November 1995 PC87306 SuperI/O TM Enhanced Sidewinder Lite Floppy Disk Controller, Keyboard Controller, Real-Time Clock, Dual UARTs, Infrared Interface, IEEE 1284 Parallel Port, and IDE Interface General Description Features The PC87306 is a single chip solution incorporating a Keyboard and PS/2É Mouse Controller (KBC), Real Time Clock (RTC) and most commonly used I/O peripherals in ISA, EISA and MicroChannelÉ based computers. In addition to the KBC and RTC, a Floppy Disk Controller (FDC), two full featured UARTs, an IEEE 1284 compatible parallel port and all the necessary control logic for an IDE interface provides support for most commonly used I/O peripherals. Standard PC-ATÉ address decoding for all the peripherals, a set of configuration registers, and two user selectable chip selects are also implemented in this highly integrated member of the SuperI/O family. The advanced features and high integration of the PC87306 result in several benefits for low cost, high performance systems. Printed circuit board space savings, fewer components on the motherboard and compatibility with the latest industry standard peripherals are only a few of the benefits of using a PC87306. The KBC is fully software compatible with the 8042AH microcontroller. It contains system timing, control logic, custom ROM program memory, RAM data memory and 18 programmable I/O lines necessary to implement dedicated control functions. It is an efficient controller which uses predominantly single byte instructions with support for binary and BCD arithmetic and extensive bit handling capabilities. (Continued) Y Y Floppy Disk Controller: Ð Software compatible with the DP8477, the 765A and the N82077 Ð 16-byte FIFO (disabled by default) Ð Burst and Non-Burst modes Ð Perpendicular recording drive support Ð High performance internal analog data separator (no external filter components required) Ð Low power CMOS with power-down mode Ð Automatic media-sense support with full IBM TDR (Tape Drive Register) implementation for PC-AT and PS/2 floppy drive types Keyboard Controller: Ð 8042AH and PC87911 software compatible Ð 8-bit Microcomputer with 2 kBytes custom ROM and 256 Bytes data RAM Ð Asynchronous access to two data registers and one status register during normal operation Ð Dedicated open drain outputs for keyboard controller application Ð Supports both interrupt and polling Ð 10 programmable I/O pins Ð 4 dedicated open-drain bidirectional pins Ð 8-bit Timer/Counter Ð Binary and BCD arithmetic (Continued) Block Diagram TL/C/12379 – 1 TRI-STATEÉ is a registered trademark of National Semiconductor Corporation. SuperI/OTM is a trademark of National Semiconductor Corporation. MicroChannelÉ, PC-ATÉ and PS/2É are registered trademarks of International Business Machines Corporation. C1995 National Semiconductor Corporation TL/C/12379 RRD-B30M115/Printed in U. S. A. PC87306 SuperI/O Enhanced Sidewinder Lite Floppy Disk Controller, Keyboard Controller, Real-Time Clock, Dual UARTs, Infrared Interface, IEEE 1284 Parallel Port, and IDE Interface PRELIMINARY General Description (Continued) Features (Continued) The RTC is a low-power design that provides a time-of-day clock, a 100-year calendar, several alarm features and 242 bytes of general purpose RAM. An external battery is used to maintain the time and contents of the general purpose RAM, when power is removed from the PC87306. The PC87306 RTC is compatible with the DS1287 and MC146818 RTC devices. The PC87306 FDC uses a high performance analog data separator eliminating need for any external filter components. The FDC is fully compatible with the PC8477 and incorporates a superset of DP8473, NEC mPD765 and N82077 floppy disk controller function. All popular 5.25× and 3.5× floppy drives, including 2.88 MB 3.5× floppy drive, are supported. Full TDR support for PC-AT and PS/2 floppy drive types is also provided. The two UARTs are fully NS16450 and NS16550 compatible. The parallel port is fully IEEE 1284 level 2 compatible. The SPP (Standard Parallel Port) is fully compatible with ISA, EISA and MicroChannel parallel ports. In addition to the SPP, EPP (Enhanced Parallel Port) and ECP (Extended Capabilities Port) modes are supported by the parallel port. All IDE control signals with DMA support, including support for Type F DMA are provided by the PC87306. Only external signal buffers are required to implement a complete IDE interface. A set of fourteen configuration registers are provided to control various functions of the PC87306. These registers are accessed using two 8-bit wide index and data registers. The ISA I/O address of the register pair can be relocated using a power-up strapping option. Two general purpose user programmable chip selects are available. These chip selects can be used to decode game port addresses. Y Y Y Y Y Y Y Y 2 Real-Time Clock: Ð DS1287, MC146818 and PC87911 compatible Ð 242 Bytes battery backed-up CMOS RAM in two banks Ð Selective lock mechanism locks any half of the RTC RAM Ð Calendar in days, day of the week, months and years with automatic leap-year adjustment Ð Time of day in seconds, minutes and hours: Ð12 or 24 hour format Ð Optional daylight savings adjustment Ð BCD or binary format for time keeping Ð Three individually maskable interrupt event flags: ÐPeriodic rates from 122 ms to 500 ms ÐTime-of-day alarm once per second to once per day Ð Separate battery pin, 2.4V operation Ð 2 mA typical power consumption Ð Double buffer time registers UARTs: Ð Software compatible with the PC16550A and NS16450 Ð IrDA Infrared, and HP SIR Interface using UART2 with dedicated pins Parallel Port: Ð EPP, ECP compatible with ECP level 2 support Ð Bi-directional data transfer under software or hardware control Ð Includes protection circuit to prevent damage to the parallel port when a connected printer is powered up or is operated at a higher voltage IDE Control Logic: Ð All IDE control signals, with DMA and support for Type F DMA. Only external signal buffers are required to implement the full IDE interface General Purpose Pins: Ð Separate pins for two user programmable chip select decoders provide ability to control a game port Ð 16 additional general purpose I/O ports Address Decoder: Ð Provides selection of all primary and secondary ISA addresses including COM1 – 4 Plug and Play: Ð Flexible IRQs and DMAs to meet the Plug and Play requirements of Microsoft’s PC ’95 Hardware Design Guide Ð Multi-programmable parallel port base address General: Ð ISA, EISA and MicroChannel compatible architecture Ð Low power CMOS technology Ð Ability to stop clocks to all modules Ð The PC87323, which includes RAM KBC, can be used as a development platform for KBC code for the PC87306 Ð Reduced pin leakage current Ð Special configuration register for power-down Ð Disable bit for RTC Ð 160-pin PQFP package Table of Contents 3.0 FDC REGISTER DESCRIPTION (Continued) 1.0 PIN DESCRIPTION 3.2.3 Status Register 2 (ST2) 2.0 CONFIGURATION REGISTERS 3.2.4 Status Register 3 (ST3) 2.1 Overview 4.0 FDC COMMAND SET DESCRIPTION 2.2 Software Configuration 4.1 Command Descriptions 2.3 Hardware Configuration 4.1.1 Configure Command 4.1.2 Dumpreg Command 4.1.3 Format Track Command 4.1.4 Invalid Command 4.1.5 Lock Command 4.1.6 Mode Command 4.1.7 NSC Command 4.1.8 Perpendicular Mode Command 4.1.9 Read Data Command 4.1.10 Read Deleted Data Command 4.1.11 Read ID Command 4.1.12 Read A Track Command 4.1.13 Recalibrate Command 4.1.14 Relative Seek Command 4.1.15 Scan Commands 4.1.16 Seek Command 4.1.17 Sense Drive Status Command 4.1.18 Sense Interrupt Command 4.1.19 Set Track Command 4.1.20 Specify Command 4.1.21 Verify Command 4.1.22 Version Command 4.1.23 Write Data Command 4.1.24 Write Deleted Data Command 2.4 Index and Data Registers 2.5 Base Configuration Registers 2.5.1 2.5.2 2.5.3 2.5.4 2.5.5 2.5.6 2.5.7 2.5.8 Function Enable Register Function Address Register Power and Test Register Function Control Register Printer Control Register KBC and RTC Control Register Power Management Control Register Tape, UARTs and Parallel Port Configuration Register 2.5.9 SuperI/O Identification Register 2.5.10 Advanced SuperI/O Configuration Register 2.5.11 Chip Select 0 Low Address 2.5.12 Chip Select 0 High Address 2.5.13 Chip Select 0 Configuration Register 2.5.14 2.5.15 2.5.16 2.5.17 2.5.18 2.5.19 2.5.20 2.5.21 2.5.22 2.5.23 Chip Select 1 Low Address Chip Select 1 High Address Chip Select 1 Configuration Register InfraRed Configuration Register General Purpose I/O Port Base Address Configuration Register SuperI/O Configuration Register 0 SuperI/O Configuration Register 1 LPT Base Address Register Plug and Play Configuration 0 Register Plug and Play Configuration 1 Register 4.2 Command Set Summary 4.3 Mnemonic Definitions for FDC Commands 5.0 FDC FUNCTIONAL DESCRIPTION 2.6 Power-Down Options 5.1 Microprocessor Interface 2.7 Power-Up Procedure and Considerations 5.2 Modes of Operation 2.7.1 UART Power-Up 2.7.2 FDC Power-Up 5.3 Controller Phases 5.3.1 Command Phase 5.3.2 Execution Phase 5.3.2.1 DMA ModeÐFIFO Disabled 5.3.2.2 DMA ModeÐFIFO Enabled 5.3.2.3 Interrupt ModeÐFIFO Disabled 5.3.2.4 Interrupt ModeÐFIFO Enabled 5.3.2.5 Software Polling 3.0 FDC REGISTER DESCRIPTION 3.1 FDC Control Registers 3.1.1 3.1.2 3.1.3 3.1.4 3.1.5 3.1.6 3.1.7 3.1.8 3.1.9 Status Register A (SRA) Read Only Status Register B (SRB) Read Only Digital Output Register (DOR) Read/Write Tape Drive Register (TDR) Read/Write Main Status Register (MSR) Read Only Data Rate Select Register (DSR) Write Only Data Register (FIFO) Read/Write Digital Input Register (DIR) Read Only Configuration Control Register (CCR) Write Only 5.3.3 Result Phase 5.3.4 Idle Phase 5.3.5 Drive Polling Phase 5.4 Data Separator 5.5 Crystal Oscillator 5.6 Perpendicular Recording Mode 3.2 Result Phase Status Registers 5.7 Data Rate Selection 3.2.1 Status Register 0 (ST0) 3.2.2 Status Register 1 (ST1) 5.8 Write Precompensation 3 Table of Contents (Continued) 5.0 FDC FUNCTIONAL DESCRIPTION (Continued) 10.0 KEYBOARD CONTROLLER AND REAL-TIME CLOCK (Continued) 5.9 FDC Low Power Mode Logic 10.1.6 Interrupts 10.1.7 Oscillator and Instruction Timing 5.10 Reset Operation 6.0 SERIAL PORTS 10.2 Real-Time Clock Function 6.1 Serial Port Registers 10.2.1 10.2.2 10.2.3 10.2.4 10.2.5 10.2.6 10.2.7 Memory Map Bus Interface Time Generation Time Keeping RAM Power Management System Bus Lock Out and Power-Up Detection 10.2.8 Oscillator 10.2.9 Interrupt Handling 10.2.10 Control Registers 6.2 Line Control Register (LCR) Read/Write 6.3 Programmable Baud Rate Generator 6.4 Line Status Register (LSR) 6.5 FIFO Control Register (FCR) 6.6 Interrupt Identification Register (IIR) 6.7 Interrupt Enable Register (IER) 6.8 MODEM Control Register (MCR) 6.9 MODEM Status Register (MSR) 6.10 Scratchpad Register (SCR) 7.0 SERIAL INFRARED WIRELESS COMMUNICATION PORT 11.0 GENERAL PURPOSE INPUT AND OUTPUT (GPIO) PORTS 8.0 PARALLEL PORT 12.0 ELECTRICAL CHARACTERISTICS 8.1 Introduction 12.1 DC Electrical Characteristics 8.2 Data Register (DTR) 12.1.1 Microprocessor, Parallel Port, and IDE Interface Pins 12.1.2 Disk Interface Pins 12.1.3 Oscillator Pin 12.1.4 Parallel Port Pins 12.1.5 GPIO Pins 12.1.6 Keyboard Controller and Real-Time Clock Pins 8.3 Status Register (STR) 8.4 Control Register (CTR) 8.5 Enhanced Parallel Port Operation 8.6 Extended Capabilities Parallel Port (ECP) 8.6.1 Introduction 8.6.2 Software Operation 8.7 Register Definitions 12.2 AC Electrical Characteristics 8.8 Software Controlled Data Transfer 12.2.1 12.2.2 12.2.3 12.2.4 12.2.5 12.2.6 12.2.7 12.2.8 8.9 Automatic Data Transfer 8.9.1 8.9.2 8.9.3 8.9.4 Forward Direction ECP Forward Write Cycle Backward Direction ECP Backward Read Cycle 8.10 FIFO Test Access 8.11 Configuration Registers Access 8.12 Interrupt Generation 9.0 INTEGRATED DEVICE ELECTRONICS INTERFACE (IDE) 12.2.9 Reset Timing 12.2.10 FDC Write Data Timing 12.2.11 FDC Read Data Timing 12.2.12 Drive Control Timing 12.2.13 IDE Timing 12.2.14 Parallel Port Timing 12.2.15 Enhanced Parallel Port Timing 12.2.16 Extended Capabilities Port Timing 12.2.16.1 Forward 12.2.16.2 Backward 9.1 Introduction 9.2 IDE Signals 10.0 KEYBOARD CONTROLLER AND REAL-TIME CLOCK 10.1 PC87306 KBC Function 10.1.1 10.1.2 10.1.3 10.1.4 10.1.5 AC Test Conditions Clock Timing Microprocessor Interface Timing Baud Out Timing Transmitter Timing Receiver Timing MODEM Control Timing DMA Timing 12.2.8.1 FDC 12.2.8.2 ECP Host System Interface Program Memory Data Memory and Registers I/O Interface Timer/Counter 12.2.17 GPIO Write Timing 12.2.18 RTC 12.2.19 Programmable Chip Select Timing 4 List of Figures FIGURE 1-1 Pins Which Utilize the Strap Function during Reset FIGURE 1-2 Multi-Function Pins FIGURE 2-1 FIGURE 2-2 FIGURE 3-1 FIGURE 4-1 FIGURE 4-2 FIGURE 5-1 FIGURE 5-2 FIGURE 5-3 FIGURE 5-4 FIGURE 6-1 FIGURE 6-2 FIGURE 7-1 FIGURE 8-1 FIGURE 8-2 FIGURE 8-3 FIGURE 8-4 FIGURE 8-5 FIGURE 8-6 FIGURE 8-7 FIGURE 9-1 FIGURE 10-1 FIGURE 10-2 FIGURE 10-3 FIGURE 10-4 FIGURE 10-5 FIGURE 10-7 FIGURE 10-6 FIGURE 10-8 FIGURE 10-9 FIGURE 10-10 FIGURE 10-11 FIGURE 10-12 FIGURE 10-13 FIGURE 10-14 FIGURE 10-15 FIGURE 10-16 FIGURE 11-1 FIGURE 12-1 FIGURE 12-2 FIGURE 12-3 FIGURE 12-4 FIGURE 12-5 FIGURE 12-6 FIGURE 12-7 FIGURE 12-8 FIGURE 12-9 FIGURE 12-10 FIGURE 12-11 FIGURE 12-12 FIGURE 12-13 FIGURE 12-14 PC87306 Configuration Registers PC87306 Four Floppy Drive Circuit FDC Functional Block Diagram FDC Command Structure IBM, Perpendicular, and ISO Formats Supported by the Format Command FDC Data Separator Block Diagram PC87306 Dynamic Window Margin Performance Read Data AlgorithmÐState Diagram Perpendicular Recording Drive R/W Head and Pre-Erase Head PC87306 Composite Serial Data Receiver FIFO Trigger Level UART2 Serial and IR Interface Block Diagram EPP 1.7 Address Write EPP 1.7 Address Read EPP Write with ZWS EPP 1.9 Address Write EPP 1.9 Address Read ECP Forward Write Cycle ECP Backward Read Cycle IDE Interface Signal Equations (Non-DMA) Keyboard Controller Functional Block Diagram Keyboard Controller to Host System Interface Status Register Fast IRQ Latching and Clearing Keyboard Controller Data Memory Map PSW Register Bits Keyboard Controller Stack Organization Active Pull-Up I/O Port Structure Using Port Pins as Inputs Timing Generation and Timer Circuit External Clock Connection Instruction Cycle Timing Oscillator Internal and External Circuitry Interrupt/Status Timing Typical Battery Configuration Typical Battery Current during Battery Backed Mode General Purpose I/O (GPIO) Ports Clock Timing Microprocessor Read Timing Microprocessor Write Timing Read after Write Operation to All Registers and RAM Timing Baud Out Timing Transmitter Timing Receiver Timing FIFO Mode Receiver Timing Timeout Receiver Timing MODEM Control Timing FDC DMA Timing ECP DMA Timing Reset Timing Write Data Timing 5 List of Figures (Continued) FIGURE 12-15 Read Data Timing FIGURE 12-16 Drive Control Timing FIGURE 12-17 FIGURE 12-18 FIGURE 12-19 FIGURE 12-20 FIGURE 12-21 FIGURE 12-22 FIGURE 12-23 FIGURE 12-24 FIGURE 12-25 FIGURE 12-26 FIGURE 12-27 IDE Timing Compatible Mode Parallel Port Interrupt Timing Extended Mode Parallel Port Interrupt Timing Typical Parallel Port Data Exchange Enhanced Parallel Port Timing ECP Parallel Port Forward Timing Diagram ECP Parallel Port Backward Timing Diagram GPIO Write Timing IRQ Release Delay MR Timing Chip Select Timing List of Tables TABLE 1-1 TABLE 2-1 TABLE 2-2 TABLE 2-3 TABLE 2-4 TABLE 2-5 TABLE 2-6 TABLE 2-7 TABLE 2-8 TABLE 2-9 TABLE 2-10 TABLE 2-11 TABLE 3-1 TABLE 3-2 TABLE 3-3 TABLE 3-4 TABLE 3-5 TABLE 3-6 TABLE 3-7 TABLE 3-8 TABLE 4-1 TABLE 4-2 TABLE 4-3 TABLE 4-4 TABLE 4-5 TABLE 4-6 TABLE 4-7 TABLE 4-8 TABLE 4-9 TABLE 4-10 TABLE 4-11 TABLE 4-12 TABLE 4-13 TABLE 4-14 TABLE 4-15 TABLE 4-16 TABLE 4-17 Pin Descriptions (Alphabetical) Default Configurations Controlled by Hardware on Reset Index and Data Register Optional Locations Encoded Drive and Motor Pin Information (FER 4 e 1) Primary and Secondary Drive Address Selection Parallel Port Addresses COM Port Selection for UART1 COM Port Selection for UART2 Address Selection for COM3 and COM4 TRI-STATE Conditions of IRQ5 and IRQ7 TRI-STATE Conditions of IRQ3 and IRQ4 Parallel Port Mode Register Description and Addresses Drive Enable Values TDR Operation Modes Media ID Bits Functions Tape Drive Assignment Values Write Precompensation Delays Default Precompensation Delays Data Rate Select Encoding Typical Format GAP3 Length Values Based on Drive Data Rate Typical Format GAP3 Length Values Based on PC Compatible Diskette Media DENSEL Default Encoding DENSEL Encoding Head Settle Time Calculation Effect of Drive Mode and Data Rate on Format and Write Commands Effect of GAP and WG on Format and Write Commands Sector Size Selection SK Effect on the Read Data Command Result Phase Termination Values with No Error SK Effect on the Read Deleted Data Command Maximum Recalibrate Scan Command Termination Values Status Register 0 Termination Codes Set Track Register Address Step Rate Time (SRT) Values Motor Off Time (MFT) Values 6 List of Tables (Continued) TABLE 4-18 Motor On Time (MNT) Values TABLE 4-19 Verify Command Result Phase TABLE 6-1 TABLE 6-2 TABLE 6-3 TABLE 6-4 TABLE 6-5 TABLE 8-1 TABLE 8-2 TABLE 8-3 TABLE 8-4 TABLE 8-5 TABLE 8-6 TABLE 8-7 TABLE 9-1 TABLE 10-1 TABLE 10-2 TABLE 12-1 TABLE 12-2 PC87306 UART PC87306 Register Summary for an Individual UART Channel PC87306 UART Reset Configuration PC87306 UART Divisors, Baud Rates and Clock Frequencies PC87306 Interrupt Control Functions Parallel Interface Standard Parallel Port SPP Data Register Read and Write Modes Parallel Port Reset States EPP Register Addresses Parallel Port Pin Out ECP Registers IDE Registers and Their ISA Addresses Summary of System Interface Operations RTC Memory Map Nominal tICP, tDRP Values Minimum tWDW Values 7 Basic Configuration TL/C/12379 – 2 8 Connection Diagram Plastic Quad Flatpak TL/C/12379 – 3 Note: Do not connect pins marked Reserved. Order Number PC87306VUL See NS Package Number VUL160A 9 1.0 Pin Description TABLE 1-1. Pin Descriptions (Alphabetical) Symbol Pin I/O 61 – 59, 56–46, 12 – 11 I Address. These address lines from the microprocessor determine which internal register is accessed. A0–A10 are don’t cares during an DMA transfer. A10 is used only during ECP operations. ACK 127 I Parallel Port Acknowledge. This input is pulsed low by a connected printer to indicate that it has received data from the parallel port. This pin has a nominal 25 kX pull-up resistor attached to it. (See DR1 and Table 7-5 for further information.) AFD 119 I/O Parallel Port Automatic Feed. When this signal is low the printer should automatically line feed after each line is printed. This pin is in a TRI-STATEÉ condition 10 ns after a 0 is loaded into the corresponding Control Register bit. The system should pull this pin high using a 4.7 kX resistor. (See DSTRB and Table 8-5 for further information.) AEN 45 I Address Enable. This input disables function selection via A15 – A0 when it is high. Access during DMA transfer is NOT affected by this pin. ASTRB 123 O Address Strobe. This signal is used in Enhanced Parallel Port (EPP) mode as an address strobe. It is active low. (See SLIN and Table 8-5 for further information.) BADDR0,1 109, 108 I Base Address. These CMOS inputs are sensed during reset to determine one of four base addresses from which the Index and Data Registers are offset (see Table 2-2). An internal pull-down resistor of 30 kX is present on this pin. Use a 10 kX resistor to pull this pin to VDD. (See RTS1, SOUT1, and BOUT1 for further information.) BOUT1,2 108, 98 O BAUD Output. This multi-function pin provides the associated serial channel Baud Rate generator output signal when test mode is selected in the Power and Test Configuration Register and the DLAB bit (LCR7) is set. After a Master Reset, this pin provides the Serial Output (SOUT) function. (See SOUT and BADDR1 for further information.) 126 I Parallel Port Busy. This pin is set high by a connected printer when it cannot accept another character. It has a nominal 25 kX pull-down resistor attached to it. (See WAIT and Table 8-5 for further information.) 84, 106 I Configuration on Power-Up. These CMOS inputs select 1 of 4 default configurations in which the PC87306 powers-up (see Table 2-1). They are provided with CMOS input buffers. An internal pull-down resistor of 30 kX is present on each pin. Use a 10 kX resistor to pull these pins to VDD. CS0, 1 6, 23 O Programmable Chip Select. CS0, 1 are programmable chip select and/or latch enable and/ or output enable signals that can be used for a game port, I/O port expander or other add-on peripheral. The decoded address and the assertion conditions are configured via the PC87306 configuration registers, 0Ah – 0Dh and 10h – 11h. CTS1,2 107, 97 I Clear to Send. When low, this indicates that the MODEM or data set is ready to exchange data. The CTS signal is a MODEM status input whose condition the CPU can test by reading bit 4 (CTS) of the MODEM Status Register (MSR) for the appropriate serial channel. Bit 4 is the complement of the CTS signal. Bit 0 (DCTS) of the MSR indicates whether the CTS input has changed state since the previous reading of the MSR. CTS has no effect on the transmitter. 33 – 39, 42 I/O DCD1,2 112, 104 I DENSEL 77 O A0 – A15 BUSY CFG0, 1 Function Note: Whenever the DCTS bit of the MSR is set an interrupt is generated if MODEM Status interrupts are enabled. D7 – 0 Data. Bi-directional data lines to the microprocessor. D0 is the LSB and D7 is the MSB. These signals all have 24 mA (sink) buffered outputs. UARTs Data Carrier Detect. When low, this signal indicates that the MODEM or data set has detected the data carrier. The DCD signal is a MODEM status input whose condition the CPU can test by reading bit 7 (DCD) of the MODEM Status Register (MSR) for the appropriate serial channel. Bit 7 is the complement of the DCD signal. Bit 3 (DDCD) of the MSR indicates whether the DCD input has changed state since the previous reading of the MSR. Note: Whenever the DDCD bit of the MSR is set, an interrupt is generated if MODEM Status interrupts are enabled. FDC Density Select. Indicates that a high FDC density data rate (500 kbps or 1 Mbps) or a low density data rate (250 kbps or 300 kbps) is selected. DENSEL’s polarity is controlled by bit 6 of the ASC register. 10 1.0 Pin Description (Continued) TABLE 1-1. Pin Descriptions (Alphabetical) (Continued) Symbol Pin I/O Function 69 O FDC Direction. This output determines the direction of the floppy disk drive (FDD) head movement (active e step in, inactive e step out) during a seek operation. During reads or writes, DIR is inactive. DR0, 1 73, 74 O FDC Drive Select 0, 1. These are the decoded Drive Select outputs that are controlled by the Digital Output Register bits D0, D1. The Drive Select outputs are gated with DOR bits 4 – 7. These are active low outputs. They are encoded with information to control four FDDs when bit 4 of the Function Enable Register (FER) is set. (See MTR0, 1 for more information.) DRATE0, 1 82, 81 O FDC Data Rate 0, 1. These outputs reflect the currently selected FDC data rate (bits 0 and 1 in the Configuration Control Register (CCR) or the Data Rate Select Register (DSR), whichever was written to last). These pins are totem-pole buffered outputs (6 mA sink, 6 mA source). DSKCHG 71 I FDC Disk Change. This input indicates if the drive door has been opened. The state of this pin is available from the Digital Input register. This pin can also be configured as the Read Gate (RGATE) data separator diagnostic input via the Mode command (see Section 4.2.6). DSR1, 2 111, 103 I UARTs Data Set Ready. When low, this signal indicates that the data set or MODEM is ready to establish a communications link. The DSR signal is a MODEM status input whose condition the CPU can test by reading bit 5 (DSR) of the MODEM Status Register (MSR) for the appropriate channel. Bit 5 is the complement of the DSR signal. Bit 1 (DDSR) of the MSR indicates whether the DSR input has changed state since the previous reading of the MSR. DSTRB 119 O Data Strobe. This signal is used in EPP mode as a data strobe. It is active low. (See AFD and Table 8-5 for further information.) DTR1, 2 106, 96 O UARTs Data Terminal Ready. When low, this output indicates to the MODEM or data set that the UART is ready to establish a communications link. The DTR signal can be set to an active low by programming bit 0 (DTR) of the MODEM Control Register to a high level. A Master Reset operation sets this signal to its inactive (high) state. Loop mode operation holds this signal to its inactive state. (See CFG0, 1 for further information.) ERR 120 I Error. A connected printer sets this input low when it has detected an error. This pin has a nominal 25 kX pull-up resistor attached to it. FDACK 28 I FDC DMA Acknowledge. Active low input to acknowledge the FDC DMA request and enable the RD and WR inputs during a DMA transfer. When in PC-AT or Model 30 mode, this signal is enabled by bit D3 of the Digital Output Register (DOR). When in PS/2 mode, FDACK is always enabled, and bit D3 of the DOR is reserved. FDACK should be held high during I/O accesses. FDRQ 27 O DMA Request. Active high output to signal the DMA controller that a FDC data transfer is needed. When in PC-AT or Model 30 mode, this signal is enabled by bit D3 of the DOR. When in PS/2 mode, FDRQ is always enabled, and bit D3 of the DOR is reserved. GPIO10–17 3-1, 158–154 I/O General Purpose I/O10 – 17. General purpose I/O pins of I/O port 1. GPIO20–27 153–150, 118–115 I/O General Purpose I/O20 – 27. General purpose I/O pins of I/O port 2. HCS0 87 O Hard Drive Chip Select 0. This output is active in the PC-AT mode when 1) the hard drive registers from 1F0– 1F7h are selected and the primary address is used or 2) when the hard drive registers from 170– 177h are selected and the secondary address is used. This output is inactive if the IDE interface is disabled via the Configuration Register. HCS0 is multiplexed with SELCS strap input. A 40 kX internal pull-up resistor is, therefore, used on this pin during reset. (See SELCS for further information.) HCS1 86 O Hard Drive Chip Select 1. This output is active in the PC-AT mode when 1) the hard drive registers from 3F6– 7 are selected and the primary address is used or 2) the hard drive registers from 376–377 are selected and the secondary address is used. This output is also inactive, if the IDE interface is disabled via the Configuration Register. DIR Note: Whenever the DDSR bit of the MSR is set, an interrupt is generated if MODEM Status interrupts are enabled. 11 1.0 Pin Description (Continued) TABLE 1-1. Pin Descriptions (Alphabetical) (Continued) Pin I/O HDSEL Symbol 62 O FDC Head Select. This output determines which side of the FDD is accessed. When Active, the head selects side 1. When inactive, the head selects side 0. IDEACK 83 I IDE DMA Acknowledge. This is the IDE DMA acknowledge input pin when bit 1 of FCR is 1. IDED7 89 I/O IDEHI 85 O IDE High Byte. This output enables the high byte data latch during a read or write to the hard drive if the hard drive returns IOCS16. This output is inactive if the IDE interface is disabled via the Configuration Register. IDELO 84 O IDE Low Byte. This output enables the low byte data latch during a read or write to the hard drive. This output is inactive if the IDE interface is disabled via the Configuration Register. (See CFG0 for further information.) INDEX 76 I FDC Index. This input signals the beginning of a FDD track. INIT 122 I/O IOCHRDY 22 O I/O Channel Ready. This is the I/O Channel Ready open drain output. When IOCHRDY is driven low, the EPP extends the host cycle. IOCS16 88 I I/O Chip Select 16-Bit. This input is driven by a connected peripheral device which can accommodate a 16-bit access. This pin is configured when bit 2 of ASC is 0. IRRX 113 I Infrared Receiver. IRTX 114 O Infrared Transmitter. IRQ1 5 I/O 19, 18 O IRQ3, 4 Function IDE Bit 7. This pin provides the data bus bit 7 signal to the IDE hard drive during accesses in the address range 1F0–1F7h, 170–177h, 3F6h and 376h. This pin is TRI-STATE during read or write accesses to 3F7h and 377h. Initialize. When this signal is low it causes a connected printer to be initialized. This pin is in a TRISTATE condition 10 ns after a 1 is loaded into the corresponding Control Register bit. The system should pull this pin high using a 4.7 kX resistor. Interrupt 1. KBC’s keyboard interrupt generated from internal P24 of the KBC. Interrupt 3 and 4. These are active high interrupts associated with the serial ports. When bit 0 of the Plug and Play register (PNP1) is 0, IRQ3 presents the signal if the serial channel has been designated as COM2 or COM4 and IRQ4 presents the signal if the serial port is designated as COM1 or COM3. The appropriate interrupt goes active whenever it is enabled via the Interrupt Enable Register (IER), the associated Interrupt Enable bit (Modem Control Register bit 3, MCR3), and any of the following conditions are active: Receiver Error, Receive Data available, Transmitter Holding Register Empty, or a Modem Status Flag is set. The interrupt signal is reset low (inactive) after the appropriate interrupt service routine is executed, after being disabled via the IER, or after a Master Reset. Either interrupt can be disabled, putting it into TRI-STATE, by setting the MCR3 bit low. When bit 0 of the PNP1 register is 1, IRQ3 and IRQ4 are selected according to bits 2 amd 6 of the PNP1 register. See Table 2-10 for IRQ3 and IRQ4 TRI-STATE conditions. IRQ5 16 I/O Interrupt 5. Active high output that indicates a parallel port interrupt. When enabled this bit follows the ACK signal input. When bit 4 in the parallel port Control Register is set and the parallel port address is designated as shown in Table 2-5, this interrupt is enabled. When it is not enabled this signal is TRI-STATE. This pin is I/O only when ECP is enabled, IRQ5 is configured, and bit 6 of PCR is 1. When bit 4 of the PNP0 register is 1, IRQ5 and IRQ7 are selected according to bit 5 of the PNP0 register. See Table 2-9 for IRQ5 and IRQ7 TRI-STATE conditions. IRQ6 15 O IRQ7 14 I/O Interrupt 6. Active high output to signal the completion of the execution phase for certain FDC commands. Also used to signal when a data transfer is ready during a Non-DMA operation. When in PC-AT or Model 30 mode, this signal is enabled by bit D3 of the DOR. When in PS/2 mode, IRQ6 is always enabled, and bit D3 of the DOR is reserved. Interrupt 7. Active high output that indicates a parallel port interrupt. When enabled this bit follows the ACK signal input. When bit 4 in the parallel port Control Register is set and the parallel port address is designated as shown in Table 2-5, this interrupt is enabled. When it is not enabled this signal is TRI-STATE. This pin is I/O only when ECP is enabled, IRQ7 is configured, and bit 6 of PCR is 1. For ECP operation, refer to the interrupt ECP Section 7.11.1 12 1.0 Pin Description (Continued) TABLE 1-1. Pin Descriptions (Alphabetical) (Continued) Pin I/O IRQ8 Symbol 13 O Function IRQ12 4 I/O Interrupt 12. KBC’s mouse interrupt generated from internal P25 of the KBC. Interrupt 8. Real-Time Clock interrupt request output. This is an open-drain output. KBCLK 94 I/O Keyboard Clock output. Connected internally to KBC’s T0. KBDAT 93 I/O Keyboard Data output. Connected internally to KBC’s P10. MCLK 91 I/O Mouse Clock output. Connected internally to KBC’s T1. MDAT 92 I/O Mouse Data output. Connected internally to KBC’s P11. MR 20 I Master Reset. Active high input that resets the controller to the idle state, and resets all disk interface outputs to their inactive states. The DOR, DSR, CCR, Mode command, Configure command, and Lock command parameters are cleared to their default values. The Specify command parameters are not affected. The Configuration Registers are set to their selected default values. MSEN0,1 79, 78 I Media Sense. These pins are Media Sense input pins when bit 0 of FCR is 0. Each pin has a 40 kX internal pull-up resistor. MTR0, 1 75, 72 O FDC Motor Select 0, 1. These are the motor enable lines for drives 0 and 1, and are controlled by bits D7–D4 of the Digital Output register. They are active low outputs. They are encoded with information to control four FDDs when bit 4 of the Function Enable Register (FER) is set. MTR0 exchanges logical motor values with MTR1 when bit 4 of FCR is set. (See DR0,1.) P12 – P17 141–146 I/O KBC I/O Port. Quasi-bidirectional port for general purpose input and output. P20, 21 147, 148 I/O KBC I/O Port. Open-drain port for general purpose input and output. PD0 – 7 136–133, 131–128 I/O Parallel Port Data. These bidirectional pins transfer data to and from the peripheral data bus and the parallel port Data Register. These pins have high current drive capability. (See DC Electrical Characteristics.) PDACK0, 1 26, 138 I Parallel Port DMA Acknowledge. Active low input to acknowledge a connected printer’s DMA request, and enable the RD and WR inputs during a DMA transfer. These inputs are valid only in Enhanced Capabilities Port (ECP) mode. At any given moment, one of these two pins is connected. The pin which is not selected is ignored. See bit 3 of SCF1 for pin selection. PDRQ0, 1 25, 70 O Parallel Port DMA Request. Active high output which signals the DMA controller that a printer data transfer is required. These outputs are valid only in ECP mode. These pins are in TRI-STATE when not selected. At any given moment, one of these two pins is connected. See bit 3 of SCF1 for pin selection. PE 125 I Parallel Port Paper End. This input is set high by a connected printer which is out of paper. This pin has a nominal 25 kX pull-down resistor attached to it. RD 44 I Read. Active low input to signal a data read by the microprocessor. RDATA 63 I FDC Read Data. This input is the raw serial data read from the floppy disk drive. 105, 95 I Ring Indicator. When low this indicates that a telephone ring signal has been received by the MODEM. The RI signal is a MODEM status input whose condition the CPU can test by reading bit 6 (RI) of the MODEM Status Register (MSR) for the appropriate serial channel. Bit 6 is the complement of the RI signal. Bit 2 (TERI) of the MSR indicates whether the RI input has changed from low to high since the previous reading of the MSR. RI1, 2 Note: When the TERI bit of the MSR is set, an interrupt is generated if MODEM Status interrupts are enabled. RTS1, 2 109, 101 O Request to Send. When low, this output indicates to the MODEM or data set that the UART is ready to exchange data. The RTS signal can be set to an active low by programming bit 1 (RTS) of the MODEM Control Register to a high level. A Master Reset operation sets this signal to its inactive (high) state. Loop mode operation holds this signal to its inactive state. (See BADDR0 for further information.) SELCS 87 I National Strap Pin. A 40 kX internal pull-up resistor. Do not pull this pin low during reset. (See CS0CF bit 7.) 13 1.0 Pin Description (Continued) TABLE 1-1. Pin Descriptions (Alphabetical) (Continued) Symbol Pin I/O Function 110, 102 I Serial Input. This input receives composite serial data from the communications link (e.g. peripheral device, MODEM, or data set). SLCT 124 I Select. When a printer is connected, it sets this input high. This pin has a nominal 25 kX pull-down resistor attached to it. SLIN 123 I/O Select Input. When this signal is low it selects the printer. This pin is a TRI-STATE condition 10 ns after a 0 is loaded into the corresponding Control Register bit. The system should pull this pin high using a 4.7 kX resistor. 108, 98 O STB 137 I/O STEP 68 O Step. This output signal issues pulses to the disk drive at a software programmable rate to move the head during a seek operation. SIN1, 2 SOUT1, 2 Serial Output. This output sends composite serial data to the communications link (peripheral device, MODEM, or data set). The SOUT signal is set to a marking state (logic 1) after a Master Reset operation. (See BOUT and BADDR1 for further information.) Data Strobe. This output indicates to the printer that valid data is available at the printer port. This pin is in a TRI-STATE condition 10 ns after a zero is loaded into the corresponding Control Register bit. The system should pull this pin high using a 4.7 kX resistor. (See WRITE for further information.) SYSCLK 21 I System Clock. This input is used as the KBC input clock when bit 7 of KRR is 1. TC 29 I Terminal Count. Control signal from the DMA controller to indicate the termination of a DMA transfer. TC is accepted only when FDACK or PDACK is active. TC is active high in PC-AT mode and active low in PS/2 mode. TRK0 65 I Track 0. This input indicates to the controller that the head of the selected floppy disk drive is at track zero. VBAT 7 VDD 160, 140, 99, 57, 41, 17 Battery. Real-Time Clock battery pin. Digital Supply. This is the 5V supply voltage for the digital circuitry. VSS 159, 149, 139, 132, 121, 100, 90, 80, 58, 40, 32, 10 Digital Ground. This is the ground for the digital circuitry. WAIT 126 I Wait. This signal is used, in EPP mode, by the parallel port device to extend its access cycle. It is active low. (See BUSY and Table 8-5 for further information.) WR 43 I Write. Active low input to signal to indicate a write from the microprocessor to the controller. WDATA 67 O Write Data. This output is the write precompensated serial data that is written to the selected floppy disk drive. Precompensation is software selectable. WGATE 66 O Write Gate. This output signal enables the write circuitry of the selected disk drive. WGATE has been designed to prevent glitches during power up and power-down. This prevents writing to the disk when power is cycled. WP 64 I Write Protect. This input indicates that the floppy disk in the selected drive is write protected. WRITE 137 O Write Strobe. This signal is used in EPP mode as a write strobe. It is active low. (See STB and Table 8-5 for further information.) X1 30 I Clock Oscillator. A TTL or CMOS compatible 24 MHz clock is connected to this pin. X1C 8 I Crystal1 Slow. Input for the internal Real-Time Clock crystal oscillator amplifier. 