TOSHIBA TLCS-90 Series TMP90CM38 CMOS 8–Bit Microcontroller TMP90CM38F/TMP90CM38T 1. Outline and Characteristics The TMP90CM38 is a high-speed, high performance 8-bit microcontroller developed for application in the control of various devices. The TMP90CM38, CMOS 8-bit microcontroller, integrates an 8-bit CPU, ROM, RAM, A/D converter, D/A converter, multi-function timer/event counter, general-purpose serial interface, signal measure circuit, timing pulse generation circuit and PWM output in a single chip, and with which external program memory and data memory can be extended up to 31KB. The TMP90CM38F uses an 80-pin flat package. The TMP90CM38T uses an 84-pin QF (PLCC) package. (6) (7) (8) (9) (10) The following are the features of TMP90CM38: (1) (2) (3) (4) (5) Highly efficient instruction set: 167 basic instructions instructions, including Division and multiplication instructions, 16-bit operation instructions and bit operation instructions Minimum instruction executing time: 250ns (at 16MHz) Built-in ROM: 32K bytes Built-in RAM: 1K bytes Memory extension capability External program memory: 31K bytes External data memory: 31K bytes (11) (12) (13) Interrupt functions: 13 internal, 5 external 8-bit A/D converter (8 channels) 8-bit D/A converter (2 channels) General-purpose serial interface mode (2 channels) • With asynchronous mode and I/O interface mode (1 channel) • With synchronous mode (1 channel) • I/O interface mode (1 channels) Timer function (1) 16-bit timer/event counter (1 channel) ----- Built-in 2 capture register and 2 comparator (2) 8-bit timer (4 channels) ----- Built-in 1 comparator in each channel (3) Watchdog timer function (WDTOUT pin having) I/O ports: Max. 66 pins HDMA function (2 channels) ----- 1 byte transmission: 1.75µs (@16.0MHz) Software standby function ----- RUN, STOP, IDLE modes Hardware standby function ----- STOP mode The information contained here is subject to change without notice. The information contained herein is presented only as guide for the applications of our products. No responsibility is assumed by TOSHIBA for any infringements of patents or other rights of the third parties which may result from its use. No license is granted by implication or otherwise under any patent or patent rights of TOSHIBA or others. These TOSHIBA products are intended for usage in general electronic equipments (office equipment, communication equipment, measuring equipment, domestic electrification, etc.) Please make sure that you consult with us before you use these TOSHIBA products in equipments which require high quality and/or reliability, and in equipments which could have major impact to the welfare of human life (atomic energy control, spaceship, traffic signal, combustion control, all types of safety devices, etc.). TOSHIBA cannot accept liability to any damage which may occur in case these TOSHIBA products were used in the mentioned equipments without prior consultation with TOSHIBA. TOSHIBA CORPORATION 1 TMP90CM38 Figure 1. TMP90CM38 Block Diagram 2 TOSHIBA CORPORATION TMP90CM38 2. Pin Assignment and Functions The assignment of input/output pins for TMP90CM38, their name and outline functions are described below. 2.1 Pin Layout Diagram Figure 2.1 (1) shows the pin assignment of TMP90CM38F. Figure 2.1 (1). Pin Assignments (Flat Package) TOSHIBA CORPORATION 3 TMP90CM38 Figure 2.1 (2) shows the pin assignment of TMP90CM38T. Figure 2.1 (2). Pin Assignments (QFJ (PLCC) Package) 4 TOSHIBA CORPORATION TMP90CM38 2.2 Pin Names and Functions The names of input/output pins and their functions are described below. Table 2.2 Shows the input/output pin names and functions. Table 2.2. Pin Names and Functions (1/3) TOSHIBA CORPORATION 5 TMP90CM38 Table 2.2. Pin Names and Functions (2/3) 6 TOSHIBA CORPORATION TMP90CM38 Table 2.2. Pin Names and Functions (3/3) TOSHIBA CORPORATION 7 TMP90CM38 3. Operation This section explains the functions and basic operations of the TMP90CM38 in blocks. 3.1 CPU The TMP90CM38F has a built-in, high performance 8 bit CPU. For the operation of the CPU, see the book TLCS 90 Series CPU Core Architecture. This section explains the CPU functions unique to the TMP90CM38 that are not explained in “that book. 3.1.1 Reset Figure 3.1 (1) shows the basic timing of reset. To reset TMP90CM38, it is required that the power supply voltage is within operating range, the internal oscillator is stably functioning, and RESET input be kept at “0” for at least 10 system clocks (10 states: 2 microseconds with a 10MHz system clock) When a reset is accepted, among I/O common ports, port 0 (address data bus A0 - A7), port 1 (address bus A8 - A15) and port 2 are set to input status (with high impedance). Output ports P30 (RD), P31 (WR), CLK, and WDTOUT (P80) are set to “1” and ALE (P83) is cleared to “0”. CPU registers and external memory are not changed. However, program counter PC and interrupt enable/disable flag IFF are cleared to “0”. The A register becomes undefined. When the reset is released, instruction execution starts from address 0000H. Figure 3.1 (1). Reset Timing of TMP90CM38 8 TOSHIBA CORPORATION TMP90CM38 3.1.2 EXF (Exchange Flag) The exchange flag EXF is inverted when the EXX instruction is executed to exchange data between the TMP90CM38 main registers and auxiliary registers. This flag is allocated to bit 7 at memory address FFE1H. 3.1.3 Wait Control For the TMP90CM38, a wait control register (WAITC) is assigned to bits 4 and 5 at memory address FFB0H. TOSHIBA CORPORATION 9 TMP90CM38 3.2 Memory Map The TMP90CM38 can provide a maximum 64K byte program memory and data memory. The program and data memories may be allocated to the addresses 0000H ~ FFFFH. (1) (2) The TMP90CM38 contains a 1K byte built-in RAM which is allocated to the addresses FBA0H ~ FF9FH. The CPU can also access some portions of the RAM (160 byte area FF00H ~ FF9FH) using short instruction codes in the direct addressing mode. Built-in ROM The TMP90CM38 has an internal 32K byte ROM. This ROM is located at addresses 0000H ~ 7FFFH. Program execution starts from address 000H after a reset operation. Addresses 0010H ~ 00C7H in the internal ROM area are used as the interrupt processing entry area. Built-in RAM (3) Built-in I/O The TMP90CM38 uses 96 bytes of the address space as a built-in I/O area. The area is allocated to the addresses FFA0H ~ FFFFH. The CPU can access the built-in I/O using short instruction codes in the direct addressing mode. Figure 3.2 shows the memory map and the access ranges of the CPU for each addressing mode. Figure 3.2. Memory Map 10 TOSHIBA CORPORATION TMP90CM38 3.3 Interrupt Functions The TMP90CM38 has a general-purpose interrupt processing routine for responding to both internal and external interrupt request, and a high-speed micro DMA (HDMA) processing mode in which the CPU automatically transfers data. Immediately after a reset is released, all responses to interrupt requests are set to the general-purpose interrupt processing mode. The high-speed DMA processing mode can be set by loading a vector value to the DMAV 0/1 register. Figure 3.3 (1) shows the interrupt response flow. Figure 3.3 (1). Interrupt Response Flow When an interrupt request is generated, this is reported to the CPU via the built-in interrupt controller. If the request is for a non-maskable interrupt or an enabled maskable interrupt, the CPU starts interrupt processing. If for a disabled maskable interrupt, the request is ignored and not received. If the interrupt is received, the CPU first reads the interrupt vector from the built-in interrupt controller to determine the source of the interrupt request. TOSHIBA CORPORATION Next, a check is made as to whether this request is for general-purpose interrupt processing, micro DMA processing or high-speed DMA (HDMA) processing, and then the corresponding processing is performed. The interrupt vector is read in an internal operation cycle so the bus cycle becomes a dummy cycle. 11 TMP90CM38 3.3.1 General-Purpose Interrupt Processing Figure 3.3 (2) shows the general-purpose interrupt processing flow. The CPU first saves the contents of the program counter PC and register AF (including the interrupt enable/disable flag IFF immediately before an interrupt) to the stack and then resets the interrupt enable/disable flag IFF to “0” (interrupt disable). Finally, the interrupt vector contents [V] are transferred to the program counter and a jump is made to the interrupt processing program. There is a 20-state overhead from the time when the interrupt is received until the jump is made to the interrupt processing program. Figure 3.3 (2). General-Purpose Interrupt Processing Flow Interrupt processing program is ended with the RETI instruction for both maskable and non-maskable interrupts. Executing this instruction restores the program counter PC and register AF contents from the stack. (Resets the interrupt enable/disable flag immediately before an interrupt.) When the CPU reads the interrupt vector, the interrupt request source confirms that the interrupt has been received and then clears the interrupt request. Non-maskable interrupts cannot be disabled by program. Maskable interrupts, however, can be enabled and disabled by program. Bit 5 of CPU reg- 12 ister F is an interrupt enable/disable flipflop (IFF). Interrupts are enabled by setting this bit to “1” with the EI (interrupt enable) instruction and disabled by resetting this bit to “0” with the DI (interrupt disable) instruction. IFF is reset to “0” by resetting and when an interrupt is received (including non-maskable interrupts). The EI instruction is actually executed after the next instruction is executed. Table 3.3 (1) shows the interrupt sources. TOSHIBA CORPORATION TMP90CM38 Table 3.3 (1) Interrupt Sources The “priority sequence” shown in Table 3.3 (1) indicates the sequence in which interrupt sources are received by the CPU when multiple interrupt requests are generated simultaneously. For example, if interrupt requests with the priority sequences 4 and 5 are generated simultaneously, the CPU will receive the interrupt request with priority sequence 4 first. When processing of the interrupt with priority sequence 4 is ended with the RETI instruction, the CPU will then receive the interrupt with priority sequence 5. If the interrupt processing program for the priority sequence 4 interrupt is interrupted by executing the EI instruction, TOSHIBA CORPORATION the CPU will receive the priority sequence 5 interrupt request. When multiple interrupt requests are generated simultaneously, the built-in interrupt controller only determines the priority sequence of the interrupt sources received by the CPU. There is no function to compare the priority sequence of the interrupt currently being processed and the interrupt currently being requested. Another interrupt can be enabled while another interrupt is being processed by resetting the interrupt enable/disable flag IFF to enable. 13 TMP90CM38 3.3.2 High-Speed Micro DMA Processing The TMP90CM38 has two built-in DMA channels called HDMA. HDMA has three times the processing capacity of µDMA and is used for high-speed data transfers. HDMA execution time (decrease the value of transfer number and the value is not “0” data) is 14 states, regardless of whether the 1-byte transfer mode or 2-byte transfer mode is used. HDMA and micro DMA (the TMP90CM38 has not the micro DMA) transfer speeds. Table 3.3 (4) shows the HDMA functions. Table 3.3 (3) Transfer Speeds Table 3.3 (4) Shows the DHMA Functions (1) HDMA Setting Registers HDMA operation. The following describes the registers required for 14 TOSHIBA CORPORATION TMP90CM38 TOSHIBA CORPORATION 15 TMP90CM38 Note: (2) 16 It is ineffective to set decrement for a destination address when a source address being increment; and to set increment for a destination address when a source address being decrement. Register Loading TOSHIBA CORPORATION TMP90CM38 (3) HDMA Start (4) HDMA Channel 0 and Channel 1 Priority Sequence HDMA can be started by any of the following TMP90CM38 maskable interrupt sources The channel where an interrupt is generated first has priority (a) Note: HDMA, regardless of an interrupt enable flag, compares the vector and the values of the DMA V0/1 register. If they match in EI mode, the HDMA starts. Do not write the vector value of the nonmaskable interrupt to the DMA V0/1 register. If doing so, the HDMA does not operate normally. To stop the HDMA from being started, set DI mode before generating the interrupt to start the HDMA, or set the DMA V0/1 register to 00H. Internal start factors • Internal I/O interrupts Assign starting of HDMA channel 0 or channel 1 to the INT0 - INT3 external interrupts, connect any of the bits of ports 0 - 8 (output mode) externally to INT0 - INT3 to genrate a start interrupt. (b) External start factors • INT0 ~ 3 pin TOSHIBA CORPORATION 17 TMP90CM38 (5) HDMA Operation Flow Figure 3.3 (6). HDMA Operation Flow 18 TOSHIBA CORPORATION TMP90CM38 (6) HDMA Operation Timing Figure 3.3 (7a). HDMA Operation Timing TOSHIBA CORPORATION 19 TMP90CM38 Figure 3.3 (7b). HDMA Operation Timing 20 TOSHIBA CORPORATION TMP90CM38 3.3.3 Interrupt Controller Figure 3.3 (9) shows an abbreviated interrupt circuit diagram. The left half of this diagram shows the interrupt controller and the right half shows the CPU interrupt request signal circuit and hold release circuit. The interrupt controller has an interrupt request flipflop and interrupt enable/disable flag for each interrupt channel (total: 18 channels), and a micro DMA enable/disable flag. The interrupt request flip-flop latches interrupt requests that arrive from the periphery. This flipflop is reset to “0” when there is a reset, when the CPU receives an interrupt and reads the vector of that interrupt channel, and when an instruction that clears the interrupt request (writes “vector value/8” to memory address FFE0H) for that channel is executed. TOSHIBA CORPORATION LD (0FFE0H), 60H/8 For example, when LD (0FFE0H), 38H/8 is executed, the interrupt request flipflop for the interrupt channel [INTT1] with the vector value 38H is reset to “0” (to clear the flipflop, also write to address FFC9H when the interrupt request flag is assigned to FFE1H and FFE2H). Table 3.3 (5) shows the “interrupt vector value/8” values. The status of the interrupt request flipflop can be determined by reading memory address FFC9H, FFCAH or FFCBH. “0” means no interrupt request and “1” means an interrupt request. Figure 3.3 (8) shows the bit layout when the interrupt request flipflop is read. 21 TMP90CM38 Table 3.3 (4) Interrupt Vector Value/8 Values Figure 3.3 (5). Interrupt Request Flipflop Read (1/2) 22 TOSHIBA CORPORATION TMP90CM38 Figure 3.3 (6). Interrupt Request Flipflop Read (2/2) TOSHIBA CORPORATION 23 TMP90CM38 Figure 3.3 (7). Interrupt Controller Block 24 TOSHIBA CORPORATION TMP90CM38 The interrupt enable/disable flags for each interrupt request channel are assigned to memory addresses FFE3H - FFE5H. Interrupt Common Terminal INT0 P81 INT1 INT2 P53 Interrupts are enabled for a channel by setting the flag to “1”. The flags are reset to “0” by reseting. Mode How to set Level INTE2 <EDGE> = 0 Rise edge INTE2 <EDGE> = 1 Rise edge T4MOD <CAPM1, 0> = 0, 0 or 0, 1 or 1, 1 Fall edge T4MOD <CAPM1, 0> = 1, 0 P54 – Rise edge INT2 P56 – Rise edge For the pulse width for external interrupt, refer to “4.7 Interrupt Operation”. TOSHIBA CORPORATION Be careful that the following five are exceptional circuits. 