PMM ( / 1024) OSC / 2048 0 TL2 7 8 TH2 15 TR2 = T2CON.2 C/T2 = T2CON.1 = 0 0 T2 = P1.0 RCAP2L 7 8 RCAP2H 15 DIVIDEBY-2 T2OE = T2MOD.1 EXF2 = T2CON.6 T2EX = P1.1 TIMER 2 INTERRUPT EXEN2 = T2CON.3 T2 FREQUENCY OUT = TIMER CLOCK INPUT / (2 x (65536—RCAP2H, RCAP2L)) Figure 11-8. Timer/Counter 2 Clock Out Mode Time–Base Selection The input clock selection is independent for each timer and the default is 12 oscillator clocks per timer tick. The control bits for the time-base selection (TxM, TxMH) are located in the clock control register (CKCON;8Eh) and the clock-mode register (CKMOD;96h). The TxM and TxMH bits for each of the timers enable input clock selections of the system clock divided by 4 and the system clock divided by 1, respectively. When TxMH is set to a logic 1, the associated TxM bit for that timer is ignored. Note that, when operating in the default system clock mode, the system clock is the same frequency as the oscillator clock. System clock mode selection is controlled by the CD1, CD0 bits of the PMR register. See the PMR register description and the CPU timing section for more information on how to modify the system clock. As described earlier, timer 2 does, however, automatically switch to two oscillator clocks per tick when configured for baud-rate generation or clock-output mode. When the time base is derived from an external source (i.e., the T0, T1, or T2 pins), the timer operates at the frequency of the external source and is not affected by the setting of the TxM or TxMH bits. The only limitation is that the external source frequency can be no faster than one-fourth of the main system clock frequency. Use of power management-mode changes the input clock to the timers in a way that does not exactly follow any of the guidelines set forth to this point. Tables 11-2 and 11-3 show the resulting timer input clock for the various system clock modes and timer control bit setting. Table 11-2 pertains only to timer 2 in the baud-rate generation or clock-output mode. Table 11-2. Timers 0, 1, 2 Input Clock Frequency SYSTEM CLOCK MODE Crystal multiply mode 4X Crystal multiply mode 2X Divide-by-1 (default) Power-management mode (divide-by-1024) PMR REGISTER BITS 4X/2X, CD1, CD0 100 000 X01, X10 X11 TIMERS 0, 1, 2 INPUT CLOCK FREQUENCY TxMH,TxM = 00 TxMH,TxM = 01 TxMH,TxM = 1x OSC / 12 OSC / 1 OSC / 0.25 OSC / 12 OSC / 2 OSC / 0.5 OSC / 12 OSC / 4 OSC / 1 OSC / 3072 OSC / 1024 OSC / 1024 ____________________________________________________________________________________________ 108 Ultra-High-Speed Flash Microcontroller User’s Guide Table 11-3. Timer 2 Baud-Rate Generation/Clock Output Mode SYSTEM CLOCK MODE PMR REGISTER BITS 4X/2X, CD1, CD0 TIMER 2 BAUD-RATE GENERATION/CLOCK OUTPUT MODE INPUT CLOCK FREQUENCY (TxMH, TxM = xx) Crystal multiply mode 4X Crystal multiply mode 2X Divide-by-1 (default) 100 000 X01, X10 OSC / 2 OSC / 2 OSC / 2 Power management mode (divide-by-1024) X11 OSC / 2048 WATCHDOG TIMER The watchdog timer reset provides CPU monitoring by requiring software to clear the timer before the user-selected interval expires. If the timer is not reset, the CPU can be reset by the watchdog. The watchdog function is optional and is described below. The watchdog timer is a user-programmable clock counter that can serve as a time-base generator, an event timer, or a system supervisor. As can be seen in Figure 11-9, the timer is driven by the main system clock that is supplied to a series of dividers. The divider output is selectable, and determines the interval between timeouts. When the timeout is reached, an interrupt flag is set, and if enabled, a reset occurs. The interrupt flag causes an interrupt to occur if its individual enable bit is set and the global interrupt enable is set. The reset and interrupt are completely discrete functions that may be acknowledged or ignored, together or separately for various applications. The watchdog timer reset function works as follows. After initializing the correct timeout interval (discussed later), software first restarts the watchdog using RWT (WDCON.0) and then enables, if desired, the reset function by setting the enable watchdog timer reset (EWT = WDCON.1) bit. Any time prior to reaching its user-selected terminal value, software can set the reset watchdog timer (RWT = WDCON.0) bit. If the watchdog timer is reset (RWT bit written to a logic 1) before the timeout period expires, the timer starts over. Hardware automatically clears the RWT after software sets it. XTAL1 RWT (WDCON.0) (RESET WATCHDOG) XTAL2 SYSTEM CLOCK MODE CONTROL DIVIDE-BY17 2 WD1 (CKCON.7) WD0 (CKCON.6) SYSCLK OUTPUT CLK MODE SYSCLK DIVIDE-BY-1 OSC / 1 2X OSC / 0.5 4X OSC / 0.25 PMM OSC / 1024 DIVIDE-BY3 2 DIVIDE-BY3 2 17 20 2 2 TIMEOUT SELECTOR 2 23 2 WDIF (WDCON.3) 26 TIMEOUT WATCHDOG INTERRUPT EWDI (EIE.4) (ENABLE WATCHDOG TIMER) 512 SYSCLK DELAY EWT (WDCON.1) (ENABLE WATCHDOG TIMER) Figure 11-9. Watchdog Timer 109 DIVIDE-BY3 2 ____________________________________________________________________________________________ RESET WTRF (WDCON.2) Ultra-High-Speed Flash Microcontroller User’s Guide If the timeout is reached without RWT being set, hardware generates a watchdog interrupt if the interrupt source has been enabled. If no further action is taken to prevent a watchdog reset in the 512 system clock cycles following the timeout, hardware has the ability to reset the CPU if EWT = 1. When the reset occurs, the watchdog timer reset flag (WTRF = WDCON.2) is automatically set to indicate the cause of the reset; however, software must clear this bit manually. The watchdog timer is a free-running timer. When used as a simple timer with both the reset and interrupt functions disabled (EWT = 0 and EWDI = 0), the timer continues to set the watchdog interrupt flag each time the timer completes the selected timer interval as programmed by WD1 (CKCON.7) and WD0 (CKCON.6). Restarting the timer using the RWT (WDCON.0) bit allows software to use the timer in a polled timeout mode. The WDIF bit is cleared by software or any reset. The watchdog interrupt is also available for applications that do not need a true watchdog reset, but a very long timer. The interrupt is enabled using the enable watchdog timer interrupt (EWDI = EIE.4) bit. When the timeout occurs, the watchdog timer sets the WDIF bit (WDCON.3), and an interrupt occurs if the global interrupt enable (EA = IE.7) is set. Note that WDIF is set 512 clocks before a potential watchdog reset. The watchdog interrupt flag indicates the source of the interrupt, and must be cleared by software. Using the watchdog interrupt during software development can allow the user to select ideal watchdog reset locations. Code is first developed without enabling the watchdog interrupt or reset functions. Once the program is complete, the watchdog interrupt function is enabled to identify the required locations in code to set the RWT (WDCON.0) bit. Incrementally adding instructions to reset the watchdog timer prior to each address location (identified by the watchdog interrupt) allows the code to eventually run without receiving a watchdog interrupt. At this point, the watchdog timer reset can be enabled without the potential of generating unwanted resets. At the same time, the watchdog interrupt may also be disabled. Proper use of the watchdog interrupt with the watchdog reset allows interrupt software to survey the system for errant conditions. When using the watchdog timer as a system monitor, the watchdog reset function should be used. If the interrupt function were used, the purpose of the watchdog would be defeated. For example, assume the system is executing errant code prior to the watchdog interrupt. The interrupt would temporarily force the system back into control by vectoring the CPU to the interrupt service routine. Restarting the watchdog and exiting by an RETI or RET would return the processor to the lost position prior to the interrupt. By using the watchdog reset function, the processor is restarted from the beginning of the program and, therefore, placed into a known state. The watchdog timeout selection is made using bits WD1 (CKCON.7) and WD0 (CKCON.6). The watchdog has four timeout selections based on the system clock frequency, as shown in the figure. Since the timeout is a function of the system clock, the actual timeout interval is dependent on both the crystal frequency and the system clock mode. Shown in Table 11-4 is a summary of the selectable watchdog timeout intervals for the various system clock modes and WD1:0 control bit settings. The watchdog reset, if enabled, is always scheduled to occur 512 system clocks following the timeout. Watchdog-generated resets last for 13 oscillator cycles. As discussed above, the watchdog timer has several SFR bits that contribute to its operation. It can be enabled to function as either a reset source, interrupt source, software polled timer, or any combination of the three. Both the reset and the interrupt have status flags. The watchdog also has a bit that restarts the timer. Table 11-5 shows the watchdog timer-related bits. Detailed bit descriptions can be found in Section 4. Table 11-4. Watchdog Timeout Intervals WATCHDOG TIMEOUT (IN NUMBER OF OSCILLATOR CLOCKS) SYSTEM CLOCK MODE PMR REGISTER BITS 4X/2X, CD1, CD0 Crystal multiply mode 4X Crystal multiply mode 2X Divide-by-1 (default) 100 000 X01, X10 WD1:0 = 00b 215 216 217 WD1:0 = 01b 218 219 220 WD1:0 = 10b 221 222 223 WD1:0 = 11b 224 225 226 Power management mode (divide-by-1024) X11 227 230 233 236 ____________________________________________________________________________________________ 110 Ultra-High-Speed Flash Microcontroller User’s Guide Table 11-5. Watchdog Timer-Related Bits BIT NAMES EWT RWT WD1,WD0 WTRF EWDI WDIF DESCRIPTION Enable watchdog timer reset Reset watchdog timer Watchdog interval control bits 1, 0 Watchdog timer reset flag Enable watchdog timer interrupt Watchdog interrupt flag REGISTER LOCATION WDCON – D8h WDCON – D8h CKCON – 8Eh WDCON – D8h EIE – E8h WDCON – D8h BIT POSITIONS WDCON.1 WDCON.0 CKCON.7,6 WDCON.2 EIE.4 WDCON.3 Section 12: SERIAL I/O The ultra-high-speed microcontroller provides two fully independent UARTs (serial ports) with framing-error detection and automatic address recognition. The two UARTs can be operated simultaneously in identical or different modes and communication speeds. In this documentation, all descriptions apply to both UARTs, unless stated otherwise. Each serial port is capable of both synchronous and asynchronous modes. In the synchronous mode, the microcontroller generates the clock and operates the UART in a half-duplex mode. In the asynchronous mode, full-duplex operation is available. Receive data is buffered in a holding register. This allows the UART to