14 1.0 Pin Description (Continued) TABLE 1-1. Pin Descriptions (Alphabetical) (Continued) Pin I/O X2C Symbol 9 O Crystal2 Slow. Output for the internal Real-Time Clock crystal oscillator amplifier. Function ZWS 24 O Zero Wait State. This pin is the Zero Wait State open drain output pin. ZWS is driven low when the EPP, or the ECP, is written, and the access can be shortened. Reserved 31 Reserved. This pin must be left unconnected. FIGURE 1-1. Pins Which Utilize the Strap Function During Reset Pin Symbols 84 IDEL0/CFG0 87 HCS0/SELCS 106 DTR1/CFG1 108 SOUT1/BOUT1/BADDR1 109 RTS1/BADDR0 FIGURE 1-2. Multi-Function Pins Pin Symbols 98 SOUT2/BOUT2 108 SOUT1/BOUT1 119 AFD/DSTRB 123 SLIN/ASTRB 126 BUSY/WAIT 137 STB/WRITE 15 2.0 Configuration Registers 2.1 OVERVIEW 2.2 SOFTWARE CONFIGURATION Eighteen registers constitute the Base Configuration Register set, and control the PC87306 setup. In general, these registers control the enabling of major functions (FDC, UARTs, parallel port, pin functionalty etc.), the I/O addresses of these functions, and whether they power-down via hardware control or not. These registers are the Function Enable Register (FER), the Function Address Register (FAR), the Power and Test Register (PTR), the Function Control Register (FCR), the Printer Control Register (PCR), the Keyboard and Real-Time Clock Control Register (KRR), the Power Management Control Register (PMC), the Tape, UARTs and Parallel Port Register (TUP), the SuperI/O Identification Register (SID), the Advanced SIO Configuration Register (ASC), the Chip Select 0 Address Low Register (CS0LA), the Chip Select 0 High Address Register (CS0HA), the Chip Select 0 Configuraton Register (CS0CF), the Chip Select 1 Low Address Register (CS1LA), the Chip Select 1 High Address Register (CS1HA), the Chip Select 1 Configuration Register (CS1CF), the Infrared Configuration Register (IRC), the General Purpose I/O Port Base Address Configuration Register (GPBA), and the SuperI/O Configuration Register 0 (SCF0). The FER, FAR, PTR, KRR, and SCF0 registers can be accessed via hardware or software. During reset, the PC87306 loads a set of default values, selected by a hardware strapping option, into the FER, FAR, and PTR Configuration Registers. The remaining 13 registers can only be accessed by software. An index and data register pair are used to read and write these registers. Each Configuration Register is pointed to by the value loaded into the Index Register. The data to be written into the Configuration Register is transferred via the Data register. A Configuration Register is read in a similar way (i.e., by pointing to it via the Index Register and then reading its contents via the Data Register). Accessing the Configuration Registers in this way requires only two system I/O addresses. Since that I/O space is shared by other devices the Index and Data Registers can still be inadvertantly accessed. To reduce the chances of an inadvertant access, a simple procedure (see Section 2.2) has been developed. To maintain compatibility with other SuperI/O chips, register bits with reserved values may not be altered. Use a readmodify-write procedure. If the system requires access to the Configuration Registers after reset, the following procedure must be used to change data in the registers. 1. Determine the PC87306 Index Register’s default location. Check the four possible locations (see Table 2-1) by reading them twice. The first byte is the ID byte 88h. The second byte read is always 00h, but read after write always brings the value of the written byte. Compare the data read with the ID byte and then 00h. A match occurs at the correct location. Note that the ID byte is only issued from the Index Register during the first read after a reset. Subsequent reads return the value loaded into the Index Register. Bits 5 – 7 are reserved and always read 0. 2. Load the Configuration Registers. A. Disable CPU interrupts. B. Write the index of the Configuration Register (00h – 0Dh) to the Index Register one time. C. Write the correct data for the Configuration Register in two consecutive write accesses to the Data Register. D. Enable CPU interrupts. 3. Load the Configuration Registers (read-modify-write). A. Disable CPU interrupts. B. Write the index of the Configuration Register (00h – 0Dh) to the Index Register one time. C. Read the configuration data in that register via the Data Register. D. Modify the configuration data. E. Write the changed data for the Configuration Register in two consecutive writes to the Data Register. The register updates on the second consecutive write. F. Enable CPU interrupts. A single read access to the Index and Data Registers can be done at any time without disabling CPU interrupts. When the Index Register is read, the last value loaded into the Index Register is returned. When the Data Register is read, the Configuration Register data pointed to by the Index Register is returned. 16 2.0 Configuration Registers (Continued) TL/C/12379 – 4 TL/C/12379 – 5 TL/C/12379 – 6 TL/C/12379 – 7 TL/C/12379 – 9 TL/C/12379 – 8 TL/C/12379 – 11 TL/C/12379 – 10 TL/C/12379 – 13 TL/C/12379 – 12 FIGURE 2-1. PC87306 Configuration Registers 17 2.0 Configuration Registers (Continued) TL/C/12379–14 TL/C/12379 – 19 TL/C/12379–15 TL/C/12379 – 20 TL/C/12379–16 TL/C/12379 – 21 TL/C/12379–17 TL/C/12379 – 22 TL/C/12379–18 TL/C/12379 – 97 FIGURE 2-1. PC87306 Configuration Registers (Continued) 18 2.0 Configuration Registers (Continued) TL/C/12379 – 23 TL/C/12379 – 25 TL/C/12379 – 24 FIGURE 2-1. PC87306 Configuration Registers (Continued) allows the registers to avoid conflicts with other adapters in the I/O address space. Table 2-2 shows the address options. 2.3 HARDWARE CONFIGURATION During reset, 1 of 4 possible sets of default values are loaded into the first five Configuration Registers. A strapping option on two pins (CFG0, 1) selects the set of values that is loaded. This allows for automatic configuration without software intervention. Table 2-1 shows the 4 possible default configurations. The default configuration can be modified by software at any time after reset by using the access procedure described in the Software Configuration Section. TABLE 2-2. Index and Data Register Optional Locations BADDR1 BADDR0 Index Addr. 0 0 398 399 0 1 26E 26F 1 0 15C 15D 1 1 2E 2F 2.4 INDEX AND DATA REGISTERS Another general aspect of the Configuration Registers is that the Index and the Data Register pair can be relocated to any one of two locations. This is controlled through a hardware strapping option on two pins (BADDR0,1) and it Data Addr. TABLE 2-1. Default Configurations Controlled by Hardware on Reset Configuration Pins Configuration Register Reset Values (Binary) Activated Functions on Reset CFG1 CFG0 FER FAR PTR KRR SCF0 0 0 01001011 00000001 00000x00 0x001101 0x110000 PRI (FDC), PRI (IDE), LPTA, COM1, GPIO KBC, RTC 0 1 1 0 00001111 00010001 00000x00 0x001101 0x000000 PRI (FDC), LPTA, COM1, COM2 (non-IR), KBC, RTC 1 1 00000000 00010000 00000x00 0x000001 0x000000 All Modules Disabled (powered-down), except KBC Reserved Mode Where: PRI is the PRImary floppy or IDE address; 3F0 – 7h or 1F0–7, 3F6, 3F7h) COM1 is the UART address at 3F8–3FFh COM2 is the UART address at 2F8–2FFh LPTA is the Parallel Port address at 3BC–3BEh Reserved Mode GPIO KBC RTC 19 is the General Purpose I/O ports 1 and 2 addresses at 78h and 79h is the Keyboard Controller function, using an X1 d 3 clock is the Real-Time Clock function 2.0 Configuration Registers (Continued) 2.5 BASE CONFIGURATION REGISTERS 2.5.1 Function Enable Register (FER, Index 00h) This register enables and disables major chip functions. Disabled functions have their clocks automatically powered down, but the data in their registers remains intact. It also selects whether the FDC and the IDE controller is located at their primary or secondary address. See Table 2-1 for the FER reset value. Bit 0 When this bit is one the parallel port can be accessed at the address specified in the FAR. Bit 1 When this bit is one, UART1 can be accessed at the address specified in the FAR. When this bit is zero, access to UART1 is blocked and it is in power-down mode. The UART1 registers retain all data in power down mode. Caution: Any UART1 interrupt that is enabled and active or becomes active after UART1 is disabled asserts the associated IRQ pin when UART1 is disabled. If disabling UART1 via software, clear the IRQ Enable bit (MCR3) to zero before clearing FER 1. This is not an issue after reset because MCR3 is zero until it is written. Bit 2 When this bit is one, UART2 can be accessed at the address specified in the FAR. When this bit is zero, access to UART2 is blocked and it is in power-down mode. The UART2 registers retain all data in power down mode. Caution: Any UART2 interrupt that is enabled and active or becomes active after UART2 is disabled asserts the associated IRQ pin when UART1 is disabled Bit 3 Bit 4 Bit 5 Bit 6 Bit 7 If disabling UART2 via software, clear the IRQ Enable bit (MCR3) to zero before clearing FER1. This is not an issue after reset because MCR3 is zero until it is written. When this bit is one, the FDC can be accessed at the address specified in the FER bits. When this bit is zero access to the FDC is blocked and it is in powerdown mode. The FDC registers retain all data in power down mode. When this bit is zero the PC87306 can control two floppy disk drives directly without an external decoder. When this bit is one the two drive select signals and two motor enable signals from the FDC are encoded so that four floppy disk drives can be controlled (see Table 2-3 and Figure 2-2 ). Controlling four FDDs requires an external decoder. The pin states shown in Table 2-3 are a direct result of the bit patterns shown. All other bit patterns produce pin states that should not be decoded to enable any drive or motor. This bit selects the primary or secondary FDC address. (See Table 2-4.) When this bit is a one the IDE drive interface can be accessed at the address specified by FER bit 7. When it is zero, bit 0 of PMC determines whether the HCS0,1 pins are inactive, or in TRI-STATE. IDEHI and IDEHLO are inactive and IDED7 is in TRISTATE. This bit selects the primary or secondary IDE address. (See Table 2-4). TABLE 2-3. Encoded Drive and Motor Pin Information (FER 4 e 1) 7 X X X 1 X X X 0 6 X X 1 X X X 0 X Digital Output Register 5 4 3 2 X 1 X X 1 X X X X X X X X X X X X 0 X X 0 X X X X X X X X X X X 1 0 0 1 1 0 0 1 1 0 0 1 0 1 0 1 0 1 MTR1 (Note) (Note) (Note) (Note) (Note) (Note) (Note) (Note) Drive Control Pins MTR0 DR1 0 0 0 0 0 1 0 1 1 0 1 0 1 1 1 1 DR0 0 1 0 1 0 1 0 1 Decoded Functions Activate Drive 0 and Motor 0 Activate Drive 1 and Motor 1 Activate Drive 2 and Motor 2 Activate Drive 3 and Motor 3 Activate Drive 0 and Deactivate Motor 0 Activate Drive 1 and Deactivate Motor 1 Activate Drive 2 and Deactivate Motor 2 Activate Drive 3 and Deactivate Motor 3 Note: When FER4 e 1, MTR1 presents a pulse that is the inverted image of the IOW strobe. This inverted pulse is active whenever an I/O write to address 3F2h or 372h takes place. This pulse is delayed by 25 ns–80 ns after the leading edge of IOW and its leading edge can be used to clock data into an external latch (e.g., 74LS175). Address 3F2h is used if the FDC is located at the primary address (FER5 e 0) and address 372h is used if the FDC is located at the secondary address (FER5 e 1). Hex Buffers ICC e 40 mA open collector FIGURE 2-2. PC87306 Four Floppy Drive Circuit 20 TL/C/12379 – 26 2.0 Configuration Registers (Continued) Bits 2–5 These bits determine which ISA I/O address range is associated with each UART (see Tables 2-6, 2-7 and 2-9). TABLE 2-4. Primary and Secondary Drive Address Selection Bit 5 Bit 7 Drive PC-AT Mode 0 X FDC Primary, 3F0– 7h 1 X FDC Secondary, 3F0– 7h X 0 IDE Primary, 1F0–7, 3F6, 3F7h X 1 IDE TABLE 2-6. COM Port Selection for UART1 FAR Bit 2 0 0 1 (3F8-F) 0 1 2 (2F8-F) 1 0 3 (Table 2-8)* 1 1 4 (Table 2-8)* Secondary 170–7, 376, 7h 2.5.2 Function Address Register (FAR, Index e 01h) This register selects the ISA I/O address range to which each peripheral function responds. See Table 2-1 for its reset value. UART1 Bit 3 COMÝ *Note: COM3 and COM4 addresses are determined by Bits 6 and 7. TABLE 2-7. COM Port Selection for UART2 FAR UART2 Bits 0,1 These bits select the parallel port address as shown in Table 2-5 and Table 2-9: Bit 5 Bit 4 COMÝ TABLE 2-5. Parallel Port Addresses 0 0 1 (3F8-F) 0 1 2 (2F8-F) 1 0 3 (Table 2-8)* 1 1 4 (Table 2-8)* PNP0 PNP0 FAR FAR Bit 4 Bit 5 Bit 1 Bit 0 0 X 0 0 Parallel Port Address LPTB (378–37F) 0 X 0 1 LPTA (3BC–3BE) (Note 2) 0 X 1 0 LPTC (278–27F) 0 X 1 1 Reserved 1 1 0 1 X X X X PC-AT Interrupt IRQ5 (Note 1) *Note: COM3 and COM4 addresses are determined by Bits 6 and 7. IRQ7 Bits 6, 7 These bits select the addresses that are used for COM3 and COM4 (see Table 2-8). IRQ5 TABLE 2-8. Address Selection for COM3 and COM4 TRI-STATE (CTR4 e 0) LPTB (378–37F) (Note 2) LTPC (278–27F) (Note 2) IRQ5 IRQ7 Bit 7 Bit 6 COM3 IRQ4 COM4 IRQ3 0 0 3E8 –Fh 2E8 –Fh 0 1 338 –Fh 238 –Fh 1 0 2E8 –Fh 2E0 –7h 1 1 220 –7h 228 –Fh Note 1: The interrupt assigned to this address can be changed to IRQ7 by setting Bit 3 of the power and test register. Note 2: The parallel port address is selected according to bits 0 and 1 of FAR or bit 6 of PNP0. TABLE 2-9. TRI-STATE Conditions of IRQ5 and IRQ7 Bit 4 of PNP0 e 0 Bit 4 of PNP0 e 1 IRQ5 ((FAR.bit1 e 0) and (FAR.bit0 e 1)) or ((FAR.bit1 e 0) and (FAR.bit0 e 0)) and (PTR.bit3 e 1)) or (CTR.bit4 e 0 and PCR.bit2 e 0) (PNP0.bit5 e 1) or (FER.bit0 e 0) or (CTR.bit4 e 0 and PCR.bit2 e 0) IRQ7 ((FAR.bit1 e 0) and (FAR.bit0 e 0)) and (PTR.bit3 e 0)) or ((FAR.bit1 e 1) and (FAR.bit0 e 0)) or (CTR.bit4 e 0 and PCR.bit2 e 0) (PNP0.bit5 e 0) or (FER.bit0 e 0) or (CTR.bit4 e 0 and PCR.bit2 e 0) 21 2.0 Configuration Registers (Continued) Set this bit to 0 for Compatible mode, Pulse Interrupt. Set this bit to 1 for Extended mode, Level Interrupt. TABLE 2-10. TRI-STATE Conditions of IRQ3 and IRQ4 IRQ3 IRQ4 Bit 0 of PNP1 e 0 Bit 0 of PNP1 e 1 According to FAR selection and bits 3, 4 of MCR [(PNP1.bit2 e 1) or (FER.bit1 e 0) or (MCR1.bit3 e 0) or (MCR1.bit4 e 1)] and [(PNP1.bit6 e 1) or (FER.bit2 e 0) or (MCR2.bit3 e 0) or (MCR2.bit4 e 1)] According to FAR selection and bits 3, 4 of MCR Note: Parallel port interrupt (Pulse/Level) in EPP and ECP modes is always pulse. This bit is ignored in ECP and EPP modes. 2.5.4 Function Control Register (FCR, Index e 03h) This register enables the ZWS option when in Enhanced Parallel Port mode. On reset the FCR is initialized to X00XXX01. Bit 0 TDR Mode Select bit. This bit selects the TDR mode when ASC2 e 0 as follows: 0: Automatic Media Sense TDR (PC87322 type). 1: PC-AT Compatible TDR (PC87312 type). This bit is ignored when ASC2 e 1. See ASC bit 2 for complete TDR mode selection. Bit 1 IDE DMA Enable Bit. When this bit is 0, the IDE DMA is disabled. When this bit is 1, the IDE DMA is enabled. Bit 2 Reserved. Bit 3 Reserved. Bit 4 Reserved. [(PNP1.bit2 e 0) or (FER.bit1 e 0) or (MCR1.bit3 e 0) or (MCR1.bit4 e 1)] and [(PNP1.bit6 e 0) or (FER.bit2 e 0) or (MCR2.bit3 e 0) or (MCR2.bit4 e 1)] 2.5.3 Power and Test Register (PTR, Index e 02h) This register determines several power-down features: the power-down method used when the power-down pin (PWDN) is asserted (crystal and clocks vs clocks only), whether hardware power-down is enabled, and provides a bit for software power-down of all enabled functions. It selects whether IRQ7 or IRQ5 is associated with LPTB. It puts the enabled UARTs into their test mode. See Table 2-1 for its reset value. Independent of this register the floppy disk controller can enter low power mode via the Mode Command or the Data Rate Select Register. Bit 0 Setting this bit causes all enabled functions to be powered down. Bit 1 Reserved. Bit 2 Reserved. Bit 3 Setting this bit associates the parallel port with IRQ7 when the address for the parallel port is 378–37Fh (LPTB). This bit is a ‘‘don’t care’’ when the parallel port address is 3BC–3BEh (LPTA) or 278–27Fh (LPTC). When bit 4 of PNP0 is 1, this bit is ignored. Bit 4 Setting this bit puts UART1 into a test mode, which causes its Baud Out clock to be present on its SOUT1 pin if the Line Control Register bit 7 is set to 1. Bit 5 Setting this bit puts UART2 into a test mode, which causes its Baud Out clock to be present on its SOUT2 pin if the Line Control Register bit 7 is set to 1. Bit 6 Setting this bit to a 1 prevents all further write accesses to the Configuration Registers. Once it is set by software it can only be cleared by a hardware reset. Bit 7 When not in EPP or ECP modes, this bit controls Compatible/Extended mode, thus controlling Pulse/ Level interrupt. Zero Wait State enable bit. If this bit is 1, ZWS is driven low when the Enhanced Parallel Port (EPP), or the ECP, can accept a short host read/ write-cycle, otherwise the ZWS open drain output is not driven. EPP ZWS operation should be configured when the system’s device is fast enough to support it. Bits 6, 7 Reserved. Use Read Modified Write to change the FCR register. Bit 5 2.5.5 Printer Control Register (PCR, Index e 04h) This register enables the EPP and ECP version modes, and interrupt options. It also enables the RTC RAM write mask bit. On reset the PCR bits are cleared to 0. The parallel port mode is software configurable as shown in Table 2-11. TABLE 2-11. Parallel Port Mode Operation Mode FER Bit 0 PTR Bit 7 PCR Bit 0 PCR Bit 2 None 0 X X X Compatible 1 0 0 0 Extended 1 1 0 0 EPP 1 X 1 0 ECP 1 X 0 1 Bit 0 Bit 1 22 EPP enable bit. When this bit is 0, the EPP is disabled, and the EPP registers are not accessible (access ignored). When this bit is 1, and bit 2 of PCR is 0, the EPP is enabled. Note that the EPP should not be configured with base address 3BCh. For further information refer to bit 5 of FCR. EPP version select bit. When this bit is 0, Version 1.7 is supported, and STB, AFD, INIT, and SLIN are open drain outputs. 2.0 Configuration Registers (Continued) Bit 2 Bit 3 Bit 4 Bit 5 Bit 6 Bit 7 When this bit is 1, Version 1.9 is supported (IEEE 1284), and STB, AFD, INIT, and SLIN are push-pull outputs. This bit has the same affect on the output buffers in ECP modes 0 and 2. ECP enabIe bit. When this bit is 0 the ECP is disabled and in power-down mode. The ECP registers are not accessible (access ignored) and the ECP interrupt and DMA are inactive. When this bit is 1 the ECP is enabled. The software should change this bit to 1 only when bits 0, 1, and 2 of the existing CTR are 1, 0 and 0 respectively. ECP clock freeze control bit. In power-down modes 2 and 3: When this bit is 0, the clock provided to the ECP is stopped; and When this bit is 1, the clock provided to the ECP is not stopped. Reserved. This bit must be set to 0. Parallel port interrupt (IRQ5 or IRQ7) polarity control bit. When this bit is 0 the interrupt polarity is level high or negative pulse. When this bit is 1 the interrupt polarity is inverted. Parallel port interrupt (IRQ5 or IRQ7) open drain control bit. When this bit is 0 the configured interrupt line (IRQ5 or IRQ7) has a totem-pole output. When this bit is 1 the configured interrupt line has an open drain output (drive low, no drive high, no internal pullup). RTC RAM write mask bit. When this bit is 0, the RTC RAM is writeable. When this bit is 1, the RTC RAM is not writeable, and writes are ignored. Bit 7 KBC clock source select bit. When this bit is 0 the KBC uses the X1 clock source. When this bit is 1 the KBC uses the SYSCLK clock source. This bit enables the KBC to operate in power-down mode, even when the X1 clock is frozen. It may be modified only when the KBC is disabled via bit 0 of KRR. See Table 2-1. 2.5.7 Power Management Control Register (PMC, Index e 06h) This register controls the TRI-STATE and input pins. The PMC register is accessed through Index 06h. The PMC Register is cleared to X00XX000 on reset. Bit 0 IDE TRI-STATE Control bit 0: When this bit is 0, it does not affect the IDE pins. 1: IDE7 and HCS0,1 are in TRI-STATE, IDEHI and IDELO are inactive when either the IDE is disabled or the chip is in power-down mode. Bit 1 FDC TRI-STATE Control bit. 0: When this bit is 0, it does not affect the FDC pins. 1: The FDC outputs, except IRQ6, are in TRI-STATE when either the FDC is disabled or the chip is in power down mode. Bit 2 UART TRI-STATE Control bit. 0: When this bit is 0, it does not affect the UART’s pins. 1: The outputs of any UART, except IRQ4 and IRQ3, are in TRI-STATE when that UART is disabled or the chip is in power-down. Bits 3, 4 Reserved. Bit 5 Selective Lock bit. Unlike bit 6 of PTR, which locks all configuration bits, this bit only enables locking of the following: Bit 5 of PMC, bit 4 of FER, bits 0 – 7 of FAR, bits 2, 3 of PTR, bit 1 of FCR, and bit 5 of KRR. Once this bit has been set by software, it can only be cleared by a hardware reset. It should be used instead of bit 6 of PTR if a configration bit should be dynamically modified by software (e.g., PMC bits). 0: No lock, except via bit 6 of PTR. 1: Any write to the above configuration bits is ignored (until a hardware reset, which clears this bit). Bit 6 Parallel Port TRI-STATE Control bit. 0: When this bit is 0, it does not affect the parallel port pins. 1: The parallel port outputs, except the configured IRQ line (IRQ5 or IRQ7), are in TRI-STATE when either the parallel port is disabled or the chip is in power-down mode. Bit 7 Reserved. 2.5.6 KBC and RTC Control Register (KRR, Index e 05h) This register enables and disables the keyboard controller (KBC) and the Real-Time Clock (RTC). It selects the clock source and operating mode of the KBC, selects different banks of CMOS RAM in the RTC, and selects the RTC test mode. When MR is high, KRR is initialized to 0X00XX01. Bits 2 and 3 are initialized according to CFG0. See Table 2-1 for initialization values upon reset. Bit 0 KBC Enable bit. When this bit is zero the KBC clock is frozen and the state of its dedicated pins cannot be altered. When this bit is one the KBC is functional. See Bit 2. Bit 1 KBC Speed control bit. Controls the KBC speed when X1 clock source is selected (KRR7 is 0). This bit is ignored when SYSCLK clock source is selected (KRR7 is 1). When this bit is 0 the KBC clock is the X1 frequency divided by three (typically 8 MHz). When this bit is 1 the KBC clock is the X1 frequency divided by two (typically 12 MHz). Bit 2 Reserved. This bit must be set to 1, otherwise the KBC will not be functional. Bit 3 RTC Enable bit. When this bit is 0 the RTC is disabled and IRQ8 is in TRI-STATE. When this bit is 1 the RTC is enabled. Bit 4 Reserved. Bit 5 RAMSREL. RTC CMOS RAM bank select. When this bit is 1 it selects the upper 128 bytes of CMOS RAM. When this bit is 0 it selects the lower 128 bytes of CMOS RAM. Bit 6 Reserved. 2.5.8 Tape, UARTs and Parallel Port Configuration Register (TUP, Index e 07h) The TUP Register is cleared to XXXXX0XX on reset. Bit 1 Reserved. 23 2.0 Configuration Registers (Continued) Bit 2 EPP Timeout Interrupt Enable bit. 2.5.13 Chip Select 0 Configuration Register (CS0CF, Index e 0Bh) This register controls the behavior of the CS0 pin. CS0 is asserted on non-DMA PIO cycles, when RD or WR is asserted. CS0 can be asserted three ways: 1) only on reads, 2) only on writes or 3) on all cycles. The register is initialized to 1X000XXX during reset. Bits 0 – 2 Reserved. Bit 3 Chip Select 0 Decode 0: Decode 16-bit 1: Decode address bits 15 – 12 (HA15 – HA12 of CS0HA) and CS0LA are ignored. Bit 4 0: Disable CS0 assertion on write cycles. When this bit is 0, the EPP timeout interrupt is masked. When this bit is 1, the EPP timeout interrupt is generated on the selected IRQ line (IRQ5 or IRQ7), according to bits [4:6] of PCR. Bits 3 – 7 Reserved. 2.5.9 SuperI/O Identification Register (SID, Index e 08h) The SID Register is accessed, like the other configuration registers, through the Index Register. This read-only register is used to identify the PC87306 chip. Bit 5 TL/C/12379–27 Bit 6 Bit 7 2.5.10 Advanced SuperI/O Configuration Register (ASC, Index e 09h) During reset bits 0 –2 and bit 5 are initialzed to 0, and bits 6– 7 are initialized to 1. Bit 0, 1 VLD0, 1. These bits determine the state of bit 5 in the FDC Tape Drive Register (TDR) when either Automatic Media Sense TDR or Enhanced TDR is configured (FCR0 e 0 or ASC2 e 1). For more details see the TDR description. Bit 2 Enhanced TDR support. 0: PC-AT Compatible TDR or Automatic Media Sense TDR; selected by FCR0. 1: Enhanced TDR. Bit 3 Reserved bit. On ASC writes, this bit must be written with a 0. On ASC reads, the value is undefined. Bit 4 Reserved. Bit 5 The value of this pin is reflected on bit 3 of CNFGA ECP register. Bit 6 DENSEL’s polarity control bit. 0: DENSEL is active low for 500 kbps or 1 Mbps data rates. 1: DENSEL is active high for 500 kbps or 1 Mbps data rates. Bit 7 System Operation Mode. The PC87306 can be configured to either PC-AT or PS/2 modes. 0: PS/2 mode 1: PC-AT mode Upon reset, this bit is initialized to 1; thus selecting PC-AT mode. 1: Enable CS0 assertion on write cycles. 0: Disable CS0 assertion on read cycles. 1: Enable CS0 assertion on read cycles. Reserved. CS0 pin enable. 0: Reserved for National use. 1: CS0 pin is enabled. 2.5.14 Chip Select 1 Low Address (CS1LA, Index e 0Ch) This register holds the low address bits of the monitored I/O address. Bit 0 holds A0. Upon reset this register is set to 0. See CS1CF for a description of CS1 enable on read and write cycles. 2.5.15 Chip Select 1 High Address (CS1HA, Index e 11h) This register holds the high address bits of the monitored I/O address. Bit 0 holds A8. Upon reset this register is set to 0. See CS1CF for a description of CS1 enable on read and write cycles. 2.5.16 Chip Select 1 Configuration Register (CS1CF, Index e 0Dh) This register controls the behavior of the CS1 pin. CS1 is asserted on non-DMA PIO cycles, when RD or WR is asserted. CS1 can be asserted three ways: 1) only on reads, 2) only on writes or 3) on all cycles. The register is initialized to XX00XXXX during reset. Bits 0 – 2 Reserved. Bit 3 Chip Select Decode. 0: Decode 16-bit (CS1LA and CS1HA). 1: Decode address bits 15 – 2. CS1LA bits 0 and 1 are ignored. Bit 4 Enable CS1 assertion on write cycles. Bit 5 Enable CS1 assertion on read cycles. Bits 6 – 7 Reserved. 2.5.11 Chip Select 0 Low Address (CS0LA, Index e 0Ah) This register holds the low address bits of the monitored I/O address. Bit 0 holds A0. Upon reset this register is set to 0. See CS0CF for a description of CS0 enable on read and write cycles. 2.5.17 InfraRed Configuration Register (IRC, Index e 0Eh) After reset, the register holds the value XX00XX00. Bit 0 UART2 Interface Mode. This bit is used for run time selection of either Normal (MODEM) Mode or IR Mode. 0: Normal (MODEM) Mode 1: InfraRed (IR) Mode. Bit 1 IR Half/Full Duplex configuration bit. 0: Full Duplex. Both transmitter and receiver are enabled simultaneously. 2.5.12 Chip Select 0 High Address (CS0HA, Index e 10h) This register holds the high address bits of the monitored I/O address. Bit 0 holds A8. Upon reset this register is set to 0. See CS0CF for a description of CS0 enable on read and write cycles. 24 2.0 Configuration Registers (Continued) UART2 is disabled or the PC87306 is in powerdown mode. 1: Half Duplex. The receiver input is blocked to 1 while the transmitter is busy; from the beginning of the start bit till the end of the stop bit(s). Bits 2 – 3 Reserved. Bit 4 Encoded IR transmitter drive control bit. 0: When IRTX is active, it is active for 1.6 ms. 1: When IRTX is active, it is active for 3/16 baud. In Normal (MODEM) Mode and in Non Encoded IR mode this bit is ignored. Bit 5 Encoded/Non Encoded IR mode. 0: Encoded Mode. UART2’s SOUT is encoded and routed to the IRTX pin and the IRRX pin is decoded and routed to UART2’s SIN. 1: Non Encoded Mode. UART2’s SOUT is inverted and routed to the IRTX pin and the IRRX is routed to UART2’s SIN. In Normal (MODEM) Mode this bit is ignored. Bits 6 – 7 Reserved. See Section 7 for further information. General Purpose I/O Port 1 (GPIO10 – 17) Enable. 0: GPIO Port 1 Disabled. Accesses to GPIO Port 1 is ignored. 1: GPIO Port 1 Enabled. GPIO Port 1 responds to non DMA read and write accesses. Bit 5 General Purpose I/O Port 2 (GPIO20 – 27) Enable. 0: GPIO Port 2 Disabled. Accesses to GPIO Port 2 is ignored. 1: GPIO Port 2 Enabled. GPIO Port 2 responds to non DMA read and write accesses. Bit 6 Reserved. Bit 7 Reserved. Upon reset the non reserved bits of SCF0 are initialized to either 0x000000 or 0x110000 according to the CFG1 strap pin. Bit 4 2.5.18 General Purpose I/O Port Base Address Configuration Register (GPBA, Index e 0Fh) 2.5.20 SuperI/O Configuration Register 1 (SCF1, Index e 18h) Upon reset the implemented bits are initialized to 0. Bit 0 Reserved. Bits 1, 2 Reported ECP DMA number, as reflected on bits 0 and 1 of the CNFGB ECP register. Bit 2 of SCF1 is reflected on bit 1 of CNFGB and bit 1 of SCF1 on bit 0 of CNFGB. Microsoft’s ECP Protocol and ISA Interface Standard defines these bits as follows (bit 1 of CNFGB is MSB): 00: Jumpered 8-bit DMA. 01: DMA channel 1 selected. 10: DMA channel 2 selected. 11: DMA channel 3 selected. Bit 3 Selected ECP DMA pins. 0: PDRQ0 and PDACK0 pins are used for ECP DMA. 1: PDRQ1 and PDACK1 pins are used for ECP DMA. This register holds address bits A2–9 of the GPIO Ports base address. Bit 0 holds A2. The other 8 bits of the 16-bit base address are 0. The base address is therefore composed as follows: A15 A10 0 A9 A2 GPBA A1 A0 0 GPIO Port 1 is at the base address and GPIO Port 2 at the base address a 1. Upon reset GPBA is initialized to 00011110, thus dictating base address e 78h. 2.5.19 SuperI/O Configuration Register 0 (SCF0, Index e 12h) See Table 2-1 for reset value. Bit 0 RTC RAM lock bit. Once this bit is set to 1, it can only be cleared by a hardware reset. 0: Read and write access to locations 38 – 3Fh of the RTC RAM is not blocked. 1: Read and write access to locations 38 – 3Fh of the RTC RAM is blocked. Writes are ignored and reads return FFh. Note that bit 7 of the PCR register also affects RTC RAM locking, even when this bit is 0. Bit 1 IRQ1 and IRQ12 latch enable. 0: Hardware latching is disabled (8042 compatible mode). 1: Hardware latching is enabled (IRQ latching mode). Bit 2 IRQ12 TRI-STATE control bit. 0: IRQ12 is driven with quasi-bidirectional buffer. 1: IRQ12 output is in TRI-STATE mode and IRQ12 input is masked to 1. Bit 3 UART2 TRI-STATE control bit. UART2 and IR outputs, except IRQ3 and IRQ4 pins, are in TRISTATE and the inputs are blocked to reduce their leakage current when this bit is 1 and either Note: (For bits 1–3.) It is the (PnP BIOS) softwares responsibility to match the DMA channel number, as reflected in the CNFGB register, to the selected channel (for the given wiring of those pins to the host DMA channel pins), i.e., to match the value written in bits 2–1 to the value of bit 3 according to the PC87306 wiring. Bits 4, 5 Bits 6, 7 Reserved. These bits are general purpose scratch bits. When bit 3 of the FCR is 0, the bits are read/ write bits. When bit 3 of the FCR is 1, the bits are read-only. After reset they are read/write bits. Once the bits are changed to be read-only, they can be changed back to be read/write bits only by assertion of reset. 2.5.21 LPT Base Address Register (LPTBA, Index e 19h) This register holds address bits A2 – 9 of the parallel port base address when LPTA is selected via bits 0, 1 of FAR, i.e., when bit 1 of FAR is 0 and bit 0 of FAR is 1. Bit 0 of LPTBA holds A2. The other eight bits of the 16-bit base address (i.e., A0 – 1 and A10 – 15) are 0. The base address is therefore composed as follows: 25 2.0 Configuration Registers (Continued) A15 A10 0 A9 A2 LPTBA A1 0: UART1’s interrupt, i.e., IRQ3 or IRQ4, is selected via bits 2 and 3 of FAR register. A0 0 UART2’s interrupt, i.e., IRQ3 or IRQ4, is selected via bits 4 and 5 of FAR register. The selection is dependent on the base I/O address selection. This is the backward-compatible mode with previous SIOs. 1: UART1’s interrupt, i.e., IRQ3 or IRQ4, is selected via bit 2 of PNP1 register. UART2’s interrupt, i.e., IRQ3 or IRQ4, is selected via bit 6 of PNP1 register. The selection is independent of the base I/O address. Bit 1 Reserved. Bit 2 UART1 interrupt mapping. UART1’s interrupt is routed to the following ISA interrupt: 0: IRQ3 1: IRQ4 When bit 0 of PNP1 register is 0 this bit is ignored, i.e., it does not determine UART1 interrupt mapping. Bits 3 – 5 Reserved. Bit 6 UART2 interrupt mapping. UART2’s interrupt is routed to the following ISA interrupt: 0: IRQ3 1: IRQ4 When bit 0 of PNP1 register is ‘‘0’’ this bit is ignored, i.e., it does not determine UART2 interrupt mapping. Bit 7 Reserved. It is the software’s responsibility to route UART1’s and UART2’s interrupts onto the ISA interrupts correctly. When bit 6 of PNP0 is 0, this register is read only and is forced to EFh. This register may be modified only when the parallel port is disabled. Bit 0 (holding A2) must be 0 when EPP is enabled and bit 6 of PNP0 is 1. Undefined results occur if a parallel port register is located at the same address as an enabled SIO register due to LPTBA configuration. Note: To modify LPTA base address, first disable the parallel port, then set bit 6 of PNP0 to 1, then update LPTBA, and then enable the parallel port. 2.5.22 Plug and Play Configuration 0 Register (PNP0, Index e 1Bh) Upon reset the implemented bits are initialized to 0. Bits 0 – 3 Reserved. Bit 4 Parallel Ports’ interrupt selection control. 0: Parallel Port’s interrupt, i.e., IRQ5 or IRQ7, is selected via bits 0 and 1 of FAR register and bit 3 of PTR register, i.e., the selection is dependent on the base I/O address selection. This is the backward-compatible mode with previous SIOs. 1: Parallel Port’s interrupt, i.e., IRQ5 or IRQ7, is selected via bit 5 of PNP0 register. The selection is independent of the base I/O address. Bit 5 Parallel Port interrupt mapping. The parallel port’s interrupt is routed to the following ISA interrupt: 0: IRQ5 1: IRQ7 When bit 4 of PNP0 register is ‘‘0’’ this bit is ignored, i.e., it does not determine Parallel Port interrupt mapping. Bit 6 Parallel Port’s LPTA base address selection control. This bit enables software configuration of LPTA base address via LPTBA configuration register. LPTA base address is selected when FAR.bit1 e 0 and FAR.bit0 e 1. There is no change in selection of LPTB and LPTC via bits 0–1 of FAR. This bit is initialized to 0 during reset. 0: LPTA is always 3BCh. This is the backward compatible mode with previous SIOs. 1: LPTA base address is selected via LPTBA configuration register. Bit 7 Reserved. 2.6 POWER-DOWN OPTIONS There are two methods for entering the power-down mode. Both methods result in one of the two possible modes. This section associates the methods of entering power-down with the resulting mode. Mode 1: The internal clock stops for a specific function (i.e., UART1 and/or UART2 and/or FDC). This mode is entered by any of the following actions: 1. Clear the FER bit for the specific function that is powered down. See Section 2.5.1 FER bits 1 – 3. 2. During reset, set certain CFG 0 – 4 pins. See Table 2-1. 3. Execute the FDC Mode Command with PTR bit 1 e 0 (XTAL/CLK). See Section 4.2.6 LOW PWR. 4. Set Data Rate Select Register bit 6, in the FDC, high, with PTR bit 1 e 0. See Section 3.6 bit 6. Mode 2: The internal clocks are stopped for all enabled functions. Note: Clocks to disabled functions are always inactive. 2.5.23 Plug and Play Configuration 1 Register (PNP1, Index e 1Ch) This mode is entered by any of the following actions: 1. Clear all FER bits for any enabled function. See Section 2.5.1 FER bits 1 – 3. 2. Clear PTR bits 1 (XTAL/CLK) and 2 (CSOUT/PWDN select). Then assert the PWDN signal low. See Section 2.5.3 PTR bits 1,2 and Section 1.0 PWDN pin. Upon reset the implemented bits are initialized to 0. Bit 0 UART1’s and UART2’s interrupt selection control. 