25 TMP90CM38 INT0 Level mode As the INT0 is not an edge type interrupt, the interrupt request flip-flop is cancelled, and thus an interrupt request from peripheral devices passes through S input of the flip-flop to become Q output. When the mode is changed over (from edge type to level type), the previous interrupt request flag will be cleared automatically. When the mode is changed from level to edge, the interrupt request flag set in the level mode is not cleared. Thus, use the following sequence to clear the interrupt request flag. DI SET 6, (0FFE5H): Switch the mode from level to edge LD (0FFE0H), 05H: Clear interrupt request flag EI INTRX1, INTRX2 26 The interrupt request flip-flop cannot be cleared only by reset operation or reading the serial channel receiving buffer, and cannot be cleared by an instruction. TOSHIBA CORPORATION TMP90CM38 Figure 3.3 (8). Interrupt Enable Flags TOSHIBA CORPORATION 27 TMP90CM38 Figure 3.3 (9). Interrupt Processing Flow Chart 28 TOSHIBA CORPORATION TMP90CM38 3.4 Standby Functions When a HALT instruction is executed, TMP90CM38 enters the RUN, IDLE1 or STOP mode according to the contents of the halt mode setting register. The features are as follows: (1) Run: Only the CPU halts, power consumption remains unchanged. (2) IDLE: Only the internal oscillators operate, while all other internal circuits halt. Power consumption is 1/10 or less than that during normal operation. (3) STOP: All internal circuits halt, including the internal oscillator. Power consumption is extremely reduced. The HALT mode setting register WDMOD <HALTM1,0> is assigned to bits 2 and 3 memory address FFECH in the built-in I/ O register area (all other bits are used to control other block functions). The RUN mode (“00”) is entered by reseting. These HALT states can be released by requesting an interrupt or resetting. Table 3.4 (2) shows how to release the HALT state. If the CPU is in the EI state for non-maskable or maskable interrupt, the interrupt will be acknowledged by the CPU and the CPU starts interrupt processing. If the CPU is in the DI state fro maskable interrupt, the CPU starts the execution from the instruction following HALT instruction, but the interrupt request flag remains at “1”. Even when HALT state is released by reset operation, the state (including the built-in RAM) just before entering the HALT can be retained. However, if HALT instruction has already been executed in the built-in RAM, the RAM contents may not be retained. Figure 3.4 (1). HALT Mode Setting Register TOSHIBA CORPORATION 29 TMP90CM38 3.4.1 RUN Mode Figure 3.4 (2) shows the timing for releasing the HALT state by an interrupt during RUN mode. In the RUN mode, the system clock inside MCU does not stop even after HALT instruction has been executed; the CPU merely stops executing instructions. Accordingly, the CPU repeats dummy cycle until HALT state, interrupt requests are sampled at the fall edge of CLK signal. Figure 3.4 (2). HALT Release Timing Using Interrupts in RUN Mode 30 TOSHIBA CORPORATION TMP90CM38 3.4.2 IDLE1 Mode Figure 3.4 (3) shows the timing used for releasing the HALT mode by interrupts in the IDLE1 mode. In the IDLE1 mode, only the internal oscillator operates, the system clock inside MCU stops and CLK signal is fixed to “1”. In the HALT state, interrupt requests are sampled asynchronously with the system clock but sampling is performed synchronously with the system clock, whereas the HALT release (restart of operation) is performed synchronously with it. Figure 3.4 (3). HALT Release Timing Using Interrupts in the IDLE1 Mode TOSHIBA CORPORATION 31 TMP90CM38 3.4.3 STOP Mode Figure 3.4 (4) shows the timing of HALT release caused by interrupts in STOP mode. In the STOP mode, all interval circuits stop, including internal oscillator. When the STOP mode is activated, all pins except special ones are put in the high-impedance state, isolated from the internal operation of MCU. Table 3.4 (1) shows the state of each pin in the STOP mode. However, if WDMOD <DRVE> (drive enable: bit 0 of memory address FFECH) of th built-in I/O register is set to “1”, the pre-halt state of the pins can be retained. The register is cleared to “0” by reset operation. When the CPU accepts an interrupt request, the internal oscillator first restarts. However, to get the stabilized oscillation, the system clock starts its output after the time set by warming up counter has passed. WDMOD <WARM> (warming up: bit 4 at memory address FFECH) is used to set up the warming up time. Warming up is executed for 214 clock oscillation time when this bit is set to “0”, while 216 clock oscillation time when set to “1”. This bit is cleared to “0” by reset operation. Figure 3.4 (4). HALT Release Timing Using Interrupts in STOP Mode 32 TOSHIBA CORPORATION TMP90CM38 The internal oscillator can also be restarted by inputting the RESET signal “0” to the CPU. However, the warming up counter remains inactive in order to make the CPU rapidly operate when the power is turned on. Accordingly, wrong operation may occur due to TOSHIBA CORPORATION unstable clocks immediately after the internal oscillator has restarted. To release the HALT state by resetting in the STOP mode, RESET signal must be kept at “0” for a sufficient period of time. 33 TMP90CM38 Table 3.4 (1) State of Pins in STOP Mode IN/OUT DRVE = 0 DRVE = 1 Input Mode Output Mode – – – Output P10 ~ P17 Input Mode Output Mode – – Input Output P20 ~ P23 Input Mode Output Mode – – Input Output P00 ~ P07 P30 ~ P31 Output Mode – Output P32 ~ P33 Input Mode Output Mode – – Input Output P40 ~ P47 Input Mode Output Mode – – Input Output P56 ~ P57 Input Mode Output Mode – – Input Output P60 ~ P67 Input Mode – – P70 ~ P77 Input Mode Output Mode – – Input Output NMI CLK X1 X2 EA Input Mode Output Mode Input Mode Output Mode Input Mode Input – – “1” Input Input – – “1” Input P80 (WDTOUT) P81 (INT0) P82 (STBY) P83 (ALE) Output Mode Input Mode Input Mode Output Mode – Input Input – Output Input Input Output P90 ~ P93 Input Mode Output Mode – Output Input Output P100, P101 Input Mode Output Mode – – Input Output – Indicates that input mode/input pin cannot be used for input and that the output mode/output pin have been set to high impedance. Input: Input is enabled. Input: The input gate is operating. Fix the input voltage at either “0” or “1” to prevent the pin floating. Ouput: Output status. 34 : TOSHIBA CORPORATION TMP90CM38 Table 3.4 (2) I/O Operation Release in HALT Mode TOSHIBA CORPORATION 35 TMP90CM38 3.4.5 Hardware Standby Function STBY input pin This pin is used for setting MCU-standby mode. When this pin is set “Low”, the oscillator stops and internal clock is frozen. The power consumption is extremely reduced. This function sets every pin to a condition as same as STOP mode which a HALT instruction is executed. Figure 3.4 (5) indicates the block diagram of standby mode. Figure 3.4 (5). Standby Mode Block Diagram 36 TOSHIBA CORPORATION TMP90CM38 3.5 Function of Ports The TMP90CM38 contains total of 66 I/O port pins. These port pins function not only as the general-purpose I/O ports but also as the I/O ports for the internal CPU and built-in I/O. Table 3.5 shows the functions of these port pins. Table 3.5 Functions of Ports These port pins function as the general-purpose I/O port pins by resetting (except P30, P31, P60 ~ P67, P80 ~ P83). The port pins, for which input or output is programmably selectable, TOSHIBA CORPORATION function as input ports by resetting. A separate program is required to use them for an internal function. 37 TMP90CM38 3.5.1 Port 0 (P00 ~ P07) Port 0 is the 8-bit general-purpose I/O port P0, each bit of which can be set independently for input or output. The control register P0CR is used to set input or output. Reset operations clear all output latch and control register bits to “0” and set port 0 to the input mode. In addition to the general-purpose I/O port function, port 0 also functions as an address/data bus (AD0 ~ AD7). When the external memory is accessed, port 0 automatically functions as the address/data bus. Figure 3.5 (1). Port 0 (P00 ~ P07) 38 TOSHIBA CORPORATION TMP90CM38 Figure 3.5 (2). Registers for Port 0 TOSHIBA CORPORATION 39 TMP90CM38 3.5.2 Port 1 (P10 ~ P17) Port 1 is the 8-bit general-purpose I/O port P1, each bit of which can be set to input or output. The port 1 control register P1CR is used to set input or output. Reset operations clear all output latch and the control register bits to “0” and sets all port 1 bits to the input mode. In addition to the general-purpose I/O port function, port 1 also functions as an address bus (A8 ~ A15). This is specified by setting the external extended specification register IRFL<EXT> to “1” and setting P1CR to the output mode. When the P1CR cleared to “0”, port 1 is set to the input mode, regardless of the external extended specification register value. IRF2 <EXT> also controls P30 (RD) and P31 (WR), P80 (ALE). Figure 3.5 (3). Port 1 (P10 ~ P17) 40 TOSHIBA CORPORATION TMP90CM38 Figure 3.5 (4). Registers for Port 1 TOSHIBA CORPORATION 41 TMP90CM38 3.5.3 Port 2 (P20 ~ P27) Port 2 is a 4-bit general-purpose I/O port, each bit of which can be set to input or output. The control register P29CR <P23C ~ P20C> is used to set for input or output. By reset operation, this control register is reset to “0”, and port 2 is placed in the input mode. This port can also be used as a stepping motor control/ pattern generation port 1 (M00 ~ M03). The function register P29FR <M0S> specifies whether it is to be used as the general-purpose I/O port or stepping motor control/pattern generation port. When reset, it becomes a general-purpose I/O port. Figure 3.5 (5). Port 2 42 TOSHIBA CORPORATION TMP90CM38 3.5.4 Port 9 (P90 ~ P93) Port 9 is a 4-bit general-purpose I/O port, each bit of which can be set for input or output. The control register P29CR <P93C ~ P90C> is used to set for input or output. When reset, control register will be cleared to “0”, placing the port 9 in the input mode. This port can also be used as a stepping motor control/ pattern generation port 1 (M10 ~ M13). The function register P29FR <M1S> specifies whether it is to be used as the general-purpose I/O port or stepping motor control/pattern generation port. When reset, it becomes a general-purpose I/O port. Figure 3.5 (6). Registers for Port 2 TOSHIBA CORPORATION 43 TMP90CM38 Figure 3.5 (7). Registers for Port 2 and Port 9 (1/2) 44 TOSHIBA CORPORATION TMP90CM38 Figure 3.5 (8). Registers for Port 2 and Port 9 (2/2) TOSHIBA CORPORATION 45 TMP90CM38 3.5.5 Port 3 (P30 ~ P33) P32, P33 are a 2-bit general-purpose I/O port. The control register P38CR <P33C, P32C> is used for input or output. P30, P31 are output ports. All bits of the output latch are set to “1” by resetting, and “1” is generated to the output port. Access of external memory makes P30, P31 function as the memory control pins (RD and WR), when set IRF2 <EXT> to “1”. When access of an internal memory makes them function, “1” is generated always. Also function register P38CR <RDE> is intended fro a pseudostatic RAM. When set IRF2 <EXT> to “1”, and set P38CR <RDE> to “1”, it always functions as RD pin. Therefore, the RD pin outputs “0” (Enable) when it is an internal memory read and internal I/O read cycle. Figure 3.5 (9). Port 3 (P30, 31) 46 TOSHIBA CORPORATION TMP90CM38 Figure 3.5 (10). Port 3 TOSHIBA CORPORATION 47 TMP90CM38 Figure 3.5 (11). Register for Port 3 48 TOSHIBA CORPORATION TMP90CM38 3.5.6 Port 4 (P40 ~ P47) Port 4 is the 8-bit general-purpose I/O port, each bit of which can be set for input or output. The control