receive an incoming word before software has read the previous value. Each UART has an associated control register (SCON0, SCON1) and each has a transmit/receive register (SBUF0, SBUF1). The SFR locations are: SCON0, 98h; SBUF0, 99h; SCON1, C0h; SBUF1, C1h. The SBUF location provides access to both transmit and receive registers. Reads are directed to the receive buffer and writes to the transmit buffer automatically. SERIAL MODE SUMMARY Each port provides four operating modes. These offer different communication protocols and baud rates. The four serial operating modes and max baud rate for each are shown in Table 12-1, followed by a brief summary of each mode. Detailed descriptions of the modes are provided later in this section. In addition, provisions for use of the serial ports in conjunction with the crystal multiplier and power-management mode are discussed later in this section. MODE 0 Mode 0 provides synchronous communication with external devices. It is commonly used to communicate with serial peripherals. Serial I/O occurs on the RXD pin. The shift clock is provided on the TXD pin. Note that, whether transmitting or receiving, the serial clock is generated by the ultra-high-speed microcontroller. Thus, any device on the serial port in mode 0 must accept the microcontroller as the master. When not using power-management mode, the baud rate in mode 0 is the system clock frequency divided by either 12 or 4, as selected by the SM2 bit for the associated UART. When set to a logic 0, the serial port runs at a divide-by-12. When set to a logic 1, the serial port runs at a divide-by-4. The SM2 bit for serial port 0 is located at SCON0.5 and the SM2 bit for serial port 1 is located at SCON1.5. With the exception of the additional new divide-by-4 selection (supported by SM2), mode 0 operation is identical to the 80C32. Table 12-1. Serial I/O Modes and Max Baud Rates MODE SYNC/ ASYNC BAUD CLOCK† START/STOP DATA BITS 9TH BIT FUNCTION MAX BAUD RATE SYSCLK = 33MHz 0 1 2 3 Sync Async Async Async 4 or 12 tCLK Timer 1 or 2* 32 or 64 tCLK Timer 1 or 2* None 1 start, 1 stop 1 start, 1 stop 1 start, 1 stop 8 8 9 9 None None 0, 1, parity 0, 1, parity 8,250,000 2,062,500 1,031,250 2,062,500 †Use of the crystal multiplier or power-management mode affects the baud clock. *Timer 2 available for serial port 0 only. 111 ____________________________________________________________________________________________ Ultra-High-Speed Flash Microcontroller User’s Guide MODE 1 This mode provides standard full-duplex asynchronous communication. A total of 10 bits are transmitted including 1 start bit, 8 data bits, and 1 stop bit. The received stop bit is stored in bit location RB8 of the relevant SCON register. In mode 1, the baud rate is a function of timer overflow. This makes the baud rate programmable by the user. One difference that exists between the two UARTs, with respect to mode 1 configuration, is that serial port 1 can use only timer 1, whereas serial port 0 can use either timer 1 or 2 to generate baud rates. If both serial ports use the same timer, they run at the same baud rate, or one can run twice as fast as the other (when baud-rate doubler bits, PCON.7 and WDCON.7, are configured differently). If the two UARTs use different timers, the baud-rate configurations, in relation to one another, are not as restrictive. Baud rates are discussed in more detail later. Mode 1 operation is identical to the standard 80C32 when timers 1 or 2 use the default oscillator divide-by-12 as an input clock. MODE 2 This mode is an asynchronous mode that transmits a total of 11 bits. These include 1 start bit, 9 data bits (the ninth data bit being programmable), and 1 stop bit. The ninth bit is determined by the value in TB8 (SCON0.3 or SCON1.3) for transmission. When the ninth bit is received, it is stored in RB8 (SCON0.2 or SCON1.2). The ninth bit can be a parity value by moving the P bit (PSW.0) to TB8. When not using the power-management mode, the baud rate for mode 2 is a function only of the oscillator frequency. It is either the oscillator input divided by 32 or 64 as programmed by the SMOD doubler bit for the associated UART. The SMOD_0 baud-rate doubler bit for serial port 0 is located at PCON.7, and the SMOD_1 baud-rate doubler bit for serial port 1 is located at WDCON.7. Mode 2 operation is identical to the standard 80C32. MODE 3 This mode has the same functionality as mode 2, but generates baud rates like mode 1. That is, this mode transmits 11 bits, but generates baud rates through the timers. Like mode 1, either timer 1 or 2 can be used for serial port 0 and timer 1 can be used for serial port 1. Mode 3 operation is identical to the standard 80C32 when timers 1 or 2 use the default oscillator divide-by-12 as an input clock. SERIAL PORT INITIALIZATION In order to use the UART function(s), the serial port must be initialized. This involves selecting the mode and time base, then initializing the baud-rate generator, if necessary. Serial communication is then available. Once the baud-rate generator is running, the UART can receive data. In mode 0, the high-speed microcontroller provides the clock. Serial reception is initiated by setting the RI bit to a logic 0 and REN to a logic 1. This generates a clock on the TXD pin and a shift in the 8 bits on the RXD pin. In the other modes, setting the REN bit to a logic 1 allows serial reception, but the external device must actually initiate it by sending a start bit. In any mode, serial transmission is initiated by writing to either the SBUF0 or SBUF1 location. Most of the serial port controls are provided by the SCON0 and SCON1 registers. For convenience, the Table 12-2 is provided, which summarizes the SFRs controlling serial port operation. Detailed bit descriptions can be found in Section 4. ____________________________________________________________________________________________ 112 Ultra-High-Speed Flash Microcontroller User’s Guide SERIAL PORT 1 SERIAL PORT 0 Table 12-2. SFR Serial Port Operation Control BIT NAMES SM0/FE_0 SM1_0 SM2_0 REN_0 TB8_0 RB8_0 TI_0 RI_0 SMOD_0 RCLK TCLK DESCRIPTION Serial mode select 0 or framing error Serial mode select 1 Serial mode select 2 Receive enable 9th transmit data bit 9th receive data bit Transmit interrupt flag Receive interrupt flag Baud-rate doubler bit Timer 2 serial receive clock enable Timer 2 serial transmit clock enable Serial data buffer Slave address Slave address mask enable REGISTER LOCATION SCON0 – 98h SCON0 – 98h SCON0 – 98h SCON0 – 98h SCON0 – 98h SCON0 – 98h SCON0 – 98h SCON0 – 98h PCON – 87h T2CON – C8h T2CON – C8h SBUF0 – 99h SADDR0 – A9h SADEN0 – B9h BIT POSITIONS SCON0.7 SCON0.6 SCON0.5 SCON0.4 SCON0.3 SCON0.2 SCON0.1 SCON0.0 PCON.7 T2CON.5 T2CON.4 SM0/FE_1 SM1_1 SM2_1 REN_1 TB8_1 RB8_1 TI_1 RI_1 SMOD_1 Serial mode select 0 or framing error Serial mode select 1 Serial mode select 2 Receive enable 9th transmit data bit 9th receive data bit Transmit interrupt flag Receive interrupt flag Baud-rate doubler bit Serial data buffer Slave address Slave address mask enable SCON1 – C0h SCON1 – C0h SCON1 – C0h SCON1 – C0h SCON1 – C0h SCON1 – C0h SCON1 – C0h SCON1 – C0h WDCON – D8h SBUF1 – C1h SADDR1 – AAh SADEN1 – BAh SCON1.7 SCON1.6 SCON1.5 SCON1.4 SCON1.3 SCON1.2 SCON1.1 SCON1.0 WDCON.7 PCON – 87h PCON.6 SMOD0 Enable framing error detection BAUD RATES Each mode has a baud-rate generator associated with it. This generator is generally the same for each UART. Several of the baud-rate generation techniques have options that are independent for the two UARTs. The following baud-rate descriptions are separated by mode. Mode 0 Mode 0 is synchronous, so the shift clock output frequency is the baud rate. Table 12-3 summarizes baud-rate generation as a function of the external oscillator frequency. The default case is divide-by-12. The user can select the shift clock frequency using the SM2 bit in the associated SCON register. For serial port 0, the SM2_0 bit is SCON0.5. For serial port 1, the SM2_0 bit is SCON1.5. When SM2 is set to a logic 0, the baud rate is fixed at a divide-by-12 of the system clock frequency, unless power-management mode is invoked. When operating in power-management mode, with the SM2 bit clear ( = 0), the serial port clock frequency is the oscillator frequency divided by 3072. When SM2 is set to a logic 1, the baud rate is generated using the system clock frequency divided by 4, unless power-management mode is invoked. When power-management mode is used with the SM2 bit set ( = 1), the serial port clock frequency tracks the system clock frequency. Note that this use of SM2 differs from a standard 80C32. In that device, SM2 had no valid use when the UART was in mode 0. Since it was generally set to a zero, for the divide-by-12, there is no compatibility problem. 113 ____________________________________________________________________________________________ Ultra-High-Speed Flash Microcontroller User’s Guide Table 12-3. Mode 0 Serial Port Clock Frequency SYSTEM CLOCK MODE PMR REGISTER BITS 4X/2X, CD1, CD0 Crystal multiply mode 4X Crystal multiply mode 2X Divide-by-1 (default) Power-management mode (/1024) 100 000 X01, X10 X11 MODE 0 SERIAL PORT CLOCK FREQUENCY SM2 = 0 OSC / 3 OSC / 6 OSC / 12 OSC / 3072 SM2 = 1 OSC / 1 OSC / 2 OSC / 4 OSC / 1024 Mode 2 In this asynchronous mode, baud rates are derived directly from the oscillator input. The following table summarizes baud-rate generation as a function of the external oscillator frequency. This mode works identically to the original 8051 family. The default case is divide-by-64. The user can effectively double the serial port clock frequency by setting the SMOD bit to a logic 1 for the associated UART. For serial port 0, the SMOD_0 bit is PCON.7. This is the original location in the 8051 family. For serial port 1, the SMOD_1 bit is WDCON.7. When operating in the power management mode (CD1:0 = 11b), the serial port clock frequency is the oscillator frequency divided by 16384 when the SMOD bit is a logic 0 and twice that frequency (OSC/8192) when the SMOD doubler bit is a logic 1. SMOD bits default to a logic 0 on all resets. Table 12-4. Mode 2 Serial Port Clock Frequency SYSTEM CLOCK MODE PMR REGISTER BITS 4X/2X, CD1, CD0 Crystal multiply mode 4X Crystal multiply mode 2X Divide-by-1 (default) Power-management mode (/1024) 100 000 X01, X10 X11 MODE 2 SERIAL PORT CLOCK FREQUENCY SMOD = 0 OSC / 64 OSC / 64 OSC / 64 OSC / 16384 SMOD = 1 OSC / 32 OSC / 32 OSC / 32 OSC / 8192 Mode 1 or 3 These asynchronous modes are commonly used for communication with PCs, modems, and other similar interfaces. The baud rates and bit timing are generated using either timer 1 or timer 2. The respective timer is placed in autoreload mode. When the timer reaches its