26 2.0 Configuration Registers (Continued) 3.0 FDC Register Description 3. Clear PTR bit 1 and then set PTR bit 0 (power-down) high. See Section 2.5.3 PTR bits 0 and 1. The floppy disk controller is suitable for all PC-AT, EISA, PS/2, and general purpose applications. The operational mode (PC-AT, PS/2, and Model 30) of the FDC is determined by bits 7, 6 of ASC register. AT mode is the default. DP8473 and N82077 software compatibility is provided. Key features include a 16-byte FIFO, PS/2 diagnostic register support, perpendicular recording mode, CMOS disk interface, and a high performance analog data separator. See Figure 3-1 . The FDC supports the standard PC data rate drives of 250/ 500 kbps, 300/500 kbps, and 1 Mbps in MFM-encoded data mode, but is no longer guaranteed through functional testing to support the older FM encoded data mode. References to the older FM mode remain in this document to clarify the true functional operation of the device. The 1 Mbps data rate is used by the high performance tape and floppy drives. The FDC supports these floppy drives which utilize high density media, and require the perpendicular recording mode format. When used with the 1 Mbps data rate, this new format allows the use of 4 MB floppy drives which format Extra Density (ED) media to 2.88 MB capacity. The high performance internal analog data separator needs no external components. It improves on the window margin performance standards of the DP8473, and is compatible with the strict data separator requirements of floppy disk and floppy-tape drives. The FDC contains write precompensation circuitry that defaults to 125 ns for 250 kbps, 300 kbps, and 500 kbps and to 41.67 ns for 1 Mbps. These values can be overridden in software to disable write precompensationn or to provide levels of precompensation up to 250 ns. 2.7 POWER-UP PROCEDURE AND CONSIDERATIONS 2.7.1 UART Power-Up The clock signal to the UARTs is controlled through the Configuration Registers (FER, PTR). In order to restore the clock signal to one or both UARTs the following conditions must exist: 1. The appropriate enable bit (FER 1,2) for the UART(s) must be set 2. and the power-down bit (PTR 0) must not be set. 2.7.2 FDC Power-Up The clock signal to the FDC is controlled through the Configuration Registers, the FDC Mode Command and the Data Rate Select Register. In order to restore the clock signal to the FDC the following conditions must exist: 1. The appropriate enable bit (FER 3) must be set 2. and the power-down bit (PTR 0) must not be set. In addition to these conditions, one of the following must be done to initiate the recovery from power-down mode: 1. Read the Main Status Register until the RQM bit (MSR7) is set or 2. Write to the Data Rate Select Register and set the Software Reset bit (DSR7) or 3. Write to the Digital Output Register, clear and then set the Reset bit (DOR2) or 4. Read the Data Register and the Main Status Register until the RQM bit is set. If the crystal has been stopped, read the RQM bit in the Main Status Register until it is set. The RQM bit is not set until the crystal has stabilized. TL/C/12379 – 28 FIGURE 3-1. FDC Functional Block Diagram 27 3.0 FDC Register Description (Continued) The FDC has internal 24 mA data bus buffers which allow direct connection to the system bus. The internal 40 mA totem-pole disk interface buffers are compatible with both CMOS drive inputs and 150X resistor terminated disk drive inputs. D1 D0 3.1.2 Status Register B (SRB) Read Only This read-only diagnostic register is part of the PS/2 floppy controller register set, and is enabled when in the PS/2 mode. The SRB can be read at any time when in PS/2 mode. In the PC-AT mode, D7 – D0 are TRI-STATE during a mP read. 3.1 FDC CONTROL REGISTERS The following FDC registers are mapped into the addresses shown in Table 3-1 and described in the following sections. The base address range is provided by the on-chip address decoder pin. For PC-AT or PS/2 applications, the diskette controller primary address range is 3F0 to 3F7h, and the secondary address range is 370 to 377h. The FDC supports three different register modes: the PC-AT mode, PS/2 mode (Micro Channel systems), and the Model 30 mode (Model 30). See Section 5.2 for more details on how each register mode is enabled. When applicable, the register definition for each mode of operation is given. If no special notes are made, then the register is valid for all three register modes. SRBÐPS/2 Mode DESC 0 0 0 0 1 1 1 1 1 1 0 0 1 1 0 0 0 1 1 1 0 1 0 1 0 0 1 0 1 1 0 0 X X X X X X X X R R R/W R/W R W R/W X R W D7 D6 D5 Register Status Register A* SRA Status Register B* SRB Digital Output Register DOR Tape Drive Register TDR Main Status Register MSR Data Rate Select Register DSR Data Register (FIFO) FIFO None (Bus TRI-STATE) Digital Input Register DIR Configuration Control Register CCR D4 D3 D2 D1 *Note: SRA and SRB are enabled by IDENT e 0 during a chip reset only. 3.1.1 Status Register A (SRA) Read Only This is a read-only diagnostic register that is part of the PS/2 floppy controller register set, and is enabled when in the PS/2 mode. This register monitors the state of the IRQ6 pin and some of the disk interface signals. The SRA can be read at any time when in PS/2 mode. In the PC-AT mode, D7 – D0 are TRI-STATE during a mP read. D0 RESET COND D7 D6 D5 D4 D3 D2 D7 D6 D5 D4 D3 D2 D1 D0 IRQ6 PEND RES STEP TRK0 HDSEL INDX WP DIR 0 N/A 0 N/A 0 N/A N/A 0 D6 1 1 D5 D4 D3 D2 D1 D0 DR0 WDATA RDATA WGATE MTR1 MTR0 0 0 0 0 0 0 Reserved: Always 1. Reserved: Always 1. Drive Select 0: Reflects the status of the Drive Select 0 bit in the DOR (address 2, bit 0). It is cleared after a hardware reset, not a software reset. Write Data: Every inactive edge transition of the WDATA disk interface output causes this bit to change states. Read Data: Every inactive edge transition of the RDATA disk interface output causes this bit to change states. Write Gate: Active high status of the WGATE disk interface output. Motor Enable 1: Active high status of the MTR1 disk interface output. Low after a hardware reset, unaffected by a software reset. Motor Enable 0: Active high status of the MTR0 disk interface output. Low after a hardware reset, unaffected by a software reset. 3.1.3 Digital Output Register (DOR) Read/Write The DOR controls the drive select and motor enable disk interface outputs, enables the DMA logic, and contains a software reset bit. The contents of the DOR is set to 00h after a hardware reset, and is unaffected by a software reset. The DOR can be written to at any time. SRAÐPS/2 Mode DESC D7 RESET N/A N/A COND TABLE 3-1. Register Description and Addresses A2 A1 A0 IDENT R/W Write Protect: Active low status of the WP disk interface input. Direction: Active high status of the DIR disk interface output. DOR D7 DESC Interrupt Pending: This active high bit reflects the state of the IRQ6 pin. Reserved. Step: Active high status of the STEP disk interface output. Track 0: Active low status of the TRK0 disk interface input. Head Select: Active high status of the HDSEL disk interface output. Index: Active low status of the INDEX disk interface input. RESET COND D7 D6 D5 D4 28 D6 D5 D4 D3 D2 D1 D0 DRIVE DRIVE MTR3 MTR2 MTR1 MTR0 DMAEN RESET SEL 1 SEL 0 0 0 0 0 0 0 0 0 Motor Enable 3: This bit controls the MTR3 disk interface output. A 1 in this bit causes the MTR3 pin to go active. Motor Enable 2: Same function as D7 except for MTR2. Motor Enable 1: Same function as D7 except for MTR1. (See bit 4 of FCR for further information.) Motor Enable 0: Same function as D7 except for MTR0. (See bit 4 of FCR for further information.) 3.0 FDC Register Description (Continued) D3 DMA Enable: This bit has two modes of operation. 3.1.4 Tape Drive Register (TDR) PC-AT mode: Writing a 1 to this bit enables the FDRQ, FDACK, TC, and IRQ6 pins. Writing a 0 to this bit disables the FDACK and TC pins and TRISTATE the FDRQ and the IRQ6 pins. This bit is a 0 after a reset when in these modes. The TDR register is the Tape Drive Register and the floppy disk controller media and drive type register. The register has three modes of operation (see Table 3-3): PC-AT Compatible mode. The register is used to assign a particular drive number to the tape drive support mode of the data separator. All other logical drives can be assigned as floppy drive support. Bits 2 – 7 are TRI-STATE during read. Automatic Media Sense mode. Bits 5 – 7 are implemented, in addition to the bits of the Compatible PC-AT TDR mode. Bits 2 – 4 are reserved. Enhanced mode. This is the PS/2 TDR mode. It uses all the register’s bits for operation with PS/2 floppy drives. The use of the TDR bits, for each of these modes, is shown in Table 3-3. D7 Extra Density: When bit 5 is 0, this media id bit is used with bit 6 to indicate the type of media currently in the active floppy drive. If bit 5 is 1, it is invalid. This bit holds MSEN1 pin value. See Table 3-4 for details regarding bits 5 – 7. D6 High Density: When bit 5 is 0, this media id bit is used with bit 7 to indicate the type of media currently in the active floppy drive. If bit 5 is 1, it is invalid. This bit holds MSEN0 pin value. See Table 3-4 for details regarding bits 5 – 7. PS/2 mode: This bit is reserved, and the FDRQ, FDACK, TC, and IRQ6 pins are always enabled. During a reset, the FDRQ, FDACK, TC, and IRQ6 lines remain enabled, and D3 is 0. D2 Reset Controller: Writing a 0 to this bit resets the controller. It remains in the reset condition until a 1 is written to this bit. A software reset does not affect the DSR, CCR, and other bits of the DOR. A software reset affects the Configure and Mode command bits (see Section 4.0 FDC Command Set Description). The minimum time that this bit must be low is 100 ns. Thus, toggling the Reset Controller bit during consecutive writes to the DOR is an acceptable method of issuing a software reset. D1, 0 Drive Select: These two bits are binary encoded for the four drive selects DR0–DR3, so that only one drive select output is active at a time. (See bit 4 of FCR for further information.) It is common programming practice to enable both the motor enable and drive select outputs for a particular drive. Table 3-2 shows the DOR values to enable each of the four drives. Read/Write TABLE 3-2. Drive Enable Values Drive DOR Value 0 1Ch 1 2D 2 4E 3 8F Note: The MTR3, MTR2, DRV3, DRV2 pins are only available in four-drive mode (bit 4 of FER is 1) and require external logic. TABLE 3-3. TDR Operation Modes Mode PC-AT Compatible Automatic Media Sense Enhanced FCR Bit 0 1 0 0 or 1 ASC Bit 2 Bit D7 D6 D5 D4 D3 D2 D1 D0 DESC X X X X X X TAPE SEL 1 TAPE SEL 0 RESET COND N/A N/A N/A N/A N/A N/A 0 0 DESC ED HD VALID DATA X X X TAPE SEL 1 TAPE SEL 0 RESET COND N/A N/A N/A N/A N/A N/A 0 0 DESC ED HD VALID DATA X SWP1 SWP0 TAPE SEL 1 TAPE SEL 0 RESET COND N/A N/A N/A 1 0 0 0 0 0 0 1 29 3.0 FDC Register Description (Continued) D5 Valid Data: 3.1.5 Main Status Register (MSR) Automatic Media Sense mode: The state of bit 5 is determined by the state of the VLD0,1 bits in the ASC Configuration Register. If this bit is 0, there is valid media id sense data in bits 7 and 6 of this register. Bit 5 holds VLD0 when drive 0 is accessed, and media sense is configurfed. It holds VLD1 when drive 1 is accessed, and media sense is configured. Otherwise, it is set to 1 to indicate that media information is not available. Valid data should be used only when accessing drives 0 and 1. See Table 3-4 for details regarding bits 5–7. PC-AT Compatible mode: When four drive encoding is used (FER4 e 1), this bit is set to 1. The read-only Main Status Register indicates the current status of the disk controller. The Main Status Register is always available to be read. One of its functions is to control the flow of data to and from the Data Register (FIFO). The Main Status Register indicates when the disk controller is ready to send or receive data through the Data Register. It should be read before each byte is transferred to or from the Data Register except during a DMA transfer. No delay is required when reading this register after a data transfer. After a hardware or software reset, or recovery from a power-down state, the Main Status Register is immediately available to be read by the mP. It contains a value of 00h until the oscillator circuit has stabilized, and the internal registers have been initialized. When the FDC is ready to receive a new command, it reports an 80h to the mP. The system software can poll the MSR until it is ready. The worst case time allowed for the MSR to report an 80h value (RQM set) is 2.5 ms after reset or power up. TABLE 3-4. Media ID Bits Functions Bit 7 Bit 6 Bit 5 Media Type X X 1 Invalid Data 0 0 0 5.25× 0 1 0 2.88M 1 0 0 1.44M 1 D4 D3, 2 D1, 0 1 0 MSR 720k Reserved. Bits 3 and 2 are read/write bits that control bits control logical drive exchange. When working with four drives encoding (bit 4 of FER is 1) the logical drive exchange is not performed. 00: No logical drive exchange. 01: Logical drive exchange between drives 0 and 1. DR1 internal signal to DR0 pin. D7 D6 D5 D4 D3 D2 D1 D0 DESC RQM DIO NON DMA CMD PROG DRV3 BUSY DRV2 BUSY DRV1 BUSY DRV0 BUSY RESET COND 0 0 0 0 0 0 0 0 D7 MTR1 internal signal to MTR0 pin. DR0 internal signal to DR1 pin. MTR0 internal signal to MTR1 pin. 10: Logical drive exchange between drives 0 and 2. The DR0 and MTR0 pins function is exchanged as follows: DR2 internal signal to DR0 pin. D6 MTR2 internal signal to MTR0 pin. 11: Reserved. Unpredictable results when 11 is configured. Tape Select 1, 0: These bits assign a logical drive number to a tape drive. Drive 0 is not available as a tape drive and is reserved as the floppy disk boot drive. See Table 3-5 for the tape drive assignment values. D4 D5 D3 TABLE 3-5. Tape Drive Assignment Values TAPESEL1 TAPESEL0 Drive Selected 0 0 None 0 1 1 1 0 2 1 1 3 Read Only D2 D1 D0 30 Request for Master: Indicates that the controller is ready to send or receive data from the mP through the FIFO. This bit is cleared immediately after a byte transfer and is set again as soon as the disk controller is ready for the next byte. During a Non-DMA Execution phase, the RQM indicates the status of the interrupt pin. Data I/O (Direction): Indicates whether the controller is expecting a byte to be written to (0) or read from (1) the Data Register. Non-DMA Execution: Indicates that the controller is in the Execution Phase of a byte transfer operation in the Non-DMA mode. This mode can be used for multiple byte transfers by the mP in the Execution Phase via interrupts or software polling. Command in Progress: This bit is set after the first byte of the Command Phase is written. This bit is cleared after the last byte of the Result Phase is read. If there is no Result Phase in a command, the bit is cleared after the last byte of the Command Phase is written. Drive 3 Busy: Set after the last byte of the Command Phase of a Seek or Recalibrate command is issued for drive 3. Cleared after reading the first byte in the Result Phase of the Sense Interrupt Command for this drive. Drive 2 Busy: Same as above, but for drive 2. Drive 1 Busy: Same as above, but for drive 1. Drive 0 Busy: Same as above, but for drive 0. 3.0 FDC Register Description (Continued) 3.1.6 Data Rate Select Register (DSR) TABLE 3-7. Default Precompensation Delays Write Only This write-only register is used to program the data rate, amount of write precompensation, power-down mode, and software reset. The data rate is programmed via the CCR, not the DSR, for PC-AT and PS/2 Model 30 and MicroChannel applications. Other applications can set the data rate in the DSR. The data rate of the floppy controller is determined by the most recent write to either the DSR or CCR. The DSR is unaffected by a software reset. A hardware reset sets the DSR to 02h, which corresponds to the default precompensation setting and 250 kbps. Data Rate 1 Mbps 41.7 ns 500 kbps 125.0 ns 300 kbps 125.0 ns 250 kbps 125.0 ns Data Rate Select 1,0: These bits determine the data rate for the floppy controller. See Table 3-8 for the corresponding data rate for each value of D1, D0. The data rate select bits are unaffected by a software reset, and are set to 250 kbps after a hardware reset. D1, 0 DSR D7 RESET COND D7 D6 D5 D4 – 2 D6 D5 D4 D3 D2 D1 D0 S/W LOW PREPREPRE0 DRATE1 DRATE0 RESET POWER COMP2 COMP1 COMP0 DESC Precompensation Delay TABLE 3-8. Data Rate Select Encoding 0 0 0 0 0 0 1 0 Data Rate Select Software Reset: This bit has the same function as the DOR RESET (D2) except that this software reset is self-clearing. Low Power: A 1 to this bit puts the controller into the Manual Low Power mode. The oscillator and data separator circuits are turned off. Manual Low Power can also be accessed via the Mode command. The chip comes out of low power after a software reset, or access to the Data Register or Main Status Register. Undefined. Should be set to 0. Precompensation Select: These three bits select the amount of write precompensation the floppy controller uses on the WDATA disk interface output. Table 3-6 shows the amount of precompensation used for each bit pattern. In most cases, the default values (Table 3-7) can be used; however, alternate values can be chosen for specific types of drives and media. Track 0 is the default starting track number for precompensation. The starting track number can be changed in the Configure command. 0.0 ns 001 41.7 ns 010 83.3 ns 011 125.0 ns 100 166.7 ns 101 208.3 ns 110 250.0 ns 000 DEFAULT MFM 1 1 1 Mbps FM 0 0 500 kbps 250 kbps 0 1 300 kbps 150 kbps 1 0 250 kbps 125 kbps Illegal 3.1.7 Data Register (FIFO) Read/Write The FIFO (read/write) is used to transfer all commands, data, and status between the mP and the FDC. During the Command Phase, the mP writes the command bytes into the FIFO after polling the RQM and DIO bits in the MSR. During the Result Phase, the mP reads the result bytes from the FIFO after polling the RQM and DIO bits in the MSR. Enabling the FIFO, and setting the FIFO threshold, is done via the Configure command. If the FIFO is enabled, only the Execution Phase byte transfers use the 16 byte FIFO. The FIFO is always disabled during the Command and Result Phases of a controller operation. A software reset will not disable enabled FIFO if the LOCK bit is set in the Lock Command. After a hardware reset, the FIFO is disabled to maintain compatibility with PC-AT systems. The 16-byte FIFO can be used for DMA, Interrupt, or software polling type transfers during the execution of a read, write, format, or scan command. In addition, the FIFO can be put into a Burst or Non-Burst mode with the Mode command. In the Burst mode, FDRQ or IRQ6 remains active until all of the bytes have been transferred to or from the FIFO. In the Non-Burst mode, FDRQ or IRQ6 is deasserted for 350 ns to allow higher priority transfer requests to be serviced. The Mode command can also disable the FIFO for either reads or writes separately. The FIFO allows the system a larger latency without causing a disk overrun/underrun error. The FIFO is typically used with multitasking operating systems and/or when running systems at or above a 1 Mbps data rate. In its default state, the FIFO is disabled and contains a zero threshold. The default state is entered after a hardware reset. Precompensation Delay 111 0 Note: FM mode is not guaranteed by functional testing. TABLE 3-6. Write Precompensation Delays Bits 4, 3, 2 Data Rate 1 31 3.0 FDC Register Description (Continued) DIRÐPS/2 Mode Data Register (FIFO) D7 D6 D5 D4 D3 DESC Data [7:0] RESET COND Byte Mode D2 D1 D0 D7 During the Execution Phase of a command involving data transfer to/from the FIFO, the system must respond to a data transfer service request based on the following formula: RESET COND N/A D0 D7 D6 D5 D4 D3 D2 D1 DSKCHG X X X X X X X RESET COND N/A N/A N/A N/A N/A N/A N/A N/A 1 D4 1 D3 1 1 N/A N/A N/A N/A D2 D1 D0 HIGH DRATE1 DRATE0 DEN N/A N/A 1 Disk Changed: Active high status of DSKCHG disk interface input. During power-down this bit is invalid, if it is read by the software. Reserved: Always 1. Data Rate Select 1,0: These bits indicate the status of the DRATE1 – 0 bits programmed through the DSR or CCR. High Density: This bit is low when the 1 Mbps or 500 kbps data rate is chosen, and high when the 300 kbps or 250 kbps data rate is chosen. This bit is independent of the IDENT value. CCRÐPC-AT and PS/2 Modes D7 D6 D5 D4 D3 D2 D1 D0 DESC 0 0 0 0 0 0 DRATE1 DRATE0 RESET COND N/A N/A N/A N/A N/A N/A 1 0 D7 – 2 D1, 0 Reserved: Should be set to 0. Data Rate Select 1,0: These bits determine the data rate of the floppy controller. See Table 3-8 for the appropriate values. 3.2 RESULT PHASE STATUS REGISTERS The Result Phase of a command contains bytes that hold status information. The format of these bytes is described below. Do not confuse these status bytes with the Main Status Register, which is a read only register that is always valid. The Result Phase status registers are read from the Data Register (FIFO) only during the Result Phase of certain commands (see Section 4.1 Command Set Summary). The status of each register bit is indicated when the bit is a 1. DIRÐPC-AT Mode DESC D5 3.1.9 Configuration Control Register (CCR) Write Only This is the write-only data rate register commonly used in PC-AT applications. This register is not affected by a software reset, and is set to 250 kbps after a hardware reset. The data rate of the floppy controller is determined by the last write to either the CCR or DSR. 3.1.8 Digital Input Register (DIR) Read Only This diagnostic register is used to detect the state of the DSKCHG disk interface input and some diagnostic signals. The function of this register depends on its mode of operation. When in the PC-AT mode, the D6–D0 are TRI-STATE to avoid conflict with the fixed disk status register at the same address. DIR is unaffected by a software reset. D6 – 0 DSKCHG D6 – 3 D2, 1 This formula is good for all data rates with the FIFO enabled or disabled. THRESH is a four bit value programmed in the Configure command, which sets the FIFO threshold. If the FIFO is disabled, THRESH is zero in the above formula. The last term of the formula, (16 c tICP) is an inherent delay due to the microcode overhead required by the FDC. This delay is also data rate dependent. See Table 9-1 for the tDRP and tICP times. See Section 12.2.2 for a description of tDRP and tICP. The programmable FIFO threshold (THRESH) is useful in adjusting the floppy controller to the speed of the system. In other words, a slow system with a sluggish DMA transfer capability uses a high value of THRESH, giving the system more time to respond to a data transfer service request (FDRQ for DMA mode or IRQ6 for Interrupt mode). Conversely, a fast system with quick response to a data transfer service request would use a low value of THRESH. D7 DESC D7 Maximum Allowable Data Transfer Service Time [(THRESH a 1) c 8 c tDRP] b (16 c tICP) D6 D0 Disk Changed: Active high status of DSKCHG disk interface input. During power-down this bit is invalid, if it is read by the software. Unused by the FDC (at TRI-STATE). The bits are used by the Hard Disk Controller Status Register. 32 3.0 FDC Register Description (Continued) D4 3.2.1 Status Register 0 (ST0) D7 D6 D5 D4 D3 D2 D1 D0 DESC IC IC SE EC 0 HDS DS1 DS0 RESET COND 0 0 0 0 0 0 0 0 D7, 6 D5 D4 D3 D2 D1, 0 D3 D2 Interrupt Code: 00 e Normal Termination of Command. 01 e Abnormal Termination of Command. Execution of command was started, but was not successfully completed. 10 e Invalid Command Issued. Command issued was not recognized as a valid command. 11 e Internal drive ready status changed state during the drive polling mode. Only occurs after a hardware or software reset. Seek End: Seek, Relative Seek, or Recalibrate command completed by the controller. (Used during a Sense Interrupt command.) Equipment Check: After a Recalibrate command, Track 0 signal failed to occur. (Used during Sense Interrupt command.) Not Used. Always 0. Head Select: Indicates the active high status of the HDSEL pin at the end of the Execution Phase. Drive Select 1,0: These two binary encoded bits indicate the logical drive selected at the end of the Execution Phase. 00 e Drive 0 selected. D1 D0 3.2.3 Status Register 2 (ST2) D7 D6 D5 D4 D3 D2 D1 D0 DESC 0 CM CD WT SEH SNS BT MD RESET COND 0 0 0 0 0 0 0 0 D7 D6 01 e Drive 1 selected. 10 e Drive 2 selected. 11 e Drive 3 selected. 3.2.2 Status Register 1 (ST1) D7 D6 D5 D4 D3 D2 D1 D0 DESC ET 0 CE OR 0 ND NW MA RESET COND 0 0 0 0 0 0 0 0 D7 D6 D5 Overrun: Controller was not serviced by the mP soon enough during a data transfer in the Execution Phase. For read operations, indicates a data overrun. For write operations, indicates a data underrun. Not Used. Always 0. No Data: Three possible problems: 1. Controller cannot find the sector specified in the Command Phase during the execution of a Read, Write, Scan, or Verify command. An address mark was found however, so it is not a blank disk. 2. Controller cannot read any Address Fields without a CRC error during a Read ID command. 3. Controller cannot find starting sector during execution of Read A Track command. Not Writable: Write Protect pin is active when a Write or Format command is issued. Missing Address Mark: If bit 0 of ST2 is clear then the controller cannot detect any Address Field Address Mark after two disk revolutions. If bit 0 of ST2 is set then the controller cannot detect the Data Field Address Mark after finding the correct Address Field. D5 D4 End of Track: Controller transferred the last byte of the last sector without the TC pin becoming active. The last sector is the End of Track sector number programmed in the Command Phase. Not Used. Always 0. CRC Error: If this bit is set and bit 5 of ST2 is clear, then there was a CRC error in the Address Field of the correct sector. If bit 5 of ST2 is also set, then there was a CRC error in the Data Field. D3 D2 D1 33 Not Used. Always 0. Control Mark: Controller tried to read a sector which contained a deleted data address mark during execution of Read Data or Scan commands. Or, if a Read Deleted Data command was executed, a regular address mark was detected. CRC Error in Data Field: Controller detected a CRC error in the Data Field. Bit 5 of ST1 is also set. Wrong Track: Only set if desired sector is not found, and the track number recorded on any sector of the current track is different from the track address specified in the Command Phase. Scan Equal Hit: ‘‘Equal’’ condition satisfied during any Scan command. Scan Not Satisfied: Controller cannot find a sector on the track which meets the desired condition during any Scan command. Bad Track: Only set if the desired sector is not found, the track number recorded on any sector on the track is FFh indicating a hard error in IBM format, and is different from the track address specified in the Command Phase. Command Phase: 3.0 FDC Register Description (Continued) D0 Missing Address Mark in Data Field: Controller cannot find the Data Field AM during a Read, Scan, or Verify command. Bit 0 of ST1 is also set. 3.2.4 Status Register 3 (ST3) 0 0 0 1 0 0 1 1 0 0 0 0 0 0 0 0 0 EIS FIFO POLL THRESH PRETRK D7 D6 D5 D4 D3 D2 D1 D0 DESC 0 WP 1 TK0 1 HDS DS1 DS0 RESET COND 0 0 1 0 1 0 0 0 Execution Phase: Internal registers written. Result Phase: None. EIS: Enable Implied Seeks. Default after a software reset. # 0 e Implied seeks disabled through Configure comNot Used. Always 0. Write Protect: Indicates active high status of the WP pin. D5 Not Used. Always 1. D4 Track 0: Indicates active high status of the TRK0 pin. D3 Not Used. Always 1. D2 Head Select: Indicates the active high status of the HD bit in the Command Phase. D1, 0 Drive Select 1,0: These two binary encoded bits indicate the DS1–DS0 bits in the Command Phase. D7 D6 mand. Implied seeks can still be enabled through the Mode command when EIS e 0. 1 e Implied seeks enabled for a read, write, scan, or verify operation. A seek and sense interrupt operation is performed prior to the execution of the read, write, scan, or verify operation. The IPS bit does not need to be set. FIFO: Enable FIFO for Execution Phase data transfers. Default after a software reset if the LOCK bit is 0. If the LOCK bit is 1, then the FIFO bit retains its previous value after a software reset. 0 e FIFO enabled for both reads and writes. 4.0 FDC Command Set Description # 1 e FIFO disabled. POLL: Disable for Drive Polling Mode. Default after a software reset. This section presents the FDC command setÐfull description in Section 4.1 and a working summary in Section 4.2. Each command contains a unique first command byte, the opcode byte, which tells the controller how many (0 or more) command bytes to expect. The information for each command is displayed using the structure shown in Figure 4-1 . If an invalid command byte is issued to the controller, it immediately enters the Result Phase and the status is 80h signifying an Invalid Command. # 0 e Enable drive polling mode. An interrupt is generated after a reset. 1 e Disable drive polling mode. If the Configure command is issued within 500 ms of a hardware or software reset, then an interrupt is not generated. In addition, the use of the four Sense Interrupt commands to clear the ‘‘Ready Changed State’’ of the four logical drives is not required. THRESH: The FIFO threshold in the Execution Phase of read and write data transfers. Programmable from 00h to 0Fh. Defaults to 00h after a software reset if the LOCK bit is 0. If the LOCK bit is 1, THRESH retains its value. A high value of THRESH is suited for slow response systems, and a low value of THRESH is better for fast response systems. PRETRK: Starting track number for write precompensation. Programmable from track 0 (‘‘00’’) to track 255 (‘‘FF’’). Defaults to track 0 (‘‘00’’) after a software reset if the LOCK bit is 0. If the LOCK bit is 1, then PRETRK retains its value. I/O Operation Opcode Command Byte 1 Command Byte 2 . . . Command Byte n FIGURE 4-1. FDC Command Structure 4.1 COMMAND DESCRIPTIONS 4.1.1 Configure Command The Configure Command controls some operation modes of the controller. It should be issued during the initialization of the FDC after power-up. These bits are set to their default values after a hardware reset. The value of each bit after a software reset is explained. The default value of each bit is denoted by a ‘‘bullet’’ to the left of each item. 4.1.2 Dumpreg Command The Dumpreg command is designed to support system runtime diagnostics, application software development and debug. This command has a one-byte command phase and a 10-byte result phase. The Result Phase returns the values of parameters set in other commands. That is, the PTR (Present Track Register) contains the least significant byte of the track the microcode has stored for each drive. The Step Rate Time, Motor Off and Motor On Times, and the DMA bit are all set in the Specify command. 34 4.0 FDC Command Set Description (Continued) 4. The Bytes per Sector code, which determines the sector size. The sixth byte of the result phase varies depending on what commands have been previously executed. If a format command has previously been issued, and no reads or writes have been issued since then, this byte contains the Sectors per track value. If a read or a write command has been executed more recently than a format command, this byte contains the End of Track value. The LOCK bit is set in the Lock command. The eighth result byte also contains the bits programmed in the Perpendicular Mode command. The last two bytes of the Dumpreg Result Phase are set in the Configure command. After a hardware or software reset, the parameters in the result bytes are set to their appropriate default values. 5. The Sector per Track parameter, which determines how many sectors are formatted on the track. 6. The Data Pattern byte, which is used as the filler byte in the Data Field of each sector. Command Phase: 0 MFM 0 0 1 1 0 1 X X X X X HD DR1 DR0 Bytes per Sector Sectors per Track Note: Some of these parameters are unaffected by a software reset, depending on the state of the LOCK bit. See the Lock Command for further information. Format Gap Data Pattern Command Phase: 0 0 0 0 1 1 1 Execution Phase: System transfers four ID bytes (track, head, sector, bytes/sector) per sector to the floppy controller via DMA or Non-DMA modes. The entire track is formatted. The data block in the Data Field of each sector is filled with the data pattern byte. 0 Execution Phase: Internal registers read. Result Phase: Result Phase: PTR Drive 0 PTR Drive 1 Status Register 0 PTR Drive 2 Status Register 1 Status Register 2 PTR Drive 3 Step Rate Time Undefined Motor Off Time Motor On Time Undefined DMA Undefined Sector per Track/End of Track (Note) LOCK 0 DC3 DC2 0 EIS FIFO POLL DC1 DC0 GAP Undefined WG THRESH To allow for flexible formatting, the mP must supply the four Address Field bytes (track, head, sector, bytes per sector code) for each sector formatted during the Execution Phase. This allows for non-sequential sector interleaving. This transfer of bytes from the mP to the controller can be done in the DMA or Non-DMA mode, with the FIFO enabled or disabled. The Format Gap byte in the Command Phase is dependent on the data rate and type of disk drive, and controls the length of GAP3. Some typical values for the programmable GAP3 are given in Table 4-1. Figure 4-2 shows the track format for each of the formats recognized by the format command. Table 4-2 shows some typical values for the Format GAP3 based on media type. The Format command terminates when the index hole is detected a second time, at which point an interrupt is generated. Only the first three status bytes in the Result Phase are significant. PRETRK Note: Sectors per Track parameter returned if last command issued was Format. End of Track parameter returned if last command issued was Read or Write. 4.1.3 Format Track Command This command formats one track on the disk in IBM, ISO, or Perpendicular format. After the index hole is detected, data patterns are written on the disk including all gaps, Address Marks, Address Fields, and Data Fields. The exact format is determined by the following parameters: 1. The MFM bit in the Opcode (first command) byte, which determines the format of the Address Marks and the encoding scheme. 2. The IAF bit in the Mode command, which selects between IBM and ISO format. 3. The WGATE and GAP bits in the Perpendicular Mode command, which select between the conventional and Toshiba Perpendicular format. 35 4.0 FDC Command Set Description (Continued) TABLE 4-1. Typical Format GAP3 Length Values Based on Drive Data Rate Sector Size (Decimal) Sector Code (Hex) EOT (Hex) Sector Gap (Hex) (Note 1) Format GAP3 (Hex) (Note 2) 250 kbps MFM 256 256 512 512 1024 2048 4096 01 01 02 02 03 04 05 12 10 08 09 04 02 01 0A 20 2A 2A 80 C8 C8 0C 32 50 50 F0 FF FF 500 kbps MFM 256 512 512 1024 2048 4096 8192 01 02 02 03 04 05 06 1A 0F 12 08 04 02 01 0E 1B 1B 35 99 C8 C8 36 54 6C 74 FF FF FF Mode Note 1: Sector Gap refers to the Intersector Gap Length parameter specified in the Command Phase of the Read, Write, Scan, and Verify commands. Although this is the recommended value, the FDC treats this byte as a don’t care in the Read, Write, Scan, and Verify commands. Note 2: Format Gap is the suggested value to use in the Format Gap parameter of the Format command. This is the programmable GAP3 as shown in Figure 4-1 . TABLE 4-2. Typical Format GAP3 Length Values Based on PC Compatible Diskette Media Media Type Sector Size (Decimal) Sector Code (Hex) EOT (Hex Sector Gap (Hex) Format GAP3 (Hex) 360k 1.2M 720k 1.44M 2.88M (Note 3) 512 512 512 512 512 02 02 02 02 02 09 0F 09 12 24 2A 1B 1B 1B 1B 50 54 50 6C 53 Note 3: The 2.88M diskette media is a Barium Ferrite media intended for use in Perpendicular Recording drives at data rates up to 1 Mbps. 36 4.0 FDC Command Set Description (Continued) Notes: TL/C/12379 – 29 CRC uses standard polynomial x16 a x12 a x5 a 1 A1* e Data Pattern of A1, Clock Pattern of 0A C2* e Data Pattern of C2, Clock Pattern of 14 Perpendicular Format GAP2 e 41 bytes for 1 Mbps and 2 Mbps. All byte counts in decimal All other data rates use GAP2 e 22 bytes All byte values in hex FIGURE 4-2. IBM, Perpendicular, and ISO Formats Supported by the Format Command ful if the system designer wishes to keep the FIFO enabled and retain the other FIFO parameter values (such as THRESH) after a software reset. After the command byte is written, the result byte must be read before continuing to the next command. The execution of the Lock command is not performed until the result byte is read by the mP. If the part is reset after the command byte is written but before the result byte is read, then the Lock command execution is not performed. This is done to prevent accidental execution of the Lock command. 4.1.4 Invalid Command If an invalid command (illegal Opcode byte in the Command Phase) is received by the controller, the controller responds with ST0 in the Result Phase. The controller does not generate an interrupt during this condition. Bits 6 and 7 in the MSR are both set to a 1, indicating to the mP that the controller is in the Result Phase and the contents of ST0 must be read. The system reads an 80h value from ST0 indicating an invalid command was received. Command Phase: Command Phase: Invalid Op Codes LOCK 0 0 1 0 1 0 0 Execution Phase: None. Execution Phase: Internal Lock register is written. Result Phase: Result Phase: Status Register 0 (80h) 0 4.1.5 Lock Command The Lock command allows the user full control of the FIFO parameters after a software reset. If the LOCK bit is set to 1, then the FIFO, THRESH, and PRETRK bits in the Configure command are not affected by a software reset. In addition, the FWR, FRD, and BST bits in the Mode command are unaffected by a software reset. If the LOCK is 0 (default after a hardware reset), then the above bits are set to their default values after a software reset. This command is use- 0 0 LOCK 0 0 0 0 4.1.6 Mode Command This command is used to select the special features of the controller. The bits for the Command Phase bytes are shown in Section 4.1, Command Set Summary, and their function is described below. These bits are set to their default values after a hardware reset. The default value of each bit is denoted by a ‘‘bullet’’ to the left of each item. The value of each parameter after a software reset is explained. 37 4.0 FDC Command Set Description (Continued) # 0 e Enable FIFO. mP write transfers druing the Exe- Command Phase: 0 0 0 0 0 TMR IAF IPS 0 FWR FRD BST R255 BFR WLD 0 0 DENSEL 0 0 0 0 1 LOW PWR 1 ETR 0 0 0 0 cution Phase use the internal FIFO. 1 e Disable FIFO. All write data transfers take place without the FIFO. FRD: FIFO Read Disable for mP read transfers from controller. Default after a software reset if LOCK is 0. If LOCK is 1, FRD retains its value after a software reset. Head Settle 0 RG 0 PU Execution Phase: Internal registers are written. Note: This bit is only valid if the FIFO is enabled in the Configure command. If the FIFO is not enabled in the Configure command, then this bit is a don’t care. Result Phase: None. TMR: Motor Timer mode. Default after a software reset. # 0 e Enable FIFO. mP read transfers during the Exe- # 0 e Timers for motor on and motor off are defined cution Phase use the internal FIFO. 1 e Disable FIFO. All read data transfers take place without the FIFO. BST: Burst Mode Disable. Default after a software reset if LOCK is 0. If LOCK is 1, BST retains its value after a software reset. for Mode 1. (See Specify command.) 1 e Timers for motor on and motor off are defined for Mode 2. (See Specify command.) IAF: Index Address Format. Default after a software reset. # 0 e The controller formats tracks with the Index Address Field included. (IBM and Perpendicular format.) Note: This bit is only valid if the FIFO is enabled in the Configure command. If the FIFO is not enabled in the Configure command, then this bit is a don’t care. 1 e The controller formats tracks without including the Index Address Field. (ISO format.) IPS: Implied Seek. Default after a software reset. # 0 e Burst mode enabled for FIFO Execution Phase data transfers. 1 e Non-Burst mode enabled. The DRQ or IRQ6 pin is strobed once for each byte to be transferred while the FIFO is enabled. R255: Recalibrate Step Pulses. The bit determines the maximum number of recalibrate step pulses the controller issues before terminating with an error. Default after a software reset. # 0 e The implied seek bit in the command byte of a read, write, scan, or verify is ignored. Implied seeks could still be enabled by the EIS bit in the Configure command. 1 e The IPS bit in the command byte of a read, write, scan, or verify is enabled so that if it is set, the controller performs seek and sense interrupt operations before executing the command. # 0 e Maximum of 85 recalibrate step pulses. If ETR e 1, controller issues 3925 recalibrate step pulses maximum. LOW PWR: Low Power mode. Default after a software reset. 