register P4CR is used to set input or output. All bits of the control register are cleared to “0” by resetting, and the port turns of input port mode (output latch is set to “1” by resetting). Figure 3.5 (12). Port 4 TOSHIBA CORPORATION 49 TMP90CM38 Figure 3.5 (13). Register for Port 4 50 TOSHIBA CORPORATION TMP90CM38 3.5.7 Port 5 (P50 ~ P57) Port 4 is the 8-bit general-purpose I/O port, each bit of which can be set for input or output. The control register P5CR is used to set input or output. By reset operation, the output latch is set to “1” and the control register is set to “0”, and port 5 is placed in the input mode. In addition to the general-purpose I/O port function, these ports function as interrupt request input, clock input for timer or event counter, or timer output, or wait input. (1) P55, P57, P51, P52 When specified by port 5 function register P5FR <TO1S ~ TO5S>, these ports become the timer output. Figure 3.5 (14). Port 5 (P55, P57, P51, P52) TOSHIBA CORPORATION 51 TMP90CM38 (2) P56 as external interrupt request input (INT3). P56 is also used as clock input (TI2) for 8-bit timer 0 Figure 3.5 (15). Port 5 (P56) 52 TOSHIBA CORPORATION TMP90CM38 (3) P53, P54 timer or event counter as well as external interrupt request input. These ports are also used as the clock input for 16-bit Figure 3.5 (16). Port 5 (P53, P54) TOSHIBA CORPORATION 53 TMP90CM38 (4) P50 Figure 3.5 (17). Port (P50) 54 TOSHIBA CORPORATION TMP90CM38 Figure 3.5 (18). Registers for Port 5 TOSHIBA CORPORATION 55 TMP90CM38 3.5.8 Port 6 (P60 - P67) Port 6 is an 8-bit general-purpose input port with fixed input function. In addition to its general-purpose input port function, these ports function as analog input pins (AN0 ~ AN7). Figure 3.5 (19). Port 6 (P60 ~ P67) Figure 3.5 (20). Registers for Port 6 56 TOSHIBA CORPORATION TMP90CM38 3.5.9 Port 7 (P70 ~ P77) Port 7 is the 8-bit general-purpose I/O port, each bit of which can be set for input or output. The control register P7CR is used to set input or output. By reset operations, all bits of the output latch are set to “0”, while all bits of control register are to “0”, and port 7 is placed in the input mode. In addition to the general-purpose I/O port function, port 7 have an internal serial interface input/output function. This is specified by function register P7FR. All bits of the function register are cleared to “0” by resetting, and the port turns to general-purpose I/O mode. Figure 3.5 (21). Port 7 TOSHIBA CORPORATION 57 TMP90CM38 Figure 3.5 (22). Registers for Port 7 58 TOSHIBA CORPORATION TMP90CM38 Figure 3.5 (23). Registers for Port 7 TOSHIBA CORPORATION 59 TMP90CM38 3.5.10 Port 8 (P80 - P83) Port 8 is the 4-bit general-purpose I/O port, P81, P82 are input-only ports. P80, P83 are output-only ports. In addition to its general-purpose input port function, or watch dog timer out output, these port function as external interrupt request input, or hardware input, or ALE output. (1) P81/INT0 P81 is the general-purpose input port, is also used as external interrupt request input (INT0). INT0 is to be used as “H” level detection interrupt or rise edge detection interrupt by control register INTE2 <EDGE>. Figure 3.5 (24). Port 8 (P81) 60 TOSHIBA CORPORATION TMP90CM38 (2) P80 P80 is used both as a general-purpose output port and for WDTOUT output. Bit 1 of the watchdog timer mode register (WDMOD: memory address FFECH) and Bit 1 of the P38 control register (P38CR: memory address FFA7H) is used to set P80 for WDTOUT output. P80 is WDTOUT output after reset. Figure 3.5 (25). Port 8 (P80) TOSHIBA CORPORATION 61 TMP90CM38 (3) P82/STBY P82 is a general purpose input port, and this port can be used also as Hardware standby. By reset operations, the control register P38CR <STBYS> is “0”, and P82 is placed in the general-purpose input port. When the control register P38CR <STBYS> is “1”, and P82 is placed in the hardware standby input pin. Figure 3.5 (26). Port 8 (P82) 62 TOSHIBA CORPORATION TMP90CM38 (4) P83 P83 is output port, and is also used as ALE pin. When P83 was 1 chip mode (EA = 1), by reset operations, the control register P38CR <ALEE> is “0”, and P83 is placed in the output port. When ALE pin uses, and <ALEE> is set to “1”. When Multi chip mode was <ALEE> is always “1”, and P83 becomes the ALE output. Figure 3.5 (27). Port 8 (P83) TOSHIBA CORPORATION 63 TMP90CM38 Figure 3.5 (28). Registers for Port 8 64 TOSHIBA CORPORATION TMP90CM38 3.5.11 Port 10 (P100, 101) All bits of the output latch are initialized to “0” by resetting, and Port 10 turns to the input mode. Port 10 is a 2-bit general-purpose I/O port. It is specified by the control register P10CR in bit basis. Figure 3.5 (29). Port 10 (P100, 101) TOSHIBA CORPORATION 65 TMP90CM38 Figure 3.5 (30). Register for Port 10 66 TOSHIBA CORPORATION TMP90CM38 3.6 Timers The TMP90CM38 contains four 8-bit timers (timers 0, 1, 2 and 3), each of which can be operated independently. The cascade connection allows these timers to be used as 16-bit timers. The following four operating modes are provided for the 8bit timers. • 8-bit interval timer mode (4 timers) • 16-bit interval timer mode (2 timers) • 8-bit programmable square wave pulse generation (PPG: variable duty with variable cycle) output mode (2 timers) • 8-bit pulse width modulation (PWM: variable duty with constant cycle) output mode (2 timers one 16-bit timer). Figure 3.6 (1) shows the block diagram of 8-bit timer (timer 0 and timer 1). 8-bit timer (timer 2, 3) are connected to the external clock pin TI2 in the timer 2 up counter input clock. Other timer 2 and timer 3 have the same circuit configuration as timer 0 and timer 1. Each interval consists of an 8-bit up-counter, 8-bit comparator, and 8-bit timer register. Besides, one timer flip-flop (TFF1 or TFF3) is provided for each pair of timer 0 and timer 1 as well as timer 2 and timer 3. Among the input clock sources for the interval timers, the internal clocks of øT1, øT4, øT16, and øT256 are obtained from the 9-bit prescaler shown in Figure 3.6 (2). The operation modes and timer flip-flops of the 8-bit timer are controlled by five control registers T01MOD, T23MOD, TFFCR, TRUN, and TRDC. The upper two can be combined (two 8-bit timers and TOSHIBA CORPORATION 67 TMP90CM38 Figure 3.6 (1). Block Diagram of 8-bit Timers (Timers 0 and 1) 68 TOSHIBA CORPORATION TMP90CM38 ➀ Prescaler This 9-bit prescaler generates the clock input to the 8-bit timers, 16-bit timer/event counters, and baud rate generators by further dividing the fundamental clock (fc) after it has been divided by 4 (fc/4). Among them, 8-bit timer uses 4 types of clock: øT1, øT4, øT16, and øT256. This prescaler can be run or stopped by the timer operation control register TRUN <PRRUN>. Counting starts when <PRRUN> is set to “1”, while the prescaler is cleared to zero and stops operation when <PRRUN> is set to “0”. Resetting clears <PRRUN> to “0”, which clears and stops the prescaler. Figure 3.6 (2). Prescaler TOSHIBA CORPORATION 69 TMP90CM38 ➁ Up-counter This is an 8-bit binary counter that counts up the input clock pulse specified by the timer 0/timer 1 mode register T01MOD and timer 2/timer 3 mode register T23MOD. The input clock pulse for timer 0 is selected from øT1 (8/fc), øT4 (32/fc) and øT16 (128/fc). Timer 2 input clock is selected from external clock (TI2 pin = P55/ INT3) and same the timer 0 in three kinds internal clock. According to the set value of T01MOD and T23MOD. The input clock of timer 1 and timer 3 differs depending on the operating mode. When set to 16-bit timer mode, the overflow output of timer 0 and timer 2 is used as the input clock. When set to any other mode than 16-bit timer mode, the input clock is selected from the internal clocks øT1 (8/fc), øT16 (128/fc), and øT256 (2048/fc) as well as the comparator output (match detection signal) of timer 0 and timer 2, according to the set value of T01MOD and T23MOD. Example:When TMOD <T01M1,0> = 01, the overflow output of timer 0 becomes the input clock of timer 1 (16-bit timer). When TMOD <T01M1,0> = 00 and T01MOD <T1CLK1,0> = 01, øT1 (8/fc) becomes the input of timer 1. Operation mode is also set by T01MOD and T23MOD. When reset, it is initialized to The counting, halt, and clear of up-counter can be controlled for each interval timer by the timer operation control register TRUN. When reset, all up-counters will be cleared to stop the timers. ➂ Timer registers This is an 8-bit register for setting an interval time. When the set value of timer register TREG0, TREG1, TREG2, and TREG3 matches the value of up-counter, the comparator match detect signal becomes active. If the set value is 00H, this signal becomes active when the up-counter overflows. Timer registers TREG0 and TREG2 are of double buffer structure, each of which makes a pair with register buffer. The TREG0 and TREG2 control whether the double buffer should be enabled or disabled through the timer register double buffer control register TRDC <TR0DE, TR2DE>. It is disabled when <TR0DE>/<TR2DE> = 0, and enabled when they are set to 1. The timing to transfer data from the register buffer to the timer register in the double buffer enable state is the moment 2n - 1 overflow occurs in PWM mode or the moment compare cycles will be equal in PPG mode. When reset, it will be initialized to <TR0DE, TR2DE> = 0 to disable the double buffer. To use the double buffer, write data in the timer register, set <TR0DE> and <TR2DE> to 1, and write the following data in the register buffer. T01MOD <T01M1, 0> = 00 and T23MOD <T23M1, 0> = 00, whereby the up-counter is placed in the 8-bit timer mode. 70 TOSHIBA CORPORATION TMP90CM38 Figure 3.6 (3). Configuation of Timer Registers 0 and 2 Note: Timer register and the register buffer are allocated to the same memory address. When <TR0DE>/<TR2DE> = 0, the same value written in the register buffer as well as the timer register, while when <TR0DE>/<TR2DE> = 1 only the register buffer is written. The memory address of each timer register is as follows. TREG0: FFC6H TREG1: FFC7H TREG2: FFC8H TREG3: FFC9H All the registers are write-only and cannot be read. TOSHIBA CORPORATION 71 TMP90CM38 Figure 3.6 (4). Timer 0/Timer 1 Mode Register (T01MOD) 72 TOSHIBA CORPORATION TMP90CM38 Figure 3.6 (5). Timer 2/Timer 3 Mode Register (T23MOD) TOSHIBA CORPORATION 73 TMP90CM38 Figure 3.6 (6). 8-Bit Timer Flip-flop Control Register (TFFCR) 74 TOSHIBA CORPORATION TMP90CM38 Figure 3.6 (7). Timer Operation Control Register (TRUN) TOSHIBA CORPORATION 75 TMP90CM38 Figure 3.6 (8). Timer Register Double Buffer Control Register (TRDC) 76 TOSHIBA CORPORATION TMP90CM38 ➃ Comparator The operation of 8-bit timers will be described below: A comparator compares the value in the up-counter with the values to which the timer register is set. When they match, the up-counter is cleared to zero and an interrupt signal (INTT0 ~ INTT3) is generated. If the timer flip-flop inversion is enabled, the timer flip-flop is inverted at the same time. (1) 8-bit timer mode Four interval timers 0, 1, 2, and 3 can be used independently as an 8-bit interval timers. All interval timers operate in the same manner, thus only the operation of timer 1 will be explained below. ➄ Timer flip-flop (timer F/F) ➀ Generating interrupts in a fixed cycle The status of the timer flip-flop is inverted by the match detect signal (comparator output) of each interval timer and the value can be output to the timer output pins TO1 (also used as P55) and TO3 (also used as P57). A timer F/F is provided for each pair of timer 0 and timer 1 as well as that of timer 2 and timer 3 and is called TFF1 and TFF3. TFF1 is output to TO1 pin, while TFF3 is output to TO3 pin. To generate timer 1 interrupt at constant intervals using using timer 1 (INTT1), first stop timer 1 then set the operation mode, input clock, and synchronization to T01MOD and TREG1, respectively. Then, enable interrupt INTT1 and start the counting of timer 1. Example: To generate timer 1 interrupt every 40 microseconds at fc = 16MHz, set each register in the following manner. Use the following table for selecting the input clock: Table 3.6 (1) 8-bit Timer Interrupt Cycle and Input Clock TOSHIBA CORPORATION Interrupt cycle (at fc = 16MHz) Resolution 0.5µs ~ 128ms 0.5µs øT1 (8/fc) 2µs ~ 512ms 2µs øT16 (32/fc) 8µs ~ 2.048ms 8µs øT256 (128/fc) 128µs ~ 32.768ms 128µs øT256 (2048/fc) Input clock 77 TMP90CM38 ➁ Generating a 50% duty square wave pulse The timer flip-flop is inverted at constant intervals, and its status is output to timer output pin (TO1). Example: To output a 3.0µs square wave pulse from TO1 pin at fc = 16MHz, set each register in the following procedures. Either timer 0 or timer 1 may be used, but this example uses timer 1. Figure 3.6 (9). Square Wave (50% Duty) Output Timing Chart 78 TOSHIBA CORPORATION TMP90CM38 ➂ Making timer 1 count up by match signal from timer 0 comparator Set the 8-bit timer mode, and set the comparator output of timer 0 as the input clock to timer 1. Figure 3.6 (10) ➃ Output inversion with software The value of timer flip-flop (timer F/F) can be inverted, independent of the timer operation. Writing “00” to TFFCR <FF1C1, 0> inverts the value of TFF1, and writing “00” into TFFCR <FF3C1,0> inverts TFF3. ➄ Initial setting of timer flip-flop (timer F/F) The value of timer F/F can be initialized to “0” or “1”, independent of timer operation. For example, write “10” in TFFCR <FF1C1,0> to clear TFF1 to “0”, while write “01” in TFFCR <FF3C1,0> to set TFF1 to “1”. (2) 16-bit timer mode A 16-bit interval timer is configured by using the pair of timer 0 and timer 1 or that of timer 2 and timer 3. As the above two pairs operate in the same manner, only the case of combining timer 0/timer 1 is discussed. To make a a 16-bit interval timer by cascade connecting timer 0 and timer 1, set timer 0/timer 1 mode register T01MOD <T01M1,0> to “0,1”. When set in 16-bit timer mode, the overflow output of timer 0 will become the input clock of timer 1, regardless of the set value of T01MOD <T1CLK1,0>. Table 3.6 (2) shows the relation between the cycle of timer (interrupt) and the selection of input clock. Note: The value of timer F/F and timer register cannot be read. TOSHIBA CORPORATION 79 TMP90CM38 Table 3.6 (2) 16-bit Timer (Interrupt) Cycle and Input Clock Interrupt cycle (at fc = 16MHz) Resolution 0.5µs ~ 32.768ms 0.5µs øT1 (8/fc) 2µs ~ 131.072ms 2µs øT16 (32/fc) 8µs ~ 524.288ms 8µs øT256 (128/fc) The lower 8 bits of the timer (interrupt) cycle are set by the timer register TREG0, and the upper 8 bits are set by TREG1. Note that TREG0 always must be set first. (Writing data into TREG0 disables the comparator temporarily, and the comparator is restarted by writing data into TREG1.) Setting example: To generate interrupt INTT1 every 0.5 seconds at fc = 16MHz, set the following values for timer registers TREG0 and TREG1. When counting by using øT16 (8µs @16MHz), 0.5s ÷ 8µs = 62500 = F424H Therefore, set TREG1 = F4H and TREG0 = 24H, respectively. 