rollover condition (FFFFh - timer 2 or FFh - timer 1), a clock is sent to the baud-rate circuit. The baud-rate circuit generates the exact baud rate by further dividing the clock by 16 or 32 (depending upon the UART baud-rate doubler bit). For serial port 0, either timer 1 or 2 can be used to generate baud rates. For serial port 1, only timer 1 can be used as the baud-rate generator. If operated in mode 1 or 3, the two UARTs may both use timer 1 for baud-rate generation, if desired. Using Timer 1 for Baud-Rate Generation To use timer 1 as the baud-rate generator, it is commonly put into the 8-bit autoreload mode. In this way, the CPU is not involved in baud-rate generation. Note that the timer interrupt should not be enabled. In the 8-bit autoreload mode (timer 1, mode 2), the reload value is stored in TH1. Thus, the combination of timer 1 input clock frequency and TH1 determine the baud rate. The timer 1 input clock, relative to the external crystal clock, can be altered in two ways: 1) changing the system clock, or 2) changing the timer input clock divide ratio. Modifying the system clock is accomplished using the clock divide bits (CD1:0) found in the PMR special function register. This procedure is discussed in Section 5. The timer 1 input clock divide ratio is configurable using the T1M (CKCON.4) and T1MH (CKMOD.4) register bits. Setting the T1MH bit to a logic 1 results in the system clock being used to clock timer 1. When T1MH is clear ( = 0), setting the T1M bit to a logic 1 provides the system clock divided by 4 input to timer 1. When both T1M and T1MH are logic 0, the Timer 1 input clock is fixed at the oscillator frequency divided by 12. When using power-management mode, setting either T1MH or T1M to a logic 1 results in the system clock (OSC/1024) being used as the input clock to timer 1. While, if both bits are clear ( = 0) in power-management mode, the system clock divided by 3 (OSC/3072) are provided to timer 1. The following table summarizes the relationship between the external crystal frequency and the timer 1 input clock for the various configurations. ____________________________________________________________________________________________ 114 Ultra-High-Speed Flash Microcontroller User’s Guide Table 12-5. Timer 1 Input Clock Frequency SYSTEM CLOCK MODE PMR REGISTER BITS 4X/2X, CD1, CD0 Crystal multiply mode 4X Crystal multiply mode 2X Divide-by-1 (default) Power-management mode (/1024) 100 000 X01, X10 X11 TIMER 1 INPUT CLOCK FREQUENCY T1MH,T1M = 00 OSC / 12 OSC / 12 OSC / 12 OSC / 3072 T1MH,T1M = 01 OSC / 1 OSC / 2 OSC / 4 OSC / 1024 T1MH,T1M = 1X OSC / 0.25 OSC / 0.5 OSC / 1 OSC / 1024 Using timer 1 in the 8-bit autoreload mode, serial port baud rates for mode 1 or 3 can be calculated using the formula below. Modes 1, 3 baud rate = 2SMOD_x x 32 Timer 1 input clock frequency (256 - TH1) Number of serial bits / Number of timer 1 rollovers Timer 1 rollover frequency Timer 1 input clock frequency can be found in the previous table, SMOD_x is the logic state of the baud-rate doubler bit for the associated UART, and TH1 is the user assigned timer 1 reload value. Often, users already know what baud rate is desired and only need to calculate the timer reload value. An equation to calculate the timer reload value, TH1, is as follows: TH1 = 256 - 2SMOD_x x timer 1 input clock frequency 32 x 5 baud rate Note that the 8-bit, autoreload mode for timer 1 is the one most commonly used for serial port applications, but that it can actually be configured in any mode, even as a counter. Using Timer 2 for Baud-Rate Generation To use timer 2 as baud-rate generator for serial port 0, the timer is configured in autoreload mode. Then, either the TCLK or RCLK bit (or both) are set to a logic 1. TCLK = 1 selects timer 2 as the baud-rate generator for the transmitter and RCLK = 1 selects timer 2 for the receiver. Thus, serial port 0 can have the transmitter and receiver operating at different baud rates by choosing timer 1 for one data direction and timer 2 for the other. RCLK and TCLK reside in T2CON.4 and TCON.5, respectively. Although the timer 2 input clock can be configured similarly to timer 1, it must be placed into a baud-rate generator mode in order to be used by serial port 0. Setting either RCLK or TCLK to a logic 1 selects timer 2 for baud-rate generation. When this is done, the timer 2 input clock becomes fixed to the oscillator frequency divided by 2. This is compatible with the 80C32. The only exception is when timer 2 is used for baud-rate generation within power-management mode. For PMM, the system clock (OSC/1024) is used as the input clock for timer 2. The timer 2 interrupt is automatically disabled when either RCLK or TCLK is set. Also, the TF2 (TCON.7) flag is not set on a timer rollover. The manual reload pin, T2EX (P1.1), does not cause a reload either. Table 12-6 illustrates this relationship. Table 12-6. Timer 2 Baud-Rate Generation 115 SYSTEM CLOCK MODE PMR REGISTER BITS 4X/2X, CD1, CD0 TIMER 2 INPUT CLOCK FREQUENCY BAUD-RATE GENERATOR MODE (RCLK OR TCLK = 1) Crystal multiply mode 4X Crystal multiply mode 2X Divide-by-1 (default) Power-management mode (/1024) 100 000 X01, X10 X11 OSC / 2 OSC / 2 OSC / 2 OSC / 1024 ____________________________________________________________________________________________ Ultra-High-Speed Flash Microcontroller User’s Guide When using timer 2 to generate baud rates, the formula is as follows. Note that the reload value is a 16-bit number, as compared with timer 1, which uses only 8 bits. A second equation is provided so that the timer 2 reload value can be calculated for a given baud rate. Modes 1, 3 baud rate = 1 16 Timer 2 input clock frequency ✕ (65536–RCAP2H,RCAP2L) Number of serial bits / Number of timer 2 rollovers RCAP2H, RCAP2L = 65536 – Timer 2 rollover frequency Timer 2 input clock frequency 16 x baud rate Timer 2 input clock frequency can be found in Table 12-6, and RCAP2H, RCAP2L is the user assigned timer 2 reload value. SERIAL I/O DESCRIPTION A detailed description and block diagram of each serial mode is given below. Note that the baud clock input (to the serial I/O control block) corresponding to the power-management mode has been omitted from each of the block diagrams. Reference the tables earlier in this section for power-management mode baud clock rates. A description of framing-error detection and multiprocessor communication follows this section. Mode 0 Mode 0 is used to communicate in synchronous, half-duplex format with devices that accept the ultra-high-speed microcontroller as a master. Figure 12-1 shows a functional block diagram and basic timing of this mode. As can be seen, there is one bidirectional data line (RXD) and one shift clock line (TXD) used for communication. The shift clock is used to shift data into and out of the microcontroller and the remote device. Mode 0 requires that the microcontroller is the master, because the microcontroller generates the serial shift clocks for both directions. As described earlier in the section, the shift clock frequency is a function of the system clock if the SM2 (SCON0.5 or SCON1.5) bit is set to a logic 1. The RXD signal is used for both transmission and reception. TXD provides the shift clock. Data bits enter and exit least-significant bit (LSb) first. The baud rate is equal to the shift clock frequency. The relevant UART begins transmitting when any instruction writes to SBUF0 or SBUF1 (address 99h or C1h). The internal shift register then begins to shift data out. The clock is activated and transfers data until the 8-bit value is complete. Data is presented just prior to the falling edge of the shift clock (TXD) so that an external device can latch the data using the rising edge. The UART begins to receive data when the REN bit in the SCON register (SCON0.4 or SCON1.4) is set to a logic 1 and the RI bit (SCON0.0 or SCON1.0) is set to a logic 0. This condition tells the UART that there is data to be shifted in. The shift clock (TXD) activates, and the UART latches incoming data on the rising edge. The external device should therefore present data on the falling edge. This process continues until 8 bits have been received. The RI bit is automatically set to a logic 1 immediately following the last rising edge of the shift clock on TXD. This causes reception to stop until the SBUF has been read, and the RI bit cleared. When RI is cleared, another byte can be shifted in. ____________________________________________________________________________________________ 116 Ultra-High-Speed Flash Microcontroller User’s Guide SBUF OUTPUT SHIFT REGISTER DIVIDEBY-12 LOAD CLOCK SYSTEM CLOCK P3.0 LATCH RXD PIN D7 D6 D5 D4 D3 D2 D1 D0 S0 DIVIDEBY-4 0 1 SM2= SM2= SCONx.5 SCONx.5 DATA BUS LDSBUF RDSBUF RECEIVE DATA BUFFER RD WR RECEIVE BUFFER LOAD DATA CLOCK R1 FLAG = SCONx.0 SI D7 D6 D5 D4 D3 D2 D1 D0 T1 FLAG = SCONx.1 SHIFT READ SERIAL BUFFER CLOCK INTS BAUD CLOCK SERIAL I/O CONTROL LOAD SERIAL BUFFER RECEIVE SHIFT REGISTER SERIAL INTERRUPT TXD PIN TRANSMIT TIMING LDSBUF WRITE TO SBUF SHIFT RXD (DATA OUT) D0 D1 D2 D3 D4 D5 D6 D7 TXD (DATA CLOCK) TI RECEIVE TIMING RDSBUF WRITE TO SCON (CLEAR RI) SHIFT RXD (DATA IN) D0 D1 D2 D3 D4 TXD (DATA CLOCK) RI Figure 12-1. Serial Port Mode 0 117 ____________________________________________________________________________________________ D5 D6 D7 Ultra-High-Speed Flash Microcontroller User’s Guide Mode 1 Mode 1 is asynchronous and full duplex, using a total of 10 bits. The 10 bits consist of a start bit (logic 0), 8 data bits, and 1 stop bit (logic 1) as illustrated in Figure 12-2. The data is transferred LSb first. As described above, the baud rates for mode 1 are generated by either a divide-by-16 of timer 1 rollover, a divide-by-16 of the timer 2 rollover, or a divide-by-32 of timer 1 rollover. The UART begins transmission after the first rollover of the divide-by-16 counter following a software write to SBUF. Transmission takes place on the TXD pin. It begins by the start bit being placed on the pin. Data is then shifted out onto the pin, LSb first. The stop bit follows. The TI bit is set by hardware after the stop bit is placed on the pin. All bits are shifted out at the rate determined by the baud-rate generator. Once the baud-rate generator is active, reception can begin at any time. The REN bit (SCON0.4 or SCON1.4) must be set to a logic 1 to allow reception. The falling edge of a start bit on the RXD pin begins the reception process. Data is shifted in at the selected baud rate. At the middle of the stop bit time, certain conditions must be met to load SBUF with the received data: • RI must = 0, and either • If SM2 = 0, the state of the stop bit does not matter, or • If SM2 = 1, the state of the stop bit must = 1. If these