1 e Maximum of 255 recalibrate step pulses. If ETR e 1, controller issues 4095 maximum recalibrate step pulses. DENSEL: Density Select Pin Configuration. This 2-bit value configures the Density Select output to one of three possible modes. The default mode configures the DENSEL pin according to the state of bit 6 of ASC after a data rate has been selected. That is, if ASC6 is high, the DENSEL pin is active high for the 500 kbps and 1 Mbps data rates. If ASC6 is low, the DENSEL pin is active low for the 500 kbps and 1 Mbps data rates. See Table 4-3. In addition to these modes, the DENSEL output can be set to always low or always high, as shown in Table 4-4. This allows the user more flexibility with new drive types. # 00 e Completely disable the low power mode. 01 e Automatic low power. For 500 kbps operation, go into low power mode 512 ms after the head unload timer times out. For 250 kbps operation the timeout period is doubled to 1s. 10 e Manual low power. Go into low power mode now. 11 e Not used. ETR: Extended Track Range. Default after a software reset. # 0 e Track number is stored as a standard 8-bit value compatible with the IBM, ISO, and Perpendicular formats. This allows access of up to 256 tracks during a seek operation. 1 e Track number is stored as a 12-bit value. The upper four bits of the track value are stored in the upper four bits of the head number in the sector Address Field. This allows access of up to 4096 tracks during a seek operation. With this bit set, an extra byte is required in the Seek Command Phase and Sense Interrupt Result Phase. FWR: FIFO Write Disable for mP write transfers to controller. Default after a software reset if LOCK is 0. If LOCK is 1, FWR retains its value after a software reset. TABLE 4-3. DENSEL Default Encoding Data Rate 250 kbps 300 kbps 500 kbps 1 Mbps* DENSEL Pin Definition ASC6 e 1 ASC6 e 0 Low Low High High High High Low Low *When TUP bit 1 e 0, a Data Rate of 1 Mbps is selected. Note: This bit is only valid if the FIFO is enabled in the Configure command. If the FIFO is not enabled in the Configure command, then this bit is a don’t care. 38 4.0 FDC Command Set Description (Continued) 4.1.7 NSC Command TABLE 4-4. DENSEL Encoding Mode Command Bit 1 Bit 0 DENSEL Pin Definition 0 0 1 1 0 1 0 1 Pin Low Pin High Undefined DEFAULT The NSC command can be used to distinguish between the FDC versions and the 82077. The Result Phase byte uniquely identifies the floppy controller as a PC87306, which returns a value of 73h. The 82077 and DP8473 return a value of 80h, signifying an invalid command. The lower four bits of this result byte are subject to change by National, and reflects the particular version of the floppy disk controller part. Command Phase: BFR: CMOS Disk Interface Buffer Enable. # 0 e Drive output signals configured as standard 4 mA push-pull outputs (actually 40 mA sink, 4 mA source). 1 e Drive output signals configured as 40 mA opendrain outputs. WLD: Scan Wild Card. 0 0 Head Settle: Time allowed for read/write head to settle after a seek during an Implied Seek operation. This is controlled as shown in Table 4-5 by loading a 4-bit value for N. (The default value for N is 8.) TABLE 4-5. Head Settle Time Calculation 250 300 500 1000 Nc8 N c 6.666 Nc4 Nc2 0–120 0–100 0 – 60 0 – 30 1 0 0 0 1 1 1 0 0 1 1 4.1.8 Perpendicular Mode Command The Perpendicular Mode command is designed to support the unique Format and Write Data requirements of Perpendicular (Vertical) Recording disk drives (4 Mbyte unformatted capacity). The Perpendicular Mode command configures each of the four logical drives as a perpendicular or conventional disk drive. Configuration of the four logical disk drives is done via the D3 – 0 bits, or with the GAP and WG control bits. This command should be issued during the initialization of the floppy controller. 1 e The Scan commands do not recognize FFh as a wildcard character. Head Settle Time (ms) 1 Result Phase: Scan command is interpreted as a wildcard character that always matches true. Multiplier (4 Bit Value) 0 Execution Phase: None. # 0 e An FFh from either the mP or the disk during a Data Rate (kbps) 0 Command Phase: 0 0 0 1 0 0 1 0 OW 0 DC3 DC2 DC1 DC0 GAP WG Execution Phase: Internal registers are written. Result Phase: None. Perpendicular Recording drives operate in ‘‘Extra High Density’’ mode at 1 Mbps and are downward compatible with 1.44 Mbyte and 720 kbyte drives at 500 kbps (High Density) and 250 kbps (Double Density) respectively. If perpendicular drives are present in the system, this command should be issued during initialization of the floppy controller, which configures each drive as perpendicular or conventional. Then, when a drive is accessed for a Format or Write Data command, the floppy controller adjusts the Format or Write Data parameters based on the data rate selected (see Table 4-6). RG: Read Gate Diagnostic. # 0 e Enable DSKCHG disk interface input for normal operation. 1 e Enable DSKCHG to act as an external Read Gate input signal to the Data Separator. This is intended as a test mode to aid in evaluation of the Data Separator. PU: PUMP Pulse Output Diagnostic. # 0 e Enable MFM output pin for normal operation. 1 e Enable the MFM output to act as the active low output of the Data Separator charge pump. This signal consists of a series of pulses indicating when the phase comparator is making a phase correction. This Pump output is active low for a pump up or pump down signal from the phase comparator, and is intended as a test mode to aid in the evaluation of the Data Separator. 39 4.0 FDC Command Set Description (Continued) TABLE 4-6. Effect of Drive Mode and Data Rate on Format and Write Commands Drive Mode GAP2 Length Written during Format Portion of GAP2 Re-Written by Write Data Command 250 kbps/300 kbps/500 kbps Conventional Perpendicular 22 Bytes 22 Bytes 0 Bytes 19 Bytes 1 Mbps Conventional Perpendicular 22 Bytes 41 Bytes 0 Bytes 38 Bytes Data Rate TABLE 4-7. Effect of GAP and WG on Format and Write Commands Mode Description GAP2 Length Written during Format 0 Conventional 22 Bytes 0 Bytes 1 Perpendicular (500 kbps) 22 Bytes 19 Bytes 1 0 Reserved (Conventional) 22 Bytes 0 Bytes 1 1 Perpendicular (1 Mbps) 41 Bytes 38 Bytes GAP WG 0 0 Portion of GAP2 Re-Written by Write Data Command Format in Table 4-1.) This GAP2 length of 41 bytes (at 1 Mbps) ensures that the Preamble in the Data Field is completely ‘‘pre-erased’’ by the Pre-Erase Head. Also, during Write Data operations to a perpendicular drive, a portion of GAP2 must be rewritten by the controller to guarantee that the Data Field Preamble has been pre-erased (see Table 4-6). Looking at the second command byte, DC3–0 corresponds to the four logical drives. A 0 written to DCn sets drive n to conventional mode, and a 1 sets drive n to perpendicular mode. The OW (Overwrite) bit offers additional control. When OW e 1, the values of DC3 – 0 (drive configuration bits) are changeable. When OW e 0, the internal values of DC3–0 are unaffected, regardless of what is written to DC3–0. The function of the DCn bits must also be qualified by setting both WG and GAP to 0. If WG and GAP are used (i.e., not set to 00), they override whatever is programmed in the DCn bits. Table 4-7 indicates the operation of the FDC based on the values of GAP and WG. Note that when GAP and WG are both 0, the DCn bits are used to configure each logical drive as conventional or perpendicular. DC3–0 is unaffected by a software reset, but WG and GAP are both cleared to 0 after a software reset. A hardware reset resets all the bits to zero (conventional mode for all drives). The Perpendicular Mode command bits may be rewritten at any time. 4.1.9 Read Data Command The Read Data command reads logical sectors containing a Normal Data Address Mark (AM) from the selected drive and makes the data available to the host mP. After the last Command Phase byte is written, the controller simulates the Motor On time for the selected drive internally. The user must turn on the drive motor directly by enabling the appropriate drive and motor select disk interface outputs with the Digital Output Register (DOR). If Implied Seeks are enabled, the controller performs a Seek operation to the track number specified in the Command Phase. The controller also issues a Sense Interrupt for the seek and waits the Head Settle time specified in the Mode command. The correct ID information (track, head, sector, bytes per sector) for the desired sector must be specified in the command bytes. SeeTable4-8SectorSizeSelectionfordetailson the bytes per sector code. In addition, the End of Track Sector Number (EOT) should be specified, allowing the controller to read multiple sectors. The Data Length byte is a don’t care and should be set to FFh. Note: When in the Perpendicular Mode for any drive at any data rate selected by the DC3–0 bits, write precompensation is set to zero. Perpendicular Recording type disk drives have a Pre-Erase Head which leads the Read/Write Head by 200 mm, which translates to 38 bytes at the 1 Mbps data transfer rate (19 bytes at 500 kbps). The increased spacing between the two heads requires a larger GAP2 between the Address Field and Data Field of a sector at 1 Mbps. (See Perpendicular 40 4.0 FDC Command Set Description (Continued) Having finished reading the sector, the controller continues reading the next logical sector unless one or more of the following termination conditions occurred: 1. The DMA controller asserted TC. The IC bits in ST0 are set to Normal Termination. 2. The last sector address (of side 1 if MT was set) was equal to EOT. The EOT bit in ST1 is set. The IC bits in ST0 are set to Abnormal Termination. This is the expected condition during Non-DMA transfers. 3. Overrun error. The OR bit in ST1 is set. The IC bits in ST0 are set to Abnormal Termination. If the mP cannot service a transfer request in time, the last correctly read byte is transferred. 4. CRC error. The CE bit in ST1 and the CD bit in ST2 are set. The IC bits in ST0 are set to Abnormal Termination. If Multi-Track Selector (MT) was set in the Opcode command byte, and the last sector of side 0 has been transferred, the controller then continues with side 1. Upon terminating the Execution Phase of the Read Data command, the controller asserts IRQ6, indicating the beginning of the Result Phase. The mP must then read the result bytes from the FIFO. The values that are read back in the result bytes are shown in Table 4-10. If an error occurs, the result bytes indicate the sector read when the error occurred. TABLE 4-8. Sector Size Selection Bytes per Sector Code Number of Bytes in Data Field 0 1 2 3 4 5 6 7 128 256 512 1024 2048 4096 8192 16384 The controller then starts the Data Separator and waits for the Data Separator to find the next sector Address Field. The controller compares the Address Field ID information (track, head, sector, bytes per sector) with the desired ID specified in the Command Phase. If the sector ID bytes do not match, then the controller waits for the Data Separator to find the next sector Address Field. The ID comparison process repeats until the Data Separator finds a sector Address Field ID that matches that in the command bytes, or until an error occurs. Possible errors are: 1. The mP aborted the command by writing to the FIFO. If there is no disk in the drive, the controller hangs up. The mP must then take the controller out of this hung state by writing a byte to the FIFO. This puts the controller into the Result Phase. 2. Two index pulses were detected since the search began, and no valid ID has been found. If the track address ID differs, the WT bit or BT bit (if the track address is FFh) is set in ST2. If the head, sector, or bytes per sector code did not match, the ND bit is set in ST1. If the Address Field AM was never found, the MA bit is set in ST1. 3. The Address Field was found with a CRC error. The CE bit is set in ST1. Once the desired sector Address Field is found, the controller waits for the Data Separator to find the subsequent Data Field for that sector. If the Data Field (normal or deleted) is not found within the expected time, the controller terminates the operation and enters the Result Phase (MD is set in ST2). If a Deleted Data Mark is found and Skip Flag (SK) was set in the Opcode command byte, the controller skips this sector and searches for the next sector Address Field as described above. The effect of SK on the Read Data command is summarized in Table 4-9. Having found the Data Field, the controller then transfers data bytes from the disk drive to the host (described in Section 5.3 Controller Phases) until the bytes per sector count has been reached, or the host terminates the operation (through TC, end of track, or implicitly through overrun). The controller then generates the CRC for the sector and compares this value with the CRC at the end of the Data Field. Command Phase: MT MFM SK 0 0 1 1 0 IPS X X X X HD DR1 DR0 Track Number Drive Head Number Sector Number Bytes per Sector End of Track Sector Number Intersector Gap Length Data Length Execution Phase: Data read from disk drive is transferred to system via DMA or Non-DMA modes. Result Phase: Status Register 0 Status Register 1 Status Register 2 Track Number Head Number Sector Number Bytes per Sector 41 4.0 FDC Command Set Description (Continued) TABLE 4-9. SK Effect on the Read Data Command SK Data Type Sector Read ? CM Bit (ST2) 0 Normal Y 0 Normal Termination Description of Results 0 Deleted Y 1 No Further Sectors Read 1 Normal Y 0 Normal Termination 1 Deleted N 1 Sector Skipped TABLE 4-10. Result Phase Termination Values with No Error ID Information at Result Phase MT HD Last Sector Track Head Sector 0 0 k EOT NC NC Sa1 NC 0 0 e EOT Ta1 NC 1 NC 0 1 k EOT NC NC Sa1 NC 0 1 e EOT Ta1 NC 1 NC 1 0 k EOT NC NC Sa1 NC 1 0 e EOT NC 1 1 NC 1 1 k EOT NC NC Sa1 NC 1 1 e EOT Ta1 0 1 NC EOT e End of Track Sector Number from Command Phase S e Sector Number last operated on by controller NC e No Change in Value T e Track Number programmed in Command Phase Bytes/Sector TABLE 4-11. SK Effect on the Read Deleted Data Command SK Data Type Sector Read ? CM Bit (ST2) Description of Results 0 Normal Y 1 No Further Sectors Read 0 Deleted Y 0 Normal Termination 1 Normal N 1 Sector Skipped 1 Deleted Y 0 Normal Termination Execution Phase: Data read from disk drive is transferred to system via DMA or Non-DMA modes. 4.1.10 Read Deleted Data Command The Read Deleted Data command reads logical sectors containing a Deleted Data AM from the selected drive and makes the data available to the host mP. This command is identical to the Read Data command, except for the setting of the CM bit in ST2 and the skipping of sectors. The effect of SK on the Read Deleted Data command is summarized in Table 4-11. See Table 4-10 for the state of the result bytes for a Normal Termination of the command. Result Phase: Status Register 0 Status Register 1 Status Register 2 Track Number Head Number Command Phase: Sector Number MT MFM SK 0 1 1 0 0 IPS X X X X HD DR1 DR0 Bytes per Sector 4.1.11 Read ID Command The Read ID command finds the next available Address Field and returns the ID bytes (track, head, sector, bytes per sector) to the mP in the Result Phase. There is no data transfer during the Execution Phase of this command. An interrupt is generated when the Execution Phase is completed. Track Number Drive Head Number Sector Number Bytes per Sector End of Track Sector Number Intersector Gap Length Data Length 42 4.0 FDC Command Set Description (Continued) 4. Multi-track and Skip operations are not allowed. SK and MT should be set to 0. The controller first simulates the Motor On time for the selected drive internally. The user must turn on the drive motor directly by enabling the appropriate drive and motor select disk interface outputs with the Digital Output Register (DOR). The Read ID command does not perform an implied seek. After waiting the Motor On time, the controller starts the Data Separator and waits for the Data Separator to find the next sector Address Field. If an error condition occurs, the IC bits in ST0 are set to Abnormal Termination, and the controller enters the Result Phase. Possible errors are: 1. The mP aborted the command by writing to the FIFO. If there is no disk in the drive, the controller hangs up. The mP must then take the controller out of this hung state by writing a byte to the FIFO. This puts the controller into the Result Phase. 2. Two index pulses were detected since the search began, and no AM has been found. If the Address Field AM was never found, the MA bit is set in ST1. 5. If there is a CRC error in the Data Field, the controller sets CE in ST1 and CD in ST2, but continues reading sectors. 6. The controller reads a maximum of EOT physical sectors. There is no support for multi-track reads. Command Phase: 0 MFM 0 0 0 0 1 0 IPS X X X X HD DR1 DR0 Track Number Drive Head Number Sector Number Bytes per Sector End of Track Sector Number Intersector Gap Length Data Length Command Phase: 0 MFM 0 0 1 0 1 0 X X X X X HD DR1 DR0 Execution Phase: Data read from disk drive is transferred to system via DMA or non-DMA modes. Result Phase: Execution Phase: Controller reads first ID Field header bytes it can find and reports these bytes to the system in the result bytes. Status Register 0 Status Register 1 Result Phase: Status Register 2 Track Number Status Register 0 Status Register 1 Head Number Status Register 2 Sector Number Track Number Bytes per Sector Head Number 4.1.13 Recalibrate Command The Recalibrate command is very similar to the Seek command. The controller sets the Present Track Register (PTR) of the selected drive to zero. It then steps the head of the selected drive out until the TRK0 disk interface input signal goes active, or until the maximum number of step pulses have been issued. See Table 4-12 for the maximum recalibrate step pulse values based on the R255 and ETR bits in the Mode command. If the number of tracks on the disk drive exceeds the maximum number of recalibrate step pulses, another Recalibrate command may need to be issued. Sector Number Bytes per Sector 4.1.12 Read A Track Command The Read A Track command reads sectors in physical order from the selected drive and makes the data available to the host. This command is similar to the Read Data command with the following exceptions: 1. The controller waits for the index pulse before searching for a sector Address Field. If the mP writes to the FIFO before the index pulse, the command enters the Result Phase with the IC bits in ST0 set to Abnormal Termination. 2. A comparison of the sector Address Field ID bytes will be performed, except for the sector number. The internal sector address is set to 1, and then incremented for each successive sector read. 3. If the Address Field ID comparison fails, the controller sets ND in ST1, but continues to read the sector. If there is a CRC error in the Address Field, the controller sets CE in ST1, but continues to read the sector. TABLE 4-12. Maximum Recalibrate Step Pulses Based on R255 and ETR 43 Maximum Recalibrate Step Pulses R255 ETR 0 0 1 0 255 0 1 3925 1 1 4095 85 (default) 4.0 FDC Command Set Description (Continued) After the last command byte is issued, the DRx BUSY bit is set in the MSR for the selected drive. The controller will simulate the Motor On time, and then enter the Idle Phase. The execution of the actual step pulses occur while the controller is in the Drive Polling Phase. An interrupt will be generated after the TRK0 signal is asserted, or after the maximum number of recalibrate step pulses are issued. There is no Result Phase. Recalibrates should not be issued on more than one drive at a time. This is because the drives are actually selected via the DOR, which can only select one drive at a time. No other command except the Sense Interrupt command should be issued while a Recalibrate command is in progress. 4.1.15 Scan Commands The Scan commands allow data read from the disk to be compared against data sent from the mP, using ones complement arithmetic, sector by sector. There are three Scan commands to choose from: 1. Scan Equal: checks to see if the scanned value of the disk data is equal to that of the mP data. The scan condition is therefore: disk data e mP data? 2. Scan Low or Equal: checks to see if the scanned value of the disk data is equal to or less than that of the mP data. The scan condition is therefore: disk data s data? 3. Scan High or Equal: checks to see if the scanned value of the disk data is equal to or greater than that of the mP data. The scan condition is therefore: disk data t mP data? The results of these comparisons are indicated in the Status Register bits 3 and 2, see Table 4-13, and the structure of the three commands follows. Each sector is compared starting with the most significant bytes first, and where the next sector is defined as the current Sector Number plus the Sector Step Size. Reading of sectors continues until either the scan condition is met, the End of Track (EOT) has been reached, or the Terminal Count (TC) is asserted. If the Wildcard mode is enabled in the Mode command, an FFh from either the disk or the mP is used as a don’t care byte that will always match equal. Read errors on the disk will have the same error conditions as the Read Data command. Additionally, if the Skip Flag (SK) bit is set, sectors with deleted data marks will be ignored. If all sectors read are skipped, the command will terminate with bit 3 of the Status Register set (mimicking a Scan Equal Hit). Command Phase: 0 0 0 0 0 1 1 1 0 0 0 0 0 0 DR1 DR0 Execution Phase: Disk drive head is stepped out to Track 0. Result Phase: None. 4.1.14 Relative Seek Command The Relative Seek command steps the selected drive in or out a given number of steps. This command will step the read/write head an incremental number of tracks, as opposed to comparing against the internal present track register for that drive. Command Phase: 1 DIR 0 0 1 1 1 1 X X X X X HD DR1 DR0 Execution Phase: Disk drive head stepped in or out a programmable number of tracks. Result Phase: None. The Relative Seek parameters are defined as follows: DIR: Read/Write Head Step Direction Control 0 e Step Head Out TABLE 4-13. Scan Command Termination Values Command Comparison Status Condition D3 D2 Scan Equal Disk Data e mP Data? 1 e Step Head In RTN: Relative Track Number. This value will determine how many incremental tracks to step the head in or out from the current track number. The controller will issue RTN number of step pulses and update the Present Track Register for the selected drive. The one exception to this is if the TRK0 disk input goes active, which indicates that the drive read/write head is at the outermost track. In this case, the step pulses for the Relative Seek are terminated, and the PTR value is set according to the actual number of step pulses issued. The arithmetic is done modulo 255. The DRx BUSY bit in the MSR is set for the selected drive. The controller will simulate the Motor On time before issuing the step pulses. After the Motor On time, the controller will enter the Idle Phase. The execution of the actual step pulses occurs in the Idle Phase of the controller. After the step operation is complete, the controller will generate an interrupt. There is no Result Phase. Relative Seeks should not be issued on more than one drive at a time. This is because the drives are actually selected via the DOR, which can only select one drive at a time. No other command except the Sense Interrupt command should be issued while a Relative Seek command is in progress. Scan Low or Equal Scan High or Equal 44 Disk Data s mP Data? Disk Data t mP Data? Condition Met? Indicated Result 1 0 Yes Disk Data e mP Data 0 1 No Disk Data mP Data 1 0 Yes Disk Data e mP Data 0 0 Yes Disk Data k mP Data 0 1 No Disk Data l mP Data 1 0 Yes Disk Data e mP Data 0 0 Yes Disk Data l mP Data 0 1 No Disk Data k mP Data i 4.0 FDC Command Set Description (Continued) SCAN LOW OR EQUAL SCAN EQUAL Command Phase: Command Phase: MT MFM SK 1 0 0 0 1 MT MFM SK 1 1 0 0 1 IPS X X X X HD DR1 DR0 IPS X X X X HD DR1 DR0 Track Number Track Number Drive Head Number Drive Head Number Sector Number Sector Number Bytes per Sector Bytes per Sector End of Track Sector Number End of Track Sector Number Intersector Gap Length Intersector Gap Length Sector Step Size Sector Step Size Execution Phase: Data transferred from system to controller is compared to data read from disk. Execution Phase: Data transferred from system to controller is compared to data read from disk. Result Phase: Result Phase: MT MFM SK 1 1 1 0 1 Status Register 0 IPS X X X X HD DR1 DR0 Status Register 1 Status Register 0 Status Register 2 Status Register 1 Track Number Status Register 2 Head Number Track Number Sector Number Head Number Bytes per Sector Sector Number 4.1.16 Seek Command The Seek command steps the selected drive in or out until the desired track number is reached. During the Execution Phase of the Seek command, the track number to seek to is compared with the present track number. The controller will determine how many step pulses to issue, and the DIR disk interface output will indicate which direction the R/W head should move. The DRx BUSY bit is set in the MSR for the appropriate drive. The controller will wait the Motor On time before issuing the first step pulse. After the Motor On time, the controller will enter the Idle Phase. The execution of the actual step pulses occurs in the Drive Polling phase of the controller. The step pulse rate is determined by the value programmed in the Specify command. An interrupt will be generated one step pulse period after the last step pulse is issued. A Sense Interrupt command should be issued to determine the cause of the interrupt. There is no Result Phase. While the internal microengine is capable of performing seek commands on 2 or more drives at the same time, software should ensure that only one drive is seeking at a time. This is because the drives are actually selected via the DOR, which can only select one drive at a time. No other command except a Sense Interrupt command should be issued while a Seek command is in progress. If the extended track range mode is enabled with the ETR bit in the Mode command, a fourth command byte should be written in the Command Phase to indicate the four most significant bits of the desired track number. Otherwise, only three command bytes should be written. Bytes per Sector SCAN HIGH OR EQUAL Command Phase: MT MFM SK 1 1 1 0 1 IPS X X X X HD DR1 DR0 Track Number Drive Head Number Sector Number Bytes per Sector End of Track Sector Number Intersector Gap Length Sector Step Size Execution Phase: Data transferred from system to controller is compared to data read from disk. Result Phase: Status Register 0 Status Register 1 Status Register 2 Track Number Head Number Sector Number Bytes per Sector 45 4.0 FDC Command Set Description (Continued) 3. The Ready Signal changed state during the polling mode for an internally selected drive. (Occurs only after a hardware or software reset.) 4. A Seek, Relative Seek, or Recalibrate command terminates. An interrupt due to reasons 1 or 2 does not require the Sense Interrupt command and is cleared automatically. This type of interrupt occurs during normal command operations and is easily discernible by the mP via the MSR. It is cleared when reading or writing information from or to the Data Register (FIFO). An interrupt caused by reasons 3 or 4 is identified with the aid of the Sense Interrupt command. This type of interrupt is cleared after the first result byte has been read. Use bits 5, 6, and 7 of ST0 to identify the cause of the interrupt as shown in Table 4-14. Issuing a Sense Interrupt command without an interrupt pending is treated as an Invalid command. If the extended track range mode is enabled, a third byte should be read in the Result Phase, which will indicate the four most significant bits of the present track number. Otherwise, only two result bytes should be read. Command Phase: 0 0 0 X X X 0 1 1 1 1 X X HD DR1 DR0 0 0 0 New Track Number MSN of Track Number 0 Note: The last Command Phase byte is required only if ETR is set in Mode Command. Execution Phase: Disk drive head is stepped in or out to a programmed track. Result Phase: None. 4.1.17 Sense Drive Status Command The Sense Drive Status command returns the status of the selected disk drive in ST3. This command does not generate an interrupt. Command Phase: 0 0 0 0 0 1 0 0 X X X X X HD DR1 DR0 Execution Phase: Disk drive status information is detected and reported. TABLE 4-14. Status Register 0 Termination Codes Status Register 0 Result Phase: Interrupt Code Status Register 3 4.1.18 Sense Interrupt Command The Sense Interrupt command is used to determine the cause of an interrupt when the interrupt is a result of the change in status of any disk drive. Command Phase: 0 0 0 0 1 0 0 Status Register 0 Present Track Number (PTR) 0 0 D7 D6 D5 1 1 0 Internal Ready Went True 0 0 1 Normal Seek Termination 0 1 1 Abnormal Seek Termination 4.1.19 Set Track Command This command is used to inspect or change the value of the internal Present Track Register. This can be useful for recovery from disk mistracking errors, where the real current track can be read through the Read ID command, and then the Set Track command can be used to set the internal Present Track Register to the correct value. If the WNR bit is a 0, a track register is to be read. In this case, the Result Phase byte contains the value in the internal register specified, and the third byte in the Command Phase is a dummy byte. If the WNR bit is a 1, data is written to a track register. In this case the third byte of the Command Phase is written to the specified internal track register, and the Result Phase byte contains this new value. The DS1 and DS0 bits select the Present Track Register for the particular drive. The internal register address depends on MSB, DS1, and DS0 as shown in Table 4-15. This command does not generate an interrupt. Result Phase: 0 Cause 0 Execution Phase: Status of interrupt is reported. MSN of PTR Seek End 0 Note: The third Result Phase byte can only be read if ETR is set in the Mode Command. Four possible causes for the interrupt are: 1. Entry into the Result Phase of any of the following commands: a. Read Data b. Read Deleted Data c. Read a Track d. Read ID e. Write Data f. Write Deleted Data g. Format h. Scan i. Verify 2. Occurrence of a data transfer in the Execution Phase while in the Non-DMA mode. 46 4.0 FDC Command Set Description (Continued) Command Phase: TABLE 4-15. Set Track Register Address DS1 DS0 0 0 0 0 1 1 1 1 MSB 0 0 1 1 0 0 1 1 0 Register Addressed 0 1 0 1 0 1 0 1 0 0 0 0 Step Rate Time PTR0 (LSB) PTR0 (MSB) PTR1 (LSB) PTR1 (MSB) PTR2 (LSB) PTR2 (MSB) PTR3 (LSB) PTR3 (MSB) 0 1 1 Motor Off Time Motor On Time DMA Execution Phase: Internal registers are written. Result Phase: None. Step Rate Time: These four bits define the time interval between successive step pulses during a seek, implied seek, recalibrate, or relative seek. The programming of this step rate is shown in Table 4-16. Command Phase: TABLE 4-16. Step Rate Time (SRT) Values 0 WNR 1 0 0 0 0 1 0 0 1 1 0 MSB DS1 DS0 Present Track Number (PTR) Execution Phase: Internal register selected by MSB of DS1 or DS0 is read or written. Result Phase: Data Rate Value Range Units 1 Mbps 500 kbps 300 kbps 250 kbps (16 b SRT)/2 (16 b SRT) (16 b SRT) c 1.67 (16 b SRT) c 2 0.5 – 8 1 – 16 1.67 – 26.7 2 –32 ms ms ms ms Motor Off Time: These four bits determine the simulated Motor Off time as shown in Table 4-17. Motor On Time: These seven bits determine the simulated Motor On time as shown in Table 4-18. DMA: This bit selects the data transfer mode in the Execution Phase of a read, write, or scan operation. 0 e DMA mode is selected Value 4.1.20 Specify Command The Specify command sets the initial values for three internal timers. The parameters of this command are undefined after power-up, and are unaffected by any reset. Thus, software should always issue a Specify command as part of an initialization routine. This command does not generate an interrupt. 1 e Non-DMA mode is selected TABLE 4-17. Motor Off Time (MFT) Values Data Rate Mode 1 (TMR e 0) Value Mode 2 (TMR e 1) Range Units Value Range 1 Mbps MFT c 8 8 – 128 MFT c 512 512 – 8192 ms 500 kbps MFT c 16 16 – 256 MFT c 512 512 – 8192 ms 300 kbps MFT c 80/3 26.7 – 427 MFT c 2560/3 853 – 13653 ms 250 kbps MFT c 32 32 – 512 MFT c 1024 1024 – 16384 ms Note: Motor Off Time e 0 is treated as MFT e 16. TABLE 4-18. Motor On Time (MNT) Values Data Rate 1 Mbps Mode 1 (TMR e 0) Mode 2 (TMR e 1) Units Value Range Value Range MNT 1 – 128 MNT c 32 32 – 4096 ms 500 kbps MNT 1 – 128 MNT c 32 32 – 4096 ms 300 kbps MNT c 10/3 3.3 – 427 MNT c 160/3 53 – 6827 ms 250 kbps MNT c 4 4 – 512 MNT c 64 64 – 8192 ms Note: Motor On Time e 0 is treated as MNT e 128. 47 4.0 FDC Command Set Description (Continued) EOT is equal to the last sector to be checked. In this case, the Data Length parameter should be set to FFh. Refer to Table 4-10 for the Result Phase values for a successful completion of the command. Also see Table 4-19 for further explanation of the result bytes with respect to the MT and EC bits. The Motor Off and Motor On timers are artifacts of the NEC mPD765. These timers determine both the delay from selecting a drive motor until a read or write operation is started, and the delay of deselecting the drive motor after the command is completed. Since the FDC enables the drive and motor select line directly through the DOR, these timers only provide some delay from the initiation of a command until it is actually started. Command Phase: 4.1.21 Verify Command The Verify command reads logical sectors containing a Normal Data AM from the selected drive without transferring the data to the host. This command is identical to the Read Data command, except that no data is transferred during the Execution Phase. The Verify command is designed for post-format or postwrite verification. Data is read from the disk, as the controller checks for valid Address Marks in the Address and Data Fields. The CRC is computed and checked against the previously stored value on the disk. The EOT value should be set to the final sector to be checked on each side. If EOT is greater than the number of sectors per side, the command will terminate with an error and no useful Address Mark or CRC data will be given. The TC pin cannot be used to terminate this command since no data is transferred. The verify command can simulate a TC by setting the EC bit to a 1. In this case, the command will terminate when SC (Sector Count) sectors have been read. (If SC e 0 then 256 sectors will be verified.) If EC e 0, then the command will terminate when MT MFM SK 1 0 1 1 0 EC X X X X HD DR1 DR0 Track Number Drive Head Number Sector Number Bytes per Sector End of Track Sector Number Intersector Gap Length Data Length/Sector Count Execution Phase: Data is read from disk but not transferred to the system. Result Phase: Status Register 0 Status Register 1 Status Register 2 Track Number Head Number Sector Number Bytes per Sector TABLE 4-19. Verify Command Result Phase MT EC SC/EOT Value (Notes 1, 2) Termination Result 0 0 DTL used (should be FFh) EOT s Ý Sectors per Side No Errors 0 0 DTL used (should be FFh) EOT l Ý Sectors per Side Abnormal Termination 0 1 SC s Ý Sectors per Side AND SC s EOT No Errors 0 1 SC l Ý Sectors Remaining OR SC l EOT Abnormal Termination 1 0 DTL used (should be FFh) EOT s Ý Sectors per Side No Errors 1 0 DTL used (should be FFh) EOT l Ý Sectors per Side Abnormal Termination 1 1 SC s Ý Sectors per Side AND SC s EOT No Errors 1 1 SC s (EOT c 2) AND EOT s Ý Sectors per Side No Errors 1 1 SC l (EOT c 2) Abnormal Termination Note 1: Ý Sectors per Side e number of formatted sectors per each side of the disk. Note 2: Ý Sectors Remaining e number of formatted sectors remaining which can be read, which includes side 1 of the disk if the MT bit is set to 1. Note 3: If MT e 1 and the SC value is greater than the number of remaining formatted sectors on side 0, verifying will continue on side 1 of the disk. 48 4.0 FDC Command Set Description (Continued) ST0 are set to Abnormal Termination. This is the expected condition during Non-DMA transfers. 4.1.22 Version Command The Version command can be used to determine the floppy controller being used. The Result Phase uniquely identifies the floppy controller version. The FDC returns a value of 90h in order to be compatible with the 82077. The DP8473 and other NEC765 compatible controllers will return a value of 80h (invalid command). 3. Underrun error. The OR bit in ST1 is set. The IC bits in ST0 are set to Abnormal Termination. If the mP cannot service a transfer request in time, the last correctly written byte will be written to the disk. If MT was set in the Opcode command byte, and the last sector of side 0 has been transferred, the controller will then continue with side 1. Command Phase: 0 0 0 1 0 0 0 0 Command Phase: Execution Phase: None. MT MFM 0 0 0 1 0 1 Result Phase: IPS X X X X HD DR1 DR0 1 0 0 1 0 0 0 Track Number 0 Drive Head Number 4.1.23 Write Data Command The Write Data command receives data from the host and writes logical sectors containing a Normal Data AM to the selected drive. The operation of this command is similar to the Read Data command except that the data is transferred from the mP to the controller instead of the other way around. The controller will simulate the Motor On time before starting the operation. If implied seeks are enabled, the seek and sense interrupt functions are then performed. The controller then starts the Data Separator and waits for the Data Separator to find the next sector Address Field. The controller compares the Address ID (track, head, sector, bytes per sector) with the desired ID specified in the Command Phase. If there is no match, the controller waits to find the next sector Address Field. This process continues until the desired sector is found. If an error condition occurs, the IC bits in ST0 are set to Abnormal Termination, and the controller enters the Result Phase. Possible errors are: 1. The mP aborted the command by writing to the FIFO. If there is no disk in the drive, the controller will hang up. The mP must then take the controller out of this hung state by writing a byte to the FIFO. This will put the controller into the Result Phase. 2. Two index pulses were detected since the search began, and no valid ID has been found. If the track address ID differs, the WT bit or BT bit (if the track address is FFh) will be set in ST2. If the head, sector, or bytes per sector code did not match, the ND bit is set in ST1. If the Address Field AM was never found, the MA bit is set in ST1. 