80 Input clock The comparator match signal is output from timer 0 each time the up-counter matches UC0, where the up-counter UC0 is not cleared. With the timer 1 comparator, the match detect signal is output at each comparator timing when up-counter UC1 and TREG1 values match. When the match detect signal is output simultaneously from both comparators of timer 0 and timer 1, the up-counters UC0 and UC1 are cleared to “0”, and the interrupt INTT1 is generated. If inversion is enabled, the value of the timer flip-flop TFF1 is inverted. TOSHIBA CORPORATION TMP90CM38 Example: When TREG1 = 04H and TREG0 = 80H Figure 3.6 (11) (3) 8-bit PPG (Programmable Pulse Generation) Mode Square wave pulse can be generated at any frequency and duty by timer 0 or timer 2. The output pulse may As an example, the case of timer 0 will be explained TOSHIBA CORPORATION be either low-active or high-active. In this mode, timer 1 and timer 3 cannot be used. Timer 0 outputs pulse to TO1 pin, (also used as P55), and timer 2 outputs to TO3 pin (also used as P57). below. (Timer 2 also functions in the same way). 81 TMP90CM38 In this mode, a programmable square wave is generated by inverting timer output each time the 8-bit upcounter (UC0) matches the timer registers TREG0 and TREG1. However, it is required that the set value of TREG0 is smaller than that of TREG1. Though the up-counter (UC1) of timer 1 cannot be used in this mode, timer 1 can be used for counting by setting TRUN <T1RUN> to “1”. Figure 3.6 (12) shows the block diagram for this mode. Figure 3.6 (12). Block Diagram of 8-bit PPG Mode When the double buffer of TREG0 is enabled in this mode, the value of register buffer will be shifted in TREG0 each time TREG1 matches UC0. 82 Use of the double buffer makes easy the handling of low duty waves (when duty is varied). TOSHIBA CORPORATION TMP90CM38 Example: Generating 1/4 duty 50KHz pulse (@fc = 16MHz) • Calculate the value to be set for timer register TOSHIBA CORPORATION To obtain the frequency 50KHz, the pulse cycle t should be: 1/50KHz = 20µs. Given øT1 = 0.5µs (@ 16MHz) 20µs ÷ 0.5µs = 40 Consequently, to set the timer register 1 (TREG1) to TREG1 = 40 = 28H and the duty to 1/4, t x 1/4 = 20µs x 1/4 5µs 5µs ÷ 0.5µs = 10 Therefore, set timer register 0 (TREG0) to TREG0 = 10 = 0AH. 83 TMP90CM38 (4) 8-bit PWM (Pulse Width Modulation) Mode This mode is valid only for timer 0 and timer 2. In this mode, maximum two PWMs of 8-bit resolution (PWM0 and PWM2) can be output. PWM pulse is output to TO1 pin (also used as P55) when using timer 0, and to TO3 pin (also used as P57) when using timer 2. Timer 1 and timer 3 can also be used as 8-bit timer. As an example, the case of timer 0 will be explained below. (Timer 2 also operates in the same way.) Timer output is inverted when up-counter (UC0) 84 matches the set value of timer register TREG0 or when 2n - (n = 6, 7, or 8; specified by T01MOD <PWM0,10>) counter overflow occurs. Up-counter UC0 is cleared when 2n - 1 counter overflow occurs. For example, when n = 6, 6-bit PWM will be output, while when n = 7, 7-bit PWM will be output. To use this PWM mode, the following conditions must be satisfied. (Set value of timer register) < (Set value of 2n - 1 counter overflow) (Set value of timer register) ≠ 0 TOSHIBA CORPORATION TMP90CM38 Figure 3.6 (13) shows the block diagram of this mode. Figure 3.6 (13). Block Diagram of 8-bit PWM Mode TOSHIBA CORPORATION 85 TMP90CM38 In this mode, the value of register buffer will be shifted in TREG0 2n - 1 overflow is detected when the double buffer of TREG0 is enabled. 86 Use of the double buffer makes easy the handling of small duty waves. TOSHIBA CORPORATION TMP90CM38 Table 3.6 (3) PWM Cycle and the Setting of 2n - 1 Counter PWM cycle (@ fc = 16MHz) Formula øT4 (32/fc) øT16 (128/fc) 26 - 1 (26 - 1) x øTn 31.5µs 126µs 504µs 7-1 (27 - 1) x øTn 63.5µs 254µs 1.01ms 8-1 (28 - 1) x øTn 127µs 510µs 2.04ms 2 2 (5) øT1 (8/fc) Table 3.6 (4) shows the list of 8-bit timer modes. Table 3.6 (4) Timer Mode Setting Registers Register name T01MOD (T23MOD) TFFCR Name of bit in register T01M (T32M) PWM1 (PWM3) T1CLK (T3CLK) T0CLK (T2CLK) FF1IS (FF3IS) Function Timer Mode PWM cycle Upper timer input clock Lower timer Input clock Timer F/F invert signal select 01 – – External clock øT1, øT16, øT256 (01, 10, 11) – External clock øT1, øT16, øT256 (01, 01, 10, 11) 0: Lower timer output 1: Upper timer output 16-bit timer mode 8-bit timer x 2 channels 00 – Lower timer match øT1, øT16, øT256 (01, 01, 10, 11) 8-bit PPG x 1 channel 10 – – External clock øT1, øT16, øT256 (01, 01, 10, 11) – 8-bit PWM x 1 channel 11 26 - 1, 27 - 1, 28 - 1 (01, 10, 11) – External clock øT1, øT16, øT256 (01, 01, 10, 11) – 8-bit timer x 1 channel 11 – øT1, øT14, øT16 (01, 10, 11) – Output disabled (Note) –: Don’t care TOSHIBA CORPORATION • Lower timer external input clock has T2CLK. But it does not have T0CLK. 87 TMP90CM38 3.7 Multi-Function 16-bit Timer/Event Counter (Timer 4) The TMP90CM38 contains one multifunctional 16-bit timer/ event counter with the following operation modes: • Frequency measurement • Pulse width measurement • Time differential measurement • 16-bit timer • 16-bit event counter • 16-bit programmable pulse generation (PPG) Figure 3.7 (1) shows the block diagram of 16-bit timer/ event counter. 88 TOSHIBA CORPORATION TMP90CM38 Figure 3.7 (1). Block Diagram of 16-Bit Timer/Event Counter (Timer 4) TOSHIBA CORPORATION 89 TMP90CM38 Timer/event counter con sists of 16-bit up-counter, two 16bit timer registers, two 16-bit capture registers, two comparators, register buffer, capture input controller, a timer flip-flop and the control circuit. Timer/event counter is controlled by 4 control registers: T4MOD, T4FFCR, TRUN, and TRDC. TRUN register includes 8-bit timer controller. For TRUN and TRDC registers, see Figure 3.6 (7) and Figure 3.6 (8). Figure 3.7 (2). 16-Bit Timer/Event Counter (Timer 4) Control/Mode Register (1/2) 90 TOSHIBA CORPORATION TMP90CM38 Figure 3.7 (2). 16-Bit Timer/Event Counter (Timer 4) Control/Mode Register (2/2) TOSHIBA CORPORATION 91 TMP90CM38 Figure 3.7 (3). 16-Bit Timer/Event Counter Timer Flip-flop Control Register 92 TOSHIBA CORPORATION TMP90CM38 ➀ Up-counter (UC16) UC16 is a 16-bit binary counter which counts up according to the input clock specified by T4MOD <T4CLK1,0> register. As the input clock, one of the internal clocks øT1 (8/ fc), øT4 (32/fc), and øTI6 (128/fc) from 9-bit prescaler (also used as 8-bit timer), and external clock from TI4 pin (commonly used as P46/INT1 pin) can be selected. When reset, it will be initialized to <T4CLK1,0> = 00 to select TI4 input mode. Counting, stop, or clearing of the counter in controlled by timer operation control register TRUN <T4RUN>. When clearing is enabled, up-counter UC16 will be cleared to zero each time it coincides matches the timer register TREG5. The “clear enable/disable” is set TREG4 timer register is of double buffer structure, which is paired with register buffer. TREG4 controls whether the double buffer should be enabled or disabled, using the timer register double buffer control register TRDC <TR4DE>; disable when <TR4DE> = 0, while enable when <TR4DE> = 1. When the double buffer is enabled, the timing to transfer data from the register buffer to the timer register is at the match between the up-counter and TREG5. When reset, it will be initialized to <TR4DE> = 0, whereby the double buffer is disabled. To use the double buffer, write data in the register buffer. TREG4 and register buffer 4 are allocated to the same TOSHIBA CORPORATION by T4MOD <CLE>. If clearing is disabled, the counter operates as a freerunning counter. ➁ Timer registers (TREG4 and TREG5) These two 16-bit registers are used to set the value of the counter. When the value of up-counter UC16 matches the set value of this timer register, the comparator match detect signal will become active. Setting data for timer register (TREG4 and TREG5) is executed using 16-bit transfer instruction od using 8bit transfer instruction twice for lower 8 bits and upper 8 bits in order. memory addresses FFCFH and FFD0H. When <TR4DE> = 1, the value is written into only the register buffer. ➂ Capture register (CAP1 and CAP2) These 16-bit registers are used to hold the values of the up-counter UC16. Data in the capture registers should be read by a 2byte data load instruction or two 1-byte data load instruction, from the lower 8 bits followed by the upper 8 bits. 93 TMP90CM38 ➃ Capture input control circuit ➄ Comparators (CP4, CP5) This circuit controls the timing to latch the value of upcounter UC16 into CAP1 and CAP2. The latch timing of capture register is controlled by register T4MOD <CAPM1,0>. These are 16-bit comparators which compare the upcounter UC16 value with the set value of TREG4 or TREG5 to detect the match. When a match is detected, the comparators generate an interrupt INTT4 and INTT5, respectively. The up-counter UC16 is cleared only when UC16 matches TREG5. (The clearing of up-counter UC16 can be disabled by setting T4MOD <CLE> = 0.) • When T4MOD <CAPM1, 0> = 00 Capture function is disabled. Disable is the default on reset. • When T4MOD <CAPM1, 0> = 01 Data is loaded to CAP1 at the rise edge of TI4 pin (commonly used as P53/INT1) input, while data is loaded to CAP2 at the rise edge of TI5 pin (commonly used as P47/INT2) input. (Time difference measurement) • When T4MOD <CAPM1, 0> = 10 Data is loaded to CAP1 at the rise edge of TI4 pin input, while to CAP2 at the fall edge. Only in this setting, interrupt INT1 occurs at fall edge. (Pulse width measurement) • When T4MOD <CAP1, 0> = 11 Data is loaded to CAP1 at the rise edge of timer flipflop TFF1, while to CAP2 at the fall edge. (Frequency measurement) Besides, the value of up-counter can be loaded to capture registers by software. Whenever “0” is written in T4MOD <CAPIN>, the current value of up-counter will be loaded to capture register CAP1. It is necessary to keep the prescaler in RUN mode (TRUN <PRRUN> to be “1”). 94 ➅ Timer flip-flop (TFF4) This flip-flop is inverted by the match detect signal from the comparators (CP4 and CP5) and the latch signals to the capture registers (CAP1 and CAP2). Disable/enable of inversion can be set for each element by T4FFCR <CAP2T4, CAP1T4, EQ5T4, EQ4T4>. TFF4 will be inverted when “00” is written in T4FFCR <TFF4C1,0>. Also, it is set to “1” when “10” is written, and cleared to “0” when “10” is written. The value of TFF4 can be output to the timer output pin TO4 (commonly used as P51). ➆ Timer flip-flop (TFF5) This flip-flop is inverted by the match detect signal from the comparator CP5 and the latch signal to the capture register CAP2. TFF5 will be inverted when “00” is written in T4FFCR <TFF5C1,0>. Also, it is set to “1” when “10” is written, and cleared to “0” when “10” is written. The value of TFF5 can be output to the timer output pin TO5 (commonly used as P52). TOSHIBA CORPORATION TMP90CM38 (1) 16-bit Timer Mode ister TREG5 to generate the interrupt INTT5. In this example, the interval time is set in the timer reg- (2) 16-bit Event Counter Mode In timer mode as described in above (1), the timer can be used as an event counter by selecting the external clock (TI4 pin input) as the input clock. To read the value of TOSHIBA CORPORATION the counter, first perform “software capture” once and read the captured value. The counter counts at the rising edge of TI4 pin input. TI4 pin can also be used as P53/INT1. 