conditions are true, then SBUF (hex address 99h or C1h) is loaded with the received byte, the RB8 bit (SCON0.2 or SCON1.2) is loaded with the stop bit, and the RI bit (SCON0.0 or SCON1.0) is set. If these conditions are false, then the received data is lost (SBUF and RB8 not loaded) and RI is not set. Regardless of the receive word status, after the middle of the stop bit time, the receiver goes back to looking for a 1 to 0 transition on the RXD pin. Each data bit received is sampled on the 7th, 8th, and 9th clock used by the divide-by-16 counter. Using majority voting, two equal samples out of the three determine the logic level for each received bit. If the start bit was determined to be invalid ( = 1), then the receiver goes back to looking for a 1 to 0 transition on the RXD pin in order to start the reception of data. Mode 2 Mode 2 uses a total of 11 bits in asynchronous full-duplex communication, as illustrated in Figure 12-3. The 11 bits consist of 1 start bit (a logic 0), 8 data bits, 1 programmable 9th bit, and one stop bit (a logic 1). Like mode 1, the transmissions occur on the TXD signal pin and receptions on RXD. For transmission purposes, the 9th bit can be stuffed as a logic 0 or 1. A common use is to put the parity bit in this location. The 9th bit is transferred from the TB8 bit position in the SCON register (SCON0.3 or SCON1.3) during the write to SBUF. Baud rates are generated as a fixed function of the crystal frequency, as described earlier in this section. Like mode 1, mode 2’s transmission begins after the first rollover of the divide-by-16 counter following a software write to SBUF. It begins by the start bit being placed on the TXD pin. The data is then shifted out onto the pin LSb first, followed by the 9th bit, and finally the stop bit. The TI bit (SCON0.1 or SCON1.1) is set when the stop bit is placed on the pin. Reception begins when a falling edge is detected as part of the incoming start bit on the RXD pin. The RXD pin is then sampled according to the baud-rate speed. The 9th bit is placed in the RB8 bit location in SCON (SCON0.2 or SCON1.2). When a stop bit has been received, the data value is transferred to the SBUF receive register (hex address 99 or C1). The RI bit (SCON0.0 or SCON1.0) is set to indicate that a byte has been received. At this time, the UART can receive another byte. Once the baud-rate generator is active, reception can begin at any time. The REN bit (SCON0.4 or SCON1.4) must be set to a logic 1 to allow reception. The falling edge of a start bit on the RXD pin begins the reception process. Data must be shifted in at the selected baud rate. At the middle of the 9th bit time, certain conditions must be met to load SBUF with the received data. • RI must = 0, and either • If SM2 = 0, the state of the 9th bit does not matter, or • If SM2 = 1, the state of the 9th bit must = 1. If these conditions are true, then SBUF is loaded with the received byte, RB8 is loaded with the 9th bit, and RI is set. If these conditions are false, then the received data is lost (SBUF and RB8 not loaded) and RI is set. Regardless of the receive word status, after the middle of the stop bit time, the receiver goes back to looking for a 1 to 0 transition on RXD. Data is sampled in a similar fashion to mode 1 with the majority voting on three consecutive samples. Mode 2 uses the sample divideby-16 counter with either the oscillator divided by 2 or 4. ____________________________________________________________________________________________ 118 Ultra-High-Speed Flash Microcontroller User’s Guide SMOD_0 = PCON.7 OR SMOD_1 = WDCON.7 0 1 0 TCLK = T2CON.4 1 D1 D0 START D4 D3 D2 D5 TXD PIN 0 DATA BUS DIVIDEBY-16 0 P3.1 LATCH S0 LDSBUF RDSBUF LOAD SERIAL BUFFER RCLK = T2CON.5 D7 D6 STOP LOAD 1 BAUD CLOCK 1 SHIFT SERIAL I/O CONTROL DIVIDEBY-2 AVAILABLE TO SERIAL PORT 0 ONLY TIMER 2 OVERFLOW TIMER 1 OVERFLOW CLOCK SBUF TRANSMIT SHIFT REGISTER SBUF READ SERIAL BUFFER RECEIVE DATA BUFFER RD WR LOAD RESET D4 D3 D2 D5 D1 D0 START SI SERIAL INTERRUPT D7 D6 R1 FLAG = SCONx.0 CLOCK T1 FLAG = SCONx.1 STOP INTS RB8 = SCONx.2 RECEIVE SHIFT REGISTER DIVIDEBY-16 RXD PIN BIT DETECTION TRANSMIT TIMING LDSBUF SHIFT TXD START D0 D1 D2 D3 D4 D5 D6 D7 STOP TI RECEIVE TIMING RXD START D0 D1 D2 D3 D4 D5 BIT DETECTOR SAMPLING SHIFT RI Figure 12-2. Serial Port Mode 1 119 ____________________________________________________________________________________________ D6 D7 STOP Ultra-High-Speed Flash Microcontroller User’s Guide 1 P3.1 LATCH S0 D4 D3 D2 D1 D0 START D5 D7 D6 LOAD OSC/2 = CRYSTAL/2 CLOCK STOP D8 SBUF TRANSMIT SHIFT REGISTER TXD PIN 0 DIVIDEBY-2 TB8 = SCONx.3 1 DATA BUS LDSBUF RDSBUF SHIFT LOAD SERIAL BUFFER SERIAL I/O CONTROL SHIFT CLOCK RESET WR LOAD D1 D0 START SI SERIAL INTERRUPT CLOCK R1 FLAG = SCONx.0 D4 D3 D2 RB8 = SCONx.2 INTS T1 FLAG = SCONx.1 RECEIVE DATA BUFFER RD D5 DIVIDEBY-16 SBUF READ SERIAL BUFFER D8 D7 D6 0 STOP SMOD_0 = PCON.7 OR SMOD_1 = WDCON.7 RECEIVE SHIFT REGISTER DIVIDEBY-16 RXD PIN BIT DETECTION TRANSMIT TIMING LDSBUF SHIFT TXD START D0 D1 D2 D3 D4 D5 D6 D7 TB8 STOP TI RECEIVE TIMING RXD START D0 D1 D2 D3 D4 D5 D6 D7 RB8 STOP BIT DETECTOR SAMPLING SHIFT RI Figure 12-3. Serial Port Mode 2 ____________________________________________________________________________________________ 120 Ultra-High-Speed Flash Microcontroller User’s Guide Mode 3 Mode 3 has the same operation as mode 2, except for the baud-rate source. As shown in Figure 12-4, mode 3 can use timer 1 or 2 for serial port 0 and timer 1 for serial port 1. The bit shifting and protocol are the same. SBUF SMOD_0 = PCON.7 OR SMOD_1 = WDCON.7 0 TCLK = T2CON.4 RCLK = T2CON.5 1 0 1 TXD PIN 0 TB8 = SCONx.3 DATA BUS LDSBUF RDSBUF LOAD SERIAL BUFFER DIVIDEBY-16 0 P3.1 LATCH S0 D4 D3 D2 D1 D0 START LOAD 1 BAUD CLOCK 1 SHIFT SBUF READ SERIAL BUFFER SERIAL I/O CONTROL DIVIDEBY-2 AVAILABLE TO SERIAL PORT 0 ONLY TIMER 2 OVERFLOW TIMER 1 OVERFLOW CLOCK STOP D8 D7 D6 D5 TRANSMIT SHIFT REGISTER RECEIVE DATA BUFFER RD WR LOAD RESET D4 D3 D2 D5 D1 D0 START SI SERIAL INTERRUPT D8 D7 D6 R1 FLAG = SCONx.0 STOP T1 FLAG = SCONx.1 CLOCK INTS RB8 = SCONx.3 RECEIVE SHIFT REGISTER DIVIDEBY-16 RXD PIN BIT DETECTION TRANSMIT TIMING LDSBUF SHIFT TXD START D0 D1 D2 D3 D4 D5 D6 D7 TB8 STOP TI RECEIVE TIMING RXD START D0 D1 D2 D3 D4 D5 D6 BIT DETECTOR SAMPLING SHIFT RI Figure 12-4. Serial Port Mode 3 121 ____________________________________________________________________________________________ D7 RB8 STOP Ultra-High-Speed Flash Microcontroller User’s Guide FRAMING ERROR DETECTION A framing error occurs when a valid stop bit is not detected. This results in the possible improper reception of the serial word. The UART can detect a framing error and notify the software. Typical causes of framing errors are noise and contention. The framing error condition is reported in the SCON register for the corresponding UART. The framing error bit, FE, is located in SCON0.7 or SCON1.7. Note that this bit normally serves as SM0 and is described as SM0/FE_0 or SM0/FE_1 in the register description. Framing error information is made accessible by the SMOD0 framing error detection enable bit located at PCON.6. When SMOD0 is set to a logic 1, the framing error information is shown in SM0/FE (SCON0.7 or SCON1.7). When SMOD0 is set to a logic 0, the SM0 function is accessible. The information for bits SM0 and FE is actually stored in different registers. Changing SMOD0 only modifies which register is accessed; not the contents of either. The FE bit is set to a 1 when a framing error occurs. It must be cleared by software. Note that the SMOD0 state must be 1 while reading or writing the FE bit. Also note that receiving a properly framed serial word does not clear the FE bit. This must be done in software. MULTIPROCESSOR COMMUNICATION The multiprocessor communication mode makes special use of the 9th data bit in modes 2 and 3. In the original 8051, the 9th bit was restricted to a 0 or 1 condition, but had no special purpose. In the 80C32 and the ultra-high-speed microcontroller, it can be used to signify that the incoming byte is an address. This allows the processor to be interrupted only if the correct address appears. If the multiprocessor mode has been enabled, the receive interrupt (signaled by the RI bit) only occurs when a recognized address is received. When a serial word is received with the 9th bit set and the appropriate SM2 = 1, the byte is assumed to be an address. The address is compared to an internally stored address. If it matches, a receive interrupt occurs. The internal address is derived from the contents of two registers. The first register specifies an absolute address. This is the user-specified address of the device. The second register is a bit-masking register that tells the comparator which address bit(s) to actually use in the comparison. This allows broadcast transmissions that reach groups of microcontrollers or all microcontrollers on a serial port. The user defines this protocol. There are two special function registers that support multiprocessor communication for each UART. These are independent, so that different addresses can be used in each. The registers are SADDR0 or SADDR1 (hex address A9h or AAh) and SADEN0 or SADEN1 (hex address B9h or BAh). The SADDR register specifies the individual processor’s address. The SADEN identifies address bits that should be ignored in matching addresses. Software writes an 8-bit address to the SADDR register. This is the microcontroller’s individual address. Any bit in SADEN that contains a logic 0 causes the corresponding bit in SADDR to be ignored in comparison. Thus, logic 0 bits in SADEN create “don’t care” bit states for address comparisons. When an address is received, each address bit that is not masked by a “don’t care” is compared to the SADDR. The microcontroller interrupts on any address that matches this comparison. Any address that meets this comparison is called a given address. The following example shows how one address can be directed to an individual processor, or two out of three: Micro 1 SADDR 11110000 SADEN 11111010 ––––––––––––––––––––––––– Given 11110x0x Micro 2 SADDR 11110001 SADEN 11111001 ––––––––––––––––––––––––– Given 11110xx1 Micro 3 SADDR 11110010 SADEN 11111010 ––––––––––––––––––––––––– Given 11110x1x Note that an address of 11110000 reaches only microcontroller 1. An address of 11110001 reaches both microcontroller 1 and microcontroller 2. An address of 11110010 reaches only microcontroller 3. ____________________________________________________________________________________________ 122 Ultra-High-Speed Flash Microcontroller User’s Guide The microcontroller also matches on any address that corresponds to the broadcast address. This is the logical OR of the