3. The Address Field was found with a CRC error. The CE bit is set in ST1. 4. If the controller detects the Write Protect disk interface input is Asserted. Bit 1 of ST1 is set. If the correct Address Field is found, the controller waits for all (conventional mode) or part (perpendicular mode) of GAP2 to pass. The controller will then write the preamble field, address marks, and data bytes to the Data Field. The data bytes are transferred to the controller by the mP. Having finished writing the sector, the controller will continue reading the next logical sector unless one or more of the following termination conditions has occurred: 1. The DMA controller asserted TC. The IC bits in ST0 are set to Normal Termination. 2. The last sector address (of side 1 if MT was set) was equal to EOT. The EOT bit in ST1 is set. The IC bits in Sector Number Bytes per Sector End of Track Sector Number Intersector Gap Length Data Length Execution Phase: Data is transferred from the system to the controller via DMA or Non-DMA modes and written to the disk. Result Phase: Status Register 0 Status Register 1 Status Register 2 Track Number Head Number Sector Number Bytes per Sector 4.1.24 Write Deleted Data The Write Deleted Data command receives data from the host and writes logical sectors containing a Deleted Data AM to the selected drive. This command is identical to the Write Data command except that a Deleted Data AM is written to the Data Field instead of a Normal Data AM. Command Phase: MT MFM 0 0 1 0 0 1 IPS X X X X HD DR1 DR0 Track Number Drive Head Number Sector Number Bytes per Sector End of Track Sector Number Intersector Gap Length Data Length Execution Phase: Data is transferred from the system to the controller via DMA or Non-DMA modes and written to the disk. 49 4.0 FDC Command Set Description (Continued) Result Phase: Result Phase: Status Register 0 Status Register 0 Status Register 1 Status Register 1 Status Register 2 Status Register 2 Track Number Undefined Head Number Undefined Sector Number Undefined Bytes per Sector Undefined 4.2 COMMAND SET SUMMARY INVALID CONFIGURE Command Phase: Invalid Op Codes Command Phase: 0 0 0 1 0 0 1 1 0 0 0 0 0 0 0 0 0 EIS FIFO POLL Execution Phase: None. Result Phase: THRESH Status Register 0 (80h) PRETRK Execution Phase: Internal registers written. LOCK Result Phase: None. Command Phase: LOCK DUMPREG Command Phase: 0 0 0 0 1 0 1 0 0 Execution Phase: Internal Lock register is written. 0 0 1 1 1 0 Result Phase: 0 Execution Phase: Internal registers read. Result Phase: 0 0 LOCK 0 0 0 0 MODE PTR Drive 0 Command Phase: PTR Drive 1 PTR Drive 2 PTR Drive 3 Step Rate Time Motor Off Time Motor On Time 0 0 0 0 0 0 0 1 TMR IAF IPS 0 LOW PWR 1 ETR FWR FRD 0 0 0 DENSEL DMA 0 BST R255 BFR WLD 0 0 0 0 Head Settle 0 RG 0 PU Sector per Track/End of Track (Note) LOCK 0 DC3 DC2 0 EIS FIFO POLL DC1 DC0 GAP Execution Phase: Internal registers are written. WG THRESH Result Phase: None. PRETRK NSC Note: Sectors per Track parameter returned if last command issued was Format. End of Track parameter returned if last command issued was Read or Write. Command Phase: 0 0 0 1 1 0 0 0 1 0 0 1 1 FORMAT TRACK Execution Phase: None. Command Phase: 0 MFM 0 0 1 1 0 1 X X X X X HD DR1 DR0 Result Phase: 0 Bytes per Sector 1 1 PERPENDICULAR MODE Sectors per Track Command Phase: Format Gap Data Pattern Execution Phase: System transfers four ID bytes (track, head, sector, bytes/sector) per sector to the floppy controller via DMA or Non-DMA modes. The entire track is formatted. The data block in the Data Field of each sector is filled with the data pattern byte. 0 0 0 1 0 0 1 0 OW 0 DC3 DC2 DC1 DC0 GAP WG Execution Phase: Internal registers are written. Result Phase: None. 50 4.0 FDC Command Set Description (Continued) Execution Phase: Controller reads first ID Field header bytes it can find and reports these bytes to the system in the result bytes. READ DATA Command Phase: MT MFM SK 0 0 1 1 0 IPS X X X X HD DR1 DR0 Result Phase: Status Register 0 Track Number Status Register 1 Drive Head Number Status Register 2 Sector Number Track Number Bytes per Sector Head Number End of Track Sector Number Sector Number Intersector Gap Length Bytes per Sector Data Length READ A TRACK Execution Phase: Data read from disk drive is transferred to system via DMA or Non-DMA modes. Command Phase: Result Phase: Status Register 0 0 MFM 0 0 0 0 1 0 IPS X X X X HD DR1 DR0 Status Register 1 Track Number Status Register 2 Drive Head Number Track Number Sector Number Head Number Bytes per Sector Sector Number End of Track Sector Number Bytes per Sector Intersector Gap Length Data Length READ DELETED DATA Execution Phase: Data read from disk drive is transferred to system via DMA or non-DMA modes. Command Phase: MT MFM SK 0 1 1 0 0 IPS X X X X HD DR1 DR0 Result Phase: Status Register 0 Track Number Drive Head Number Status Register 1 Sector Number Status Register 2 Bytes per Sector Track Number Head Number End of Track Sector Number Sector Number Intersector Gap Length Bytes per Sector Data Length RECALIBRATE Execution Phase: Data read from disk drive is transferred to system via DMA or Non-DMA modes. Command Phase: Result Phase: Status Register 0 0 0 0 0 0 1 1 1 0 0 0 0 0 0 DR1 DR0 Status Register 1 Track Number Execution Phase: Disk drive head is stepped out to Track 0. Result Phase: None. Head Number RELATIVE SEEK Sector Number Command Phase: Status Register 2 Bytes per Sector READ ID 1 DIR 0 0 1 1 1 1 X X X X X HD DR1 DR0 RELATIVE TRACK NUMBER Command Phase: 0 MFM 0 0 1 0 1 0 X X X X X HD DR1 DR0 51 4.0 FDC Command Set Description (Continued) Execution Phase: Disk drive head stepped in or out a programmable number of tracks. Result Phase: Status Register 0 Result Phase: None. Status Register 1 SCAN EQUAL Status Register 2 Track Number Command Phase: Head Number MT MFM SK IPS X X 1 0 0 0 1 X X HD DR1 DR0 Sector Number Bytes per Sector Track Number SCAN LOW OR EQUAL Drive Head Number Sector Number Command Phase: Bytes per Sector End of Track Sector Number MT MFM SK 1 1 0 0 1 IPS X X X X HD DR1 DR0 Intersector Gap Length Track Number Sector Step Size Drive Head Number Execution Phase: Data transferred from system to controller is compared to data read from disk. Sector Number Bytes per Sector Result Phase: End of Track Sector Number MT MFM SK 1 1 1 0 1 Intersector Gap Length IPS X X X X HD DR1 DR0 Sector Step Size Status Register 0 Execution Phase: Data transferred from system to controller is compared to data read from disk. Status Register 1 Status Register 2 Result Phase: Track Number Status Register 0 Head Number Status Register 1 Sector Number Status Register 2 Bytes per Sector Track Number SCAN HIGH OR EQUAL Head Number Command Phase: Sector Number MT MFM SK IPS X X 1 1 1 0 1 X X HD DR1 DR0 Bytes per Sector SEEK Track Number Command Phase: Drive Head Number Sector Number Bytes per Sector 0 0 0 0 1 1 1 1 X X X X X HD DR1 DR0 0 0 0 New Track Number End of Track Sector Number MSN of Track Number 0 Intersector Gap Length Note: The last Command Phase byte is required only if ETR is set in Mode Command. Sector Step Size Execution Phase: Data transferred from system to controller is compared to data read from disk. Execution Phase: Disk drive head is stepped in or out to a programmed track. Result Phase: None. 52 4.0 FDC Command Set Description (Continued) VERIFY SENSE DRIVE STATUS Command Phase: Command Phase: 0 0 0 0 0 1 0 0 MT MFM SK 1 0 1 1 0 X X X X X HD DR1 DR0 EC X X X X HD DR1 DR0 Track Number Execution Phase: Disk drive status information is detected and reported. Drive Head Number Sector Number Result Phase: Bytes per Sector Status Register 3 End of Track Sector Number Intersector Gap Length SENSE INTERRUPT Data Length/Sector Count Command Phase: 0 0 0 0 1 0 0 Execution Phase: Data is read from disk but not transferred to the system. 0 Execution Phase: Status of interrupt is reported. Result Phase: Result Phase: Status Register 0 Status Register 0 Status Register 1 Present Track Number (PTR) MSN of PTR 0 Status Register 2 0 0 0 Track Number Head Number Note: The third Result Phase byte can only be read if ETR is set in the Mode Command. Sector Number SET TRACK Bytes per Sector Command Phase: VERSION 0 WNR 1 0 0 0 0 1 0 0 1 1 0 MSB DS1 DS0 Command Phase: 0 Present Track Number (PTR) 0 0 1 0 0 0 0 1 0 0 0 0 Execution Phase: None. Execution Phase: Internal register selected by MSB of DS1 or DS0 is read or written. Result Phase: Result Phase: 1 0 0 Value WRITE DATA Command Phase: SPECIFY Command Phase: 0 0 0 0 Step Rate Time 0 0 1 1 MFM 0 0 0 1 0 1 X X X X HD DR1 DR0 Track Number Motor Off Time Motor On Time MT IPS Drive Head Number DMA Sector Number Execution Phase: Internal registers are written. Bytes per Sector Result Phase: None. End of Track Sector Number Intersector Gap Length Data Length 53 4.0 FDC Command Set Description (Continued) Execution Phase: Data is transferred from the system to the controller via DMA or Non-DMA modes and written to the disk. Execution Phase: Data is transferred from the system to the controller via DMA or Non-DMA modes and written to the disk. Result Phase: Result Phase: Status Register 0 Status Register 0 Status Register 1 Status Register 1 Status Register 2 Status Register 2 Track Number Track Number Head Number Head Number Sector Number Sector Number Bytes per Sector Bytes per Sector WRITE DELETED DATA Command Phase: MT MFM 0 0 1 0 0 1 IPS X X X X HD DR1 DR0 Track Number Drive Head Number Sector Number Bytes per Sector End of Track Sector Number Intersector Gap Length Data Length 54 4.0 FDC Command Set Description (Continued) MFM 4.3 MNEMONIC DEFINITIONS FOR FDC COMMANDS Symbol Description BFR Buffer enable bit used in the Mode command. Enabled open-collector output buffers. Burst Mode disable control bit used in Mode command. Selects the Non-Burst FIFO mode if the FIFO is enabled. Drive Configuration 0–3. Used to set DC1a drive to conventional or perpendicular DC2 mode. Used in Perpendicular Mode DC3 command. Density Select control bits used in the Mode command. Direction control bit used in Relative Seek command to indicate step in or out. DMA mode enable bit used in the Specify command. Drive Select 0–1 bits used in most commands. Selects the logical drive. Data Length parameter used in the Read, Write, Scan and Verify commands. Enable Count control bit used in the Verify command. When this bit is 1, the DTL parameter becomes SC (Sector Count). Enable Implied Seeks. Used in the Configure command. End of Track parameter set in the Read, Write, Scan, and Verify commands. Extended Track Range used with the Seek command. First-In First-Out buffer. Also a control bit used in the Configure command to enable or disable the FIFO. FIFO Read disable control bit used in the Mode command. FIFO Write disable control bit used in the Mode command. GAP2 control bit used in the Perpendicular Mode command. Head Select control bit used in most commands. Selects Head 0 or 1 of the disk. Index Address Field control bit used in the Mode command. Enables the ISO Format during the Format command. Implied Seek enable bit used in the Mode, Read, Write, and Scan commands. Lock enable bit in the Lock command. Used to make certain parameters be unaffected by a software reset. Low Power control bits used in the Mode command. BST DC0 – 3 DENSEL DIR DMA DR0 – 1 DTL EC EIS EOT ETR FIFO FRD FWR GAP HD IAF IPS LOCK LOW PWR MFT MNT MT OW POLL PRETRK PTR PU R255 RG RTN SC SK SRT ST0 – 3 THRESH TMR WG WLD 55 Modified Frequency Modulation control bit used in the Read, Write, Format, Scan and Verify commands. Selects MFM or FM data encoding. Motor Off Time programmed in the Specify command. Motor On Time programmed in the Specify command. Multi-Track enable bit used in the Read, Write, Scan and Verify commands. Overwrite control bit used in the Perpendicular Mode command. Enable Drive Polling bit used in the Configure command. Precompensation Track Number used in the Configure command. Present Track Register. Contains the internal track number for one of the four logical disk drives. Pump diagnostic enable bit used in the Mode command. Recalibrate control bit used in Mode command. Sets maximum recalibrate step pulses to 255. Read Gate diagnostic enable bit used in the Mode command. Relative Track Number used in the Relative Seek command. Sector Count control bit used in the Verify command. Skip control bit used in read and scan operations. Step Rate Time programmed in the Specify command. Determines the time between step pulses for seek and recalibrates. Status Register 0 – 3. Contains status ST1 information about the execution of an ST2 command. Read in the Result Phase of some ST3 commands. FIFO threshold parameter used in the Configure command. Timer control bit used in the Mode command. Affects the timers set in the Specify command. Write Gate control bit used in the Perpendicular Mode command. Wildcard bit in the Mode command used to enable or disable the wildcard byte (FF) during Scan commands. 5.0 FDC Functional Description microprocessor. In addition, when no command is in progress, the controller is in the Idle Phase or Drive Polling Phase. The PC87306 is software compatible with the DP8473 and 82077 floppy disk controllers. Upon a power on reset, the 16-byte FIFO will be disabled. Also, the disk interface outputs will be configured as active push-pull outputs, which are compatible with both CMOS inputs and open-collector resistor terminated disk drive inputs. The FIFO can be enabled with the Configure command. The FIFO can be very useful at the higher data rates, with systems that have a large amount of DMA bus latency, or with multi-tasking systems such as the EISA or MicroChannel bus structures. The FDC will support all the DP8473 Mode command features as well as some additional features. Additional features include control over the enabling of the FIFO for reads and writes, a Non-Burst mode for the FIFO, a bit that will configure the disk interface outputs as open-drain outputs, and programmability of the DENSEL output. 5.3.1 Command Phase During the Command Phase, the mP writes a series of bytes to the Data Register. The first command byte contains the opcode for the command, and the controller knows how many more bytes to expect based on this opcode byte. The remaining command bytes contain the particular parameters required for the command. The number of command bytes varies for each particular command. All the command bytes must be written in the order specified in the Command Description Table. The Execution Phase starts immediately after the last byte in the Command Phase is written. Prior to performing the Command Phase, the Digital Output Register should be set and the data rate should be set with the Data Rate Select Register or Configuration Control Register. The Main Status Register controls the flow of command bytes, and must be polled by the software before writing each Command Phase byte to the Data Register. Prior to writing a command byte, the RQM bit (D7) must be set and the DIO bit (D6) must be cleared in the MSR. After the first command byte is written to the Data Register, the CMD PROG bit (D4) is also set and remains set until the last Result Phase byte is read. If there is no Result Phase, the CMD PROG bit will be cleared after the last command byte is written. A new command may be initiated after reading all the result bytes from the previous command. If the next command requires selecting a different drive or changing the data rate, the DOR and DSR or CCR should be updated. If the command is the last command, the software should deselect the drive. 5.1 MICROPROCESSOR INTERFACE The FDC interface to the microprocessor consists of the A9 – A3, AEN, RD, and WR lines, which access the chip for reads and writes; the data lines D7–0; the address lines A2 – 0, which select the appropriate register (see Table 3-1); the IRQ6 signal, and the DMA interface signals DRQ, DACK, and TC. It is through this microprocessor interface that the floppy controller receives commands, transfers data, and returns status information. 5.2 MODES OF OPERATION The FDC has two modes of operation: PC-AT mode and PS/2 mode, which are determined by the state of bits 7 and 6 of the ASC register. See Section 3.0 Register Description for more details on the register set used for each mode of operation. PC-AT ModeÐBits 7, 6 of ASC register are 11 (default): The PC-AT register set is enabled. The DMA enable bit in the Digital Output Register becomes valid (IRQ6 and DRQ can be TRI-STATE). TC and DENSEL become active high signals (defaults to a 5.25× floppy drive). PS/2 ModeÐBits 7, 6 of ASC register are 01: This mode supports the PS/2 Models 50/60/80 configuration and register set. The DMA enable bit in the Digital Output Register becomes a don’t care (IRQ6 and DRQ signals are always valid). TC and DENSEL become active low signals (default to 3.5× floppy drive). Note: As a general rule, the operation of the controller core is independent of how the mP updates the DOR, DSR, and CCR. The software must ensure that the manipulation of these registers is coordinated with the controller operation. 5.3.2 Execution Phase During the Execution Phase, the disk controller performs the desired command. Commands that involve data transfers, (e.g., read, write, or format operation) require the mP to write or read data to or from the Data Register at this time. Some commands such as a Seek or Recalibrate control the read/write head movement on the disk drive during the Execution Phase via the disk interface signals. Execution of other commands does not involve any action by the mP or disk drive, and consists of an internal operation by the controller. If there is data to be transferred between the mP and the controller during the Execution, there are three methods that can be used, DMA mode, interrupt transfer mode, and 5.3 CONTROLLER PHASES The FDC has three separate phases of a command, the Command Phase, the Execution Phase, and the Result Phase. Each of these controller phases determine how data is transferred between the floppy controller and the host 56 5.0 FDC Functional Description (Continued) For both the Burst and Non-Burst modes, when the last byte in the FIFO has been read, DRQ goes inactive. DRQ is then reasserted when the FIFO trigger condition is satisfied. After the last byte of a sector has been read from the disk, DRQ is again generated even if the FIFO has not yet reached its threshold trigger condition. This guarantees that all the current sector bytes are read from the FIFO before the next sector byte transfer begins. software polling mode. The last two modes are called the Non-DMA modes. The DMA mode is used if the system has a DMA controller. This allows the mP to do other tasks while the data transfer takes place during the Execution Phase. If the Non-DMA mode is used, an interrupt is issued for each byte transferred during the Execution Phase. Also, instead of using the interrupt during Non-DMA mode, the Main Status Register can be polled by software to indicate when a byte transfer is required. All of these data transfer modes work with the FIFO enabled or disabled. Write Data Transfers Whenever the number of bytes in the FIFO is less than or equal to THRESH, a DRQ is generated. This is the trigger condition for the FIFO write data transfers from the mP to the floppy controller. Burst Mode. DRQ remains active until enough bytes have been written to the controller to completely fill the FIFO. Non-Burst Mode. DRQ is deasserted after each write transfer. If the FIFO is not full, DRQ is reasserted after a 350 ns delay. This deassertion of DRQ allows other higher priority DMA transfers to take place between floppy transfers. The FIFO has a byte counter which monitors the number of bytes being transferred to the FIFO during write operations for both Burst and Non-Burst modes. When the last byte of a sector is transferred to the FIFO, DRQ is deasserted even if the FIFO has not been completely filled. Thus, the FIFO is cleared after each sector is written. Only after the floppy controller has determined that another sector is to be written is DRQ asserted again. Also, since DRQ is deasserted immediately after the last byte of a sector is written to the FIFO, the system does not need to tolerate any DRQ deassertion delay and is free to do other work. 5.3.2.1 DMA ModeÐFIFO Disabled The DMA mode is selected by writing a 0 to the DMA bit in the Specify command and by setting the DMA enabled bit (D3) in the DOR. With the FIFO disabled, a DMA request (DRQ) is generated in the Execution Phase when each byte is ready to be transferred. The DMA controller should respond to the DRQ with a DMA acknowledge (DACK) and a read or write strobe. The DRQ is cleared by the leading edge of the active low DACK input signal. After the last byte is transferred, an interrupt is generated, indicating the beginning of the Result Phase. During DMA operations the chip select input (CS) must be held high. The DACK signal acts as the chip select for the FIFO in this case, and the state of the address lines A2–A0 is a don’t care. The Terminal Count (TC) signal can be asserted by the DMA controller to terminate the data transfer at any time. Due to internal gating, TC is only recognized when DACK is low. PC-AT Mode. When in the PC-AT interface mode with the FIFO disabled, the controller is in single byte transfer mode. That is, the system has one byte time to service a DMA request (DRQ) from the controller. DRQ will be deasserted between each byte. PS/2 Mode. When in the PS/2 mode, DMA transfers with the FIFO disabled are performed differently. Instead of a single byte transfer mode, the FIFO is actually enabled with THRESH e 0Fh. Thus, DRQ is asserted when one byte has entered the FIFO during reads, and when one byte can be written to the FIFO during writes. DRQ is deasserted by the leading edge of the DACK input, and is reasserted when DACK goes inactive high. This operation is very similar to Burst mode transfer with the FIFO enabled except that DRQ is deasserted between each byte. Read and Write Data Transfers The DACK input signal from the DMA controller may be held active during an entire burst or it may be strobed for each byte transferred during a read or write operation. When in the Burst mode, the floppy controller deasserts DRQ as soon as it recognizes that the last byte of a burst was transferred. If DACK is strobed for each byte, the leading edge of this strobe is used to deassert DRQ. If DACK is strobed, RD or WR are not required. This is the case during the ReadVerify mode of the DMA controller. If DACK is held active during the entire burst, the trailing edge of the RD or WR strobe is used to deassert DRQ. DRQ is deasserted within 50 ns of the leading edge of DACK, RD, or WR. This quick response should prevent the DMA controller from transferring extra bytes in most applications. 5.3.2.2 DMA ModeÐFIFO Enabled Read Data Transfers Whenever the number of bytes in the FIFO is greater than or equal to (16 b THRESH), a DRQ is generated. This is the trigger condition for the FIFO read data transfers from the floppy controller to the mP. Burst Mode. DRQ remains active until enough bytes have been read from the controller to empty the FIFO. Non-Burst Mode. DRQ is deasserted after each read transfer. If the FIFO is not completely empty, DRQ is reasserted after a 350 ns delay. This allows other higher priority DMA transfers to take place between floppy transfers. In addition, this mode allows the controller to work correctly in systems where the DMA controller is put into a read verify mode, where only DACK signals are sent to the FDC, with no RD pulses. This read verify mode of the DMA controller is used in some PC software. The FIFO Non-Burst mode allows the DACK input from the DMA controller to be strobed, which correctly clocks data from the FIFO. Overrun Errors An overrun or underrun error terminates the execution of the command if the system does not transfer data within the allotted data transfer time (see Section 3.7), which puts the controller into the Result Phase. During a read overrun, the mP is required to read the remaining bytes of the sector before the controller asserts IRQ6, signifying the end of execution. During a write operation, an underrun error terminates the Execution Phase after the controller has written the remaining bytes of the sector with the last correctly written byte to the FIFO and generated the CRC bytes. Whether there is an error or not, an interrupt is generated at the end of the Execution Phase, and is cleared by reading the first Result Phase byte. 57 5.0 FDC Functional Description (Continued) DACK asserted alone without a RD or WR strobe is also counted as a transfer. If RD or WR are not being strobed for each byte, DACK must be strobed for each byte so that the floppy controller can count the number of bytes correctly. A new command, the Verify command, has been added to allow easier verification of data written to the disk without the need of actually transferring the data on the data bus. 5.3.2.3 Interrupt ModeÐFIFO Disabled If the Interrupt (Non-DMA) mode is selected, IRQ6 is asserted instead of DRQ when each byte is ready to be transferred. The Main Status Register should be read to verify that the interrupt is for a data transfer. The RQM and NON DMA bits (D7 and D5) in the MSR are set. The interrupt is cleared when the byte is transferred to or from the Data Register. CS and RD or CS and WR must be used to transfer the data in or out of the Data Register (A2–A0 must be valid). CS asserted by itself is not significant. CS must be asserted with RD or WR for a read or write transfer to be recognized. The mP should transfer the byte within the data transfer service time (see Section 3.7). If the byte is not transferred within the time allotted, an Overrun Error will be indicated in the Result Phase when the command terminates at the end of the current sector. An interrupt is also be generated after the last byte is transferred. This indicates the beginning of the Result Phase. The RQM and DIO bits (D7 and D6) in the MSR is set, and the non-DMA bit (D5) is cleared. This interrupt is cleared by reading the first Result Phase byte. 5.3.3 Result Phase During the Result Phase, the mP reads a series of bytes from the data register. These bytes indicate the status of the command. This status may indicate whether the command executed properly, or contain some control information (see the Command Description and Status Register Description). These Result Phase bytes are read in the order specified for that particular command. Some commands will not have a result phase. Also, the number of result bytes varies with each command. All of the result bytes must be read from the Data Register before the next command can be issued. Like the Command Phase, the Main Status Register controls the flow of result bytes, and must be polled by the software before reading each Result Phase byte from the Data Register. The RQM bit (D7) and DIO bit (D6) must both be set before each result byte can be read. After the last result byte is read, the COM PROG bit (D4) in the MSR is cleared, and the controller is ready for the next command. 5.3.4 Idle Phase After a hardware or software reset, or after the chip has recovered from the power-down mode, the controller enters the Idle Phase. Also, when there are no commands in progress the controller is in the Idle Phase. The controller waits for a command byte to be written to the Data Register. The RQM bit is set and the DIO bit cleared in the MSR. After receiving the first command (opcode) byte, the controller enters the Command Phase. When the command is completed the controller again enters the Idle Phase. The Data Separator remains synchronized to the reference frequency while the controller is idle. While in the Idle Phase, the controller will periodically enters the Drive Polling Phase (see Section 5.3.5). 5.3.2.4 Interrupt ModeÐFIFO Enabled The Interrupt (Non-DMA) mode with the FIFO enabled is very similar to the Non-DMA mode with the FIFO disabled. In this case, IRQ6 is asserted instead of DRQ under the exact same FIFO threshold trigger conditions. The MSR should be read to verify that the interrupt is for a data transfer. The RQM and NON DMA bits (D7 and D5) in the MSR is set. CS and RD or CS and WR must be used to transfer the data in or out of the Data Register (A2–A0 must be valid). CS asserted by itself is not significant. CS must be asserted with RD or WR for a read or write transfer to be recognized. The Burst mode may be used to hold the IRQ6 pin active during a burst, or the Non-Burst mode may be used to toggle the IRQ6 pin for each byte of a burst. The Main Status Register is always valid from the mP point of view. For example, during a read command, after the last byte of data has been read from the disk and placed in the FIFO, the MSR still indicates that the Execution Phase is active, and that data needs to be read from the Data Register. Only after the last byte of data has been read by the mP from the FIFO does the Result Phase begin. The same overrun and underrun error procedures from the DMA mode apply to the Non-DMA mode. Also, whether there is an error or not, an interrupt is generated at the end of the Execution Phase, and is cleared by reading the first Result Phase byte. 5.3.5 Drive Polling Phase The National FDC supports the polling mode of the old generation 8-inch drives as a means of monitoring any change in status for each disk drive present in the system. This mode is supported for the sole purpose of providing backward compatibility with software that expects its presence. While in the Idle Phase the controller enters a Drive Polling Phase every 1 ms (based on the 500 kbps data rate). While in the Drive Polling Phase, the controller interrogates the Ready Changed status for each of the four logical drives. The internal Ready line for each drive is toggled only after a hardware or software reset, and an interrupt is generated for drive 0. At this point, the software must issue four Sense Interrupt commands to clear the Ready Changed State status for each drive. This requirement can be eliminated if drive polling is disabled via the POLL bit in the Configure command. The Configure command must be issued within 500 ms (worst case ) of the hardware or software reset for drive polling to be disabled. Even if drive polling is disabled, drive stepping and delayed power-down occurs in the Drive Polling Phase. The controller checks the status of each drive and if necessary it issues a step pulse on the STEP output with the DIR signal at the appropriate logic level. Also, the controller uses the Drive Polling Phase to control the Automatic Low Power mode. When the Motor Off time has expired, the controller waits 512 ms based on the 500 kbps and 1 Mbps data rate before powering down if this function is enabled via the Mode command. 5.3.2.5 Software Polling If the Non-DMA mode is selected and interrupts are not suitable, the mP can poll the MSR during the Execution Phase to determine when a byte is ready to be transferred. The RQM bit (D7) in the MSR reflects the state of the IRQ6 signal. Otherwise, the data transfer is similar to the Interrupt Mode described above. This is true for the FIFO enabled or disabled. 58 5.0 FDC Functional Description (Continued) TL/C/12379 – 30 FIGURE 5-1. FDC Data Separator Block Diagram tomatically switched into the data separator circuit when the data rate is selected via the Data Rate Select or Configuration Control Register. These filters have been optimized through lab experimentation, and are designed into the controller to reduce the external component cost associated with the floppy controller. The FDC has a dynamic window margin and lock range performance capable of handling a wide range of floppy disk drives. Also, the data separator works well under a variety of conditions, including the high motor speed fluctuations of floppy compatible tape drives. If a new command is issued when the FDC is in the middle of a polling routine, the MSR will not indicate a ready status for the next parameter byte until the poling sequence completes the loop. This can cause a delay between the first and second bytes of up to 500 ms at 250 kbps. 5.4 DATA SEPARATOR The internal data separator consists of an analog PLL and its associated circuitry. The PLL synchronizes the raw data signal read from the disk drive. The synchronized signal is used to separate the encoded clock and data pulses. The data pulses are deserialized into bytes and then sent to the mP by the controller. The main PLL consists of five main components, a phase comparator, a charge pump, a filter, a voltage controlled oscillator (VCO), and a programmable divider. The phase comparator detects the difference between the phase of the divider’s output and the phase of the raw data being read from the disk. This phase difference is converted to a current by the charge pump, which either charges or discharges one of three filters which is selected based on the data rate. The resulting voltage on the filter changes the frequency of the VCO and the divider output to reduce the phase difference between the input data and the divider’s output. The PLL is ‘‘locked’’ when the frequency of the divider is exactly the same as the average frequency of the data read from the disk. A block diagram of the data separator is shown in Figure 5-1 . To ensure optimal performance, the data separator incorporates several additional circuits. The quarter period delay line is used to determine the center of each bit cell, and to disable the phase comparator when the raw data signal is missing a clock or data pulse in the MFM or FM pattern. A secondary PLL is used to automatically calibrate the quarter period delay line. The secondary PLL also calibrates the center frequency of the VCO. To eliminate the logic associated with controlling multiple data rates, the FDC supports each of the four data rates (250, 300, 500 kbps, and 1 Mbps) with a separate, optimized internal filter. The appropriate filter for each data rate is au- Figure 5-2 shows the floppy disk controller dynamic window margin performance at the four different data rates. Dynamic window margin is the primary indicator of the quality and performance level of the data separator. This measurement indicates how much motor speed variation (MSV) of the drive spindle motor and bit jitter (or window margin) can be tolerated by the data separator. MSV is shown on the x-axis of the dynamic window margin graph. MSV is translated directly to the actual data rate of the data as it is read from the disk by the data separator. That is, a faster than nominal motor will result in a higher frequency in the actual data rate. The dynamic window margin performance curves also indicate how much bit jitter (or window margin) can be tolerated by the data separator. This parameter is shown on the y-axis of the graphs. Bit jitter is caused by the magnetic interaction of adjacent data pulses on the disk, which effectively shifts the bits away from their nominal positions in the middle of the bit window. Window margin is commonly measured as a percentage. This percentage indicates how far a data bit can be shifted early or late with respect to its nominal bit position, and still be read correctly by the data separator. If the data separator cannot correctly decode a shifted bit, then the data is misread and a CRC results. The dynamic window margin performance curves contain two pieces of information: 1) the maximum range of MSV (also called ‘‘lock range’’) that the data separator can handle with no read errors, and 2) the maximum percentage of 59 5.0 FDC Functional Description (Continued) 250 kbps 300 kbps TL/C/12379 – 32 TL/C/12379–31 500 kbps 1 Mbps TL/C/12379–33 TL/C/12379 – 34 FIGURE 5-2. PC87306 Dynamic Window Margin Performance (Typical performance at VDD e 5.0V, 25§ C) An external oscillator circuit must have a duty cycle of at least 40% – 60% and minimum input levels of 2.4V and 0.4V. The controller should be configured so that the external oscillator clock is input into the X1/OSC pin. window margin (or bit jitter) that the data separator can handle with no read errors. Thus, the area under the dynamic window margin curves in Figure 5-2 is the range of MSV and bit jitter that the FDC can handle with no read errors. The FDC internal analog data separator has a much better performance than comparable digital data separator designs, and does not require any external components. 5.6 PERPENDICULAR RECORDING MODE The FDC is fully compatible with perpendicular recording mode disk drives at all data rates. These perpendicular mode drives are also called 4 Mbyte (unformatted) or 2.88 Mbyte (formatted) drives, which refers to their maximum storage capacity. Perpendicular recording will orient the magnetic flux changes (which represent bits) vertically on the disk surface, allowing for a higher recording density than the conventional longitudinal recording methods. With this increase in recording density comes an increase in the data rate of up to 1 Mbps, thus doubling the storage capacity. In addition, the perpendicular 2.88M drive is read/write compatible with 1.44M and 720k diskettes (500 kbps and 250 kbps respectively). Note: The dynamic window margin curves were generated using a FlexStar FS-540 Floppy Disk Simulator and a proprietary dynamic window margin test program written by National Semiconductor. The controller takes best advantage of the internal analog data separator by implementing a sophisticated read algorithm. This ID search algorithm, shown in Figure 5-3 , enhances the PLL’s lock characteristics by forcing the PLL to relock to the crystal reference frequency any time the data separator attempts to lock to a non-preamble pattern. This algorithm ensures that the PLL is not thrown way out of lock by write splices or bad data fields. 5.5 CRYSTAL OSCILLATOR The FDC is clocked by a single 24 MHz signal. 