95 TMP90CM38 (3) 16-bit Programmable Pulse Generation (PPG) Mode The PPG mode is entered by inversion of the timer flipflop TFF4 that is to be enabled bu the match of the up- When the double buffer of TREG4 is enabled in this mode, the value of register buffer 4 will be shifted in TREG4 at match with TREG5. This feature makes easy the handling of low duty waves (when duty rate is varied). (4) Application examples of capture function The loading of up-counter (UC16) vaules into the capture registers CAP1 and CAP2, the timer flip-flop TFF4 inversion due to the match detection by comparators CP4 and CP5, and the output of the TFF4 status to TO4 pin can be enabled or disabled. Combined with interrupt function, they can be applied in many ways, for example. ➀ One-shot pulse output by using external trigger pulse ➁ Frequency measurement ➂ Pulse width measurement ➃ Time difference measurement 96 counter UC16 with the timer register TREG4 or 5 and to be output to TO4 (also used as P51). In this mode, the following conditions must be satisfied. ➀ One-shot pulse output from the rising edge of external trigger pulse. Set the up-counter UC16 in free-running mode with the internal input clock, input the external trigger pulse from TI4 pin, and load the value of up-counter into capture register CAP1 at the rise edge of the TI4 pin. Then set to T4MOD <CAPM1,0> = 01. When the interrupt INT1 is generated at the rise edge of TI4 pin, set the CAP1 value (c) plus a delay time (d) to TREG4 (= c + d), and set the above set value (c + d) plus a one-shot pulse width (p) to TREG5 (= c + + d + p). When the interrupt INT1 occurs the T4FFCR register should be set that the TFF4 inversion is enabled only when the up-counter value matches TREG4 or 5. When interrupt INTT5 occurs, this inversion will be disabled. TOSHIBA CORPORATION TMP90CM38 Figure 3.7 (4). One-Shot Pulse Output (with Delay) TOSHIBA CORPORATION 97 TMP90CM38 When delay time is unnecessary, invert timer flip-flop TFF4 when the up-counter value is loaded into loaded into capture register 1 (CAP1), and set the CAP1 value (c) plus the one-shot pulse width (p) to TREG5 when the interrupt INT1 occurs. The TFF4 inversion should be enabled when the up-counter (UC16) value matches TREG5, and disabled when generating the interrupt INTT5. Figure 3.7 (5). One-Shot Pulse Output (without Delay) 98 TOSHIBA CORPORATION TMP90CM38 ➁ Frequency measurement The frequency of the external clock can be measured in this mode. The clock is input through the TI4 pin, and its frequency is measured by the 8-bit timers (Timer 0 and Timer 1) and the 16-bit timer/event counter (Timer 4). The TI4 pin input should be selected for the input clock of Timer 4. The value of the up-counter is loaded into the capture register CAP1 at the rise edge of the timer flip-flop TFF1 of 8-bit timers (Timer 0 and Timer 1), and CAP2 at its fall edge. The frequency is is calculated by the difference between the loaded values in CAP1 and CAP2 when the interrupt (INTT0 or INTT1) is generated by either 8bit timer. Figure 3.7 (6). Frequency Measurement For example, if the value for the level “1” width of TFF1 of the 8-bit timer is set to 0.5 sec. and the difference TOSHIBA CORPORATION between CAP1 and CAP2 is 100, the frequency will be 100/0.5 [sec.] = 200[Hz]. 99 TMP90CM38 ➂ Pulse width measurement This mode allows to measure the “H” level width of an external pulse. While keeping the 16-bit timer/event counter counting (free-running) with the internal clock input, the external pulse is input though the TI4. Then the capture function is used to load the UC16 values into CAP1 and CAP2 at the rising edge and falling edge of the external trigger pulse, respectively. The interrupt INT1 occurs at the falling edge of TI4. The pulse width is obtained from the difference between the values of CAP1 and CAP2 and the internal clock cycle. For example, if the internal clock is 0.8 microseconds and the difference between CAP1 and CAP2 is 100, the pulse width will be 100 x 0.8 = 80 microseconds. Figure 3.7 (7). Pulse Width Measurement Note: 100 Only in this pulse width measuring mode (T4MOD <CAPM1, 0> = 10), external interrupt INT1 occurs at the falling edge of TI4 pin input. In other modes, it occurs at the rising edge. The width of “L” level can be measured from the difference between the first C2 and the second C1 at the second INT1 interrupt. TOSHIBA CORPORATION TMP90CM38 ➃ Time difference measurement This mode is used to measure the difference in time between the rising edges of external pulses input through TI4 and TI5. Keep the 16-bit timer/event counter (Timer 4) counting (free-running) with the internal clock, and load the UC16 value into CAP1 at the rising edge of the input pulse to TI4. Then the interrupt INT1 is generated. Similarly, the UC16 value is loaded into CAP2 at the rising edge of the input pulse to TI5, generating the interrupt INT2. The time difference between these pulses can be obtained from the difference between the time counts at which loading the up-counter value into CAP1 and CAP2 has been done. Figure 3.7 (8). Time Difference Measurement TOSHIBA CORPORATION 101 TMP90CM38 3.8 Stepping Motor Control/Pattern Generation Ports (P2 and P9) TMP90CM38 contains 2 channels (M0 and M1) of 4-bit hardware stepping motor control/pattern generation ports (herein after called SMC) which actuate in synchronization with the (8bit/16-bit) timers. The SMC’s ports (M0 and M1) are shared by 4-bit I/O ports P2 and P9. Channel 0 (M0) is synchronous with 8-bit timer 0 or timer 1, and channel 1 (M1) is synchronous with 8-bit timer 2 or timer 3 or 16-bit timer 4, to update the output. The SMC ports are controlled by three control registers P29CR, P29FR, and TRDC, and can select either stepping motor control mode or pattern generation mode. write mode or 4-bit write mode. In 4-bit write mode, writing the SMC is executed only on the 4-bit shift alternate register, and SMC functions as a pattern generation port. To use SMC as a stepping motor control port, select the method of excitation and the method to control the direction of rotation by P29FR <M0M>/<M1M> and P29FR <CCW0>/<CCW1>, respectively. (3) With TRDC <M1T>, select a time which synchronizes SMC channel 1 (M1). When <M1T> = 0, M1 is synchronous with timer 2 or timer 3, while when <M1T> = 1 it is synchronous with timer 4. 3.8.1 Control Registers (1) Ports 2 and 9 I/O Selection Register (P29CR) (4) This register specifies either input or output for each bit of the 4-bit I/O ports 6 and 7. When reset, all bits of P29CR are cleared to “0”, so that port 2 and port 9 function as input ports. To use port 6 and port 7 as SMC, set all bits of P29CR to “1”, specifying them as output pins. P29CR is a write-only register and so cannot be read. Port 9 Port 2 and 9 function control register (P29FR) This register is used for setting port 2 and port 9 as SMC. To use port 2 and port 9 as SMC, set P29FR <M0S>/<M1S> to “1”. With P29 <PAT0>/<PAT1>, set SMC in either 8-bit 102 Port 2 This is a 4-bit I/O port allocated to address FFA4. The lower 4 bits are assigned as port 6, while the upper 4 bits function as the shifter alternate register SA6 which is used in pattern generation mode or to drive the stepping motor by 1-2 excitation. (5) (2) Selecting the channel 1 synchronizing time (TRDC) This is a 4-bit I/O port allocated to address FFB3. The lower 4 bits are assigned as port 7, while the upper 4 bits function as the shifter alternate register SA7 which is used in pattern generation mode or to drive the stepping motor by 1-2 excitation. TOSHIBA CORPORATION TMP90CM38 Figure 3.8 (1). Port 2 and Port 9 I/O Selection Register (P29CR) TOSHIBA CORPORATION 103 TMP90CM38 Figure 3.8 (2a). Port 2 and Port 9 Function Control Register (P29FR) 104 TOSHIBA CORPORATION TMP90CM38 Figure 3.8 (2b). Port 2 and 9 Function Control Register (P29FR) TOSHIBA CORPORATION 105 TMP90CM38 Figure 3.8 (3). Timer Register Double Buffer Control Register (TRDC) Figure 3.8 (4). Port 2 and Port 9 106 TOSHIBA CORPORATION TMP90CM38 3.8.2 Pattern Generation Mode SMC functions as a pattern generation port according to the setting of P29FR <PAT1>/<PAT0>. In this mode, becuase writing from CPU is executed only on the shifter alternate register, writing ports 2 and 9 can be done during the interrupt opera- tion of the timer for shift trigger and a pattern can be output, synchronous with the timer . In this mode, P29FR <M1M>/<M0M> must be always set to “1”. Figure 3.8 (5) shows the block diagram of this mode. Figure 3.8 (5). Pattern Generation Mode Block Diagram (Port 2) In this pattern generation mode, only writing the output latch is disabled by hardware, but other functions do the same operation as 1-2 excitation in stepping motor control port TOSHIBA CORPORATION mode. Accordingly, the data shifted by trigger signal from a timer must be written before the next trigger signal is output. 107 TMP90CM38 3.8.3 Stepping Motor Control Mode (1) 4-Phase 1-Step/2-Step Excitation Figure 3.8 (6) and Figure 3.8 (7) show the output waveforms of 4-phase 1 excitation and 4-phase 2 excitation, respectively when channel 0 is selected. Figure 3.8 (6). Output Waveforms of 4-Phase 1-step Excitation (Normal Rotation and Reverse Rotation) 108 TOSHIBA CORPORATION TMP90CM38 Figure 3.8 (7). Output Waveforms of 4-Phase 2-step Excitation (Normal Rotation) The operation when channel 0 is selected is explained below. The output latch of M0 (also used as P2) rotates at the rising edge of the TFF1 trigger pulse (that inverts the value of TFF1) and is output to the port. The direction of shift is specified by P29FR <CCW0>: Normal rotation (M00→M01→M02→M03), when <CCW0> is set to “0”; reverse rotation (M00←M01←M02←M03) when “1”. 4-phase 1-step excitation can be selected when only one bit is set to “1” during the initialization of port 6, while 4-phase 2-step excitation will be selected when two consecutive bits are set to “1”. Figure 3.8 (8). shows the block diagram. Figure 3.8 (8). Block Diagram of 4-Phase 1-step/2-step Excitation (Normal Rotation) TOSHIBA CORPORATION 109 TMP90CM38 (2) 4 Phase 1-2 Step Excitation 1-2 step excitation when channel 0 is selected. Figures 3.8 (9) shows the output waveforms of 4-phase Figure 3.8 (9). Output Waveforms of 4-Phase 1-2 step Excitation (Normal Rotation and Reverse Rotation) The initialization for 4-phase 1-2 step excitation is as follows. By rearranging the initial value “b7 b6 b5 b4 b3 b2 b1 b0” to “b3 b7 b2 b6 b1 b5 b0 b4”, the consecutive 3 bits are set to “1” and other bits are set to “0” (positive logic). For example, if b3, b7 and b2 are set to “1” provided the initial value becomes “10001100”, obtaining the waveforms as shown in Figure 3.8 (9). To get an output waveform of negative logic, set values 1’s and 0’s of the initial value should be inverted. For example, 110 to change the output waveform shown in Figure 3.8 (9) into negative logic, change the initial value to “011110011”. The operation will be explained below for channel 0. The output latch of M0 (shared by P2) and the shifter alternate register (SA6) for stepping motor control are shifted at the rising edge of trigger signal from the timer to be output to the port. The direction of shift is set by P29FR <CCW0>. Figure 3.8 (10) shows the block diagram. TOSHIBA CORPORATION TMP90CM38 Figure 3.8 (10). Block Diagram of 4-Phase 1-2 Step Excitation (Normal Rotation) TOSHIBA CORPORATION 111 TMP90CM38 Setting Example: To drive channel 0 (M0) by 4-phase 1-2 step excitation (normal rotation) when 3.8.4 Trigger Signal From Timer The trigger signal from the timer which is used by SMC is not equal to the reverse trigger signal of each timer flip-flop (TFF1, timer 0 is selected, set each register as follows. TFF3, TFF4, and TFF5) and differs as shown in Table 3.8 (1) depending on the operation mode of the timer. Table 3.8 (1) 8-Bit Timers 0 and 1 (Same for timers 2 and 3) TFF1 inversion SMC shift 8-bit timer mode Selected by TFFCR0 (FF1IS) when the up-counter value matches TREG0 or TREG1 value ← 16-bit timer mode When the up-counter value matches with both TREG0 and TREG1 values (The value of up-counter = TREG1*28 + TREG0) ← PPG output mode When the up-counter value matches with both TREG0 and TREG1 When the up counter value matches TREG1 value (PPG cycle) PWM output mode When the up-counter value matches TREG0 value at PWM cycle. Trigger signal for SMC shift is not output. Note: To shift SMC, TFFCR <FF1IE> must be set to “1” to enable TFF1 inversion. Channel 1 of SMC can by synchronized with the 16-bit timer. In this case, the SMC shift register trigger signal from the 112 16-bit timer is output only when the up-counter value matches TREG5. Set either T4FFCR <EQ5T4> or T4MOD <EQ5T5> to “1”. TOSHIBA CORPORATION TMP90CM38 3.8.5 Application of SMC and Timer Output As explained in 3.8.4 “Trigger signal from timer”, the timing to shift SMC and invert TFF differs depending on the mode of timer. An application to operate SMC while operating an 8-bit timer PPG mode will be explained below. To drive a stepping motor, in addition to the value of each phase (SMC output), synchronizing signal is often required at the timing when excitation is changed over. In this application, noting this fact, port 6 is used as a stepping motor control port to output a synchronizing signal to the TO1 pin (shared by P55). Output Waveforms of 4-Phase 1-Step Excitation TOSHIBA CORPORATION 113 TMP90CM38 3.9 Serial Channels The TMP90CM36 contains three serial channels (SIO0,1, 2).The three serial channels have the following operation modes. In mode 1 and mode 2, parity bit can be added. Mode 3 has a wake-up function for making the master controller start slave controllers in serial link (multi-controller system). Figure 3.9 (1) shows the data format (1 frame) for each mode. Figure 3.9 (1). Data Formats 114 TOSHIBA CORPORATION TMP90CM38 The serial channel has a buffer register for transmitting and receiving operations, in order to temporarily store transmitted or received data, so that transmitting and receiving operations can be done independently (full duplex). However, in I/O interface mode, SCLK (serial clock) pin is commonly used for both transmission and receiving, the channel becomes half-duplex. The receiving buffer register is of a double buffer structure to prevent the occurrence of overrun error and provides one frame of margin before CPU reads the received data. Namely, the one buffer stores the already received data while the other buffer receives the next frame data. In the UART mode, a check function is added not to start the receiving operation by error start bits due to noise. The channel starts receiving data only when the start bit is detected to be normal at least twice in three samplings. TOSHIBA CORPORATION When the transmission buffer becomes empty and requests the CPU to send the next transmission data, or when data is stored in the transmission buffer and the CPU is requested to read the data, INTTX or INTRX interrupt occurs. Besides, if an overrun error, parity error, or framing error occors during receiving operation, flag SCCR <OERR, PERR, FERR> will be set. In the I/O interface mode, it is possible to input synchronous signals as well as to transmit or receive data by an external clock. The SIO0 or SIO1 includes a special baud rate generator, which can set any baud rate can be set by dividing by the frequency of 4 clocks (øT0, øT2, øT8, and øT32) from the internal prescaler (shared by 8-bit/16-bit timer) by the value of 2 to 16. Internal clock (SIO2) is able to select in speed from øT0, øT1, øT4, and øT16. 