SADDR and SADEN registers, with any zeros defined as don’t cares. In most cases, the broadcast address is FFh. The multiprocessor communication is always enabled. However, the SADEN registers default to 00h, which means all address bits are “don’t care,” so all match. Thus, if no multiprocessor communication is used, these registers can be ignored. SECTION 13: TIMED-ACCESS PROTECTION The ultra-high-speed microcontroller uses a protection feature called timed access to prevent accidental writes to critical SFR bits. These bits could cause a system failure or prevent the watchdog timer from doing its job if improperly written. The timed access involves opening a timing window during which the protected bit can be modified. If the window is opened correctly, it remains open long enough to alter one protected register. This section explains which bits are protected, why, and how to use the timed-access feature. PROTECTED BITS Bits that are protected by the timed-access feature are shown below. Only critical function bits unique to the ultra-high-speed microcontroller are protected, ensuring code compatibility with the original 80C51 or 80C52. A full description of the function of each bit is provided in Section 4: WDCON.0 WDCON.1 WDCON.3 WDCON.6 EXIF.0 ACON.5 ACON.6 ACON.7 ROMSIZE.0 ROMSIZE.1 ROMSIZE.2 ROMSIZE.3 FCNTL.0 FCNTL.1 FCNTL.2 FCNTL.3 RWT EWT WDIF POR BGS PAGES0 PAGES1 PAGEE RMS0 RMS1 RMS2 RMS3 FC0 FC1 FC2 FC3 Reset watchdog timer Watchdog reset enable Watchdog interrupt flag Power-on reset flag Bandgap select Page mode select bit 0 Page mode select bit 1 Page mode enable Program memory size select bit 0 Program memory size select bit 1 Program memory size select bit 2 Program RAM enable Flash command bit 0 Flash command bit 1 Flash command bit 2 Flash command bit 3 PROTECTION SCHEME Each bit mentioned above is protected against an accidental write by requiring the software to perform a procedure before writing the bit. Timed access requires the software to write two specific values to the timed-access register during two consecutive instruction cycles. The values AAh, then 55h, must be written in consecutive instructions to the TA register at SFR location C7h. If the writes are performed correctly, the write-access window opens for three memory cycles. During this window, the software may modify a protected bit. The suggested code to open a timed-access window is: MOV 0C7h, #0AAh MOV 0C7h, #55h The procedure to modify a timed-access-protected bit begins by writing the value AAh to the timed-access register (TA;C7h). The value 55h must then be written to the timed-access register within three memory cycles of writing AAh. This opens a three memory-cycle window, after the write of 55h, during which any timed-access protected bits may be modified. Failure to complete any of the required steps also requires the procedure to begin again, starting with the write of AAh to the timed-access register. Attempts to modify timedaccess-protected bits after the window has closed are ignored. This is regardless of whether any bits were modified. Figure 13-1 illustrates a number of examples of correct and incorrect use of the timed-access procedure. 123 ____________________________________________________________________________________________ Ultra-High-Speed Flash Microcontroller User’s Guide VALID TIMED-ACCESS PROCEDURES Three Memory Cycles MOV 0C7h, #0AAh Three Memory Cycles MOV 0C7h, #55h Three Memory Cycles SETB EWT Three Memory Cycles MOV 0C7h, #0AAh Three Memory Cycles MOV 0C7h, #55h One Memory Cycle NOP Three Memory Cycles MOV 0C7h, #0AAh Three Memory Cycles MOV 0C7h, #55h Three Memory Cycles MOV WDCON, #02h Two Memory Cycles SETB EWT INVALID TIMED-ACCESS PROCEDURES Three Memory Cycles One Memory Cycle Three Memory Cycles Two Memory Cycles MOV 0C7h, #0AAh NOP MOV 0C7h, #55H SETB EWT *Second write to TA register does not occur within 3 cycles of first write. Three Memory Cycles Three Memory Cycles One Memory Cycle Three Memory Cycles MOV 0C7h, #0AAh MOV 0C7h, #55H NOP MOV WDCON, #02h *Modification of protected bit did not occur with 3 cycles of second write to TA register. Three Memory Cycles Three Memory Cycles Two Memory Cycles Two Memory Cycles MOV 0C7h, #0AAh MOV 0C7h, #55h SETB EWT SETB EWT *Modification of second protected bit did not complete within 3 cycles of second write to TA register. Figure 13-1. Timed-Access Examples TIMED-ACCESS PROTECTS WATCHDOG Any microcontroller-based system can be faced with environmental conditions that are beyond its designed abilities. These include external signal transients due to component failure, fluctuating power conditions, massive electrostatic discharge (ESD), and other unexpected system events. When a microcontroller is exposed to such conditions, program execution can become corrupted. The ultra-high-speed microcontroller incorporates a watchdog timer that can initiate a reset to recover from these conditions. The primary function of the timed-access feature is to protect against accidental disabling of the watchdog timer by an “out-of-control” device. This allows the watchdog timer to reset the system in the event of program execution failure. The following hypothetical example demonstrates how a single bit change can corrupt program execution. The timed-access procedure protects against an accidental write to the EWT bit by the errant code, allowing the watchdog timer reset function to reset the device. While this is a purely fictitious example, it illustrates how the watchdog timer and timed-access feature allow the ultra-highspeed microcontroller to minimize the effect of accidental code corruption. Note: Timed access is not optional and must be supported if the protected bits are used. This example helps explain the category of problem that the timed access prevents. EXAMPLE: A TRANSIENT CAUSES THE WATCHDOG TO BE DISABLED: TABLE_READ: C2D2 90 0A 00 C2D5 79 FF C2D7 78 90 C2D9 C2DA C2DB C2DC C2DD E0 F6 06 A3 D9 C2 D9 MOV MOV MOV LOOP: MOVX MOV INC INC DJNZ DPTR, 0A00H R1, #0FFH R0, #90H ;LOAD TABLE POINTER ;LOAD COUNTER ;DESTINATION POINTER A, @DPTR @R0, A R0 DPTR R1, LOOP ;READ DATA BYTE ;STORE IT IN RAM ;NEXT TABLE LOCATION ;NEXT DATA VALUE ;NEXT BYTE OR DONE ? ____________________________________________________________________________________________ 124 Ultra-High-Speed Flash Microcontroller User’s Guide A transient occurs while the op code is being fetched for the first instruction. The transient causes 1 bit of the op code in the first instruction to be read as a 0 instead of 1. The resulting program is what the microcontroller would actually execute: TABLE_READ: C2D2 80 0A 00 C2D5 79 FF C2D7 78 90 C2D9 C2DA C2DB C2DC C2DD SJMP MOV MOV LOOP: MOVX MOV INC INC DJNZ E0 F6 06 A3 D9 C2 D9 0BH R1, #0FFH R0, #90H ;RELATIVE JUMP BY 10 LOCATIONS ;LOAD COUNTER ;DESTINATION POINTER A, @DPTR @R0, A R0 DPTR R1, LOOP ;READ DATA BYTE ;STORE IT IN RAM ;NEXT TABLE LOCATION ;NEXT DATA VALUE ;NEXT BYTE OR DONE ? The resulting jump is to address C2DE. This is not even a real op code, but would be treated as such. The resulting fetch is the value C2 D9. This is the op code for CLR D9h. The bit-addressable location D9h corresponds to the EWT. If the timed-access procedure did not prevent it, this errant instruction would disable the watchdog. Note that the program execution is completely lost now. Real op codes are being replaced by operands, data, and garbage. In the ultra-high-speed microcontroller, the watchdog recovers from this state as soon as it times out, since it could not have been disabled in this way. In the ultra-high-speed microcontroller it is very hard to contrive a situation that accidentally disables the watchdog. Note, the timed access prevents accidentally writing a bit. It can not prevent accidentally calling the correct code that writes a bit. This is much more unlikely, however. SECTION 14: INSTRUCTION SET DETAILS Details of flags modified by each instruction are located in Section 4. MNEMONIC ADD A, Rn DEC @Ri D7 0 0 a7 0 0 d7 0 0 a7 0 0 d7 1 1 a7 1 1 d7 0 0 0 a7 0 1 0 0 0 a7 0 D6 0 0 a6 0 0 d6 0 0 a6 0 0 d6 0 0 a6 0 0 d6 0 0 0 a6 0 0 0 0 0 a6 0 MUL AB 1 0 DIV AB 1 0 ADD A, direct ADD A, @Ri ADD A, #data ADDC A, Rn ADDC A, direct ADDC A, @Ri ARITHMETIC OPERATION ADDC A,#data SUBB A, Rn SUBB A, direct SUBB A, @Ri SUBB A, #data INC A INC Rn INC direct INC @Ri INC DPTR DEC A DEC Rn DEC direct 125 INSTRUCTION CODE D5 D4 D3 D2 1 0 1 n2 1 0 0 1 a5 a4 a3 a2 1 0 0 1 1 0 0 1 d5 d4 d3 d2 1 1 1 n2 1 1 0 1 a5 a4 a3 a2 1 1 0 1 1 1 0 1 d5 d4 d3 d2 0 1 1 n2 0 1 0 1 a5 a4 a3 a2 0 1 0 1 0 1 0 1 d5 d4 d3 d2 0 0 0 1 0 0 1 n2 0 0 0 1 a5 a4 a3 a2 0 0 0 1 1 0 0 0 0 1 0 1 0 1 1 n2 0 1 0 1 a5 a4 a3 a2 0 1 0 1 1 0 D1 n1 0 a1 1 0 d1 n1 0 a1 1 0 d1 n1 0 a1 1 0 d1 0 n1 0 a1 1 1 0 n1 0 a1 1 BYTE 1 CYCLE 1 2 2 (A) = (A) + (direct) 1 2 (A) = (A) + ((Ri)) 2 2 (A) = (A) + #data 1 1 (A) = (A) + (C) + (Rn) 2 2 (A) = (A) + (C) + (direct) 1 2 (A) = (A) + (C) + ((Ri)) 2 2 (A) = (A) + (C) + #data 1 1 (A) = (A) - (C) - (Rn) 2 2 (A) = (A) - (C) - (direct) 1 2 (A) = (A) - (C) - ((Ri)) 2 2 (A) = (A) - (C) - #data 1 1 1 1 (A) = (A) + 1 (Rn) = (Rn) + 1 2 2* (direct) = (direct) + 1 n0 1 a0 i HEX 28-2F 25 Byte 2 26-27 24 Byte 2 38-3F 35 Byte 2 36-37 34 Byte 2 98-9F 95 Byte 2 96-97 94 Byte 2 04 08-0F 05 Byte 2 06-07 A3 14 18-1F 15 Byte 2 16-17 1 1 1 1 2 1 1 1 ((Ri)) = ((Ri)) + 1 (DPTR) = (DPTR) + 1 (A) = (A) - 1 (Rn) = (Rn) - 1 2 2* (direct) = (direct) -1 D0 n0 1 a0 i 0 d0 n0 1 a0 i 0 d0 n0 1 a0 i 0 d0 0 n0 1 a0 i 1 0 EXPLANATION (A) = (A) + (Rn) 1 2 ((Ri)) = ((Ri)) - 1 0 0 1 0 0 A4 1 9 (B15–8 ), (A7–0 ) = (A) x (B) 0 0 1 0 0 84 1 10 (B15–8 ), (A7–0 ) = (A) / (B) ____________________________________________________________________________________________ Ultra-High-Speed Flash Microcontroller User’s Guide ARITHMETIC OPER MNEMONIC DA A D6 1 1 0 0 a7 0 0 d7 0 a7 0 a7 d7 0 0 a7 0 0 d7 1 1 a6 1 1 d6 1 a6 1 a6 d6 1 1 a6 1 1 d6 0 INSTRUCTION CODE D5 D4 D3 D2 D1 D0 HEX BYTE CYCLE EXPLANATION Contents of accumulator are BCD, IF [[(A3-0 ) > 9] OR [(AC) = 1]] THEN (A3-0 ) = (A3-0 ) + 6 AND IF [[(A7-4 ) > 9] OR [(C) = 1]] THEN (A7-4 ) = (A7-4 ) + 6 0 1 0 1 0 0 D4 1 2 1 1 a4 1 1 d4 1 a4 1 a4 d4 0 0 a4 0 0 d4 1 0 a3 0 0 d3 0 a3 0 a3 d3 1 0 a3 0 0 d3 n2 1 a2 1 1 d2 0 a2 0 a2 d2 n2 1 a2 1 1 d2 n1 0 a1 1 0 d1 1 a1 1 a1 d1 n1 1 a1 1 0 d1 n0 i a0 i 0 d0 0 a0 1 a0 d0 n0 I a0 i 0 d0 1 (A) = (A) AND (Rn) 2 2 (A) = (A) AND (direct) 1 2 (A) = (A) AND ((Ri)) 2 2 (A) = (A) AND #data 2 2* (direct) = (direct) AND A 3 3 (direct) = (direct) AND #data 1 0 0 0 0 1 0 XRL A, @ Ri 0 a7 d7 0 0 a7 0 1 a6 d6 1 1 a6 1 0 a5 d5 1 1 a5 1 0 a4 d4 0 0 a4 0 0 a3 d3 1 0 a3 0 0 a2 d2 n2 1 a2 1 1 a1 d1 n1 0 a1 1 1 a0 d0 n0 1 a0 i XRL A, #data 0 1 1 0 0 1 0 0 XRL direct, A 0 1 1 0 0 0 1 0 0 a7 d7 1 1 1 a6 d6 1 1 1 a5 d5 