60 5.0 FDC Functional Description (Continued) pre-erase head is located at the beginning of the Data Field preamble. This means that WGATE should be asserted when the read/write head is at least 38 bytes (at 1 Mbps) before the preamble. See Table 4-5 for a description of the WGATE timing for perpendicular drives at the various data rates. Because of the 38 byte spacing between the read/write head and the pre-erase head at 1 Mbps, the GAP2 length of 22 bytes used in the standard IBM disk format is not long enough. There is a new format standard for 2.88M drives at 1 Mbps called the Perpendicular Format, which increases the GAP2 length to 41 bytes (see Figure 4-1 ). The Perpendicular Mode command will put the floppy controller into perpendicular recording mode, which allows it to read and write perpendicular media. Once this command is invoked, the read, write and format commands can be executed in the normal manner. The perpendicular mode of the floppy controller will work at all data rates, adjusting the format and write data parameters accordingly. See Section 4.2.6 for more details. The 2.88M drive has unique format and write data timing requirements due to its read/write head and pre-erase head design (see Figure 5-4 ). Unlike conventional disk drives which have only a read/write head, the 2.88M drive has both a pre-erase head and read/write head. With conventional disk drives, the read/write head by itself is able to rewrite the disk without problems. For 2.88M drives, a preerase head is needed to erase the magnetic flux on the disk surface before the read/write can write to the disk surface. The pre-erase head is activated during disk write operations only, i.e., Format and Write Data commands. In 2.88M drives, the pre-erase head leads the read/write head by 200 mm, which translates to 38 bytes at 1 Mbps (19 bytes at 500 kbps). For both conventional and perpendicular drives, WGATE is asserted with respect to the position of the read/write head. With conventional drives, this means that WGATE is asserted when the read/write head is located at the beginning of the Data Field preamble. With the 2.88M drives, since the preamble must be pre-erased before it is rewritten, WGATE should be asserted when the TL/C/12379 – 35 FIGURE 5-3. Read Data AlgorithmÐState Diagram TL/C/12379 – 36 FIGURE 5-4. Perpendicular Recording Drive R/W Head and Pre-Erase Head 61 5.0 FDC Functional Description (Continued) function of the manual low power mode is a logical OR function between the DSR low power bit and the Mode command manual low power bit setting. Automatic low power mode will switch the controller into low power 500 ms (at the 500 kbps MFM data rate) after it has entered the idle state. Once the auto low power mode is set, it does not have to be set again, and the controller will automatically go into low power mode after it has entered the idle state. Automatic low power mode can only be set with the Mode command. There are two ways the FDC section can recover from the power-down state. 1) The part will power up after a software reset via the DOR or DSR. Since a software reset requires reinitialization of the controller, this method can be undesirable. 2) The part will also power up after a read or write to either the Data Register or Main Status Register. This is the preferred method of power up since all internal register values are retained. It may take a few milliseconds for the oscillator to stabilize, and the mP will be prevented from issuing commands during this time through the normal Main Status Register protocol. That is, the RQM bit in the MSR will be a 0 until the oscillator has stabilized. When the controller has completely stabilized from power up, the RQM bit in the MSR is set to 1 and the controller can continue where it left off. The Data Rate Select, Digital Output, and Configuration Control Registers are unaffected by the power-down mode. They will remain active. It is up to the user to ensure that the Motor and Drive Select signals are turned off. 5.7 DATA RATE SELECTION The data rate can be chosen two different ways with the FDC. For PC compatible software, the Configuration Control Register at address 3F7h is used to program the data rate for the floppy controller. The lower bits D1 and D0 are used in the CCR to set the data rate. The other bits should be set to zero. See Table 3-7 for the data rate select encoding. The data rate can also be set using the Data Rate Select Register at address 4. Again, the lower two bits of the register are used to set the data rate. The encoding of these bits is exactly the same as those in the CCR. The remainder of the bits in the DSR are used for other functions. Consult the Register Description (Section 3.1.6) for more details. The data rate is determined by the last value that is written to either the CCR or the DSR. In other words, either the CCR or the DSR can override the data rate selection of the other register. When the data rate is selected, the microengine and data separator clocks are scaled appropriately. Also, the DRATE0 and DRATE1 output pins will reflect the state of the data select bits that were last written to either the CCR or the DSR. 5.8 WRITE PRECOMPENSATION Write precompensation is a way of preconditioning the WDATA output signal to adjust for the effects of bit shift on the data as it is written to the disk surface. Bit shift is caused by the magnetic interaction of data bits as they are written to the disk surface, and has the effect of shifting these data bits away from their nominal position in the serial MFM or FM data pattern. Data that is subject to bit shift is much harder to read by a data separator, and can cause soft read errors. Write precompensation predicts where bit shift could occur within a data pattern. It then shifts the individual data bits early, late, or not at all such that when they are written to the disk, the resultant shifted data bits will be back in their nominal position. The FDC supports software programmable write precompensation. Upon power up, the default write precomp values will be used (see Table 3-6). The programmer can choose a different value of write precomp with the DSR register if desired (see Table 3-5). Also on power up, the default starting track number for write precomp is track zero. This starting track number for write precomp can be changed with the Configure command. Note: If the power to an external oscillator driving the PC87306 is to be independently removed during the FDC low power mode, it must not be done until 2 ms after the FDC low power command is issued. 5.10 RESET OPERATION The floppy controller can be reset by hardware or software. Hardware reset is enacted by pulsing the Master Reset input pin. A hardware reset will set all of the user addressable registers and internal registers to their default values. The Specify command values will be don’t cares, so they must be reinitialized. The major default conditions are: FIFO disabled, FIFO threshold e 0, Implied Seeks disabled, and Drive Polling enabled. A software reset can be performed through the Digital Output Register or Data Rate Select Register. The DSR reset bit is self-clearing, while the DOR reset bit is not self-clearing. If the LOCK bit in the Lock command was set to a 1 previous to the software reset, the FIFO, THRESH, and PRETRK parameters in the Configure command will be retained. In addition, the FWR, FRD, and BST parameters in the Mode command will be retained if LOCK is set to 1. This function eliminates the need for total reinitialization of the controller after a software reset. After a hardware (assuming the FDC is enabled in the FER) or software reset, the Main Status Register is immediately available for read access by the mP. It will return a 00h value until all the internal registers have been updated and the data separator is stabilized. When the controller is ready to receive a command byte, the MSR will return a value of 80h (Request for Master bit is set). The MSR is guaranteed to return the 80h value within 2.5 ms after a hardware or software reset. All other user addressable registers other than the Main Status Register and Data Register (FIFO) can be accessed at any time, even while the part is in reset. 5.9 FDC LOW POWER MODE LOGIC The FDC section of the PC87306 supports two low power modes described here in detail. Other low power modes of the PC87306 are described in Section 2.5. Details concerning entering and exiting low power mode via setting Date Rate Select Register bit 6 or by executing the FDC Mode Command are covered below and in Section 3.1.6 and Section 4.1.6. The microcode is driven from the clock, so it will be disabled while the clock is off. The FDC clock is always disabled upon entering this mode, however, the oscillator is only disabled when PTR1 e 1. Upon entering the powerdown state, the RQM (Request For Master) bit in the MSR will be cleared. There are two modes of low power in the floppy controller: manual low power and automatic low power. Manual low power is enabled by writing a 1 to bit 6 of the DSR. The chip will go into low power immediately. This bit will be cleared to 0 after the chip is brought out of low power. Manual low power can also be accessed via the Mode command. The 62 6.0 Serial Ports Bits 0,1 These two bits specify the number of data bits in each transmitted or received serial character. The encoding of bits 0 and 1 is as follows: Each of these serial ports functions as a serial data input/ output interface in a microcomputer system. The system software determines the functional configuration of the UARTs via a 8-bit bidirectional data bus. The UARTs are completely independent. They perform serial-to-parallel conversion on data characters received from a peripheral device or a MODEM, and parallel-to-serial conversion on data characters received from the CPU. The CPU can read the complete status of either UART at any time during the functional operation. Status information reported includes the type and condition of the transfer operations being performed by the UART, as well as any error conditions (parity, overrun, framing, or break interrupt). The UARTs have programmable baud rate generators that are capable of dividing the internal reference clock by divisors of 1 to (216 –1), and producing a 16x clock for driving the transmitter logic. Provisions are also included to use this 16x clock to drive the receiver logic. The UARTs have complete MODEM-control capability and a prioritized interrupt system. Interrupts can be programmed to the user’s requirements, minimizing the computing required to handle the communications link. Bit 2 Bit 3 6.1 SERIAL PORT REGISTERS Two identical register sets, one for each channel, are in the PC87306. All register descriptions in this section apply to the register sets in both channels. See Table 6-1. Bit 4 TABLE 6-1. PC87306 UART Register Addresses (AEN e 0) DLAB A2 A1 A0 Selected Register 0 0 0 0 Receiver Buffer (Read), Transmitter Holding (Write) 0 0 0 1 Interrupt Enable 0 0 1 0 Interrupt Identification (Read) FIFO Control (Write) X 0 1 1 Line Control X 1 0 0 MODEM Control X 1 0 1 Line Status X 1 1 0 MODEM Status X 1 1 1 Scratch 1 0 0 0 Divisor Latch (Least Significant Byte) 1 0 0 1 Divisor Latch (Most Significant Byte) Bit 5 Bit 6 Bit 1 Bit 0 Data Length 0 0 5 Bits 0 1 6 Bits 1 0 7 Bits 1 1 8 Bits This bit specifies the number of Stop bits transmitted with each serial character. If it is 0, one Stop bit is generated in the transmitted data. If it is 1, when a 5-bit data length is selected, one and a half Stop bits are generated. If it is logic 1, when either a 6-, 7-, or 8-bit word length is selected, two Stop bits are generated. The receiver checks the first Stop bit only, regardless of the number of Stop bits selected. This bit is the Parity Enable bit. When it is 1, a Parity bit is generated (transmit data) or checked (receive data) between the last data bit and Stop bit of the serial data. (The Parity bit is used to produce an even or odd number of 1s when the data bits and the Parity bit are summed.) This bit is the Even Parity Select bit. When parity is enabled and bit 4 is 0, an odd number of logic 1s is transmitted or checked in the data word bits and Parity bit. When parity is enabled and bit 4 is 1, an even number of logic 1s is transmitted or checked. This bit is the Stick Parity bit. When parity is enabled it is used in conjunction with bit 4 to select Mark or Space Parity. When LCR bits 3, 4 and 5 are 1 the Parity bit is transmitted and checked as a 0 (Space Parity). If bits 3 and 5 are 1 and bit 4 is a 0, then the Parity bit is transmitted and checked as 1 (Mark Parity). If bit 5 is 0 Stick Parity is disabled. This bit is the Break Control bit. It causes a break condition to be transmitted to the receiving UART. When it is set to 1, the serial output (SOUT) is forced to the Spacing state (0). The break is disabled by setting bit 6 to 0. The Break Control bit acts only on SOUT and has no effect on the transmitter logic. Note that this feature enables the CPU to alert a terminal. If the following sequence is used, no erroneous characters will be transmitted because of the break. 1. Wait for the transmitter to be idle, (TEMT e 1). 2. Set break for the appropriate amount of time. If the transmitter will be used to time the break duration, then check that TEMT e 1 before clearing the Break Control bit. 3. Clear break when normal transmission has to be restored. During the break, the Transmitter can be used as a character timer to accurately establish the break duration by sending characters and monitoring THRE and TEMT. 6.2 LINE CONTROL REGISTER (LCR) Read/Write The system programmer uses the Line Control Register (LCR) to specify the format of the asynchronous data communications exchange and set the Divisor Latch Access bit. This is a read and write register. Table 6-2 shows the contents of the LCR. Details on each bit follow. TL/C/12379 – 37 FIGURE 6-1. PC87306 Composite Serial Data 63 6.0 Serial Ports (Continued) TABLE 6-2. PC87306 Register Summary for an Individual UART Channel Register Address Bit No. A0 –2 e 0 DLAB e 0 A0 –2 e 0 DLAB e 0 A0 – 2 e 1 DLAB e 0 2 2 3 4 5 6 7 Receiver Buffer Register (Read Only) Transmitter Holding Register (Write Only) Interrupt Enable Register Interrupt Ident. Register (Read Only) FIFO Control Register (Write Only) Line Control Register MODEM Control Register Line Status Register MODEM Status Register Scratch Pad Register A0– 2 e 0 A0– 2 e 1 DLAB e 1 DLAB e 1 Divisor Latch (LSB) Divisor Latch (MSB) RBR THR IER IIR FCR LCR MCR LSR MSR SCR DLL DLM 0 Data Bit 0 (Note 1) Data Bit 0 Enable Received Data Available Interrupt ‘‘0’’ if Interrupt Pending FIFO Enable Word Length Select Bit 0 Data Terminal Ready (DTR) Data Ready (DR) Delta Clear to Send Bit 0 Bit 0 Bit 8 1 Data Bit 1 Data Bit 1 Enable Transmitter Holding Register Empty Interrupt Interrupt ID Bit RCVR FIFO Reset Word Length Select Bit 1 Request to Send (RTS) Overrun Error (OE) Delta Data Set Ready Bit 1 Bit 1 Bit 9 2 Data Bit 2 Data Bit 2 Enable Receiver Line Status Interrupt Interrupt ID Bit XMIT FIFO Reset Number of Stop Bits Out 1 Bit (Note 3) Parity Error (PE) Trailing Edge Ring Indicator Bit 2 Bit 2 Bit 10 3 Data Bit 3 Data Bit 3 Enable MODEM Status Interrupt Interrupt ID Bit (Note 2) Reserved Parity Enable IRQ Enable Framing Error (FE) Delta Data Carrier Detect Bit 3 Bit 3 Bit 11 4 Data Bit 4 Data Bit 4 0 0 Reserved Even Parity Select Loop Break Interrupt (BI) Clear to Send Bit 4 Bit 4 Bit 12 5 Data Bit 5 Data Bit 5 0 0 Reserved Stick Parity 0 Transmitter Holding Register (THRE) Data Set Ready Bit 5 Bit 5 Bit 13 6 Data Bit 6 Data Bit 6 0 FIFOs Enabled (Note 2) RCVR Trigger (LSB) Set Break 0 Transmitter Empty (TEMT) Ring Indicator Bit 6 Bit 6 Bit 14 7 Data Bit 7 Data Bit 7 0 FIFOs Enabled (Note 2) RCVR Trigger (MSB) Divisor Latch Access Bit (DLAB) 0 Error in RCVR FIFO (Note 2) Data Carrier Detect Bit 7 Bit 7 Bit 15 Note 1: Bit 0 is the least significant bit. It is the first bit serially transmitted or received. Note 2: These bits are always 0 in the NS16450 Mode. Note 3: This bit no longer has a pin associated with it. 64 6.0 Serial Ports (Continued) TABLE 6-3. PC87306 UART Reset Configuration Register or Signal Reset Control Reset State Interrupt Enable Master Reset (MR) 0000 0000 (Note 1) Interrupt Identification Master Reset 0000 0001 FIFO Control Master Reset 0000 0000 Line Control Master Reset 0000 0000 MODEM Control Master Reset 0000 0000 Line Status Master Reset 0110 0000 MODEM Status Master Reset XXXX 0000 (Note 2) SOUT Master Reset High INTR (RCVR Errs) Read LSR/MR Low/TRI-STATE INTR (RCVR Data Ready) Read RBR/MR Low/TRI-STATE INTR (THRE) Read IIR/Write THR/MR Low/Low/TRI-STATE INTR (Modem Status Changes) Read MSR/MR Low/TRI-STATE Interrupt Enable Bit Master Reset Low RTS Master Reset High DTR Master Reset High RCVR FIFO MR or (FCR1 e 1 and FCR0 e 1) or Change in FCR0 All Bits Low XMIT FIFO MR or (FCR2 e 1 and FCR0 e 1) or Change in FCR0 All Bits Low Note 1: Boldface bits are permanently low. Note 2: Bits 7–4 are driven by the input signals. Bit 7 This bit is the Divisor Latch Access Bit (DLAB). It must be set high (logic 1) to access the Divisor Latches of the Baud rate Generator during a Read or Write operation or to have the Baud Out (BOUT) signal appear on the BOUT pin. It must be set low (logic 0) to access any other register. TABLE 6-4. PC87306 UART Divisors, Baud Rates and Clock Frequencies 24 MHz Input Divided to 1.8462 MHz 6.3 PROGRAMMABLE BAUD RATE GENERATOR The PC87306 contains two independently programmable Baud rate Generators. The 24 MHz oscillator frequency input is divided by 13, resulting in a frequency of 1.8462 MHz. This is sent to each Baud rate Generator and divided by the divisor of the associated UART. The output frequency of the Baud rate Generator (BOUT1,2) is 16 c the baud rate. divisor Ý e (frequency input) (baud rate c 16) The output of each Baud rate Generator drives the transmitter and receiver sections of the associated serial channel. Two 8-bit latches per channel store the divisor in a 16-bit binary format. These Divisor Latches must be loaded during initialization to ensure proper operation of the Baud rate Generator. Upon loading either of the Divisor Latches, a 16-bit Baud Counter is loaded. Table 6-4 provides decimal divisors to use with crystal frequencies of 24 MHz. The oscillator input to the chip should always be 24 MHz to ensure that the Floppy Disk Controller timing is accurate and that the UART divisors are compatible with existing software. Using a divisor of zero is not recommended. Baud Rate Decimal Divisor for 16x Clock Percent Error (Note) 50 75 110 134.5 150 300 600 1200 1800 2000 2400 3600 4800 7200 9600 19200 38400 57600 115200 2304 1536 1047 857 768 384 192 96 64 58 48 32 24 16 12 6 3 2 1 0.1 0.4 0.5 Note: The percent error for all baud rates, except where indicated otherwise is 0.2%. 65 6.0 Serial Ports (Continued) Bit 5 6.4 LINE STATUS REGISTER (LSR) This 8-bit register provides status information to the CPU concerning the data transfer. Table 6-2 shows the contents of the Line Status Register. Details on each bit follow: Bit 0 This bit is the receiver Data Ready (DR) indicator. It is set to 1 whenever a complete incoming character has been received and transferred into the Receiver Buffer Register or the FIFO. It is reset to 0 by reading the data in the Receiver Buffer Register or the FIFO. Bit 1 This bit is the Overrun Error (OE) indicator. It indicates that data in the Receiver Buffer Register was not read by the CPU before the next character was transferred into the Receiver Buffer Register, thereby destroying the previous character. The OE indicator is set to 1 upon detection of an overrun condition, and reset whenever the CPU reads the contents of the Line Status Register. If the FIFO mode data continues to fill the FIFO beyond the trigger level, an Overrun error will occur only after the FIFO is completely full and the next character has been received in the shift register. OE is indicated to the CPU as soon as it happens. The character in the shift register is overwritten, but is not transferred to the FIFO. Bit 2 This bit is the Parity Error (PE) indicator. It indicates that the received data character does not have the correct parity, as selected by the even-parity select bit. The PE bit is set to 1 upon detection of a parity error and is reset to 0 whenever the CPU reads the contents of the Line Status Register. In the FIFO mode this error is associated with the particular character in the FIFO that it applies to. This error is revealed to the CPU when its associated character is at the top of the FIFO. Bit 3 This bit is the Framing Error (FE) indicator. It indicates that the received character did not have a valid Stop bit. It is set to 1 whenever the Stop bit following the last data bit or parity bit is a 0 (Spacing level). The FE indicator is reset whenever the CPU reads the contents of the Line Status Register. In the FIFO mode this error is associated with the particular character in the FIFO that it applies to. This error is revealed to the CPU when its associated character is at the top of the FIFO. The UART will try to resynchronize after a framing error by assuming that the error was due to the next start bit. It samples this ‘‘start’’ bit twice and then takes in the bits following it as the rest of the frame. Bit 4 This bit is the Break Interrupt (BI) indicator. It is set to 1 whenever the received data input is held in the Spacing (0) state for longer than a full word transmission time (i.e., the total time of Start bit a data bits a Parity a Stop bits). The BI indicator is reset whenever the CPU reads the contents of the Line Status Register. In the FIFO mode this error is associated with the particular character in the FIFO that it applies to. This error is revealed to the CPU when its associated character is at the top of the FIFO. When break occurs only one character is loaded into the FIFO. To Restart after a break is received, the SIN pin must be 1 for at least (/2 bit time. Bit 6 Bit 7 This bit is the Transmitter Holding Register Empty (THRE) indicator. It indicates that the UART is ready to accept a new character for transmission. In addition, it causes the UART to issue an interrupt to the CPU when the Transmit Holding Register Empty Interrupt enable is set high. The THRE bit is set to 1 when a character is transferred from the Transmitter Holding Register into the Transmitter Shift Register. The bit is reset to 0 whenever the CPU loads the Transmitter Holding Register. In the FIFO mode it is set when the XMIT FIFO is empty; it is cleared when at least 1 byte is written to the XMIT FIFO. This bit is the Transmitter Empty (TEMT) indicator. It is set to 1 whenever the Transmitter Holding Register (THR) and the Transmitter Shift Register (TSR) are both empty. It is reset to 0 if either the THR or TSR contains a data character. In the FIFO mode this bit is set to 1 whenever the transmitter FIFO and the shift register are both empty. In the NS16450 Mode this is 0. In the FIFO Mode this bit is set when there is at least one parity error, framing error or break indication in the FIFO. It is cleared when the CPU reads the LSR, if there are no subsequent errors in the FIFO. Note: The Line Status Register is intended for read operations only. Writing to this register is not recommended as this operation is only used for factory testing. In the FIFO mode the software must load a data byte in the Rx FIFO via the Loopback Mode in order to write to LSR2–LSR4. LSR0 and LSR7 can’t be written to in FIFO Mode. 6.5 FIFO CONTROL REGISTER (FCR) This is a write-only register at the same location as the IIR (the IIR is a read-only register). This register is used to enable the FIFOs, clear the FIFOs and to set the RCVR FIFO trigger level. Bit 0 Writing a 1 to FCR0 enables both the XMIT and RCVR FIFOs. Resetting FCR0 clears all bytes in both FIFOs. When changing from FIFO Mode to NS16450 Mode and vice versa, data is automatically cleared from the FIFOs. This bit must already be 1 when other FCR bits are written to or they will not be programmed. Bit 1 Writing 1 to FCR1 clears all bytes in the RCVR FIFO and resets its counter logic to 0. The shift register is not cleared. The 1 that is written to this bit position is self-clearing. Bit 2 Writing 1 to FCR2 clears all bytes in the XMIT FIFO and resets its counter logic to 0. The shift register is not cleared. The 1 that is written to this bit position is self-clearing. Bit 3 Writing to FCR3 does not change UART operations. Bits 4, 5 FCR4 and FCR5 are reserved for future use. Bits 6, 7 The combination of FCR6 and FCR7 are used to designate the interrupt trigger level. See Figure 6-2 . When the number of bytes in the RCVR FIFO equals the designated interrupt trigger level, a Received Data Available Interrupt is activated. This interrupt must be enabled by setting the Interrupt Enable Register (IER) bit 0. Note: Bits 1 through 4 are the error conditions that produce a Receiver Line Status interrupt whenever any of the corresponding conditions are detected and the interrupt is enabled. 66 6.0 Serial Ports (Continued) FCR Bits 7 6 RCVR FIFO Trigger Level (Bytes) 0 0 01 0 1 04 1 0 08 1 1 14 Bit 3 This bit enables the MODEM Status Interrupt when set to logic 1. Bits 4 – 7 These four bits are always logic 0. 6.8 MODEM CONTROL REGISTER (MCR) This register controls the interface with the MODEM or data set (or a peripheral device emulating a MODEM). The contents of the MODEM Control Register (MCR) are indicated in Table 6-2 and are described as follows: Bit 0 This bit controls the Data Terminal Ready (DTR) output. When it is set to 1, the DTR output is forced to a logic 0. When it is reset to 0, the DTR output is forced to 1. In Local Loopback Mode, this bit controls bit 5 of the MODEM Status Register. FIGURE 6-2. Receiver FIFO Trigger Level 6.6 INTERRUPT IDENTIFICATION REGISTER (IIR) In order to provide minimum software overhead during data character transfers, the UART prioritizes interrupts into four levels and records these in the Interrupt Identification Register. The four levels of interrupt conditions in order of priority are Receiver Line Status; Received Data Ready; Transmitter Holding Register Empty; and MODEM Status. When the CPU accesses the IIR, the UART freezes all interrupts and indicates the highest priority pending interrupt to the CPU. While this CPU access is occurring, the UART records new interrupts, but does not change its current indication until the current access is complete. Table 6-2 shows the contents of the IIR. Details on each bit follow. Bit 0 This bit can be used in an interrupt environment to indicate whether an interrupt condition is pending. When it is 0, an interrupt is pending and the IIR contents may be used as a pointer to the appropriate interrupt service routine. When it is 1, no interrupt is pending. See Table 6-5. Bits 1, 2 These two bits of the IIR are used to identify the highest priority interrupt pending as indicated in Table 6-5. Bit 3 In the 16450 mode this bit is 0. In the FIFO mode it is set along with bit 2 when a time-out interrupt is pending. See Table 6-5. Bits 4, 5 These bits of the IIR are always 0. Bits 6, 7 These two bits are set when FCR0 e 1. (FIFO Mode enabled.) Note: The DTR and RTS output of the UART may be applied to an EIA inverting line driver (such as the DS1488) to obtain the proper polarity input at the MODEM or data set. This bit controls the Request to Send (RTS) output. Its effect on the RTS output is identical to that described above for bit 0. In Local Loopback Mode, this bit controls bit 4 of the MODEM Status Register. Bit 2 This bit is the OUT1 bit. It does not have an output pin associated with it. It can be written to and read by the CPU. In Local Loopback Mode, this bit controls bit 6 of the MODEM Status Register. Bit 3 This bit enables the interrupt when set. No external pin is associated with this bit other than IRQ3, 4. In Local Loopback Mode, this bit controls bit 7 of the MODEM Status Register. Bit 4 This bit provides a Local loopback feature for diagnostic testing of the UART. When it is set to 1, the following changes take place: the transmitter Serial Output (SOUT) is set to the Marking (1) state; the receiver Serial Input (SIN) is disconnected; the output of the Transmitter Shift Register is ‘‘looped back’’ (connected) to the Receiver Shift Register; the four MODEM Control inputs (DSR, CTS, RI and DCD) are disconnected; and the DTR, RTS, OUT1, IRQ ENABLE bits in MCR are internally connected to DSR, CTS, RI and DCD in MSR, respectively. The MODEM Control output pins are forced to their high (inactive) states. In the Loopback Mode, data that is transmitted is immediately received. This feature allows the processor to verify the transmit-and-received-data paths of the serial port. In the Loopback Mode, the receiver and transmitter interrupts are fully operational. The MODEM Status Interrupts are also operational, but the interrupts’ sources are the lower four bits of MCR instead of the four MODEM control inputs. Writing a 1 to any of these 4 MCR bits will cause an interrupt. In Loopback Mode the interrupts are still controlled by the Interrupt Enable Register. The IRQ3 and IRQ4 pins will be at TRI-STATE in the Loopback Mode. Bits 5 – 7 These bits are permanently set to 0. Bit 1 6.7 INTERRUPT ENABLE REGISTER (IER) This register enables the five types of UART interrupts. Each interrupt can individually activate the appropriate interrupt (IRQ3 or IRQ4) output signal. It is possible to totally disable the interrupt system by resetting bits 0 through 3 of the Interrupt Enable Register (IER). Similarly, setting bits of this register to 1, enables the selected interrupt(s). Disabling an interrupt prevents it from being indicated as active in the IIR and from activating the interrupt output signal. All other system functions operate in their normal manner, including the setting of the Line Status and MODEM Status Registers. Table 6-2 shows the contents of the IER. Details on each bit follow. See MODEM Control Register bit 3 for more information on enabling the interrupt pin. Bit 0 When set to 1 this bit enables the Received Data Available Interrupt and Timeout Interrupt in the FIFO Mode. Bit 1 This bit enables the Transmitter Holding Register Empty Interrupt when set to 1. Bit 2 This bit enables the Receiver Line Status Interrupt when set to logic 1. 67 6.0 Serial Ports (Continued) TABLE 6-5. PC87306 Interrupt Control Functions Interrupt Identification Register Interrupt Set and Reset Functions Bit 3 (FIFO Mode Only) Bit 2 Bit 1 Bit 0 0 0 0 1 Ð None None 0 1 1 0 Highest Receiver Line Status Overrun Error, Parity Error, Framing Error or Break Interrupt Reading the Line Status Register 0 1 0 0 Second Received Data Available Receiver Data Available Read Receiver Buffer 1 1 0 0 Second (FIFO Mode Only) Character Time-Out Indication No Characters Have Been Removed from or input to the RCVR FIFO During the Last 4 Character Times and there is at least 1 Character in it during this Time. Reading the Receiver Buffer Register 0 0 1 0 Third Transmitter Holding Register Empty Transmitter Holding Register Empty Reading the IIR Register (if Source of Interrupt) or Writing the Transmitter Holding Register 0 0 0 0 Fourth MODEM Status Clear to Send or Data Set Ready or Ring Indicator or Data Carrier Detect Reading the MODEM Status Register Priority Level Interrupt Type Bit 6 6.9 MODEM STATUS REGISTER (MSR) This register provides the current state of the control lines from the MODEM (or peripheral device) to the CPU. In addition to this current-state information, four bits of the MODEM Status Register provide change information. These bits are set to a logic 1 whenever a control input from the MODEM changes state. They are reset to logic 0 whenever the CPU reads the MODEM Status Register. Table 6-2 shows the contents of the MSR. Details on each bit follow. Bit 0 This bit is the Delta Clear to Send (DCTS) indicator. Bit 0 indicates that the CTS input to the chip has changed state since the last time it was read by the CPU. Bit 1 This bit is the Delta Data Set Ready (DDSR) indicator. It indicates that the DSR input to the chip has changed state since the last time it was read by the CPU. Bit 2 This bit is the Trailing Edge of Ring Indicator (TERI) detector. It indicates that the RI input to the chip has changed from a low to a high state. Bit 3 This bit is the Delta Data Carrier Detect (DDCD) indicator. It indicates that the DCD input to the chip has changed state. Bit 7 Bit 5 Interrupt Reset Control Ð This bit is the complement of the Ring Indicator (RI) input. If bit 4 of the MCR is set to a 1, this bit is equivalent to OUT1 in the MCR. This bit is the complement of the Data Carrier Detect (DCD) input. If bit 4 of the MCR is set to a 1, this bit is equivalent to IRQ ENABLE in the MCR. 6.10 SCRATCHPAD REGISTER (SCR) This 8-bit Read/Write Register does not control the UART in any way. It is intended as a scratchpad register to be used by the programmer to hold data temporarily. 7.0 Serial Infrared Wireless Communication Port The PC87306 supports a serial InfraRed (IR) wireless communication port. An IR interface connects a UART’s serial in and serial out signals via one or more transmitter LED and receiver photo diode pairs. The IR port of the PC87306 is hooked to UART2. SOUT of UART2 is also encoded by the PC87306 and routed to the InfraRed Transmitter (IRTX) pin. The InfraRed Receiver pin (IRRX) is decoded by the PC87306 and routed to the SIN signal of UART2. There are two UART2 interface modes: Note: Whenever bit 0, 1, 2, or 3 is set to logic 1, a MODEM Status Interrupt is generated. Bit 4 Interrupt Source # Normal (MODEM) mode: The normal NS16550A UART This bit is the complement of the Clear to Send (CTS) input. If bit 4 (loopback) of the MCR is set to a 1, this bit is equivalent to RTS in the MCR. This bit is the complement of the Data Set Ready (DSR) input. If bit 4 of the MCR is set to a 1, this bit is equivalent to DTR in the MCR. interface. # IR mode: InfraRed mode, IrDA compliant. A system board can have a PC87306 configured with two UARTs and one IR interface because these functions utilize separate pins on the device. The selection of UART2 interface mode is a run-time selection that yields either two 68 7.0 Serial Infrared Wireless Communication Port (Continued) PMC bit 2 for UART TRI-STATE control. An additional register, the IRC (InfraRed Configuration), is provided to configure the IR interface. Also, UART2’s IRQ mapping can be modified using the PNP1 register. The state of UART2 pins for both modes follow. In IR mode (IRC.0 e 1) the default values for the following UART2 pins are: RI2 and DCD2 are 1. DSR2 and CTS2 are 0. SOUT2 is driven high (inactive). In Normal (MODEM) mode (IRC.0 e 0) the default value for the IR port pins are: IRRX is 1. IRTX is driven low (inactive). Float Control: UART2 pins (except IRQ3 and IRQ4) and IR pins IRTX and IRRX are controlled by the UART2 TRISTATE control bit (bit 3 of SCF0). The outputs are in TRISTATE, and the inputs are blocked to reduce the leakage current, when the UART2 TRI-STATE control bit is 1, and either UART2 is disabled or the PC87306 is in power-down mode. UARTs, or one UART plus one IR port, simultaneously. Normal (MODEM) mode implies two UARTs. IR mode implies one UART plus one IR port. The IR port is either Full Duplex or Half Duplex where: # Full Duplex e Both transmitter and receiver are enabled simultaneously. # Half Duplex e The receiver input is blocked to ‘‘1’’ while the transmitter is busy; from the beginning of the start bit till the end of the stop bit(s). The IR port is either in Encoded mode or Non-Encoded mode where: # Encoded mode e UART2’s SOUT pin is encoded and routed to the IRTX pin, and the IRRX pin is decoded and routed to UART2’s SIN pin. # Non-Encoded mode e UART2’s SOUT pin is inverted and routed to the IRTX pin, and the IRRX is routed to UART2’s SIN pin. When in Encoded mode, the IR port’s output signal (IRTX) is active for intervals of either 1.6 ms or 3/16 baud. Configuration of UART2 is still handled via the FER, FAR, and PTR registers. However, SCF0 bit 3 is used instead of TL/C/12379 – 38 FIGURE 7-1. UART2 Serial and IR Interface Block Diagram 69 8.0 Parallel Port TABLE 8-3. SPP Data Register Read and Write Modes 8.1 INTRODUCTION This parallel interface is designed to provide all of the signals and registers needed to communicate through a standard parallel printer port as found in the IBM PC-XT, PC-AT, PS/2 and Centronics systems. This parallel port supports three standard modes of operation: SPP, EPP, ECP. The Standard Parallel Port (SPP) is a software based protocol with performance of up to 150 kbps. The Enhanced Parallel Port (EPP) is a hardware protocol which offers up to 2 Mbps. The Extended Capabilities Port (ECP) is also a hardware protocol with up to 2 Mbps transfer rate. In addition, the ECP has FIFO’s for receive and transmit, and DMA support, to reduce the CPU overhead. The ECP mode 0 is in fact compatible with the SPP mode. The ECP specification defines the AC/DC parameters of the signals to allow fast communication without termination problems. All the above standards are incorporated into the 1284 IEEE specifications. The address decoding of the registers in SPP mode utilizes A0 and A1 as shown in Table 8-1. (Address decoding of registers in EPP and ECP modes is shown in Table 8-5 and 8-7 respectively.) Table 8-3 shows the Reset states of Parallel port registers and pin signals. These registers are shown in Section 8.2 to Section 8.4. A0 Address 0 0 0 Data Register Read/Write 0 1 1 Status Read 1 0 2 Control Read/Write 1 1 3 TRI-STATE Access 0 1 0 Data Written to PD0 – 7 Result 0 X 0 1 Data Read from the Output Latch 1 0 1 0 Data Written to PD0 – 7 1 1 1 0 Data Written is Latched 1 0 0 1 Data Read from the Output Latch 1 1 0 1 Data Read from PD0 – 7 This is a bidirectional data port that transfers 8-bit date. The direction is determined by the PTR7 and the CTR5 bits. When PTR7 is high, the CTR5 bit will determine the data direction in conjunction with the Read and Write strobes. When PTR7 bit is low, the parallel port operates in the output mode only. See PTR7 bit and CTR5 bit for further information. The reset value of this register is 0. See Table 8-3. PTR7 Extended WR 1 TL/C/12379 – 39 TABLE 8-2. Standard Parallel Port Modes Selection Port Function RD X 8.2 DATA REGISTER (DTR) Special circuitry provides protection against damage that might be caused when the printer is powered but the PC87306 is not. There are two Standard Parallel Port (SPP) modes of operation (Compatible and Extended, see Table 8-2), two Enhanced Parallel Port (EPP) modes of operation and one Extended Capabilities Port (ECP) mode to complete a full IEEE 1284 parallel port. Compatible CTR5 0 To support a high transfer rate, the I/O buffers of the parallel port signals require improved impedance matching. IEEE P1284 defines two types of I/O buffers, Level 1 (standard I/O buffersÐopen-drain with weak pull-up), and Level 2 (new I/O buffersÐpush-pull with driver impedance and slew rate according to new standards). IEEE P1284, Draft 2.00, September 10, 1993, Microsoft’s ECP Protocol, and ISA Interface Standard, revision 1.14, July 7, 1993 provide detailed descriptions of Level 2 I/O buffers. When in EPP 1.9 or ECP mode 3, or ECP mode 2 with bit 1 of PCR equal to 1, the following parallel port pins have Level 2 I/O buffers: PD0 – PD7, STB, AFD, INIT, SLIN, ACK, ERR, SLCT, PE and BUSY. (See DC Electrical Characteristics, Section 11.1 for further information.) TABLE 8-1. Parallel Interface Register Addresses A1 PTR7 8.3 STATUS REGISTER (STR) In Compatible mode a write operation causes the data to be presented on pins PD0–PD7. A read operation in this mode causes the Data Register to present the last data written to it by the CPU. See Table 8-3. In the Extended Mode a write operation to the data register causes the data to be latched. If the Data Port Direction bit (CTR5) is 0, the latched data is presented to the pins; if it is 1 the data is only latched. When Data Port Direction bit (CTR5) is 0, a read operation from this register allows the CPU to read the last data it wrote to the port. In the Extended Mode with the Data Port Direction bit set to 1 (read), a read from this register causes the port to present the data on pins PD0 – PD7. TL/C/12379 – 40 This register provides status for the signals listed below. It is a read only register. Writing to it is an invalid operation that has no effect. 