115 TMP90CM38 (1) Serial Channel (SIO0) (3.8.4 ~ 3.8.6) SCMOD1, SCCR1, BRGCR1, and P7FR. Transmitted and received data are stored in register SCBUF1. 3.9.1 Control Registers The serial channel SIO1 is controlled by 4 control registers Figure 3.9 (2). Serial Channel Mode Register (SCMOD1) 116 TOSHIBA CORPORATION TMP90CM38 Figure 3.9 (3). Serial Channel Mode Register (SCCR1) Figure 3.9 (4). Serial Transmission/Receiving Buffer Registers (SCBUF1) TOSHIBA CORPORATION 117 TMP90CM38 Figure 3.9 (5). Baud Rate Generator Control Registers (BRGCR1) 118 TOSHIBA CORPORATION TMP90CM38 Figure 3.9 (6). Port 7 Function Register TOSHIBA CORPORATION 119 TMP90CM38 3.9.2 Configuration Figure 3.9 (7) shows the block diagram of the serial channel. Figure 3.9 (7). Serial Channel (SIO1) Block Diagram 120 TOSHIBA CORPORATION TMP90CM38 ➀ Baud rate generator The relation between the input clock and the source clock (fc) is as follows. Baud rate generator comprises of a circuit that generates transmission and receiving clocks that determine the transfer rate of the serial channel. The input clock to the baud rate generator, øT0 (fc/4), øT2 (fc/16), øT8 (fc/64), or øT32 (fc/256) is generated by the 9-bit prescaler which is shared by the timers. One of these input clocks is selected by the baud rate genorator control register BRGCR1 <BG11,01>. The baud rate generator includes a 4-bit frequency divider, which divides frequency by 2 to 16 values to determine the transfer rate. How to calculate a transfer rate when the baud rate generator is used is explained below. øT0 øT2 øT8 øT32 = fc/4 = fc/16 = fc/64 = fc/256 Accordingly, when source clock fc is 12.288MHz, input clock øT2 (fc/16), and frequency divisor is 5, the transfer rate in UART mode becomes as follows. • UART mode = 12.288 x 106/16/5/16 = 9600 (bps) Table 3.9 (1) shows an example of the transfer rate in UART mode. Also with 8-bit timer 0, the serial channel (SIO1) can get a transfer rate. Table 3.9 (1) shows an example of baud rate using timer 2. Table 3.9 (1) Selection of Transfer Rate (1) (When Baud Rate Generator is Used) Unit: Kbps Input clock Frequency clock øT0 (fc/4) øT2 (fc/16) øT8 (fc/64) øT32 (fc/256) 9.8304MHz – 2457.600 614.400 153.600 38.400 – 2 76.800 19.200 4.800 1.200 – 4 38.400 9.600 2.400 0.600 – 8 19.200 4.800 1.200 0.300 – 16 9.600 2.400 0.600 0.150 12.288MHz – 3072.000 786.000 192.000 48.000 – 5 38.400 9.600 2.400 0.600 Source clock (fc) – 10 19.200 4.800 1.200 0.300 14.7456MHz – 3686.400 921.600 230.400 57.600 – 3 76.800 19.200 4.800 1.200 – 6 38.400 9.600 2.400 0.600 – 12 19.200 4.800 1.200 0.300 TOSHIBA CORPORATION 121 TMP90CM38 Table 3.9 (2) Selection of Transfer Rate (2) (When Timer 0 (Input Clock øT1) is Used)) Unit: Kbps fc TREG2 12.288 MHz 12 MHz 9.8304 MHz 8 MHz 6.144 MHz 1H 96 – 76.8 62.5 48 2H 48 – 38.4 31.25 24 3H 32 31.25 – – 16 4H 24 – 19.2 – 12 5H 19.2 – – – 9.6 8H 12 – 9.6 – 6 AH 9.6 – – – 4.8 10H 6 – 4.8 – 3 14H 4.8 – – – 2.4 How to calculate the transfer rate (when timer 2 is used0) ➂ Receiving counter Input clock of timer 2 øT1 = fc/8 øT4 = fc/32 øT16 = fc/128 ➁ Serial clock generation circuit This circuit generates the basic clock for transmitting and receiving data. 1) I/O interface mode When in SCLK1 output mode with the setting of SCCR1 <IOC1> = “0”, the basic clock will be generated by dividing by 2 the output of the baud rate generator described before. When in SCLK1 input mode with the setting of SCCR1 <IOC1> = “1”, the rising edge or the falling edge will be detected according to the setting of SCCR1 <SCLK1> register to generate the basic clock. 2) Asynchronous communication (UART) mode 122 According to the setting of SCMOD1 <SC11,01>, the above baud rate generator clock, internal clock ø1 (fc/ 2) (312.5 Kbaud at 10MHz), or the match detect signal from timer 2 will be selected to generate the basic clock SIOCK. The receiving counter is a 4-bit binary counter used in asynchronous communication (UART) mode and counts up by SIOCLK1 clock. 16 pulses of SIOCLK1 are used for receiving 1 bit of data, and the data is sampled three times at the 7th, 8th and 9th clock. With the three samples, the received data is evaluated by the rule of majority. For example, if the sampled data is “1”, “0” and “1” at 7th, 8th and 9th clock, respectively, the received data is evaluated as “1”. The sampled data “0”, “0”, and “1” is evaluated that the received data is “0”. ➃ Receiving control 1) I/O interface mode When in SCLK1 output mode with the setting of SCCR1 <IOC1> = “0”, RxD1 signal will be sampled at the rising edge of shift clock which is output to SCLK pin. TOSHIBA CORPORATION TMP90CM38 When in SCLK1 input mode with the setting of SCCR1 <IOC1> = “1”, RxD1 signal will be sampled at the rising edge or falling edge of SCLK1 input according to the setting of SCCR1 <SCLK1> register. 2) Asynchronous communication (UART) mode The receiving control has a circuit for detecting the start bit by the rule of majority. When two or more “0” are detected during 3 samples, it is recognized as normal start bit and the receiving operation is started. Data being received are also evaluated by the rule of majority. ➄ Receiving buffer To prevent overrun from occurring, the receiving buffer has a double structure. Received data are stored one bit by one bit in the receiving buffer 1 (shift register type). When 7 bits or 8 bits of data rae stored in the receiving buffer 1, the stored data are transferred to another receiving buffer 2 (SCBUF1), generating an interrupt INTRX1. The CPU reads only receiving buffer 2 (SCBUF1). Even before the CPU reads the receiving buffer 2 (SCBUF1), the received data can be stored in ➆ Transmission controller 1) I/O interface mode In SCLK1 output mode with the setting of SCCR1 <IOC1> = “0”, the data in the transmission buffer are output bit by bit to TxD1 pin at the rising edge of shift clock which is output from SCLK1 pin. In SCLK1 input mode with the setting of SCCR1 <IOC1> = “0”, the data in the transmission buffer are output bit by bit to TxD1 pin at the rising edge or falling edge of SCLK input according to the setting of SCCR1 <SCLK1> register. 2) Asynchronous communication (UART) mode When the transmission data are written in the trans- TOSHIBA CORPORATION receiving buffer 1 . However, unless the receiving buffer 2 (SCBUF1) is read before all bits of the next data are received by the receiving buffer 1, an overrun error occurs. If an overrun error occurs, the contents of receiving buffer 1 will be lost, although the contents of receiving buffer 2 and SCCR1 <RB81> is still preserved. The parity bit added in 8-bit UART mode and the most significant bit (MSB) in 9-bit UART mode are stored in SCCR0 <RB81>. When in the 9-bit UART, the wake-up function of the slave controllers is enabled by setting SCMOD1 <WU1> to “1”, and interrupt INTRX occurs only when SCCR1 <RB81> is set to “1”. ➅ Transmission counter Transmission counter is a 4-bit binary counter which is used in asynchronous communication (UART) mode and, like a receiving counter, counts by SIOCLK1 clock, generating TxDCLK1 every 16 clock pulses. mission buffer sent from the CPU, transmission starts at the rising edge of the next TxDCLK1, generating a transmission shift clock TxDSFT1. Hand-shake function The TMP90CM38 supports a hand-shake function by the connection of CTS of one TMP90CM38 and RTS0 of the other device. The hand-shake function allows receiving/transmitting data on a frame basis to prevent overrun errors. This function is enabled or disabled by the control register SCCR <CTSE>. When the last bit (parity bit or MSB) of 1-frame data is received by the receiving unit, the RTS pin turns to the “H” level to request the transmission unit to halt transmission. When the CTS pin turned to the “H” level, the trans123 TMP90CM38 mission unit halts transmission, after completing the current data transmission, until the pin turns to the “L” level. At this time, the interrupt INTTX is generated, to request the CPU to transfer data. Then the data is written into the transmission buffer, and the transmission unit is placed in the standby until the CTS pin turned to the “L” level. When the received data are read by the CPU, the RTS pin returns to the “L” level, requesting that the transmission is restarted. Figure 3.9 (8). Hand-shake Function Figure 3.9 (9). Hand-shake CTS (clear to send) Signal 124 TOSHIBA CORPORATION TMP90CM38 ⑧ Transmission buffer 2) Parity error (SCCR1 <PERR1>) Transmission buffer SCBUF1 shifts out and sends the transmission data written from the CPU from the least significant bit (LSB) in order, using transmission shift clock TxDSFT1 which is generated by the transmission control. When all bits are shifted out, the transmission buffer becomes empty and generates INTTX1 interrupt. The parity generated for the data shifted in receiving buffer 2 (SCBUF1) is compared with the parity bit received from the RxD1 pin. If they are not equal, a parity error occurs. 3) Framing error (SCCR1 <FERR1>) ⑨ Parity control circuit When serial channel control register SCCR1 <PE1> is set to “1” , it is possible to transmit and receive data with parity. However, parity can be added only in 7-bit UART or 8-bit UART mode. With SCCR1 <EVEN1> register, even (odd) parity can be selected. For transmission, parity is automatically generated according to the data written in the transmission buffer SCBUF, and data are transmitted after being stored in SCBUF1 <TB71> when in 7-bit UART mode while in SCMOD1 <TB81> in 8-bit UART mode. <PE1> and <EVEN1 > must be set before transmission data are written in the transmission buffer. For receiving, data are shifted in the receiving buffer 1, and parity is added after the data are transferred in the receiving buffer 2 (SCBUF1), and then compared with <RB71> of SCBUF1 when in 7-bit UART mode and with SCCR1 <RB81> when in 8-bit UART mode. If they are not equal, a parity error occurs, and SCCR1 <PERR1> flag is set. The stop bit of received data is sampled three times around the center. If the majority results is “0”, a framing error occurs. 11 1) UART mode Receiving Mode Three error flags are provided to increase the reliability of receiving data. 1) Overrun error (SCCR1 <OERR1>) If all bits of the next data are received in receiving buffer 1 while valid data are still stored in receiving buffer 2 (SCBUF1), an overrun error will occur. TOSHIBA CORPORATION 9 Bit 8 Bit + Parity 8 Bit, 7 Bit + Parity, 7 Bit Interrupt timing Center of last bit (Bit 8) Center of last bit (Parity Bit) Center of stop bit Framing error timing Center of stop bit Center of stop bit Center of stop bit Parity error timing Center of last bit (Bit 8) Center of last bit (Parity Bit) Center of stop bit Over-run error timing Center of last bit (Bit 8) Center of last bit (Parity Bit) Center of stop bit Note: ➉ Error flag Generation Timing Framing error occurs after an interrupt has occurred. Therefore, to check for framing error during interrupt operation, it is necessary to wait for 1 bit period of time. Transmitting Mode Interrupt timing 9 Bit Just before last bits transmitted 8 Bit + Parity 8 Bit, 7 Bit + Parity, 7 Bit ← ← 125 TMP90CM38 3.9.3 Operational Description (1) Mode 0 (I/O Interface Mode) This mode is used to increase the number of I/O pins of TMP90CM36 for transmitting or receiving data to or from the external shifter register. This mode incudes SCLK1 output mode to output synchronous clock SCLK1 and SCLK1 input mode to input external synchronous clock SCLK1. Figure 3.9 (10). I/O Interface Mode 126 TOSHIBA CORPORATION TMP90CM38 ➀ Transmission In SCLK1 output mode, 8-bit data and synchronous clock are output from TxD1 pin and SCLK1 pin, respectively, each time the CPU writes data in the transmission buffer. When all data is output, IRF2 <IRFTX1> will be set to generate INTTX1 interrupt. Figure 3.9 (11). Transmitting Operation in I/O Interface Mode (SCLK1 Output Mode) In SCLK1 output mode, 8-bit data are output from TxD1 pin and SCLK1 input becomes active while data are written in the transmission buffer by CPU. Figure 3.9 (12). Transmitting Operation in I/O Interface Mode (SCLK1 Input Mode) TOSHIBA CORPORATION 127 TMP90CM38 ➁ Receiving In SCLK1 output mode, received data are read by the CPU, and synchronous clock is output from SCLK1 pin and the next data are shifted in the receiving buffer 1 whenever the receive interrupt flag IRF2 <IRFRX1> is cleared. When 8-bit data are received, the data will be transferred in the receiving buffer 2 (SCBUF1), and <IRFRX1> will be set again to generate INTRX1 interrupt. Figure 3.9 (13). Receiving Operation in I/O Interface Mode (SCLK1 Output Mode) In SCLK1 input mode, received data are read by the CPU, and the next data is shifted in the receiving buffer 1 when SCLK1 input becomes active while the receive interrupt flag <IRFRX> is cleared. When 8-bit data are received, the data will be shifted in the receiving buffer 2 (SCBUF1), and <IRFRX1> will be set again to generate INTRX1 interrupt. Figure 3.9 (14). Receiving Operation in I/O Interface Mode (SCLK1 Input Mode) 128 TOSHIBA CORPORATION TMP90CM38 (2) Mode 1 (7-Bit UART Mode) The 7-bit mode can be set by setting serial channel mode register SCMOD1 <SM11, 01> to “01”. In this mode, a parity bit can be added, and the addi- TOSHIBA CORPORATION tion of a parity bit can be enabled or disabled by serial channel control register. SCCR1 <PE1>, and even parity or odd parity is selected by SCCR1 <EVEN1> when <PE1> is set to “1” (enable). 