1 1 0 a4 d4 0 1 0 a3 d3 0 0 0 a2 d2 1 1 58-5F 55 Byte 2 56-57 54 Byte 2 52 Byte 2 53 Byte 2 Byte 3 48-4F 45 Byte 2 46-47 44 Byte 2 42 Byte 2 43 Byte 2 Byte 3 68-6F 65 Byte 2 66-67 64 Byte 2 62 Byte 2 63 Byte 2 Byte 3 E4 F4 1 0 a5 0 0 d5 0 a5 0 a5 d5 0 0 a5 0 0 d5 ANL A, Rn ANL A, direct ANL A, @Ri ANL A, #data ANL direct, A ANL direct, #data ORL A, Rn ORL A, direct LOGICAL OPERATION D7 ORL A, @Ri ORL A, #data ORL direct, A ORL direct, #data XRL A, Rn XRL A, direct XRL direct, #data CLR A CPL A 0 1 a1d1 0 0 1 a0 d0 0 0 1 1 2 2 1 2 (A) = (A) OR (Rn) (A) = (A) OR (direct) (A) = (A) OR ((Ri)) 2 2 (A) = (A) OR #data 2 2* (direct) = (direct) OR (A) 3 3 (direct) = (direct) OR #data 1 1 2 2 1 2 2 2 2 2* 3 3 (direct) = (direct) XOR #data 1 1 1 1 (A) = 0 (A) = (A) (A) = (A) XOR (Rn) (A) = (A) XOR (direct) (A) = (A) XOR ((Ri)) (direct) = (A) XOR #data (direct) = (direct) XOR (A) ____________________________________________________________________________________________ 126 Ultra-High-Speed Flash Microcontroller User’s Guide MNEMONIC RL A D7 D6 0 0 INSTRUCTION CODE D5 D4 D3 D2 1 0 0 0 D1 D0 HEX BYTE CYCLE 1 1 23 1 1 EXPLANATION A7 A6 A5 A4 A3 A2 A1 A0 The contents of the accumulator are rotated left by 1 bit. LOGICAL OPERATION RLC A RR A 1 1 0 0 1 1 33 1 1 C A7 A6 A5 A4 A3 A 2 A1 A0 0 0 0 0 0 0 1 1 03 1 1 A7 A6 A5 A4 A3 A 2 A0 A1 The contents of the accumulator are rotated right by 1 bit. RRC A MOV A, direct MOV A, @Ri MOV A, #data DATA TRANSFER 0 Rotated left through the carry by 1 bit. SWAP A MOV A, Rn MOV Rn, A MOV Rn, direct MOV Rn, #data MOV direct, A MOV direct, Rn MOVdirect1, direct2 MOV direct, @Ri 127 0 0 1 1 1 a7 1 0 d7 1 1 a7 0 d7 1 a7 1 a7 1 a7 a7 1 a7 0 1 1 1 a6 1 1 d6 1 0 a6 1 d6 1 a6 0 a6 0 a6 a6 0 a6 0 1 0 1 1 a5 1 1 d5 1 1 a5 1 d5 1 a5 0 a5 0 a5 a5 0 a5 0 0 0 a5 0 1 d4 1 0 a5 1 d4 1 a4 0 a4 0 a4 a4 0 a4 0 0 1 0 a3 0 0 d3 1 1 a3 1 d3 0 a3 1 a3 0 a3 a3 0 a3 0 1 n2 1 a2 1 1 d2 n2 n2 a2 n2 d2 1 a2 n2 a2 1 a2 a2 1 a2 1 1 0 n1 0 a1 1 0 d1 n1 n1 a1 n1 d1 0 a1 n1 a1 0 a1 a1 1 a1 0 n0 1 a0 i 0 d0 n0 n0 a0 n0 d0 1 a0 n0 a0 1 a0 a0 i a0 13 C4 E8-EF E5 Byte 2 E6-E7 74 Byte 2 F8-FF A8-AF Byte 2 78-7F Byte 2 F5 Byte 2 88-8F Byte 2 85 Byte 2 Byte 3 86-87 Byte 2 1 1 C A7 A6 A5 A4 A3 A2 A0 A1 1 1 1 1 Rotated right through the carry by 1 bit. (A3-0 ) _ (A7-4 ) (A) = (Rn) 2 2 (A) = (direct) 1 2 (A) = ((Ri)) 2 2 (A) = #data 1 1 (Rn) = (A) 2 2 (Rn) = (direct) 2 2 (Rn) = #data 2 2* (direct) = (A) 2 2* (direct) = (Rn) 3 3* (direct1) = (direct2) (source) (destination) 2 2* (direct) = ((Ri)) ____________________________________________________________________________________________ Ultra-High-Speed Flash Microcontroller User’s Guide MNEMONIC MOV direct, #data MOV @Ri, A MOV @Ri, direct MOV @Ri, #data DATA TRANSFER MOV DPTR, #data16 MOVC A, @A + DPTR MOVC A, @A + PC MOVX A, @Ri MOVX @DPTR MOVX @Ri, A MOVX @DPTR,A PUSH direct POP direct XCH A, Rn XCH A, direct XCH A, @Ri XCHD A, @Ri MNEMONIC CLR C BOOLEAN VARIABLE MANIPULATION CLR bit SETB C SETB bit CPL C CPL bit ANL C, bit ANL C, bit ORL C, bit ORL C, bit MOV C, bit MOV bit, C D7 0 a7 d7 1 1 a7 0 d7 1 d7 d7 D6 1 a6 d6 1 0 a6 1 d6 0 d6 d6 INSTRUCTION CODE D5 D4 D3 D2 1 1 0 1 a5 a4 a3 a2 d5 d4 d3 d2 1 1 0 1 1 0 0 1 a5 a4 a3 a2 1 1 0 1 d5 d4 d3 d2 0 1 0 0 d5 d4 d3 d2 d5 d4 d3 d2 D1 0 a1 d1 1 1 a1 1 d1 0 d1 d1 D0 1 a0 d0 i i a0 i d0 0 d0 d0 HEX 75 Byte 2 Byte 3 F6-F7 A6-A7 Byte 2 76-77 Byte 2 90 Byte 2 Byte 3 BYTE CYCLE 3 3 EXPLANATION (direct) = #data 1 1 ((Ri)) = A 2 2 ((Ri)) = (direct) 2 2 ((Ri)) = #data 3 3 (DPTR) = #data15-0 (DPH) = #data15-8 (DPL) = #data7-0 1 0 0 1 0 0 1 1 93 1 3 (A) = ((A) + (DPTR)) 1 1 0 1 0 1 0 0 0 0 0 0 1 1 1 i 83 E2-E3 1 1 3 2 (A) = ((A) + (PC)) (A) = ((Ri)) 1 1 1 0 0 0 0 0 E0 1 2 (A) = ((DPTR)) 1 1 1 1 0 0 1 i F2-F3 1 2 ((Ri)) = (A) 1 1 1 1 0 0 0 0 F0 1 2 1 a7 1 a7 1 1 a7 1 1 1 a6 1 a6 1 1 a6 1 1 0 a5 0 a5 0 0 a5 0 0 0 a4 1 a4 0 0 a4 0 1 0 a3 0 a3 1 0 a3 0 0 0 a2 0 a2 n2 1 a2 1 1 0 a1 0 a1 n1 0 a1 1 1 0 a0 0 a0 n0 1 a0 i i C0 Byte 2 D0 Byte 2 C8-CF C5 Byte 2 C6-C7 D6-D7 D7 1 1 b7 1 1 b7 1 1 b7 1 b7 1 b7 1 b7 1 b7 1 b7 1 b7 D6 1 1 b6 1 1 b6 0 0 b6 0 b6 0 b6 1 b6 0 b6 0 b6 0 b6 INSTRUCTION CODE D5 D4 D3 D2 0 0 0 0 0 0 0 0 b5 b4 b3 b2 0 1 0 0 0 1 0 0 b5 b4 b3 b2 1 1 0 0 1 1 0 0 b5 b4 b3 b2 0 0 0 0 b5 b4 b3 b2 1 1 0 0 b5 b4 b3 b2 1 1 0 0 b5 b4 b3 b2 1 0 0 0 b5 b4 b3 b2 1 0 0 0 b5 b4 b3 b2 0 1 0 0 b5 b4 b3 b2 D1 1 1 b1 1 1 b1 1 1 b1 1 b1 0 b1 1 b1 0 b1 1 b1 1 b1 D0 1 0 b0 1 0 b0 1 0 b0 0 b0 0 b0 0 b0 0 b0 0 b0 0 b0 HEX C3 C2 Byte 2 D3 D2 Byte 2 B3 B2 Byte 2 82 Byte 2 B0 Byte 2 72 Byte 2 A0 Byte 2 A2 Byte 2 92 Byte 2 ((DPTR)) = (A) 1 2 (SP) = (SP) + 1 ((SP)) = (direct) (direct) = ((SP)) (SP) = (SP) - 1 (A) = (Rn) 2 3 (A) = (direct) 1 1 3 3 (A) = ((Ri)) (A3-0) = ((Ri3-0 )) 2 2 2 2* BYTE 1 CYCLE 1 (C) = 0 EXPLANATION 2 2* (bit) = 0 1 1 (C) = 1 2 2* (bit) = 1 1 1 (C) = (C) 2 2* (bit) = (bit) 2 2 (C) = (C) AND (bit) 2 2 (C) = (C) AND (bit) 2 2 (C) = (C) OR (bit) 2 2 (C) = (C) OR (bit) 2 2 (C) = (bit) 2 2 (bit) = (C) ____________________________________________________________________________________________ 128 Ultra-High-Speed Flash Microcontroller User’s Guide D7 D6 INSTRUCTION CODE D5 D4 D3 D2 a10 a7 a9 a6 a8 a5 0 a15 a7 0 a14 a6 RET 0 RETI MNEMONIC ACALL addr 11 PROGRAM BRANCHING D0 HEX BYTE CYCLE 0 a3 0 a2 0 a1 1 a0 Byte 1 Byte 2 2 2 EXPLANATION (PC) = (PC) + 2 (SP) = (SP) + 1 ((SP)) = (PC7-0 ) (SP) = (SP) + 1 ((SP)) = (PC15-8 ) (PC) = page address (PC) = (PC) + 3 (SP) = (SP) + 1 ((SP)) = (PC7-0 ) (SP) = (SP) + 1 ((SP)) = (PC15-8 ) (PC) = addr15-0 (PC15-8 ) = ((SP)) (SP) = (SP) - 1 (PC7-0 ) = ((SP)) (SP) = (SP) - 1 (PC15-8 ) = ((SP)) (SP) = (SP) - 1 (PC7-0 ) = ((SP)) (SP) = (SP) - 1 (PC) = (PC) + 2 (PC10-0 ) = page address 0 a13 a5 1 a12 a5 0 a11 a3 0 a10 a2 1 a9 a1 0 a8 a0 12 Byte 2 Byte 3 3 3 0 1 0 0 0 1 0 22 1 3 0 0 1 1 0 0 1 0 32 1 3 a10 a7 0 a15 a7 1 r7 a9 a6 0 a14 a6 0 r6 a8 a5 0 a13 a5 0 r5 0 a4 0 a12 a4 0 r4 0 a3 0 a11 a3 0 r3 0 a2 0 a10 a2 0 r2 0 a1 1 a9 a1 0 r1 1 a0 0 a8 a0 0 r0 Byte 1 Byte 2 02 Byte 2 Byte 3 80 Byte 2 2 2 3 3 (PC) = addr15-0 2 3 (PC) = (PC) + 2 (PC) = (PC) + rel JMP @A + DPTR 0 1 1 1 0 0 1 1 73 1 3 (PC) = (A) + (DPTR) JZ rel 1 r7 0 r6 0 r5 0 r4 0 r3 0 r2 0 r1 0 r0 60 Byte 2 2 3 JNZ rel 1 r7 0 r6 0 r5 0 r4 0 r3 0 r2 0 r1 0 r0 70 Byte 2 2 3 LCALL addr 16 AJMP addr 11 LJMP addr 16 SJMP rel 129 1 a8 D1 ____________________________________________________________________________________________ (PC) = (PC) + 2 IF (A) = 0 THEN (PC) = (PC) + rel (PC) = (PC) + 2 IF (A) ≠ 0 THEN (PC) = (PC) + rel Ultra-High-Speed Flash Microcontroller User’s Guide D7 D6 INSTRUCTION CODE D5 D4 D3 D2 D1 D0 HEX BYTE CYCLE JC rel 0 r7 1 r6 0 r5 0 r4 0 r3 0 r2 0 r1 0 r0 40 Byte 2 2 3 JNC rel 0 r7 1 r6 0 r5 1 r4 0 r3 0 r2 0 r1 0 r0 50 Byte 2 2 3 0 b7 r7 0 b7 r7 0 b6 r6 0 b6 r6 1 b5 r5 0 b5 r5 0 b4 r4 1 b4 r4 0 b3 r3 0 b3 r3 0 b2 r2 0 b2 r2 0 b1 r1 0 b1 r1 0 b0 r0 0 b0 r0 20 Byte 2 Byte 3 30 Byte 2 Byte 3 3 4 3 4 JBC bit, rel 0 b7 r7 0 b6 r6 0 b5 r5 1 b4 r4 0 b3 r3 0 b2 r2 0 b1 r1 0 b0 r0 10 Byte 2 Byte 3 3 4* CJNE A, direct, rel 0 a7 r7 0 a6 r6 0 a5 r5 1 a4 r4 0 a3 r3 0 a2 r2 0 a1 r1 0 a0 r0 B5 Byte 2 Byte 3 3 5 CJNE A, #data, rel 1 d7 r7 0 d6 r6 1 d5 r5 1 d4 r4 0 d3 r3 1 d2 r2 0 d1 r1 0 d0 r0 B4 Byte 2 Byte 3 3 4 CJNE Rn, #data, rel 1 d7 r7 0 d6 r6 1 d5 r5 1 d4 r4 1 d3 r3 n2 d2 r2 n1 d1 r1 n0 d0 r0 B8-BF Byte 2 Byte 3 3 4 CJNE @Ri, #data, rel 1 d7 r7 0 d6 r6 1 d5 r5 1 d4 r4 0 d3 r3 1 d2 r2 1 d1 r1 i d0 r0 B6-B7 Byte 2 Byte 3 3 5 DJNZ Rn, rel 1 r7 1 r6 0 a5 r5 1 r4 1 r3 n2 r2 n1 r1 n0 r0 D8-Df Byte 2 3 4 DJNZ direct,rel 1 a7 r7 1 a6 r6 1 d5 r5 1 a4 r4 0 a3 r3 1 a2 r2 0 a1 r1 1 a0 r0 D5 Byte 2 Byte 3 3 5 NOP 0 0 0 0 0 0 0 0 00 1 1 MNEMONIC JB bit, rel PROGRAM BRANCHING JNB bit, rel EXPLANATION (PC) = (PC) + 2 IF (C) = 1 THEN (PC) = (PC) + rel (PC) = (PC) + 2 IF (C) ≠ 0 THEN (PC) = (PC) + rel (PC) = (PC) + 3 IF (bit) = 1 THEN (PC) = (PC) + rel (PC) = (PC) + 3 IF (bit) = 0 THEN (PC) = (PC) + rel (PC) = (PC) + 3 IF (bit) = 1 THEN (bit) = 0 and (PC) = (PC) + rel (PC) = (PC) + 3 IF (direct) < (A) THEN (PC) = (PC) + rel and (C) = 0 OR IF (direct) > (A) THEN (PC) = (PC) + rel and (C) = 1 (PC) = (PC) + 3 IF #data < (A) THEN (PC) = (PC) + rel and (C) = 0 OR IF #data > (A) THEN (PC) = (PC) + rel and (C) = 1 (PC) = (PC) + 3 IF #data < (Rn) THEN (PC) = (PC) + rel and (C) = 0 OR IF #data > (Rn) THEN (PC) = (PC) + rel and (C) = 1 (PC) = (PC) + 3 IF #data < ((Ri)) THEN (PC) = (PC) + rel and (C) = 0 OR IF #data > ((Ri)) THEN (PC) = (PC) + rel and (C) = 1 (PC) = (PC) + 2 (Rn) = (Rn) - 1 IF (Rn) ≠ 0 THEN (PC) = (PC) + rel (PC) = (PC) + 3 (direct) = (direct) - 1 IF (direct) ≠ 0 THEN (PC) = (PC) + rel (PC) = (PC) + 1 * Note: One additional clock cycle is required if the PSW, SP, DPS, IE, EIE, IP0, IP1, EIP0, or EIP1 register is accessed by certain direct addressing instructions marked with an *. Additionally, the JBC bit instruction requires one additional clock cycle to clear a bit if the jump is actually taken. ____________________________________________________________________________________________ 130 Ultra-High-Speed Flash Microcontroller User’s Guide SECTION 15: PROGRAM LOADING The ultra-high-speed flash microcontroller family can perform program loading or reloading in a number of ways. First, ROM loader mode can be invoked to create a serial communication channel, which permits in-system program/erase of the internal and external program memory. Secondly, parallel programming mode allows programming and erasure of the internal flash memory using industrystandard EPROM or flash parallel programmers. ROM LOADER MODE The ultra-high-speed flash microcontroller defaults to the normal operating (nonloader) mode without external hardware. ROM loader mode can be invoked at any time, as described later in this section. Once the loader session is complete, the device performs a hardware reset and begin operation. This is identical to an external reset, except that the ROM loader during the loader session may modify locations in scratchpad RAM in order to execute properly. The Table 15-1 shows which areas of scratchpad RAM are guaranteed preserved and which ones are of indeterminate state after exiting the loader. Table 15-1. Preserved and Indeterminate Scratchpad Memory TYPE Guaranteed Preserved Indeterminate ULTRA-HIGH-SPEED FLASH MICROCONTROLLER SCRATCHPAD MEMORY 80h–FFh 00h–7Fh The guaranteed preserved locations are areas in scratchpad RAM that are not changed by the ROM loader. The indeterminate area contains various stacks and buffers used by the loader, and a given byte in this area may or may not be modified by the loader. As such, the user should not rely on the loader preserving any data in this area. It should also be noted that the loader, upon being invoked, clears the EWT bit (WDCON.1) so that the watchdog timer is prevented from generating an