70 8.0 Parallel Port (Continued) Bit 0 When in EPP mode, this is the timeout status bit. When this bit is 0, no timeout. When this bit is 1, timeout occurred on EPP cycle (minimum 10 ms). It is cleared to 0 after STR is read, i.e., consecutive reads (after the first read) always return 0. It is also cleared to 0 when EPP is enabled (bit 0 of PCR is changed from 0 to 1). When not in EPP mode, this bit is 1. Bit 1 Reserved, this bit is always 1. Bit 2 In the compatible mode (PTR7 bit is 0), or in ECP and EPP modes, this bit is always one. In the Extended Mode (PTR7 bit is 1), or in ECP and EPP with bit 4 of PCR e 1, this bit is the IRQ STATUS bit. In the Extended mode if CTR4 e 1, then this bit is latched low when the ACK signal makes a transition from low to high. Reading this bit sets it to a one. Bit 3 This bit represents the current state of the printer error signal (ERROR). The printer sets this bit low when there is a printer error. This bit follows the state of the ERR pin. Bit 4 This bit represents the current state of the printer select signal (SLCT). The printer sets this bit high when it is selected. This bit follows the state of the SLCT pin. Bit 5 This bit represents the current state of the printer paper end signal (PE). The printer sets this bit high when it detects the end of the paper. This bit follows the state of the PE pin. Bit 6 This bit represents the current state of the printer acknowledge signal (ACK). The printer pulses this signal low after it has received a character and is ready to receive another one. This bit follows the state of the ACK pin. Bit 7 This bit (BUSY) represents the current state of the printer busy signal. The printer sets this bit low when it is busy and cannot accept another character. This bit is the inverse of the (BUSY/WAIT) pin. Bit 2 This bit (INIT) directly controls the signal to initialize the printer via the INIT pin. Setting this bit to low initializes the printer. This bit follows the INIT pin. Note: This pin must be set to 1 before enabling the EPP or ECP modes via bits 0 or 2 of the PCR register. Bit 3 This bit (SLIN) directly controls the select in signal to the printer via the SLIN pin. Setting this bit high selects the printer. It is the inverse of the SLIN pin. Bit 4 This bit controls the parallel port generated by the ACK signal. Its function changes slightly depending on the parallel port mode selected. In ECP mode this bit should be set to 0. In the following description, IRQx indicates either IRQ5 or IRQ7 (based upon PTR3). Compatible mode: when bit 4 e 0 IRQx is floated when bit 4 e 1 IRQx follows ACK transitions Extended mode: when bit 4 e 0 IRQx is floated when bit 4 e 1 IRQx becomes active on ACK trailing edge EPP mode: when bit 4 e 0 IRQx is floated when bit 4 e 1 IRQx follows ACK transitions, or an EPP timeout occurs. Bit 5 This bit determines the parallel port direction when bit 7 of PTR is 1. The default condition results in the parallel port being in the output mode. This is a Read/Write bit in EPP mode, and the software should write a 1 to it. In SPP mode it is a write only bit; a read from it will return 1. See Table 8-3 for further details. Bits 6, 7 Reserved. These bits are always 1. Normally when the Control Register is read, the bit values are provided by the internal output data latch. These bit values can be superseded by the logic level of the STB, AFD, INIT, and SLIN pins, if these pins are forced high or low by an external voltage. In order to force these pins high or low the corresponding bits should be set to their inactive state (e.g., AFD e STB e SLIN e 0, INIT e 1). See Table 8-4. TABLE 8-4. Parallel Port Reset States 8.4 CONTROL REGISTER (CTR) Signal TL/C/12379 – 41 This register provides all output signals to control the printer. Except for bit 5, it is a read and write register. Bit 0 This bit (STB) directly controls the data strobe signal to the printer via the STB pin. This bit is the inverse of the STB pin. Bit 1 This bit (AFD) directly controls the automatic feed XT signal to the printer via the AFD pin. Setting this bit high causes the printer to automatically feed after each line is printed. This bit is the inverse of the AFD pin. Reset Control State after Reset SLIN MR TRI-STATE INIT MR Zero AFD MR TRI-STATE STB MR TRI-STATE IRQ5, 7 MR TRI-STATE 8.5 ENHANCED PARALLEL PORT OPERATION EPP mode provides for greater throughput, and more complexity, than Compatible or Extended modes by supporting faster transfer times and a mechanism that allows the host to address peripheral device registers directly. Faster transfers are achieved by automatically generating the address and data strobes. EPP is compatible with both Compatible and Extended mode parallel-port devices. It consists of eight (0 – 7) single-byte registers (See Table 8-5). 71 8.0 Parallel Port (Continued) TABLE 8-5. EPP Register Addresses A2 A1 A0 Address Register Access Description 0 0 0 0 Data (DTR) R/W A write to this register sets the state of the eight data pins on the 25-pin Dshell connector. 0 0 1 1 Status (STR) R A read from this register presents the system microprocessor with the real-time status of five pins on the 25-pin D-shell connector, and the IRQ. 0 1 0 2 Control (CTR) R/W A write operation to this port sets the state of four pins on the 25-pin Dshell connector, and controls both the parallel port interrupt enable and direction. 0 1 1 3 Address R/W A write operation to this port initiates an EPP device/register selection operation. 1 0 0 4 Data Port 0 R/W Accesses to this port initiate device read or write operations with bits 0 – 7. 1 0 1 5 Data Port 1 R/W This port is only accessed to transfer bits 8 to 15 of a 16-bit read or write to data port 0. 1 1 0 6 Data Port 2 R/W This port is only accessed to transfer bits 16 to 23 of a 32-bit read or write to data port 0. 1 1 1 7 Data Port 3 R/W This port is only accessed to transfer bits 24 to 31 of a 32-bit read or write to data port 0. 4. When IOCHRDY goes high it causes WR to go high. If WAIT is high during the host write cycle then the EPP does not pull IOCHRDY to low. 5. When WR goes high it causes the EPP to pull WRITE and ASTRB to high. Only when WRITE and ASTRB are high can the EPP change PD0 – 7. There are two EPP modes: EPP rev 1.7 is supported when bit 0 of PCR is 1, and bit 1 of PCR is 0. EPP rev 1.9 (IEEE 1284) is supported when bit 0 of PCR is 1, and bit 1 of PCR is 1. EPP is supported for a parallel port whose base address is 8-bytes aligned (A0, 1, 2 e 0, 0, 0). Otherwise unpredictable results occur. There are four EPP transfer operations: address write, address read, data write and data read. An EPP transfer operation is composed of a host read or write cycle (from or to an EPP register) and an EPP read or write cycle (from a peripheral device to an EPP register, or from an EPP register to a peripheral device). The software must write zero to bits 0, 1 and 3 of the CTR register, before accessing the EPP registers, since the pins controlled by these bits are controlled by hardware during EPP access. Once these bits are written with zero, the software may issue multiple EPP access cycles. When EPP is enabled, the software should write a 1 to bit 5 of CTR. To meet the EPP 1.9 specifications, the software should write to DTR only when bit 7 of STR is 1. EPP 1.7 Address Write The following procedure selects a peripheral device or register. See also Figure 8-1 . 1. The host writes a byte to the EPP address register. WR goes low to latch D0–D7 into the address register. The latch drives the address register onto PD0–7 and the EPP pulls WRITE low. 2. The EPP pulls ASTRB low to indicate that data has been sent. 3. If WAIT is low during the host write cycle, IOCHRDY goes low. When WAIT goes high, the EPP pulls IOCHRDY high. TL/C/12379 – 42 FIGURE 8-1. EPP 1.7 Address Write EPP 1.7 Address Read The following procedure reads from the address register. See also Figure 8-2 . 1. The host reads a byte from the EPP address register. RD goes low to gate PD0 – 7 into D0 – 7. 2. The EPP pulls ASTRB low to signal the peripheral to start sending data. 72 8.0 Parallel Port (Continued) 5. If the peripheral is fast enough to pull WAIT low before the host terminates the write cycle, the EPP pulls IOCHRDY to low, but does not pull ZWS to low, thus carrying out a normal (non-ZWS EPP 1.7) write operation. 3. If WAIT is low during the host read cycle, then the EPP pulls IOCHRDY low. When WAIT goes high, the EPP stops pulling IOCHRDY to low. 4. When IOCHRDY goes high it causes RD to go high if WAIT is high during the host read cycle then the EPP does not pull IOCHRDY to low. 5. When RD goes high, it causes the EPP to pull ASTRB high. Only when ASTRB is high can the EPP change PD0 – 7. After ASTRB goes high, the EPP puts D0–7 in TRISTATE. Note: A read operation is similar, except for the data direction, and the activation of RD instead of WR. EPP Zero Wait State (ZWS) Data Write/Read Operation (both 1.7 and 1.9) An EPP 1.7 and 1.9 Zero Wait State data write/read operation is similar to the EPP Zero Wait State address write/read operation with the exception that the data strobe (DSTRB signal), and a data register, replace the address strobe (ASTRB signal) and the address register, respectively. See Figure 8-3 . TL/C/12379 – 43 FIGURE 8-2. EPP 1.7 Address Read EPP 1.7 Data Write and Data Read This procedure writes to the selected peripheral device or register. See also Figure 8-3 . An EPP 1.7 data write operation is similar to the EPP 1.7 address write operation, and an EPP 1.7 data read operation is similar to the EPP 1.7 address read operation, except that the data strobe (DSTRB signal), and a data register, replace the address strobe (ASTRB signal) and the address register respectively. TL/C/12379 – 44 FIGURE 8-3. EPP Write with ZWS EPP 1.9 Address Write The following procedure selects a peripheral or register. See also Figure 8-4 . 1. The host writes a byte to the EPP address register. 2. The EPP pulls IOCHRDY low, and waits for WAIT to go low. 3. When WAIT goes low the EPP pulls WRITE to low drives the latched byte onto PD0 – 7. If WAIT was already low, steps 2 and 3 occur concurrently. 4. The EPP pulls ASTRB low and waits for WAIT to go high. EPP Zero Wait State (ZWS) Address Write/Read Operation (both 1.7 and 1.9) The following procedure performs a short write to the selected peripheral device or register. ZWS should be configured as follows: bit 5 of FCR is 1 and bit 6 of FCR is 0. 1. The host writes a byte to the EPP address register. WR goes low to latch D0–7 into the data register. The latch drives the data register onto PD0–7. 2. The EPP first pulls WRITE low, and then pulls ASTRB low to indicate that data has been sent. 3. If WAIT is high during the host write cycle, ZWS goes low and IOCHRDY goes high. 4. When the host pulls WR high, the EPP pulls ASTRB, ZWS and WRITE to high. Only when WRlTE and ASTRB are high can the EPP change PD0–7. 5. When WAIT goes high, the EPP stops pulling IOCHRDY low, pulls ASTRB high, and waits for WAIT to go low. 6. Only if no EPP write is pending, when WAIT goes low, (or when bit 7 of PTR is 1, and the direction is changed to Backwards by setting bit 5 of CTR to 1), the EPP pulls WRITE to high. If an EPP write is pending WRITE remains low, and the EPP may change PD0 – 7. 73 8.0 Parallel Port (Continued) 4. When RD goes high, the EPP puts D0 – 7 in TRI-STATE. EPP 1.9 Data Write and (Backward) Data Read This procedure writes to the selected peripheral drive or register. See Figure 8-5 . EPP 1.9 data read and write operations are similar to EPP 1.9 address read and write operations, respectively, except that the data strobe (DSTRB signal) and a data register replace the address strobe (ASTRB signal) and the address register. Table 8-6 shows the standard 25-pin, D-type connector definition for various parallel port operations. TL/C/12379–45 FIGURE 8-4. EPP 1.9 Address Write EPP 1.9 Address Read The following procedure reads from the address register. See also Figure 8-5 . 1. The host reads a byte from the EPP address register. When RD goes low, the EPP pulls IOCHRDY low, and waits for WAIT to go low. 2. When WAIT goes low, the EPP pulls ASTRB low and waits for WAIT to go high. If wait was already low, steps 2 and 3 occur concurrently. 3. When WAIT goes high, the EPP stops pulling IOCHRDY low, latches PD0–7, and pulls ASTRB high. TL/C/12379 – 46 FIGURE 8-5. EPP 1.9 Address Read TABLE 8-6. Parallel Port Pin Out Connector Pin No. Chip Pin No. SPP, ECP Mode Pin Direction EPP Mode Pin Direction 1 137 STB I/O WRITE I/O 2 136 PD0 I/O PD0 I/O 3 135 PD1 I/O PD1 I/O 4 134 PD2 I/O PD2 I/O 5 133 PD3 I/O PD3 I/O 6 131 PD4 I/O PD4 I/O 7 130 PD5 I/O PD5 I/O 8 129 PD6 I/O PD6 I/O 9 128 PD7 I/O PD7 I/O 10 127 ACK I ACK I 11 126 BUSY I WAIT I 12 125 PE I PE I 13 124 SLCT I SLCT I 14 119 AFD I/O DSTRB I/O 15 120 ERR I ERR I 16 122 INIT I/O INIT I/O 17 123 SLIN I/O ASTRB I/O 74 8.0 Parallel Port (Continued) ward direction when bit 5 of DCR is 1). All DMA transfers are to or from these registers. The ECP does not assert DMA request for more than 32 consecutive DMA cycles. The ECP stops requesting DMA when TC is detected during an ECP DMA cycle. Writing into a full FIFO, and reading from an empty FIFO, are ignored. The written data is lost, and the read data is undefined. The FIFO empty and full status bits are not affected by such access. Some registers are not accessible in all modes of operation, or may be accessed in one direction only. Accessing a non accessible register has no effect: Data read is undefined, data written is ignored, the FIFO does not update. The PV87323VF Parallel Port registers (DRT, STR and CTR) are not accessible when ECP is enabled. To improve noise immunity in ECP cycles, the state machine does not examine the control handshake response lines until the data has had time to switch. IN ECP mode: DATAR replaces DTR of SPP/EPP modes DSR replaces SPR of SPP/EPP modes DCR replaces CTR of SPP/EPP modes A detailed description of the various modes follows in Sections 8.8 – 8.11. 8.6 EXTENDED CAPABILITIES PARALLEL PORT (ECP) 8.6.1 Introduction The ECP support includes a 16-byte FIFO that can be configured for either direction, command/data FIFO tags (one per byte), a FIFO threshold interrupt for both directions, FIFO empty and full status bits, automatic generation of strobes (by hardware) to fill or empty the FIFO, transfer of commands and data, and a Run Length Encoding (RLE) expanding (decompression) as explained below. The Extended Capabilities Port (ECP) is enabled when bit 2 of PCR is 1. Once enabled, its mode is controlled via the mode field of ECRÐbits 5,6,7 of ECR register. The ECP has ten registers. See Table 8-7. The AFIFO, CFIFO, DFIFO and TFIFO registers access the same ECP FIFO. The FlFO is accessed at Base a 000h, or Base a 400h, depending on the mode field of ECR and the register. FIFO can be accessed by host DMA cycles, as well as host PIO cycles. When DMA is configured and enabled (bit 3 of ECR is 1 and bit 2 of ECR is 0) the ECP automatically (by hardware) issues DMA requests to fill the FIFO (in the forward direction when bit 5 of DCR is 0) or to empty the FIFO (in the back- TABLE 8-7. ECP Registers A10 A1 A0 Offset Address Access Size Mode Ý ECR (5 – 7) Function 0 0 0 0 DATAR R/W Byte 000,001 Parallel Port Data Register 0 0 0 0 0 1 0 AFIFO W Byte 011 ECP Address FIFO 1 DSR R Byte ALL Status Register 0 1 1 0 0 2 DCR R/W Byte ALL Control Register 0 3 CFIFO W Byte 010 1 Parallel Port Data FIFO 0 0 3 DFIFO R/W Byte 011 ECP Data FIFO 1 0 0 3 TFIFO R/W Byte 110 Test FIFO 1 0 0 3 CNFGA R Byte 111 Configuration Register A 1 0 1 4 CNFGB R Byte 111 Configuration Register B 1 1 0 5 ECR R/W Byte ALL Extended Control Register Register The Base address is stored in bits A2–A9. It is 278h, 378h or 3BCh, as specified in the FAR register. 75 8.0 Parallel Port (Continued) Reading this register has no effect and the data read is undefined. Writes to this register during backward direction (bit 5 of DCR is 1) have no effect and the data is ignored. DSR: Read only. Same as the current STR register, except for bit 2, which is reserved. Writes to this register have no effect and the data is ignored. 8.6.2 Software Operation Software operation is detailed in the document Extended Capabilities Port Protocol and ISA Interface Standard. To highlight the ECP usage some software operations are detailed below: 1. The software should enable ECP (bit 2 of PCR is 1) after bits 0 – 4 of the Parallel Port Control Register (CTR) are 00100. 2. When ECP is enabled, and the software wishes to switch modes, it should switch only through modes 000 or 001. 3. When ECP is enabled, the software should change direction only in mode 001. 4. The software should switch from mode 010, or 011 to mode 000, or 001, only when the FIFO is empty. 5. The software should switch to mode 011 when bits 0 and 1 of DCR are 0. 6. The software should switch to mode 010 when bit 0 of DCR is 0. 7. The software should disable ECP (bit 2 of PCR is 0) only when in mode 000 or 001. Software may switch from mode 011 backward direction to modes 000 or 001 when there is an on-going ECP read cycle. In this case the read cycle is aborted by deasserting AFD. The FIFO is reset (empty) and a potential byte expansion (RLE) is automatically terminated since the new mode is 000 or 001. The ZWS signal is asserted by the ECP when ECP is enabled, and an ECP register is accessed by host PIO instructions, thus using a host zero wait cycle. The ECP uses the X1/OSC clock. This clock can be frozen (a power-down mode). When this power-down mode occurs, the DMA is disabled, all interrupts (except ACK) are masked, and the FIFO registers are not accessible (access is ignored). The other ECP registers are always accessible when the ECP is enabled. During this period the FIFO status and contents are not lost, although the host reads bit 2 of ECR as 0, bit 1 of ECR as 1 and bit 0 of ECR as 1, regardless of the actual values of these bits. When the clock starts toggling again these bits resume their original functions (values). When the clock is frozen, an on-going ECP cycle may be corrupted but next ECP cycle will not start. This is true even if in forward direction the FIFO is not empty, and in backward direction the FIFO is not full. If the ECP clock starts or stops toggling during a host cycle that accesses the FIFO, the cycle may yield wrong data. Note: The FDC has a register of the same name (DSR). DCR: Same as the SPP CTR register, with the following exceptions: 1. When bit 5 of the DCR is 0 the ECP is in forward direction, and when bit 5 is high (1) the ECP is in backward direction. 2. The ECP drives the PD0 – PD7 pins in the forward direction but does not drive them in the backward direction. 3. The direction bit, bit 5, is readable and writable. 4. When the control bit is read, the bit values are provided by the internal data latch. These bit values are not superseded by the logic level of the STB, AFD, INIT and SLIN pins. 5. Bit 4 of the DCR enables the ACK deassertion interrupt event (1 e enable, 0 e mask). If a level interrupt is configured (bit 4 of PCR is 1) clearing this bit clears the ACK pending interrupt request. This bit does not float the IRQ pin. In mode 011, when bit 1 of DRC is 0, AFD is controlled by ECP hardware. When bit 1 of DCR is 1, AFD is driven low. In mode 011 or 010, when bit 0 of DCR is 0, STB is controlled by the ECP hardware. When bit 0 of DCR is 1, STB is driven low. CFIFO: Parallel Port FIFO Register. Write only. A byte written, or DMAed, to this register is pushed into the FIFO and tagged as data. Reading this register has no effect and the data read is undefined. DFIFO: ECP Data FIFO Register. In the forward direction (bit 5 of DCR is 0) a byte written, or DMAed, to this register is pushed into the FIFO and tagged as data. Reading this register has no effect and the data read is undefined. In the backward direction (bit 5 of DCR is 1) the ECP automatically issues ECP read cycles to fill the FIFO. Reading this register pops a byte from the FIFO. Writing this register has no effect and the data written is ignored. TFIFO: Test FlFO Register. A byte written into this register is pushed into the FIFO. A byte read from this register is popped from the FIFO. The ECP does not issue a ECP cycle to transfer the data to or from the peripheral device. The TFIFO is readable and writable in both directions. In the forward direction (bit 5 of DCR is 0) PD0 – PD7 are driven, but the data is undefined. Note 1: The ECP outputs are inactive when the ECP is disabled. Note 2: Only the FIFO/DMA/RLE are not functional when the clock is frozen. All other registers are accessible and functional. The FIFO/ DMA/RLE are affected by ECR modifications. i.e., they are reset even when exits from modes 010/011 are carried out while the clock is frozen. 8.7 REGISTER DEFINITIONS DATAR: Same as DTR register, except that read always returns the values of the PD0–PD7 pins (not the register latched data). AFIFO: ECP Address FIFO Register. Write Only. In the forward direction (bit 5 of DCR is 0) a byte written into this register is pushed into the FIFO and tagged as command. 76 8.0 Parallel Port (Continued) The FIFO does not stall when overwritten or underrun (access is ignored). Bytes are always read from the top of the FIFO, regardless of the direction bit (bit 5 of DCR). For example if 44h, 33h, 22h, 11h is written into the FIFO, reading the FIFO returns 44h, 33h, 22h, 11h (in the same order it was written). CNFGA: Configuration Register A. Read only. Reading this register always returns 0001x100. Writing this register has no effect and the data is ignored. Bit 3 of CNFGA reflects the value of bit 5 of the ASC register. CNFGB: Configuration Register B. Read only. Reading this register returns the configured parallel port interrupt line, and its state, as follows. Bit 7 This bit is always 0. Bit 6 Holds the (non-inverted) value on the configured IRQ pin. Bits 5, 4 These bits are 1 when IRQ5 is configured, and 0 when IRQ7 is configured. Bit 3 This bit is always 1. Bit 2 This bit is always 0. Writing this register has no effect and the data is ignored. Bits 1, 0 These bits reflect the values of bit 2 and 1, respectively, of the SCF1 register. ECR: Extended Control Register. This register controls the ECP and parallel port functions. Upon reset this register is initialized to 00010101. IOCHRDY is driven low on ECR read when the ECR status bits are not holding updated data. Bits 7 – 5 (Mode) Bit 7 is the MSB of the three bits value. 000: Standard mode. Write cycles are performed under software control. STB, AFD, INIT, and SLIN are open-drain outputs. Bit 5 of DCR is ignored (assumes forward direction) and PD0–PD7 are driven. The FIFO is reset (empty). 001: PS/2 mode. Read and write cycles are performed under software control. The FIFO is reset (empty). STB, AFD, INIT, and SLIN are open-drain outputs. 010: Parallel Port FIF0 mode. Write cycles are performed under hardware control (STB is controlled by hardware). Bit 5 of DCR is ignored (assumes forward direction) and PD0–PD7 are driven. STB, AFD, INIT, and SLIN are open-drain output buffers when bit 1 of PCR is 0. When bit 1 of PCR is 1, the output buffers are push-pull. 011: ECP FIFO mode. The FIFO direction is controlled by bit 5 of DCR. Read and write cycles to the device are performed under hardware control (STB and AFD are controlled by hardware). STB, AFD, INIT, and SLIN are push-pull output buffers. 100: Reserved. 101: Reserved. Bit 4 Bit 3 110: FIFO test mode. The FIFO is accessible via the TFIFO register. The ECP does not issue ECP cycles to fill/ empty the FIFO. 111: Configuration mode. The CNFGA and CNFGB registers are accessible in this mode. ECP Interrupt Mask bit. When this bit is 0 an interrupt is generated on ERR assertion (the high-to-low edge of ERR). An interrupt is also generated when ERR is asserted while this bit is changed from 1 to 0; this prevents the loss of an interrupt between ECR read and ECR write. When this bit is 1, no interrupt is generated. ECP DMA Enable bit. When this bit is 0, DMA is disabled and the PDRQ pin is in TRI-STATE. When this bit is 1, DMA is enabled and DMA starts when bit 2 of ECR is 0. Note: PDACK is assumed inactive when this bit is 0. Bit 2 Bit 1 Bit 0 77 ECP Service bit. When this bit is 0, and one of the following three interrupt events occur, an interrupt is generated and this bit is set to 1 by hardware. 1. Bit 3 of ECR is 1, and terminal count is reached during DMA. 2. Bit 3 of ECR is 0 and bit 5 of DCR is 0, and there are eight or more bytes free in the FIFO. 3. Bit 3 of ECR is 0 and bit 5 of DCR is 1, and there are eight or more bytes to be read from the FIFO. When this bit is 1, DMA and the above three interrupts are disabled. Writing 1 to this bit does not cause an interrupt. When the ECP clock is frozen this bit is read as 0, regardless of its actual value (even though the bit may be modified by software when the ECP clock is frozen). FIFO Full bit. Read only. This bit is 0 when the FIFO has at least one free byte. This bit is 1 when the FIFO is full. This bit continuously reflects the FIFO state, and therefore can only be read. Data written to this bit is ignored. When the ECP clock is frozen this bit is read as 1, regardless of the actual FIFO state. FIFO Empty bit. Read only. This bit is 0 when the FiFO has at least one byte of data. This bit is 1 when the FIFO is empty. This bit continuously reflects the FIFO state, and therefore can only be read. Data written to this bit is ignored. When the ECP clock is frozen this bit is read as 1, regardless of the actual FIFO state. 8.0 Parallel Port (Continued) 8.8 SOFTWARE CONTROLLED DATA TRANSFER (Modes 000 and 001) deasserts STB to terminate the write cycle. In 011 mode the ECP waits for BUSY to be high before it deasserts STB. Software controlled data transfer is supported in modes 000 and 001. The software generates peripheral-device cycles by modifying the DATAR and DCR registers and reading the DSR, DCR and DATAR registers. The negotiation phase and nibble mode transfer, as defined in the IEEE 1284 standard, are performed in these modes. In these modes the FIFO is reset (empty) and is not functional, the DMA and RLE are idle. Mode 000 is for the forward direction only; the direction bit is forced to 0 and PD0–7 are driven. Mode 001 is for both the forward and backward directions. The direction bit controls whether PD0 –7 are driven. When BUSY is high the ECP deasserts STB and changes AFD and PD0 – 7 only after BUSY is low. 8.9 AUTOMATIC DATA TRANSFER (Modes 010 and 011) Automatic data transfer (ECP cycles generated by hardware) is supported only in modes 010 and 011. Automatic DMA access to fill or empty the FIFO is supported in modes 010, 011 and 110. Mode 010 is for the forward direction only. The direction bit is forced to 0 and PD0–7 are driven. Mode 011 is for both the forward and backward directions. The direction bit controls whether PD0–7 are driven. Automatic Run Length Expanding (RLE) is supported in the backward direction. 8.9.3 Backward Direction (bit 5 of DCR is 1) When the ECP is in the backward direction, and the FIFO is not full (bit 1 of ECR is 0), the ECP issues a read cycle from the peripheral device and monitors the BUSY signal. If BUSY is high the byte is a data byte and it is pushed into the FIFO. If BUSY is low the byte is a command byte. The ECP checks bit 7 of the command byte, if it is high the byte is ignored, if it is low the byte is tagged as an RLC byte (not pushed into the FIFO but used as a Run Length Count to expand the next byte read). Following an RLC read the ECP issues a read cycle from the peripheral device to read the data byte to be expanded. This byte is considered a data byte, regardless of its BUSY state (even if it is low). This byte is pushed into the FIFO (RLC a 1) times (i.e., RLC e 0: push the byte once, RLC e 127: push the byte 128 times). When the ECP is in the backward direction, and the FIFO is not empty (bit 0 of ECR is 0), the FIFO can be emptied by software reads from the FIFO register (only DFIFO in mode 011, no AFIFO and CFIFO read). When DMA is enabled (bit 3 of ECR is 1 and bit 2 of ECR is 0) the ECP automatically issues DMA requests to empty the FIFO (only in mode 011). TL/C/12379 – 47 FIGURE 8-6. ECP Forward Write Cycle Note 1: FIFO-full condition is checked before every expanded byte push. Note 2: A pending DMA request is removed, and a pending RLE expansion is aborted, when you switch from modes 010 or 011 to other modes. Note 3: The two FIFO ports are neither synchronized nor linked together, except via the empty and full FIF0 status bits. The FIFO shall not delay the push and pop operations, even when they are performed concurrently. Care must be taken not to corrupt PD0–PD7 or D0–7 while the other FIFO port is accessed. Note 4: In the forward direction, the empty bit is updated when the ECP cycle is completed, not right after the last byte is popped out of the FIFO (valid cleared on cycle end). Note 5: ZWS is not asserted for DMA cycles. Note 6: The one-bit command/data tag is used only in forward direction. 8.9.1 Forward Direction (bit 5 of DCR e 0) When the ECP is in forward direction and the FIFO is not full (bit 1 of ECR is 0) the FIFO can be filled by software writes to the FlFO registers (AFIFO and DFIFO in mode 011, and CFIFO in mode 010). When DMA is enabled (bit 3 of ECR is 1 and bit 2 of ECR is 0) the ECP automatically issues DMA requests to fill the FIFO with normal data byte. When the ECP is in forward direction and the FIFO is not empty (bit 0 of ECR is 0) the ECP pops a byte from the FIFO and issues write cycle to the peripheral device. The ECP drives AFD according to the operation mode (ECR bits 5–7) and according to the tag of the popped byte as follows: In Parallel Port FIFO mode (mode 010) AFD is controlled by bit 1 of DCR. In ECP mode (mode 011) AFD is controlled by the popped tag. AFD is driven high for normal data byte and driven low for command byte. 8.9.4 ECP Backward Read Cycle An ECP read cycle starts when the ECP drives AFD low. The peripheral device drives BUSY high for a normal data read cycle, or drives BUSY low for a command read cycle, and drives the byte to be read onto PD0 – 7. When ACK is asserted the ECP drives AFD high. When AFD is high the peripheral device deasserts ACK. The ECP reads the PD0 – 7 byte, then drives AFD low. When AFD is low the peripheral device may change BUSY and PD0 – 7 states in preparation for the next cycle. 8.9.2 ECP Forward Write Cycle An ECP write cycle starts when the ECP drives the popped tag onto AFD and the popped byte onto PD0–7. When BUSY is low the ECP asserts STB. In 010 mode the ECP TL/C/12379 – 48 FIGURE 8-7. ECP Backward Read Cycle 78 9.2 IDE SIGNALS 8.0 Parallel Port (Continued) Using ’LS244 devices in the IDE interface provides buffering of the control and address lines. Four control signals, IDEHI, IDELO, HCS0, HCS1, one status signal, IOCS16, and one data signal, IDED7, are required by the IDE interface. The PC87306 provides all of these signals. They are summarized below. IDEHI enables an ’LS245 octal bus transceiver for the upper data lines (D15 – 8) during 16-bit read and write operations at addresses 1F0 – 1F7. IDEHI will activate the ’LS245 only if the IOCS16 output from the hard drive is active. IDELO enables another ’LS245 octal bus transceiver for the lower data lines (D7 – 0) during all (1F0 – 1F7, 3F6 and 3F7) reads and writes. The IDED7 signal insures that the D7 data bus signal line is disabled for address 3F7 (this bit is used for the Disk Changed register on the floppy disk controller at that address). The two ’LS245 chips are used to enable or TRI-STATE the data bus signals. In the PC-AT mode the PC87306 provides the two hard disk chip selects (HCS0, HCS1) for the IDE interface. The HCS0 output is active low when the 1F0 – 1F7h I/O address space is chosen and corresponds to the 1FX signal on the IDE header. The HCS1 output is active low when the 3F6 or 3F7 I/O addresses are chosen and corresponds to 3FX on the IDE header. These are the two address blocks used in the PC-AT hard disk controller. Table 9-1 summarizes the addresses used by the PC-AT hard disk controller. 8.10 FIFO TEST ACCESS (Mode 110) Mode 110 is for testing the FIFO in PIO and DMA cycles. Both read and write operations (pop and push) are supported, regardless of the direction bit. In the forward direction PD0–PD7 are driven, but the data is undefined. This mode can be used to measure the hostECP cycle throughput, usually with DMA cycles. This mode can also be used to check the FIFO depth and its interrupt threshold, usually with PIO cycles. 8.11 CONFIGURATION REGISTERS ACCESS (Mode 111) The two configuration registers, CNFGA and CNFGB, are accessible only in this mode. 8.12 INTERRUPT GENERATION An interrupt pulse is generated when any of the following events occur: 1. When bit 2 of ECR is 0, bit 3 of ECR is 1 and TC is asserted during ECP DMA cycle. 2. When bit 2 of ECR is 0, bit 3 of ECR is 0, bit 5 of DCR is 0 and there are eight or more bytes free in the FIFO. It includes the case when bit 2 of ECR is cleared to 0 and there are already eight or more bytes free in the FIFO (modes 010, 011 and 110 only). 3. When bit 2 of ECR is 0, bit 3 of ECR is 0, bit 5 of DCR is 1 and there are eight or more bytes to be read from the FIFO. It includes the case when bit 2 of ECR is cleared to 0 and there are already eight or more bytes to be read from the FIFO (modes 011 and 110 only). 4. When bit 4 of ECR is 0 and ERR is asserted (high to low edge) or ERR is asserted when bit 4 of ECR is modified from 1 to 0. 5. When bit 4 of DCR is 1 and ACK is deasserted (low-tohigh edge). The interrupt is generated according to bits 5 and 6 of PCR. TABLE 9-1. IDE Registers and Their ISA Addresses Address Note: Interrupt events Ý2, Ý3 and Ý4 are level events, thus they are shaped as interrupt pulses. These interrupts are masked (inactive) when the ECP clock is frozen. Interrupt event Ý1 is a pulse event. The last interrupt event behaves as in the normal SPP mode: the IRQ signal follows the ACK signal transition (when bit 5 of PCR is 0 and bit 6 of PCR is 0). Note that interrupt event Ý4 may be lost when the ECP clock is frozen. Read Function Write Function 1F0 Data Data 1F1 Error Features (Write Precomp) 1F2 Sector Count Sector Count 1F3 Sector Number Sector Number 1F4 Cylinder Low Cylinder Low 9.0 Integrated Device Electronics Interface (IDE) 1F5 Cylinder High Cylinder High 1F6 Drive/Head Drive/Head 9.1 INTRODUCTION Another key interface for PC design is facilitated through the use of the PC87306 IDE (Integrated Drive Electronics) Hard Disk interface. Only three buffer chips are required to construct the IDE Hard Disk Interface circuit. The IDE interface is essentially the PC-AT bus ported to the hard drive. The hard disk controller resides on the hard drive itself. So the IDE interface circuit must provide the PC-AT bus signals, including data bits D15–0, address lines A3 – 0, as well as the common control signals. 