129 TMP90CM38 (3) Mode 2 (8-Bit UART Mode) The 8-bit UART mode can be set by setting serial channel mode register SCMOD1 <SM11, 01> to “10”. In this mode, a parity bit can be added, the addition of a parity bit is enabled or disabled by SCCR10 <PE1>. SCCR1 <PE1>, and even parity or odd parity is 130 selected by SCCR1 <EVEN1> when <PE1> is set to “1” (enable). Setting example: When receiving data with the following format, the control registers should be set as described below. TOSHIBA CORPORATION TMP90CM38 (4) Mode 3 (9-Bit UART Mode) Wake-up function The 9-bit UART mode can be specified by setting SCMOD1 <SM11, 11> to “11”. In this mode, parity bit cannot be added. For transmission, the MSB (9th bit) is written in SCMOD1 <TB81>, while in receiving it is stored in SCCR1 <RB81> . For writing or reading the buffer, the MSB is read or written first then SCBUF1. In 9-bit UART mode, the wake-up function of slave controllers is enabled by setting SCMOD1 <WU1> to “1”. The interrupt INTRX1 occurs only when SCCR1 <RB81> = 1. Figure 3.9 (15). Serial Link Using the Wake-up Function TOSHIBA CORPORATION 131 TMP90CM38 Protocol cleared to “0”. The MSB (bit 8) SCMOD1 <TB81> is set to “0”. ➀ Select the 9-bit UART mode for the master and slave controllers. ➁ Set the SCMOD1 <WU1> bit of each slave controller to “1” to enable data receiving. ➂ The master controller transmits one-frame, including the 8-bit select code of the slave controllers. The MSB (bit 8) SCMOD1 <TB81> is set to “1”. ➃ Each slave controller receives the above frame, and clears WU bit to “0” if the above select code matches its own select code. ➄ The master controller transmits data to the specified slave controller whose SCMOD1 <WU1> bit is 132 ➅ The other slave controllers (with SCMOD1 <WU1> bit remaining at “1”) ignore the receiving data because their MSBs (bit 8 or SCCR1 <RB81>) are set to “0” to disable the interrupt INTRX1. When the WU1 bit is cleared to “0”, the interrupt INTRX1 occurs, so that the slave controller can read the receiving data. The slave controllers (WU1 = 0) transmit data to the master controller, and it is possible to indicate the end of data receiving to the master controller by this transmission. Setting example: To link two slave controllers serially with the master controller, and use the internal clock ø/1 (fc/2) as the transfer clock . TOSHIBA CORPORATION TMP90CM38 TOSHIBA CORPORATION 133 TMP90CM38 3.9.4 Configuration The serial channels are connected to external circuits through three-pin serial ports: SCLK2 (P76), TXD2 (P77) and RXD2 (P75). Figure 3.9 (16). Block Diagram of Serial Channels (SIO2) 134 TOSHIBA CORPORATION TMP90CM38 Serial clock SIO2 pulses make the following selections through the serial channel mode register SCMOD2. b. (External clock) ➀ Clock Source Selection <SCLK2> uses the clock pulse externally supplied to the SCLK2 pin as the serial clock pulse. <SCLKS2> selects either an internal or external clock as the clock source. ➁ Shift Edge Selection a. (Internal clock) SCMOD2 <CLK1, CLK0> selects the speed of either øT1 (fc/4), øT1 (fc/8), øT4 (fc/32), or øT16 (fc/128) serial clock. The serial clock pulse is externally output from the SCLK2 pin. The serial clock automatically stops after it ends the “1-frame” serial operation. It waits until next serial operation. TOSHIBA CORPORATION a. Rising edge shift Data shifts on the serial clock pulses’s rising edge (falls at the SCLK2 pin). b. Falling edge shift Data shifts on the serial clock pulses’s falling edge (rises at the SCLK2 pin or no falling edge shift in send mode) 135 TMP90CM38 3.9.5 Explanation of Operations The send, receive and simultaneous send-receive modes for SCMOD2 <SMD1, SMD0>. (1) Send Mode The first send data is written into buffer registers SCBUF after the send mode is set in the command register. (Data will not be written into the buffers if the command register is not in send mode.) Then, storing “1” into serial transfer control registers SCMOD2 <SIOE> starts the send. As the send starts, the send data is synchronized with serial clock pulses and sequentially output from the TxD2 pin on the LSB side. At the same time, the send data is transferred from the buffer registers to the shift registers. Since the buffer registers are empty, buffer empty interrupt INTTX2 is generated to request the next send data. When the interrupt service program writes the next send data into the buffer register, the interrupt request signal isn’t cleared to “0”. (Internal clock pulse) (External clock pulses) In the internal clock operation, data must be stored in the buffer registers before the next data shift operation begins. The transfer speed in an interrupt service program is determined by the maximum delay time from the interrupt request generation to buffer register data write. To end a send, the buffer empty interrupt service program disables (clears to “0”) serial transfer control register SCMOD2 <SIO0E> instead of writing the next send data. When serial transfer control is disabled, the serial transfer ends when the send data now being shifted out is finished being sent. The end of send can be determined by the status of serial transfer monitor flag SCMOD2 <FFSI>. In the external clock operation, the serial transfer control register SCMOD2 <SIOE> must be disabled before starting the next send data shift operation. If the serial transfer control register SCMOD2 <SIOE> is not disabled before the shift operation begins, operations stop after sending the next send data (dummy). In the internal clock operations, if all data is sent and no subsequent data is stored in the register, the serial clock pulse stops and a wait begins. 136 TOSHIBA CORPORATION TMP90CM38 Figure 3.9 (17). Chart of Serial Channel 0 Send Mode Timing TOSHIBA CORPORATION 137 TMP90CM38 (2) Receive Mode (3) Setting the command register to receive mode, then setting serial serial transfer control SCMOD2 <SIOE> to enable makes receive possible. Shift data is synchronized with serial clock pulses and fetched from the RxD2 pin. When data is fetched, it is transferred from the shift register to the buffer register and the buffer-full interrupt INTRX2 is generated to request a read of receive data. When the interrupt service program read the next receive data from the buffer register, the interrupt request signal is cleared. The following data continues to be fetched after the interrupt is generated. After the interrupt request is cleared, data is transferred from the shift register to the buffer register when data is fetched. The first send data is written into buffer registers SCBUF2 after the send-receive mode is set by the command register. Setting the serial transfer control register SCMOD2 <SIOE> to 1 enables receiving or sending data. Send data is output from the TxD2 pin on the rising edge of the serial clock pulse, while receive data is fetched from the RxD2 pin on the falling edge of the serial clock pulse. When data is fetched, data is transferred from the shift registers to the buffer registers and buffer-full interrupt INTRX2 is genrated to request receive data read. When the interrupt service program reads the next receive data from the buffer register, the interrupt request signal is cleared. (Internal clock pulses) (Internal clock pulses) In the internal clock operation, if the previous receive data has not been read from the buffer register after the next data is fetched, the serial clock stops and waits until the previous data is read. In the internal clock operation, a wait begins until the received data is read and the next send data is written. (External clock pulses) In the external operation, shift operations are synchronized with externally supplied clock pulses. The data is read before the next receive data is transferred into the buffer register. If the previous data has not been read, the receive data will not be transferred into the buffer registers and all subsequently input receive data will be cancelled. The maximum transfer speed of the external clock operation is determined by the maximum delay time from interrupt request generation to receive data read. Rising and falling edge shifts can be selected in the receive mode. Because data is fetched on the serial clock pulses’s rising edge in the rising edge shift, the first shift data must already be input to the RxD2 pin when the initial serial clock pulses are applied at transfer start. 138 Send-Receive Mode (External clock pulses) In the external clock operation, the receive data must be read and the next send data written before starting the next shift operation, because the shift operation is synchronized with the external supplied clock pulses. The maximum transfer speed of the external clock operation is determined by the maximum delay time from interrupt request generation to send data fetch and receive data write. Because the same buffer registers are used for send and receive, always ensure that send data is written after 8 bits of receive data are fetched. To end send-receive, disable the serial transfer control register. When the serial transfer control register is disabled, send-receive ends afetr receive data is organized and transferred to the buffer register. The program checks the end of send-receive by reading serial transfer monitor flags SCMOD2 <FFSI>. TOSHIBA CORPORATION TMP90CM38 Figure 3.9 (18) - 1. Chart of Serial Channel ø Send-Receive Mode (Falling Edge Shift) Timing TOSHIBA CORPORATION 139 TMP90CM38 Figure 3.9 (18) - 2. Chart of Serial Channel ø Send-Receive Mode (Falling Edge Shift) Timing 140 TOSHIBA CORPORATION TMP90CM38 Figure 3.9 (19) - 1. Serial Channel Control Register TOSHIBA CORPORATION 141 TMP90CM38 Figure 3.9 (19) - 2. Serial Channel Buffer Registers 142 TOSHIBA CORPORATION TMP90CM38 3.10 Analog/Digital Converter The TMP90CM38 contains a high-speed, high-accuracy analog/digital converter (A/D converter) with 8-channel analog input that features 8-bit sequential comparison. Figure 3.10 (1) shows the block diagram of the A/D converter. 8-channel analog input pins (AN7 to AN0) are shared by input-only port P6 and so can be used as input port. Figure 3.10 (1). Block Diagram of A/D Converter TOSHIBA CORPORATION 143 TMP90CM38 3.10.1 Control Registers Figure 3.10 (2). A/D Conversion Mode Register (ADMOD) 144 TOSHIBA CORPORATION TMP90CM38 Figure 3.10 (3). A/D Conversion Result Register (ADREG0 ~ 3) Figure 3.10 (4). A/D Converter Channel Select Register TOSHIBA CORPORATION 145 TMP90CM38 3.10. 2 Operation (1) Analog Reference Voltage High analog reference voltage is applied to the AVCC pin, and the low analog voltage is applied to AVSS pin. The reference voltage between AVCC and AVSS is divided by 256 using ladder resistance, and compared with the analog input voltage for A/D conversion. (2) (3) (5) There are two A/D conversion speed modes: high speed mode and low speed mode. The selection is executed by ADMOD <ADCS> register. When reset, ADMOD <ADCS> will be initialized to “0”, so that high speed conversion mode will be selected. (6) • A/D conversion single mode Analog input channel is selected by ADMOD <ADCH1,0>, ADMOD <ADCH2>. However, which channel to select depends on the operation mode of the A/D converter. In fixed analog input mode, one channel is selected by ADMOD <ADCH1,0>, ADMOD <ADCH2> among three pins: AN0 to AN3. In analog input channel scan mode, the number of channels to be scanned from AN0 is specified by ADMOD <ADCH1,0>, ADMOD <ADCH2>, such as AN0 ➱ AN1, AN0 ➱ AN1 ➱ AN2, and AN0 ➱ AN1 ➱ AN2 ➱ AN3 or the number of channels to be scanned from AN4 is such as AN4 ➱ AN5, AN4 ➱ AN5 ➱ AN6, AN4 ➱ AN5 ➱ AN6 ➱ AN7. When reset, A/D conversion channel register will be initialized to ADMOD <ADCH1,0> = 00, ADCH <ADCH2> = 0 so that AN0 pin will be selected. The pins which are not used as analog input channel can be used as ordinary input port P5. ADMOD <EOCF> for A/D conversion end will be set to “1”, ADMOD <ADBF> flag will be reset to “0”, and INTAD interrupt will be enabled when A/D conversion of specified channel ends in fixed conversion channel mode or when A/D conversion of the last channel ends in channel scan mode. Interrupt requesting flip-flop is cleared only by resetting operation or reading the A/D conversion result storing register and cannot be cleared by instruction. Starting A/D Conversion • A/D conversion repeat mode For both fixed conversion mode and conversion channel scan mode, INTAD will be disabled when in repeat mode. Always leave the INTE0 <ADIS> flag at “0”. Write “0” to ADMOD <REPET> to end the repeat mode. Then, the repeat mode will be exited as soon as the conversion in progress is completed. (7) Storing the A/D Conversion Result The results of A/D conversion are stored ADREG04 to ADREG37 registers for each channel. In repeat mode, the registers are updated whenever conversion ends. ADREG04 to ADREG37 are read-only registers. (8) Reading the A/D Conversion Result A/D Conversion Mode Both fixed A/D conversion channel mode and A/D conversion channel scan mode have two conversion modes, i.e., single and repeat conversion modes. In fixed channel repeat mode, conversion of specified one channel is executed repeatedly. In scan repeat mode, scanning from AN0 ⋅⋅⋅➱ AN3 or from AN4, ⋅⋅⋅➱ AN7 is executed repeatedly. A/D conversion mode is selected by ADMOD <REPET, SCAN>. 