internal reset during the loader session. Invoking the ROM Loader Mode The ROM loader mode is invoked by simultaneously applying a logic 1 to the RST pin, a logic 0 to the EA pin, and driving the PSEN pin to a logic 0 level. If power were to cycle while the required input stimuli were present, the loader would be invoked on power-up. When the ROM loader mode is invoked, the device awaits an incoming <CR> character (0Dh) on serial port 0 at a baud rate that can be detected by the autobaud routine. The autobaud routine is described later in this section. The autobaud routine receives and transmits data only on serial port 0, ignoring activity on serial port 1. Upon successful baud-rate detection, the ROM loader transmits a banner similar to the one shown below, signaling to the host that loader mode has successfully been invoked. The banner is followed by a “>” prompt, which indicates the device is ready to receive a command. The command set recognizable by the ROM loader is also detailed later in this section. The flow of these conditions is shown in Figure 15-1. Application Note 3262: In-System Programming with 8051-Based Microcontrollers contains more information on the use of the ROM loader as well as tips on debugging potential problems. Exiting the Loader To exit ROM loader mode, first float the PSEN signal, and then float or drive the RST pin low. The RST pin has an internal pulldown. The PSEN signal is an output and drives itself high. When the loader stimulus is removed, the processor performs a hardware reset and begin execution at location 0000h. Note that both of these conditions must occur, or the loader is exited. The flow of these conditions is shown in Figure 15-1. Serial Program Load Operation Program loading through a serial port is a convenient method of loading application software into the flash memory or external memory. Communication is performed over a standard, asynchronous serial communications port using a terminal emulator program with 8-N-1 (8 data bits, no parity, 1 stop bit) protocol settings. A typical application would use a simple RS-232 serial interface to in-system program the device as part of a final production procedure. The hardware configuration for the serial program load operation is illustrated in Figure 15-2. A variety of crystals can be used to produce standard baud rates. The ROM is designed to operate across a 3-wire interface from a standard UART. The receive, transmit, and ground wires are all that are necessary to establish communication with the device. The ROM implements an easy-to-use command line interface, which allows an Intel hex file to be loaded and read back from the device. Intel hex is the standard format output by 8051 cross-assemblers. 131 ____________________________________________________________________________________________ Ultra-High-Speed Flash Microcontroller User’s Guide PROGRAM EXECUTION INTERNAL MEMORY DISABLED EA\ =0 INTERNAL MEMORY ENABLED EA\ =1 N RST = 1 Y N EA = 0 DEVICE IN RESET (INTERNAL MEMORY ENABLED) Y N PSEN = 0 DEVICE IN RESET (INTERNAL MEMORY DISABLED) Y AUTOBAUD ROUTINE AWAITING <CR> CHARACTER ON RXD OF SERIAL PORT 0 N SERIAL COMMUNICATION WITH THE ROM LOADER CAN PROCEED AS ESTABLISHED ON SERIAL PORT 0 N POWER-ON RESET PSEN TOGGLED 0 Y 1 Y Figure 15-1. Invoking and Exiting the Loader on the Ultra-High-Speed Flash Microcontroller ____________________________________________________________________________________________ 132 Ultra-High-Speed Flash Microcontroller User’s Guide AUTOBAUD-RATE DETECTION The ROM loader can automatically detect, within certain limits, the external baud rate and configure itself to that speed. The loader controls serial port 0 in mode 1 (asynchronous, 1 start bit, 8 data bits, no parity, 1 stop bit, full duplex), using timer 1 in 8-bit autoreload mode with the serial port 0 doubler bit (PCON.7) set. For these settings, an equation to calculate possible baud rates is provided as a function of crystal frequency and timer reload value. Table 15-1 shows baud rates generated using the equation: ROM Loader_Baud rate = Crystal Frequency 192 x (256-Timer Reload) ** Timer reload values attempted by the loader: FF, FE, FD, FC, FB, FA, F8, F6, F5, F4, F3, F0, EC, EA, E8, E6, E0, DD, D8, D4, D0, CC, C0, BA, B0, A8, A0, 98, 80, 60, 40 When communicating with a PC COM port having a standard 8250/16450 UART, attempt to match the loader baud rate and PC COM port baud rate within 3% in order to maintain a reliable communication channel. If baud rates cannot be matched exactly, it is suggested configuring the loader to the faster baud rate to avoid the possibility of overflowing the microcontroller serial input buffer. COMMAND LINE INTERFACE The ROM loader uses an easy-to-use command line interface that responds to alphabetic commands that are summarized in Table 15-3. A detailed description of each command can be found in the following pages. ULTRA-HIGH-SPEED FLASH MICROCONTROLLER TO PC TD ROOUT RXDO TOOUT DS232A TOIN TXDO ROIN RD DTR R1IN T1OUT R1OUT EA T1IN VCC RST PSEN HC/AC125 Figure 15-2. Serial Load Hardware Configuration 133 ____________________________________________________________________________________________ Ultra-High-Speed Flash Microcontroller User’s Guide Table 15-2. ROM Loader Baud Rates vs. Crystal Frequency CRYSTAL FREQUENCY (MHz) 32.0000 29.4912 24.5760 24.0000 22.1184 20.0000 16.0000 11.0592 1.84320 TIMER RELOAD LOADER BAUD RATE ERROR (%) PC UART BAUD RATE PC UART RELOAD F6 F3 EC FC F8 F4 F6 F5 F4 F3 F0 E6 FF FE FD F8 F0 E8 FB F6 F4 FF FE FD FF FE FD 16667 12821 8333 38400 19200 12800 12800 11636 10667 9615 7812 4808 115200 57600 38400 13021 6510 4340 16667 8333 6944 57600 28800 19200 9600 4800 3200 -1.3 -0.2 -1.3 0.0 0.0 0.0 0.0 -1.0 -1.8 -0.2 -1.7 -0.2 0.0 0.0 0.0 -1.7 -1.7 -1.7 -1.3 -1.3 -2.4 0.0 0.0 0.0 0.0 0.0 0.0 16457 12800 8229 38400 19200 12800 12800 11520 10473 9600 7680 4800 115200 57600 38400 12800 6400 4267 16457 8229 6776 57600 28800 19200 9600 4800 3200 7 9 14 3 6 9 9 10 11 12 15 24 1 2 3 9 18 27 7 14 17 2 4 6 12 24 36 Note: Only a few possible timer reload/PC UART reload values are shown per crystal frequency. This table, by no means, is an exhaustive list of acceptable configurations for each crystal frequency, nor should it be considered a list of the allowable crystal frequencies. Table 15-3. Alphabetic Commands COMMAND B C CX D DX K L LB LE FUNCTION Self-CFC of internal ROM code. CRC-16 of flash memory range (inhibited if either LB2 or LB3 set). CRC-16 of external RAM range. Dump Intel hex from internal flash memory range. Dump Intel hex from external RAM range. Klear-Erase entire flash memory range. Load flash memory (Oh–3FFFh). Load flash memory blindly (Oh–3FFFh)—no verify or precheck. Load encryption vector (0–3Fh). ____________________________________________________________________________________________ 134 Ultra-High-Speed Flash Microcontroller User’s Guide Table 15-3. Alphabetic Commands (continued) COMMAND LX R V VE VX W ^C FUNCTION Load external RAM (Oh–FFFFh). Read configuration. Verify flash memory against incoming hex. Verify encryption vector against incoming hex. Verify external RAM against incoming hex. Write register(s) Reset loader. Selected commands require arguments and some commands have optional arguments. In all cases, arguments are expected to be hexadecimal numbers. In addition, an ASCII control-C character (^C) causes the ROM loader to terminate any function currently being executed and display the command line prompt. An incoming break character (defined as a received null character (00h) with the stop bit = 0) causes the ROM loader to be restarted and the baud rate redetermined. COMMAND LINE SYNTAX Single-letter ASCII characters are recognized as commands by the ROM loader. Arguments are represented by hexadecimal numbers. A hexadecimal number is any sequence of hexadecimal characters. A hexadecimal character may be a digit, 0 through 9, or one of the letters A through F. A byte is always the right-most two digits of a hexadecimal number. An address is always the right-most four digits of a hexadecimal number. BYTE CONVERSION A → 0AH AB → 0ABH ABC → 0BCH ABCD → 0CDH ADDRESS CONVERSION A → 000AH AB → 00ABH ABC → 0ABCH ABCD → 0ABCDH ABCDE → 0BCDEH The C, CX, D, and DX commands allow optional addresses to be entered. The syntax [begin-address [end-address]] is used to convey the following meanings: • No arguments: Begin-address is set to 0 and end-address is set to the range. • One argument: Begin-address is set to the argument and end-address is set to the range. • Two arguments: Begin-address is set to the first argument and end-address is set to the second argument. This second address must not exceed the address value specified by the range. In the second and third bullets above, an error message is generated if the end address is less than the begin address, either implicitly or explicitly. Error messages are transmitted as soon as errors are detected. All messages are preceded by the two characters E:, and followed by a mnemonic description. Commands are not processed until an entire command line is entered and terminated with a <CR>. No command line may be greater than 17 bytes. Since a command line is not processed until a <CR> is entered, the <delete> character can be used to make edits. Lines longer than 17 characters return an error message and no action is taken for that command line. Only legal characters are echoed back by the loader. The legal characters are: 0123456789, <:>, <space>, ABCDEFGHIJKLMNOPQRSTUVWXYZ, and <delete>. Backspaces (<BS>) are converted to delete characters. The horizontal tab character is converted to space. Lowercase alphabetic characters are converted to uppercase alphabetic. The <delete> character is executed as a <BS> <space> <BS> when possible in command mode. This causes the character to be overprinted on a hard-copy device. The <CR> character generates a <CR> <LF> pair. 135 ____________________________________________________________________________________________ Ultra-High-Speed Flash Microcontroller User’s Guide COMMAND SUMMARIES B Return the CRC-16 (cyclic redundancy check) of the internal ROM code. This self-CRC computation should always return 0000h. C [begin-address [end-address]] Return the CRC-16 (cyclic redundancy check) of the flash memory. This computation is performed over the range unless optional start and end addresses are given. The CRC-16 algorithm is commonly used in data communications. A blank device will return the following checksum values: DS89C420/430: B001; DS89C440: 4040; DS89C450: 4400. CX [begin-address [end-address]] Return the CRC-16 (cyclic redundancy check) of the external RAM. This computation is performed over the range unless optional start and end addresses are given. The CRC-16 algorithm is commonly used in data communications. D [begin-address [end-address]] Dump flash memory in Intel hex format. An optional address range may be specified. Each record contains up to 32 data bytes. The last line printed is the end-of-data record. DX [begin-address [end-address]] Dump external memory in Intel hex format. This command functions in the same manner as the D command with the exception of the target memory being external. K Perform an erasure of the entire flash memory range, including security block, option control register, and bank-select bit. L Load standard ASCII Intel hex formatted data into flash memory. No lock bits may be set when attempting to load the internal flash program memory. Only record types 00 and 01 are processed. Each record of the file is treated the same way. All characters are discarded before the header character <:> is read. The rest of the record, defined by the length byte, is then processed. All characters following the record checksum and prior to the next <:> are discarded. Control returns to the command prompt after an Intel end record is encountered. Prior to programming each byte, the loader performs a preread of that flash memory location to assess whether the new hex value can be written. After programming each byte, the loader reads the flash memory location again to verify proper programming. The processing of each record is confirmed by an ACK/NAK response. New records should not be transmitted until the ACK/NAK byte associated with the previous record has been received. The ACK/NAK responses from the ROM loader are as follows: A–Invalid address in Intel Hex record: Intel Hex record contains an address that exceeds the maximum internal flash memory or encryption vector address (3Fh). F–Flash controller error: An error was detected within the flash memory of the target device. Contact Dallas Semiconductor if the device responds with repeated error messages of this type. G–Good record: Record was received and programmed without error. H–Invalid Intel Hex record format: Intel Hex record contains a nonhex character. L–Invalid Intel Hex record length: Intel Hex record length exceeds allowable length [20 bytes (type 0); 0 bytes (type 1 EOF)]. P–Failure to write 1s to 0s during programming: Target flash memory was not erased prior to loading, or is programming a byte more than once. Be sure to use the ROM loader K command before attempting to load new software. Also be sure that the Intel Hex file does not reference more than one byte at the same location. R–Invalid Intel Hex record type: ROM loader only accepts Intel Hex record types 00 and 01 in standard Intel Hex format; make sure the assembler/compiler is not configured for Intel Extended Hex or HEX-386 format. S–Invalid checksum in Intel Hex record: Intel Hex record contains a checksum that does not correspond to its hex record. This error is caused by manual edits to the Intel Hex file or a compiler error. V–Verify Error: Target flash memory does not match expected value. Erase memory and reprogram. ____________________________________________________________________________________________ 136 Ultra-High-Speed Flash Microcontroller User’s Guide LB Load blind of internal flash memory—Loads standard Intel hex-formatted data into internal flash memory. This command functions in the same manner as the L command, except that the preprogramming assessment and postprogramming verification of the flash memory are not executed by the loader. When using this command, the P and V NAK responses are not returned by the loader. All other ACK/NAK responses are still generated by the loader. LE Load encryption vector—Loads standard Intel hex-formatted data into flash security block. This command functions in the same manner as the L command, except that it operates on the flash security block (0–3Fh). LX Load external memory—Loads standard Intel hex-formatted data into external memory. This command functions similar to the L command, except that it operates on the external memory (0–FFFFh) and can write without restriction to any address location in the range. If an external page mode or MOVX stretch cycle different from the default setting is desired, the ACON or CKCON registers should be modified prior to execution of the LX command. R Read—Displays the values of the lock bits, option control register, address control register, clock control register, power management register, Ports 0, 1, 2, 3, and the flash control register in the following format. Although displayed, the FCNTC register is not used in the DS89C420. LB:XX OCR:XX ACON:XX CKCON:XX P0:XX P1:XX P2:XX P3:XX FCNTL:XX V Verify current contents of flash memory vs. the received Intel hex. This command operates similar to the load command, except that it does not write to the flash memory; it compares the data byte(s) in the flash memory to the byte(s) in the data stream. The same ACK/NAK response scheme is used during verify operations as is used for load operations. VE Verify encryption vector—Verifies current contents of the flash security block vs. the received Intel hex. This command works in the same manner as the V command, but operates on the flash security block (0–3Fh). VX Verify external memory—Verifies current contents of external memory versus the received Intel hex. This command works in the same manner as the ‘V’ command, but operates on external memory (0–FFFFh). W [LB | OCR | ACON | CKCON | P0 | P1 | P2 | P3] byte Writes byte to the requested register. Valid entries for LB are 1 (enable LB1), 3 (enable LB1, LB2), and 7 (enable LB1, LB2, LB3). ACON register writes modify only bits 7, 6, 5. CKCON register writes modify only bits 2,1, 0. PMR register writes modify only bit 0. P3 register writes modify only P3.7–P3.2 (P3.1 and P3.0 are not altered). ^C Interrupt whatever is going on, clear all the buffers, put up a prompt, and wait for the next command. Anything in the type-ahead buffer is removed. All output is stopped. ERROR MESSAGES E:ARGREQ An argument or arguments are required for this command. E:BADCMD An invalid command letter was entered. E:BADREG This message is printed if a register other than OCR, ACON, CKCON, PMR, P0, P1, P2, or P3 is used as the argument for the W command. 137 ____________________________________________________________________________________________ Ultra-High-Speed Flash Microcontroller User’s Guide E:BADVAL The requested value cannot be programmed into the OCR register because it contains 1’s in bit position(s) where 0’s have already been programmed. E:EXTARG Extra data was encountered on the command line when it was not needed. Reenter the command. E:ILLOPT The optional parameters given were in error. If the start address is greater than the end address, either implicitly or explicitly, then an error is printed. The range bit implicitly determines the maximum range. E:LOCK BITS ENABLED The requested operation cannot be performed due to the current lock bit settings. E:LOCK BITS ALREADY SET The requested lock bit setting cannot be programmed because a higher order lock bit has already been programmed. E:NOTHEX A nonhexadecimal character was found when expecting a hexadecimal character. E:VALUE MUST BE 1, 3 OR 7 A value other than 1, 3, or 7 was entered when trying to write the lock bits. PARALLEL PROGRAMMING MODE The parallel program load mode is compatible with most industry-standard parallel programmers (Figure 15-3). The data sheet contains the most comprehensive information relating to the parallel programming mode. ADDRESS 7:0 P1.0 P1.1 P1.2 P1.3 P1.4 P1.5 P1.6 P1.7 DRIVE HIGH RST ERROR P3.0 P3.1 P3.2 P3.3 P3.4 P3.5 P3.6 P3.7 BUSY/READY CONTROL CONTROL 4MHz TO 6MHz XTAL2 XTAL1 P0.0 P0.1 P0.2 P0.3 P0.4 P0.5 P0.6 P0.7 µC EA ALE PROG PSEN P2.7 P2.6 P2.5 P2.4 P2.3 P2.2 P2.1 P2.0 DATA DRIVE HIGH CONTROL DRIVE LOW CONTROL CONTROL Figure 15-3. Parallel Load Hardware Configuration ____________________________________________________________________________________________ 138 Ultra-High-Speed Flash Microcontroller User’s Guide USER CODE IN-APPLICATION PROGRAMMING MODE The data sheet contains the most comprehensive information relating to the in-application programming mode. Additional supporting information can be found in the SFR definitions of FCNTL (D5h) and FDATA (F6h) of this user’s guide. In-application prpgramming mode is not supported on the DS89C420. INTEL HEX FILE FORMAT Assemblers that are 8051-compatible produce an absolute output file in Intel hex format. These files are composed of a series of records. Records in an Intel hex file have the following format: <header><hex information><record terminator> The specific record elements are detailed as follows: : II aaaa tt dddddd ... dd xx where: : Indicates a record beginning II Indicates the record length aaaa Indicates the 16-bit load address tt Indicates the record type dd Indicates hex data xx Indicates the checksum = (two’s complement (II+aa+a+tt+dd+dd+...dd) Record type 00 indicates a data record and type 01 indicates an end record. An end record appears as :00 00000 01 FF. These are the only valid record types for a NIL hex file. Spaces are provided for clarity. The following is a short Intel hex file. The data bytes begin at 01 and count up to 2F. Notice the record’s length, beginning address, and record type at the start of each line and the checksum at the end: :200000000102030405060708090A0B0C0D0E0F101112131415161718191A1B1C1D1E1F20D0 :0F0020002122232425262728292A2B2C2D2E2F79 :00000001FF REVISION HISTORY January 24, 2001. Original Issue October 3, 2002. Revision 1 December 2, 2002. Revision 2 Changed title to reflect “flash” and removed “DS89C420” reference. August 21, 2003. Revision 3 Made documument universal to all Dallas ultra-high-speed microcontrollers. February 19, 2004. Revision 4 Corrected cycle times for ADDC A, Rn instruction; clarified that CTM bit is cleared in Stop mode; clarified that PMR SFR is not displayed in ROM loader. August 6, 2004. Revision 5 Added FCNTL and FDATA to Special-Function Register Locations table. December 1, 2004. Revision 6 Changed the ACK/NAK responses in the Command Summaries section (p.136). Maxim cannot assume responsibility for use of any circuitry other than circuitry entirely embodied in a Maxim product. No circuit patent licenses are implied. Maxim reserves the right to change the circuitry and specifications without notice at any time. Maxim Integrated Products, 120 San Gabriel Drive, Sunnyvale, CA 94086 408-737-7600 _____________________________________________________ 139 © 2004 Maxim Integrated Products Printed USA is a registered trademark of Dallas Semiconductor Corporation. is a registered trademark of Maxim Integrated Products.