146 A/D Conversion End and Interrupt Analog Input Channels A/D conversion starts when A/D conversion register ADMOD <ADS> is written “1”. When A/D conversion starts, A/D conversion busy flag ADMOD <ADBF> which indicates “A/D conversion is in progress” will be set to “1”. (4) A/D Conversion Speed Selection The results of A/D conversion are stored ADREG04 to ADREG37 registers. When the contents of one of ADREG0 to ADREG3 registers are read, ADMOD <EOCF> will be cleared to “0”. Setting example: ➀ When the analog input voltage of the AN3 pin pin is A/ D converted and the results are read in the memory at FF10H by A/D interrupt INTAD routine. TOSHIBA CORPORATION TMP90CM38 TOSHIBA CORPORATION 147 TMP90CM38 3.11 Watchdog Timer (Looping Detection Timer) The purpose of the watchdog timer (WDT) is to detect the start of CPU misoperation due to noise, etc., and bring it back to normal. 3.11 .1 Configuration The TMP90CM38 multiplexes the watchdog timer output (WDTOUT) and P80. P80 (output port) is switched to the WDTOUT pin and RESET is returned inside the chip by setting bit WDMOD <RESCR> = “1” of the watchdog timer mode register at address #FFD0H to “1”. Figure 3.11 (1) shows the WDT block diagram. Figure 3.11 (1). Watchdog Timer Block Diagram 148 TOSHIBA CORPORATION TMP90CM38 The watchdog timer is a 22-stage binary counter that uses (fc/2) as the input clock. The binary counter outputs are 216/fc, 218/fc, 220/fc and 22 2 /fc. One of these outputs is used for watchdog timer output WDTOUT. WDTOUT outputs “0” to reset the peripheral devices 3.11.2 Control Registers The watchdog timer (WDT) is controlled by two control registers (WDMODE and WDCR). (1) Watchdog Timer Mode Register (WDMOD ➀ Watchdog timer detection time setting (WDTP) This is a 2-bit flag used to set the watchdog timer interrupt time for looping (runaway) detection. This flag is initialized to WDMOD <WDTP0, 1> = 00 by resets, which results in a value of 216/fc [sec]. (The number of states is approximately 32,768.) ➁ Watchdog timer enable/disable control (WDTE) This bit is initialized to WDTE = 1 by resets, which enables the watchdog timer. To disable the watchdog timer, it is only necessary to clear this bit to “0” and write the disable code (B1H) to the WDCR register. It is difficult for the watchdog timer to be disabled by looping. To disable the watchdog timer after it has been enabled, it is only necessary to write “1” to the <WDTE> bit. TOSHIBA CORPORATION when the watchdog timer overflows. WDTOUT also is connected to RESET inside the WDTOUT TMP90CM38. In this case, WDTOUT outputs “0” in a 32/fxtal = 2.0µsec (fxtal = 16MHz) cycle and simultaneously resets the TMP90CM38. ➂ Watchdog timer out reset connection (RESCR) This flag is used to set whether or not the TMP90CM38 will be reset when looping is detected and whether or not to output WDTOUT. <RESCR> is set to “1” by reset operations; therefore, P80 is set as the WDTOUT pin and connected internally to the RESET pin. P80 can be set as either the WDTOUT pin or port pin by overwriting <RESCR>. However, caution is required because a redundant structure is used to prevent misoperation. The <RESCR> bit is linked to the P38CR <WDTOUTC> therefore, it is always necessary to write “1” the <WDTOUTC> when the <RESCR> bit is overwritten. The <RESCR> bit is set only after “1” is written to the <WDTOUTC> and then either “0” or “1” is written to the RESCR bit. Writing to the <RESCR> bit automatically clears the <WDTOUTC> to “0”; therefore, when resetting the <RESCR> bit, again write “1” to the <WDTOUTC>. 149 TMP90CM38 Figure 3.11 (2). Flowchart of P80/WDTOUT Pin Switching 150 TOSHIBA CORPORATION TMP90CM38 Figure 3.11 (3). Watchdog Timer Mode Register TOSHIBA CORPORATION 151 TMP90CM38 (2) Watchdog Timer Control Register (WDCR) This register enables and disables the watchdog timer, and clears the binary counter. • Enable control Set WDMOD <WDTE> to “1”. • Binary counter clear control • Disable control The watch timer is disabled by clearing WDMOD <WDTE> to “0” and then writing the disable code (B1H) to the WDCR register. The binary counter is cleared and restarts counting when the clear code (4EH) is written to WDCR. Figure 3.11 (4). Watchdog Timer Control Register 152 TOSHIBA CORPORATION TMP90CM38 3.11.3 Operation The watchdog timer is a timer that outputs “0” level from the watchdog timer output pin (WDTOUT) after the detection time set with WDMOD <WDRP1, 0> elapses. The watchdog timer binary counter is cleared to “0” before an overflow occurs. If the CPU misoperates (loops) due to noise, etc., the binary counter will overflow unless the parity counter clear instruction is executed. The CPU can be returned to normal operation by TOSHIBA CORPORATION resetting internally. A reset can be applied to both the TMP90CM36 and CPU by connecting the WDTOUT pin to the RESET pins of the peripheral devices. The watchdog timer again starts operating immediately after the reset is canceled. The watchdog timer stops during the IDLE mode and STOP mode and operates during the RUN mode. The watchdog timer can also be disabled when entering the RUN mode. 153 TMP90CM38 4. Electrical Characteristics TMP90CM38F/TMP90CM38T 4.1 Maximum Ratings Symbol VCC Item Power supply voltage Rating Unit -0.5 ~ + 7 V VIN Input voltage -0.5 ~ VCC + 0.5 V ∑IOL Output current (Total) 100 mA ∑IOH Output current (Total) -70 mA PD F 500 Power dissipation (Ta = 85°C) mW T 600 260 °C TSTG Storage temperature -65 ~ 150 °C TOPR Operating temperature -20 ~ 70 °C TSOLDER Soldering temperature (10s) 4.2 DC Electrical Characteristics VCC = 5V ± 10%, TA = -20 ~ 70°C (1 ~ 16MHz) Typical Values are for TA = 25°C, VCC = 5V. Symbol Item Min Max Unit Conditions -0.3 0.8 V – VIL Input Low Voltage (P0) VIL1 P1, P2, P3, P4 , P5, P5, P6, P7, P9, P10 -0.3 0.3VCC V – VIL2 RESET, P81 (INT0), P82 (STBY), NMI -0.3 0.25VCC V – VIL3 EA -0.3 0.3 V – VIL4 X1 -0.3 0.2VCC V – VIH Input Low Voltage (P0) 2.2 VCC + 0.3 V – VIH1 RESET, P81 (INT0), P82 (STBY), NMI 0.7VCC VCC + 0.3 V – VIH2 RESET, P81 (INT0), P82 (STBY) 0.75VCC VCC + 0.3 V – VIH3 EA VCC - 0.3 VCC + 0.3 V – VIH4 X1 0.8VCC VCC + 0.3 V VOL Output Low Voltage – 0.45 V IOL = 1.6mA VOH VOH1 VOH2 Output High Voltage 2.4 0.75VCC 0.9VCC – – V V V IOH = -400µA IOH = -100µA IOH = -20µA IDAR Darlington Drive Current (8 I/O Pins max) -1.0 -3.5 mA VEXT = 1.5V REXT = 1.1kΩ – ILI Input Leakage Current 0.02 (Typ) ±5 µA 0.0 ≤ Vin ≤ VCC ILO Output Leakage Current 0.05 (Typ) ± 10 µA 0.2 ≤ Vin ≤ VCC - 0.2 Operating Current (RUN) Idle 35 (Typ) 1.5 (Typ) 50 5 mA mA tosc = 16MHz 0.2 (Typ) 40 10 µA µA 0.2 ≤ Vin ≤ VCC - 0.2 VIL1 = 0.2Vcc, VIL2 = 0.8Vcc ICC STOP (TA = -20 ~ 70°C) STOP (TA = 0 ~ 50°C) VSTOP Power Down Voltage of (@STOP) (RAM back up) 2.0 6.0 V RRST RESET Pull Up Register 50 150 KΩ – 10 pF 0.4 1.0 (Typ) V CIO Pin Capacitance VTH Schmitt width (RESET, P81, P82) – testfreq = 1MHz – Note: IDAR is guaranteed for up to 8 optional ports. 154 TOSHIBA CORPORATION TMP90CM38 4.3 AC Electrical Characteristics VCC = 5V ± 10% TA = -20 ~ 70°C Variable Symbol tOSC 12.5MHz Clock 16MHz Clock Item Oscillation cycle ( = X) Unit Min Max Min Max Min Max 62.5 1000 80 – 62.5 – ns tAL A0 ~ A7 effective address→ALE fall 0.5x - 15 – 25 – 16 – ns tLA ALE fall→A0 ~ A7 hold 0.5x - 15 – 25 – 16 – ns tLL ALE pulse width x - 40 – 40 – 23 – ns tLC ALE fall→RD/WR fall 0.5x - 30 – 10 – 1 – ns tCL RD/WR rise→ALE rise 0.5x - 20 – 20 – 11 – ns tACL A0 ~ A7 effective address→RD/WR fall x - 25 – 55 – 38 – ns tACH Upper effective address→RD/WR fall 1.5x - 50 – 70 – 44 – ns tCA RD/WR fall→Upper address hold 0.5x - 20 – 20 – 11 – ns tADL A0 ~ A7 effective address→Effective data input – 3.0x - 35 – 205 – 153 ns tADH Upper effective address→Effective data input – 3.5x - 55 – 225 – 164 ns tRD RD fall→Effective data input – 2.0x - 50 – 110 – 75 ns 2.0x - 40 – 120 – 85 – ns 0 – 0 – 0 – ns tRR RD Pulse width tHR RD rise→Data hold tRAE RD rise→Address enable x - 15 – 65 – 48 – ns tWW WR pulse width 2.0x - 40 – 120 – 85 – ns tDW Effective data→WR rise 2.0x - 50 – 100 – 65 – ns tWD WR rise→Effective data hold 0.5x - 10 – 30 – 21 – ns AC Measurement Conditions • Output level: High 2.2V/Low 0.8V,CL = 50pF (However, CL = 100pF for AD0 ~ 7, A8 ~ 15, ALE, RD, WR) • Input level High 2.4V/Low 0.45V (AD0 ~ AD7) High 0.8VCC/Low 0.2VCC (excluding AD0 ~ AD7) TOSHIBA CORPORATION 155 TMP90CM38 4.4 A/D Conversion Characteristics VCC = 5V ± 10% TA = -20 ~ 70°C f = 1 ~ 16MHz Symbol Item Min Typ Max AVCC Analog reference voltage Vcc - 1.5 Vcc Vcc AGND Analog reference voltage Vss Vss Vss VAIN Analog input voltage range Vss – Vcc IREF Analog reference voltage power supply current – 0.5 1.0 Total error (TA = 25°C, Vcc = VREF = 5.0V) – – 1.0 Total error – – 2.5 Error (Quantize error of ± 0.5 LSB not included) Unit V mA LSB 4.5 Timer/Counter Input Clock (TI2, TI4) VCC = 5V ± 10% TA = -20 ~ 70°C f = 1 ~ 16MHz Variable Symbol 12.5MHz Clock 16MHz Clock Item Unit Min Max Min Max Min Max tVCK Clock cycle 8x + 100 – 740 – 600 – ns tVCKL Low clock pulse width 4x + 40 – 360 – 290 – ns tVCKH High clock pulse width 4x + 40 – 360 – 290 – ns 4.6 Interrupt Operation VCC = 5V ± 10% TA = -20 ~ 70°C f = 1 ~ 16MHz Variable Symbol 12.5MHz Clock 16MHz Clock Item Unit Min Max Min Max Min Max 4x – 320 – 250 – ns 4x – 320 – 250 – ns 8x + 100 – 740 – 600 – ns 8x + 100 – 740 – 600 – ns INT0 Low level pulse width tINTAL tINTAH INT0 High level pulse width INT1, INT2 Low level pulse width tINTBL INT1, INT2 High level pulse width tINTBH 156 TOSHIBA CORPORATION TMP90CM38 4.7 Serial Channel SIO1 Timing - I/O Interface Mode (1) SCLK1 Input Mode VCC = 5V ± 10% TA = -20 ~ 70°C f = 1 ~ 16MHz Variable Symbol 12.5MHz Clock 16MHz Clock Item Unit Min Max Min Max Min Max 16x – 1.28 – 1 – µs tSCY/2 - 5x - 50 – 190 – 137 – ns 5x - 100 – 300 – 212 – ns tSCY SCLK1 cycle tOSS Output data →Rising edge of SCLK tOHS SCLK1 rising edge→Output data hold tHSR SCLK1 rising edge→Input data hold 0 – 0 – 0 – ns tSRD SCLK1 rising edge→ Effective data input – tSCY - 5x - 100 – 780 – 587 ns (2) SCLK1 Output Mode Variable Symbol 12.5MHz Clock 16MHz Clock Item Unit Min Max Min Max Min Max tSCY SCLK cycle (programmable) 16x 8192x 1.28 655.4 1 512 µs tOSS Output data setup→SCLK rising edge tSCY - 2x - 50 – 970 – 725 – ns tOHS SCLK rising edge→Output data hold 2x - 80 – 80 – 45 – ns tHSR SCLK rising edge→Input data hold 0 – 0 – 0 – ns tSRD SCLK rising edge→ Effective data input – tSCY - 2x - 150 – 970 – 725 ns TOSHIBA CORPORATION 157 TMP90CM38 4.8 Serial Channel SIO2 Timing 10MHz Clock Symbol tSCR Serial port clock cycle time tSCL SCLK2 Low width tSCH SCLK2 High width tSKDO tSRD tHSR 158 Item VariableClock Condition SCLK2 → TXD2 (Output data) delay time SCLK2 Rising edge to input Unit Min Max Min Max Internal 800 12800 8x 128x External 1600 – 16x – Internal * * * * External * * * * Internal * * * * External * * * * Internal * – * – External * – * – Internal * – * – data valid External * – * – Internal * – * – rising edge External * – * – Input data hold after SCLK2 ns ns ns ns ns ns TOSHIBA CORPORATION TMP90CM38 4.9 Timing Chart TOSHIBA CORPORATION 159 TMP90CM38 4.10 Serial Channel SIO1 I/O Interface Mode Timing Chart 4.11 Serial Channel SIO2 Timing Chart 160 TOSHIBA CORPORATION TMP90CM38 5. Special Function Register List The special function registers (SFR) are the input/output ports, peripheral control registers. These SFR are assigned to 96byte address areas from 0FFA0H ~ 0FFFFH. (1) (2) (3) (4) (5) Input/Output Port Input/Output Port Control Timer/Event Counter Control A/D Converter Control Interrupt Control TOSHIBA CORPORATION (6) (7) (8) (9) (10) (11) (12) (13) HDMA Control WDT Control Serial Channel Control Time Base Counter Control Timing Pulse Generation Control Capture Control D/A Converter PWM Control 161 TMP90CM38 TMP90CM36 Special Function Register List 162 TOSHIBA CORPORATION TMP90CM38 (1) I/O Port TOSHIBA CORPORATION 163 TMP90CM38 (2) I/O Port Control (1/2) 164 TOSHIBA CORPORATION TMP90CM38 (2) I/O Port Control (2/2) TOSHIBA CORPORATION 165 TMP90CM38 (3) Timer/Event Counter Control (1/3) 166 TOSHIBA CORPORATION TMP90CM38 (3) Timer/Event Counter Control (2/3) TOSHIBA CORPORATION 167 TMP90CM38 (3) Timer/Event Counter Control (3/3) 168 TOSHIBA CORPORATION TMP90CM38 (4) A/D Converter Control TOSHIBA CORPORATION 169 TMP90CM38 (5) Interrupt Control 170 TOSHIBA CORPORATION TMP90CM38 (6) HDA Control (7) WDT Control TOSHIBA CORPORATION 171 TMP90CM38 (8) Serial Channel Control 172 TOSHIBA CORPORATION