1F7 Status Command 3F6 Alternate Status Device Control 3F7 Drive Address (Note) Not Used. Data Bus TRI-STATE Note: Data bus bit D7 is dedicated to the floppy disk controller at this address. When reading this address the floppy disk controller disk change status will be provided by bit D7. There is no write function at this address in the IDE associated with this bit. 79 9.0 Integrated Device Electronics Interface (IDE) (Continued) b) IDED7 is functional (i.e., read: IDED7 to D7, write D7 to IDED7) when IDEACK is asserted and either RD or WR is asserted. c) HCS1,0 are not asserted on DMA cycles. The equations show in Figure 9-1 define the signals of the PC87306 IDE pins. The IDE Interface Circuit has some additional IDE DMA support. When bit 1 of FCR is 1, the PC87306 IDE responds to a DMA acknowledgement on the IDEACK input pin as follows: a) IDEL0 is asserted when IDEACK is asserted and either RD or WR is asserted. Equations HCS0 e A9*A8*A7*A6*A5*A4*A3* AEN HCS1 e A9*A8*A7*A6*A5*A4*A3*A2*A1* AEN IDELO e [HCS0*(RD a WR)] a ÀHCS1*[(WR*A0) a RD] Ó IDEHI e IOCS16*HCS0*(RD a WR) IDED7 (read) e (HCS0*RD) a [(HCS1* A0)*RD] IDED7 (write) e WR*[HCS0 a (HCS1* A0)] Comments Active at 1F0 – 1F7 Active at 3F6, 3F7 Write 1F0 – 1F7, 3F6; Read 1F0 – 1F7, 3F6, 3F7 Read or Write 1F0 – 1F7 in AT Mode Provides D7 during Read 1F0 – 1F7 and 3F6 Provides D7 during Write 1F0 – 1F7 and 3F6 FIGURE 9-1. IDE Interface Signal Equations (Non-DMA) 80 10.0 Keyboard Controller and Real-Time Clock vices are available by special order. For applications which require program development capabilities, the RAM-based PC87323 SuperI/O Sidewinder is recommended. The keyboard controller (KBC) is a general purpose 8-bit microcontroller. It consists of 256 Bytes of data memory and 2 kBytes of Read-Only Memory (ROM), two 8-bit I/O ports, a three register system interface, an 8-bit timer/counter, and facilities for both binary and Binary Coded Decimal (BCD) arithmetic. Figure 10-1 shows the relationships of these functional blocks and Figure 10-2 displays the interface between the PC87306 and a host system. The ROM may be custom programmed during manufacture according to customer requirements. These versions of de- The KBC is enabled when bit 0 of KRR is 1. It is guaranteed to work with SYSCLK up to 16 MHz (when KRR7 e 1). The KBC is software compatible with the 8042AH industry standard keyboard controller as well as National’s PC87911. The PC87306 can execute code previously written for an 8042 without further development. TL/C/12379 – 49 FIGURE 10-1. Keyboard Controller Functional Block Diagram 81 10.0 Keyboard Controller and Real-Time Clock (Continued) The STATUS register holds status information related to the system interface. Figure 10-3 shows the bit definition. It is read-only by the system. Section 10.1 will describe in detail the operation of the functional blocks shown in Figure 10-1 . The RTC supports a 100-year calendar as well as a time-ofday clock with alarm features and battery backup. Section 10.2 will describe the RTC in detail. ST7 bit 10.1 PC87306 KBC FUNCTION 7 ST6 ST5 ST4 F1 F0 6 5 4 3 2 FIGURE 10-3. Status Register 10.1.1 Host System Interface IBF OBF 1 0 OBF . Output Buffer Full. A 1 indicates that data is written into the DBBOUT register. It is cleared by a system read operation. Bit 1 IBF . Input Buffer Full. When a write operation is performed by the host system, it will be set to 1. Upon executing an IN A,DBB instruction, it will be cleared. Bit 2 F0 . A general purpose flag that can be cleared or toggled by the keyboard controller software. Bit 3 F1 . Command/Data Flag. This flag holds the state of A2 when the host system performs a write operation. It is typically used to distinguish between commands and data coming from the host system. For example; when A2 e 1, F1 e 1 and indicates a command was written by the host. When A2 e 0, F1 e 0, indicating data was written by the host. Bits 4-7 ST4 – ST7 . General purpose flags. They can be written by a MOV STS,A instruction. Table 10-1 shows the register address decoding utilized by the keyboard control system interface. Bit 0 Common Signals The keyboard controller is interfaced to the host system through a common system interface. The interface consists of the address bus A0–A15, Address Enable (AEN) signal, data bus D0 – D7, and control signals I/O Read (IOR), and I/O Write (IOW) (see Figure 10-2 ). Address bus A0 –A15 connects to the system address SA0 – SA15 of the system. The PC87306 uses the system address and the AEN signal to decode the access to the configuration register. Data bus D0 – D7 connects to the peripheral bus XD0–XD7 of the system. The IOR and IOW inputs connect to the IOR and IOW lines of the system. All read and write operations to the PC87306 are I/O operations. The PC87306 decodes the keyboard controller chip-select from A0 – A15 (60h or 64h). System Interface Registers The keyboard controller consists of three 8-bit registers: Data Byte Buffer Output (DBBOUT), Data Byte Buffer Input (DBBIN), and STATUS. See Figure 10-1 . The DBBOUT register is used to transfer data from the keyboard controller to the host system. It is written by the keyboard controller using the OUT DBB,A instruction. A data read operation by the host system reads its content. The DBBIN register is used to transfer data from the host system to the keyboard controller. It is written by the host system. It is read by the keyboard controller using an IN A, DBB instruction. TABLE 10-1. Summary of System Interface Operations RD WR A2 A0 – 15* 0 1 0 0 Read DBBOUT Operation 1 0 0 0 Write DBBIN, F1 Clear 0 1 1 0 Read STATUS 1 0 1 0 Write DBBIN, F1 Set X X X 1 No Operation *A0–A15 e 00000000011000XX. TL/C/12379 – 50 FIGURE 10-2. Keyboard Controller to Host System Interface 82 10.0 Keyboard Controller and Real-Time Clock (Continued) TL/C/12379 – 53 FIGURE 10-4. Fast IRQ Latching and Clearing IRQ1 and IRQ2 When the KBC is enabled, the keyboard and mouse IRQ lines (IRQ1 and IRQ12) are either identical to, or a function of, the 8042’s P24 and P25; as detailed in Figure 10-4. IRQ12 can be placed in TRI-STATE by setting the IRQ12 TRI-STATE control bit (bit 2 of SCF0) to 1. In this case IRQ12 input is blocked to 1. NOTE: EN FLAGS command (used for routing OBF and IBF onto P24 and P25) will cause unpredictable results and should not be issued. 10.1.2 Program Memory The keyboard controller of the PC87306 has a 2k x 8 ROM based program memory. An 11-bit program counter allows direct access to every location of the program memory. Figure 10-5 shows the memory map. There are three special locations associated with hardware functions. 1. 000h After the keyboard controller is reset, the program counter is initialized to 000h. 2. 003h When the input buffer of the host interface (DBBlN) is full, and the IBF interrupt is enabled, the CPU makes an interrupt call to this location. 3. 007h When the timer overflows and the timer interrupt is enabled, the CPU makes an interrupt call to this location. 10.1.3 Data Memory and Registers The keyboard controller has 256 Bytes of data memory, including 2 banks of registers, 8 registers each, and an 8-level stack. Figure 10-5 shows the data RAM organization. TL/C/12379 – 51 FIGURE 10-5. Keyboard Controller Data Memory Map 83 10.0 Keyboard Controller and Real-Time Clock (Continued) RAM locations 0h –7h are used as register bank0. They are designated as R0–R7 respectively. Register bank1 (R0Ê – R7Ê ) are located in 18h–1Fh of the address map. Bank0 is the default register bank when the chip comes out from reset. Bank switching is accomplished by using ‘‘register bank select’’ instructions (SEL RB0, SEL RB1). Locations 8h– 17h are reserved for the stack. Each stack entry consists of 2 bytes. Figure 10-6 shows the organization of the stack. The stack pointer in the PSW register points to the top of the stack. A 0 in the stack pointer corresponds to RAM locations 8h and 9h. Bit 5 F0. Flag 0. A general purpose software flag. Bit 4 BS. The current active register bank. 0 e bank 0, 1 e bank 1. Reserved. User should not change its power-up value. Bits 2-0 Stack Pointer. 3-bit Stack Pointer for 8-level stack. When the CPU performs a subroutine call or interrupt call, the PC and the upper 4 bits of the PSW are pushed into the stack. Upon return, the PSW can be restored in option. If a RETR Return instruction is executed, PSW is restored. PSW is not restored if an RET Return instruction is executed. See Figure 10-6 . Bit 3 10.1.4 I/O Interface The keyboard controller of the PC87306 provides 16 general purpose I/O lines, two open-collector output lines and two input lines as shown in the Functional Block Diagram in Figure 10-1 . General Purpose I/O The 16 general purpose I/O lines, P12 – P17 and P20 – P21, are mapped to Port 1 and Port 2 respectively. These I/O lines are quasi-bidirectional because the output buffer cannot be turned off even if the I/O line is intended for input. When a 0 is written to an I/O line, it behaves like an output. When a 1 is written to an I/O line, it behaves like an input. On reset, all I/O lines are inputs. Figure 10-8 illustrates the structure of the I/O line. Q1 and Q2 are the normal output transistors. When the I/O line is intended for output, 0 is written into the flip-flop, which latches, and Q2 turns on. When 1 is written, Q1 turns on briefly because it is gated by a short port-write pulse also. Q1 charges up the pad to near VCC and then turns off. An active pull-up transistor Q3 turns on also, but it provides a large pull-up resistance because it is a weak transistor. At this stage, the I/O line can be used as input: a low impedance low voltage can override the pin and the input buffer reads a low level input. If the pin is driven high or undriven, a high level is read. Therefore, to use a port pin as input, a logic 1 must first be written to it. When the keyboard controller is reset, all port lines will be initialized to logic 1. Because Q1 turns on momentarily when a write to the port is performed there is potential for a current surge. A series resistor connected to those port lines used as inputs, is recommended to limit the potential surge (Figure 10-9) . TL/C/12379–52 FIGURE 10-6. Keyboard Controller Stack Organization Program Status Word Register (PSW) This 8-bit register holds the program execution status. The bit definition of the PSW register is shown in Figure 10-7 . CY bit Bit 7 Bit 6 7 AC F0 BS res 6 5 4 3 FIGURE 10-7. PSW Register Bits SP k2:0l 2-0 CY. Carry flag of the accumulator. AC. Auxiliary Carry flag of the accumulator, i.e., carry from bit 3 to 4 of the ALU. TL/C/12379 – 54 FIGURE 10-8. Active Pull-Up I/O Port Structure 84 10.0 Keyboard Controller and Real-Time Clock (Continued) R: current limiting resistor A small-value series current limiting resistor is recommended when PORT PINS are used as inputs. TL/C/12379 – 55 FIGURE 10-9. Using Port Pins as Inputs When the port line is used as an output, an active pull-up technique is used (instead of a traditional pull-up resistor), which turns on Q3 only when Q2 is off, hence eliminating a steady current flow. Timer Operation The counter can be set to the timer operation mode by connecting its clock source to the internal timing generator. The clock frequency to the timer is equal to the oscillator frequency divided by 480. The initial value of the timer is programmable. After the timer is started (by STRT T instruction), it counts up continuously until it is stopped or the keyboard controller is reset. The Timer Overflow Flag is set when the count value overflows from FFh back to 00h. The Timer Overflow Flag can be tested by a conditional jump instruction (JTF). This instruction also resets the flag. When the timer interrupt is enabled, an interrupt occurs when the timer overflows. This is discussed more detail in the interrupt section. Open-Collector Outputs In order to reduce the glue logic used in a PC-AT compatible environment, four dedicated open-collector outputs are provided: KBCLK, KBDAT, MCLK and MDAT. KBCLK is the complement of P26. KBDAT is the complement of P27. MCLK is the complement of P23. MDAT is the complement of P22. These four drivers can drive 16 mA, making them suitable for driving keyboard and mouse cables. TEST0 and TEST1 are internally connected to KBCLK and KBDAT, respectively, in the PC-AT compatible applications. P10 and P11 are connected internally to KBDAT and MDAT, respectively. Event Counter Operation When the clock input of the counter is switched to the external input (MCLK), it essentially becomes an event counter. The falling edge of the signal on the MCLK pin causes the counter to increment. Timer Overflow Flag and Timer interrupt operate in the same way as they do in the timer mode. Test Inputs KBCLK and MCLK are two dedicated input pins. Conditional jump instructions directly check the level of these two pins. MCLK also serves as the event counter input. 10.1.6 Interrupts The keyboard controller of the PC87306 provides 2 different internal interrupts. They are the Input Buffer Full (IBF) interrupt and Timer Overflow interrupt. These two interrupts can be independently enabled or disabled by software. Both of them are disabled when the chip comes out from reset. 10.1.5 Timer/Counter The keyboard controller is equipped with an 8-bit counter which can be used as a timer or an event counter. Figure 10-10 shows the two different clock sources for the counter. The clock source is selected by software. TL/C/12379 – 56 FIGURE 10-10. Timing Generation and Timer Circuit 85 10.0 Keyboard Controller and Real-Time Clock (Continued) Timer Interrupt 10.2 REAL-TIME CLOCK FUNCTION If the timer interrupt is enabled, upon the timer overflow, an interrupt occurs. It causes the program to perform a subroutine call to program address 007h. The interrupt is cleared by this subroutine call. The current Program Counter (PC) and the upper 4 bits of the PSW is pushed into the stack before the call occurs. At the end of the timer interrupt service routine, an RETR instruction restores the PSW and the PC. Because the timer interrupt has a lower priority than the IBF interrupt, simultaneous IBF and timer interrupts cause the timer interrupt to be pending. It is served as soon as the program returns from the IBF interrupt service routine. The RTC in the PC87306 is a low-power clock that provides a time-of-day clock and 100-year calendar with alarm features and battery operation. Other features include three maskable interrupt sources and 242 bytes of general purpose RAM. Valid RAM and time can be maintained through the use of an external battery source. It is software compatible with the DS1287 and MC146818. 10.2.1 Memory Map The RTC contains 14 time and control registers. It also contains 242 bytes of general purpose battery backed static RAM in two banks. RAMSEL is used to select the active bank. Refer to Table 10-2. Input Buffer Full (IBF) Interrupt If the lBF interrupt is enabled, when there is a host write operation to the keyboard controller (IOW e 0), an interrupt occurs. The processor saves the current Program Counter and the upper 4 bits of the PSW into the stack and performs a subroutine call to the program address 003h. Upon entering the interrupt service routine, further interrupts are held off. The interrupt is re-enabled upon the execution of the RETR instruction. 10.2.2 Bus Interface The RTC function is mapped to I/O locations 70h (index) and 71h (data). This decode is done internal to the PC87306. 10.2.3 Time Generation The Time Generation function divides the 32.768 kHz from the external clock pins (X1C, X2C) down to a one Hertz signal. The divider chain is controlled by bits 6 – 4 of Control Register A. Bits 3 – 0 of Control Register A select one of fifteen taps from the divider chain to be used as a Periodic Interrupt. See Control Register A in the Control Register and Interrupt section for divider configurations and rate selections. During divider reset (bits 6 – 4 of Control Register A e 11x), the divider chain is reset to 0. An update will occur 500 ms after the divider chain is activated into normal operational mode (bits 6 – 4 of Control Register A e 010). The periodic flag also becomes active 1/2 of the programmed value when the divider chain is activated. 10.1.7 Oscillator and Instruction Timing The oscillator pin of the keyboard controller is X1, when bit 7 of KRR is 0. When bit 7 of KRR is 1, the Keyboard Controller clock is SYSCLK. Figures 10-11 shows the connections for external clock configuration via the X1 pin. The oscillator clock is divided by 3 to generate the state timing, then is further divided down by 5 to generate the instruction timing (Figure 10-12) . Hence each instruction cycle consists of 5 states and 15 clock cycles. Most of the keyboard controller instructions require only one instruction cycle. The others require two cycles. Refer to the instruction set for details. Figure 10-13 represents the internal and external circuitry that comprise the oscillator. The oscillator input may be driven from an external source. If this is desired, the input should be driven rail to rail and be approximately a 50% duty cycle. The oscillator output should be open in this case. The external capacitor values should be chosen to provide the manufacturer’s specified load capacitance for the crystal when combined with the parasitic capacitance of the trace, socket, and package, which can vary from 0 pF to 8 pF. The rule of thumb in choosing these capacitors is: CL e (C1 * C2)/(C1 a C2) a Cparasitic C2 l C1 C1 can be trimmed to obtain the 32768.0 Hz. TL/C/12379–57 FIGURE 10-11. External Clock Connection TL/C/12379 – 58 FIGURE 10-12. Instruction Cycle Timing 86 10.0 Keyboard Controller and Real-Time Clock (Continued) TABLE 10-2. RTC Memory Map Index (RAMSEL e 0) 00 01 02 03 04 05 06 07 08 09 0A 0B 0C 0D 0E – 7F (RAMSEL e 1) 00 – 7F Function BCD Format Seconds Seconds Alarm Minutes Minutes Alarm Hours Hours Alarm Day of Week Date of Month Month Year Control Register A Control Register B Control Register C Control Register D Battery-Backed RAM (114 Bytes) 00 – 59 00 – 59 00 – 59 00 – 59 12 hr. e 12 hr. e 24 hr. e 12 hr. e 12 hr. e 24 hr. e 01 – 07 01 – 31 01–12 00 – 99 01 – 12 (AM) 81 – 92 (PM) 00 –23 01 – 12 (AM) 81 – 92 (PM) 00 –23 Battery-Backed RAM (128 Bytes) Comments 00 – 3b 00 – 3b 00 – 3b 00 – 3b 01 – 0c (AM) 81 – 8c (PM) 00 –17 01 – 0c (AM) 81 – 8c (PM) 00 –17 01 – 07 (Sunday e 1) 01 – 1f 01 – 0c 00 – 63 R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W (Bit 7 is Read Only) R/W (Bit 3 is Read Only) R R R/W R/W Daylight savings and leap year exceptions are handled by the time keeping function. When bit 0 of Control Register A (DSE) is a ‘‘1’’, time advances from 1:59:59 AM to 3:00:00 on the first Sunday in April. On the last Sunday of October time changes from 1:59:59 to 1:00:00 when daylight savings is enabled. On leap year February is extended to 29 days. The time is updated once per second. If a read of the timing registers coincides with an update, data read may not be valid. Also, writes to time registers during an update have undefined results. To avoid accessing invalid data, the RTC in the PC87306 provides a user copy of the time registers. The sequence used to update the time registers is shown in Figure 10-14. Four methods are available for reading and writing time to ensure that the correct time is written or read. They are as follows: Method 1ÐSet the SET bit. (CRB7) Set the SET bit to 1. This takes a snapshot of the internal time registers and loads it into the user copy. Then when the SET bit is written with 0, the user copy will update the internal registers. Method 2ÐAccess after detection of an Update Ended Interrupt. This interrupt indicates that an update has just completed and the next update will occur after 999 ms. Method 3ÐPoll update-in-progress (bit 7 in Control Register A). Rext e 5 kX 15 pF C1 e 12 C2 e 15 pF 8 pF Cparasitic e 0 x x TL/C/12379–96 FIGURE 10-13. Oscillator Internal and External Circuitry The start-up time of this oscillator may vary from two to seven seconds and is due to the high ‘‘Q’’ of the crystal. The parameters below describe the crystal to be used: Parallel Resonant, tuning fork (N cut) or XY Bar Q l 35k Load Capacitance (CL) Accuracy Temperature Coefficient Binary Format 9 pF to 13 pF User Choice User Choice 10.2.4 Time Keeping The time is kept in BCD or binary format. The format is determined by bit 2 of Control Register B (DM). Either 12 or 24 hour representation for the hours can be maintained as determined by bit 1 of Control Register B (24/12). Note that when changing the format the time registers must be re-initialized to the corresponding data format. 87 10.0 Keyboard Controller and Real-Time Clock (Continued) TL/C/12379 – 59 where: UIP e Update In Progress Bit UF e Update Ended Flag (Update Ended Interrupt if Enabled) PF e Periodic Flag (Periodic Interrupt if Enabled) AF e Alarm Flag (Alarm Interrupt if Enabled) Flags (and IRQ) are reset at the conclusion of Control Register C read or by RESETC low. A B C P Update in Progress Bit High before Update Occurs e 244 ms Periodic interrupt to Update e P/2 a 244 ms e Update to Alarm Interrupt e 30.5 ms e Period is programmed by RS3–0 of control Register A. e e FIGURE 10-14. Interrupt/Status Timing The update occurs 244 ms after the update-in-progress bit goes high. Therefore if a 0 is read, there is a minimum of 244 ms in which the time is guaranteed to remain stable. Method 4ÐUse a Periodic interrupt to determine if an update cycle is in progress. The periodic interrupt is first set to a desired period. The program can then use the periodic interrupt to signify that there is (Periodic Interrupt/2 a 244 ms) remaining until another update occurs. The Alarm condition is also generated by the Time Keeping function. After each update, the seconds, minutes, and hours are compared with the seconds alarm, minutes alarm, and hours alarm. If equal, the Alarm lag is set in Control Register C. This causes an interrupt condition (IRQZ e 0) if the Alarm Interrupt Enable bit is set in Control Register B. If both bit 7 and bit 6 of any alarm byte (seconds alarm, minutes alarm, hours alarm) are 1, then that alarm byte is a ‘‘don’t care’’. 10.2.6 Power Management The Power Management function provides power to the RTC in the PC87306. During system operation, power from the system is used. When system voltage falls below battery voltage, the Power Management function switches the RTC cell to battery power. For proper operation, a 500 mV differential is needed between VCC and VBAT. Figure 10-15 represents a typical battery configuration and Figure 10-16 represents typical battery current during battery backed mode. TL/C/12379 – 60 FIGURE 10-15. Typical Battery Configuration 10.2.5 RAM The RAM data is accessed at locations 0E–7F when RAMSEL (bit 2) of the configuration registers is 0, and locations 00 – 7F when RAMSEL is 1. Battery backed power enables the RAM to retain information during system powerdown. Bit 0 of SCF0 locks RTC RAM cells at addresses 38-3Fh. The bit is 0 upon reset, thus enabling read and write access. When the bit is 1, writes to RTC RAM cells at addresses 38 – 3Fh are ignored, and reads return FFh. Once set to 1 the bit can only be cleared by a hardware reset. TL/C/12379 – 61 FIGURE 10-16. Typical Battery Current during Battery Backed Mode 88 10.0 Keyboard Controller and Real-Time Clock (Continued) Bits 0 – 3 RS3 – 0ÐPeriodic Interrupt Rate Select (Read/Write) 10.2.7 System Bus Lock Out and Power-Up Detection As the RTC switches to battery power all inputs are locked out so that the internal registers can not be modified. This lockout condition continues for 62 ms (min) to 125 ms (max) after the RTC switches from battery power to system power. The 62 ms–125 ms lockout during power transition from battery to system power depends on the following conditions: 1. If the Divider Chain Control (bits 6–4 in Control Register A) is in any mode but Normal Operation (010), all inputs are enabled immediately upon detection of system voltage above that of battery voltage. 2. When Battery voltage is below 1V and MR is 1, all inputs are enabled immediately upon detection of system voltage above that of battery voltage. This also initializes registers 00 through 0D to 00. 3. If the VRT bit (bit 7 in Control Register D) is 0, all inputs are enabled immediately upon detection of system voltage above that of battery voltage. These bits control the rate of the periodic interrupt. Reset has no effect on these bits. RS3 –0 0000 10.2.8 Oscillator When power is applied to the RTC in the PC87306, the Oscillator is operational with the following exceptions: 1. The VRT bit (bit 7 in Control Register D) is a 0. The oscillator is disabled after initial power-up. This reduces power consumption during the time between when the battery is initially installed and when the RTC is initialized for timekeeping. 2. The Divider Chain Control (bits 6–4 in Control Register A) is in Oscillator Disabled modes (000 and 001). This provides a means for the user to ‘‘shut-down’’ the oscillator and reduce the power consumption of the RTC cell. The RAM remains functional when the oscillator is disabled. Bit 7 Control Register A DV2 DV1 DV0 RS3 RS2 RS1 RS0 7 6 5 4 3 2 1 0 3.90625 ms 0010 7.8125 ms 0011 122.070 ms 0100 244.141 ms 0101 488.281 ms 0110 976.562 ms 0111 1.953125 ms 1000 3.90625 ms 1001 7.8125 ms 1010 15.625 ms 1011 31.25 ms 1100 62.5 ms 1101 125 ms 1110 250 ms 1111 500 ms DV2 –0 10.2.10 Control Registers The four Control Registers used to configure timing interrupts can be accessed at any time during non-battery backed operation. These Control Registers are located at addresses 0Ah, 0Bh, 0Ch and 0Dh. Their descriptions follow. UIP None 0001 Bits 4 – 6 DV2 – 0ÐDivider Chain Control (Read/Write) These bits control the configuration of the divider chain in the Timing Generation function. Reset has no effect on these bits. 10.2.9 Interrupt Handling The Periodic, Alarm, and Update ended Interrupts are generated (IRQZ is driven low) when the respective enable bits in Control Register B are set and an interrupt condition occurs. A read from Register C clears the active interrupt. If a second interrupt condition occurs (other than that which caused the interrupt) during a read from Register C, IRQZ remains active (low). Thus, it is recommended that when multiple interrupts are enabled, the interrupt service routine continues to read (and service interrupts) until bit 7 of Control Register C (IRQ Flag) returns to a 1. Note that if an interrupt is not serviced before a second condition of the same interrupt occurs, the second interrupt event is lost. bit Periodic Interrupt Rate 89 Divider Chain Configuration 000 Oscillator Disabled 001 Oscillator Disabled 010 Normal Operation 011 TEST 100 TEST 101 TEST 110 Divider Chain RESET 111 Divider Chain RESET UIPÐUpdate in Progress (Read Only) 1: Signifies that the timing registers will be updated within 244 ms. 0: Signifies that an update will not occur before 244 ms. This bit reads 0 when bit 7 of Control Register B (SET) is a 1. Reset has no effect on this bit. 10.0 Keyboard Controller and Real-Time Clock (Continued) 0: Disables generation of the Periodic interrupt. Control Register B bit Bit 0 Bit 1 Bit 2 Bit 3 Bit 4 SET PIE AIE UIE 0 DM 24/12 DSE 7 6 5 4 3 2 1 0 Reset forces this bit to 0. Note: It is recommended that the software clears bits 4–6 of Control Register B, and reads Control Register C, after reset. DSEÐDaylight Savings Enable (Read/Write) 1: Enables daylight savings. Two conditions apply as follows: Daylight Savings Spring: Time advances from 1:59:59 to 3:00:00 on the first Sunday in April. Daylight Savings Fall: Time advances from 1:59:59 to 1:00:00 on the last Sunday in October. 0: Disables the daylight savings feature. Reset has no effect on this bit. 24/12Ð24 or 12 Hour Mode (Read/Write) 1: Enables 24 hour format. 0: Enables 12 hour format. Reset has no effect on this bit. DMÐData Mode (Read/Write) 1: Enables BINARY format. 0: Enables BCD format. Reset has no effect on this bit. This bit is defined as ‘‘Square Wave Enable’’ by the MC146818 and is not supported by the RTC Cell. This bit is always read as 0. UIEÐUpdate Ended Interrupt Enable (Read/Write) 1: Enables generation of the Update Ended interrupt. This interrupt is generated at the time an update occurs. 0: Disables generation of the Update Ended interrupt. Reset forces this bit to 0. Bit 7 SETÐSet mode (Read/Write) 1: The user copy of time is ‘‘frozen’’ allowing the time registers to be accessed without regard for an occurrence of an update. 0: The timing updates occur normally. Reset has no effect on this bit. Control Register C IRQF Note: It is recommended that the software clears bits 4–6 of Control Register B, and reads Control Register C, after reset. Bit 5 AIEÐAlarm Interrupt Enable (Read/Write) 1: Enables generation of the Alarm interrupt. The Alarm interrupt is generated immediately after a time update in which the Seconds, Minutes, and Hours time matches with their respective alarm counterparts. 0: Disables generation of the alarm interrupt. Reset forces this bit to 0. AF UF 0 0 0 0 VRT 0 0 0 0 0 0 0 bit 7 6 5 4 3 2 1 0 Bits 0 – 6 Reserved (Always 0) Bit 7 VRTÐValid RAM and Time (Read Only) 1: This bit indicates that the contents of the RTC are valid. 0: This bit indicates that the battery source is low and that the RTC data and RAM data are questionable. This bit is set to 1 at the conclusion of a read from this register. Note: It is recommended that the software clears bits 4–6 of Control Register B, and reads Control Register C, after reset. Bit 6 PF bit 7 6 5 4 3 2 1 0 Bits 0 – 3 Reserved (Always 0) Bit 4 UFÐUpdate Ended Interrupt Flag (Read Only) 1: When the time registers are updated. This bit is reset to 0 at the conclusion of a read from this register. Reset forces this bit to 0. Bit 5 AFÐAlarm Interrupt Flag (Read Only) 1: When an alarm condition is detected. This bit is reset to 0 at the conclusion of a read from this register. Reset forces this bit to 0. Bit 6 PFÐPeriodic Interrupt Flag (Read Only) 1: When a transition occurs on the selected tap of the divider chain. This bit is reset to 0 at the conclusion of a read from this register. Reset forces this bit to 0. Bit 7 IRQFÐInterrupt Request Flag (Read Only) 1: When bit pairs PIE and PF; AIE and AF, and UIE and UF all equal 1. The IRQF bit mirrors the value on the IRQZ output. When lRQZ is active (low), IRQF is 1. Reset forces this bit to 0. Control Register D PIEÐPeriodic Interrupt Enable (Read/Write) 1: Enables generation of the Periodic interrupt. Bits 3–0 of Control Register A determine the rate of the Periodic interrupt. 90 11.0 General Purpose Input and Output (GPIO) Ports The GPIO ports have open-drain outputs with internal pullups and TTL inputs. The GPIO behave as follows: 1. Writes are latched internally. A latched 0 drives the pin low; a latched 1 pulls the pin high via the internal pullup. 2. Reads return the pins’ values. A driven-low pin is always read as 0. A pulled-up pin is read as either 0 or 1, according to the pin voltage as dictated by the pullup or external active driver (TTL input buffers are used). Upon reset the write latches are initialized to FFh. A GPIO port must not be enabled at the same address as another accessible PC87306 register. Undefined results will occur if a GPIO is configured in this way. The PC87306 supports two identical General Purpose I/O (GPIO) ports. See Figure 11-1. The PC87306 can wake up with both GPIO Ports either enabled or disabled, according to the CFG1 strap pin. GPIO Port 1 is at offset 0 and GPIO Port 2 is at offset 1 from the 16-bit GPIO Port base address. The GPIO port’s base address is software configurable. GPBA is an 8-bit configuration register that holds address bit A2-9 (bit 0 holds A2). The other eight bits of the 16-bit GPIO port base address are 0. Upon reset GPBA is initialized to 00011110, thus dictating base address e 78h). TL/C/12379 – 65 FIGURE 11-1. General Purpose I/O (GPIO) Ports 91 12.0 Electrical Characteristics ABSOLUTE MAXIMUM RATINGS (Notes 1 and 2) RECOMMENDED OPERATING CONDITIONS If Military/Aerospace specified devices are required, please contact the National Semiconductor Sales Office/Distributors for availability and specifications. b 0.5V to a 7.0V Supply Voltage (VDD) Supply Voltage (VDD) Input Voltage (VI) Output Voltage (VO) Storage Temperature (TSTG) Power Dissipation (PD) Lead Temperature (TL) (Soldering, 10 seconds) 4.5V to 5.5V 0§ C to a 70§ C 1500V Operating Temp. Range (TA) ESD Tolerance CZAP e 100 pF (Note 3) RZAP e 1.5 kX (Note 3) b 0.5V to VDD a 0.5V b 0.5V to VDD a 0.5V Note 3: Value based on test complying with NSC SOP5-028 human body model ESD testing using the ETS-910 tester. b 65§ C to a 165§ C 1W a 260§ C Note 1: Absolute Maximum Ratings are those values beyond which damage to the device may occur. Note 2: Unless otherwise specified, all voltages are referenced to ground. CAPACITANCE TA e 25§ C, f e 1 MHz Symbol Parameter Min Typ Max Units CIN Input Pin Capacitance 5 7 pF CIN1 Clock Input Capacitance 8 10 pF CIO I/O Pin Capacitance 10 12 pF CO Output Pin Capacitance 6 8 pF 12.1 DC ELECTRICAL CHARACTERISTICS Under Recommended Operating Conditions Symbol Max Units Input High Voltage 2.0 VDD V VIL Input Low Voltage b 0.5 0.8 mA ICC VDD Average Supply Current VIH ICCSB IIL Parameter VDD Quiescent Supply Current in Low Power Mode Input Leakage Current (Note 4) Conditions Min Typ VIL e 0.5V VIH e 2.4V No Load 32 50 mA VIL e VSS VIH e VDD No Load 1.3 1.7 mA VIN e VDD 10 mA VIN e VSS b 10 mA Note 4: During reset the MFM pin is rated for 10 mA; b 150 mA leakage is due to an internal pull-up resistor. The RTS1, 2; SOUT1, 2; DTR1, 2; HCS0, 1; IDEHI; VLD0, 1 and IDEHLO are rated for 100 mA; b 10 mA leakage is due to internal pull-down resistors. During normal operation the BUSY, PE, and SLCT pins are rated for 100 mA; b 10 mA leakage is due to internal pull down resistors. The ACK and ERR pins are rated for 10 mA; b 100 mA leakage is due to internal pull-up resistors. 92 12.0 Electrical Characteristics (Continued) 12.1 DC ELECTRICAL CHARACTERISTICS (Continued) 12.1.1 Microprocessor, Parallel Port, and IDE Interface Pins Symbol VOH VOL IOZ Parameter Output High Voltage Output Low Voltage Conditions Min IOH e b15 mA on: D0 – 7, IDED7, IRQ3 – 7, DRQ IOH e b6 mA on: DTR, RTS, SOUT, MFM, DRATE, CSOUT, IDEHI, IDELO, HCS0, HCS1, CS0, CS1 2.4 Max V IOL e 24 mA on: D0 – 7, IDED7, IRQ3 – 7, IRQ8, DRQ, ZWS, IOCHRDY IOL e 12 mA on: DTR, RTS, SOUT, HCS, IDEHI, IDELO IOL e 6 mA on: DRATE, CS0, CS1 Output TRI-STATE Leakage Current (D7–0, IRQ3–7, DRQ) Units 0.4 V VIN e VDD 50 mA VIN e VSS b 50 mA Note 1: Pins IOCHRDY and ZWS are open drain pins. 12.1.2 Disk Interface Pins Symbol Parameter Conditions Min Typ Max 250 Units VH Input Hysteresis VOH Output High Voltage (Note 1) IOH e b4 mA mV VOL Output Low Voltage IOL e 40 mA 0.4 V ILKG Output High Leakage Current (Note 1) VIN e VDD 10 mA VIN e VSS b 10 mA 2.4 V Note 1: VOH and ILKG for the disk interface pins are valid for CMOS buffered outputs only. 12.1.3 Oscillator Pin (X1) Symbol Parameter VIH X1 Input High Voltage IXLKG X1 Leakage Conditions Min Typ Max 2.0 Units V VIN e VDD 400 mA VIN e VSS b 400 mA 12.1.4 Parallel Port Pins Parameter Conditions IOH Symbol High Level Output Current (Note 1) VOH e 2.4V Min Typ 12 Max Units mA IOL Low Level Output Current VOL e 0.4V 12 mA Note 1: AFD, INIT, SLIN and STB are open drain pins when the PC87306 parallel port is operating in one of the following modes: Compatible, Extended, EPP1.7, ECP mode 0 or 1, or ECP mode 2 if PCR1 e 0. 4.7 kX resistors should be used. 12.1.5 GPIO Pins Symbol VOL Parameter Conditions Output Low Voltage (Note 1) IOL e 2.0 mA Note 1: GPIO pins are open drain outputs, with internal pull-up resistors between 10 and 26 kX. 93 Min Typ Max Units 0.4 V 12.0 Electrical Characteristics (Continued) 12.1.6 Keyboard Controller and Real-Time Clock Pins 0§ C –70§ C Symbol VOH VOL IIN Parameter Conditions Minimum High Level Output Voltage # P10–P17, P20–P27 (Note 1) # All other push-pull outputs, KBCLK, KBDAT, MCLK, MDAT (Note 2) VIN e VIH or VIL # IOH e b2.0 mA # IOH e b4.0 mA Maximum Low Level Output Voltage # P10–P17, P20–P27 # KBDAT, KBCLK, MCLK, MDAT # All Other Outputs VIN e VIH or VIL # IOL e 2.0 mA # IOL e 16.0 mA # IOL e 4.0 mA Maximum Input Current # Inputs only: A0–A9, AEN, CS, EA, IOR, IOW, PWRGOOD, T0, T1 # Inputs with Resistors: SS, TRI STROBE IBAT VBAT Quiescent Supply Current 2.4 2.4 Units V V V V V # VIN e VCC or VSS 10 mA # VIN e VCC 10 400 10 10 400 mA mA mA mA mA 8 mA 0.4 mA VCC b 0.5 V e e e e VIL VCC or VSS VCC VIL VBAT e 3.0V VIN e VCC e VSS e 0V Real-Time Clock f e 32.768 kHz Input Leakage VBAT Max 0.4 0.4 0.4 VIN # VIN # VIN VIN # I/Os Only: D0–D7 # I/Os with Resistors: P10–P17, P20–P27, SYNC, Min VCC e 5.0V, VBAT e 3.0V Battery Voltage 2.4 Note 1: IOH is driven for 10 ns on switch from low to high. Note 2: KBCLK, KBDAT, MCLK and MDAT are open drain outputs with 10k minimum pull-up resistors. 12.2 AC ELECTRICAL CHARACTERISTICS 12.2.1 AC Test Conditions TA e 0§ C to a 70§ C, VDD e 5.0V g 10% LOAD CIRCUIT (Notes 1, 2, 3) AC TESTING INPUT, OUTPUT WAVEFORM TL/C/12379 – 67 TL/C/12379–66 Note 1: CL e 100 pF, includes jig and scope capacitance. Note 2: S1 e Open for push-pull outputs. S1 e VDD for high impedance to active low and active low to high impedance measurements. S1 e GND for high impedance to active high and active high to high impedance measurements. RL e 1.0 kX for mP interface pins. Note 3: For the FDC Open Drive Interface Pins S1 e VDD and RL e 150X. 94 12.0 Electrical Characteristics (Continued) 12.2.2 Clock Timing Parameter Min tCH Symbol Clock High Pulse Width 16 Max Units tCL Clock Low Pulse Width 16 tCP Clock Period 40 tICP Internal Clock Period (Table 12-1) ns tDRP Data Rate Period (Table 12-1) ns ns ns 43 ns TABLE 12-1 Nominal tICP, tDRP Values MFM Data Rate tDRP tICP Value Units 1 Mbps 1000 3 x tCP 125 ns 500 kbps 2000 3 x tCP 125 ns 300 kbps 3333 5 x tCP 208 ns 250 kbps 4000 6 x tCP 250 ns TL/C/12379 – 68 FIGURE 12-1. Clock Timing 12.2.3 Microprocessor Interface Timing Symbol Parameter Min Max Units tAR Valid Address to Read Active 18 tAW Valid Address to Write Active 18 ns tDH Data Hold 0 ns tDS Data Setup 18 ns tHZ Read to Floating Data Bus 13 tPS Port Setup 10 ns tRA Address Hold from Inactive Read 0 ns tRCU Read Cycle Update 45 ns tRD Read Strobe Width 60 ns tRDH Read Data Hold 10 tRI Read Strobe to Clear IRQ6 55 ns tRVD Active Read to Valid Data 55 ns tWA Address Hold from Inactive Write 0 ns tWCU Write Cycle Update 45 ns tWI Write Strobe to Clear IRQ6 55 ns tWO Write Data to Port Update 60 ns tWR Write Strobe Width 60 ns RC Read Cycle e tAR a tRD a tRCU 123 ns WC Write Cycle e tAW a tWR a tWCU 123 ns tWRR RD Low after WR High 80 ns 95 ns 25 ns ns 12.0 Electrical Characteristics (Continued) TL/C/12379 – 69 FIGURE 12-2. Microprocessor Read Timing TL/C/12379 – 70 FIGURE 12-3. Microprocessor Write Timing TL/C/12379 – 72 FIGURE 12-4. Read after Write Operation to All Registers and RAM Timing 96 12.0 Electrical Characteristics (Continued) 12.2.4 Baud Out Timing Symbol Parameter Conditions N Baud Divisor tBHD Baud Output Positive Edge Delay tBLD Baud Output Negative Edge Delay Min Max Units 1 65535 ns CLK e 24 MHz/2, 100 pF Load 56 ns CLK e 24 MHz/2, 100 pF Load 56 ns TL/C/12379 – 73 FIGURE 12-5. Baud Out Timing 12.2.5 Transmitter Timing Symbol Parameter Min Max Units 1.6 ms */16 BAUD OUT Cycles 40 ns tIRTXW IRTX Pulse Width tHR Delay from WR (WR THR) to Reset IRQ tIR Delay from RD (RD IIR) to Reset IRQ (THRE) 55 ns tIRS Delay from Initial IRQ Reset to Transmit Start 8 24 Baud Out Cycles tSI Delay from Initial Write to IRQ 16 24 Baud Out Cycles tSTI Delay from Start Bit to IRQ (THRE) 8 Baud Out Cycles TL/C/12379 – 74 Note 1: See Microprocessor Write cycle timing, Figure 12-3 . Note 2: See Microprocessor Read cycle timing, Figure 12-2 . FIGURE 12-6. Transmitter Timing 97 12.0 Electrical Characteristics (Continued) 12.2.6 Receiver Timing Symbol Parameter Min Max Units 1.6 ms 6/16 BAUD OUT Cycles Delay from Active Edge of RD to Reset IRQ 78 ns Delay from Inactive Edge of RD (RD LSR) to Reset IRQ 55 ns tIRRXW IRRX Pulse Width tRAI tRINT tSCD Delay from RCLK to Sample Time tSINT Delay from Stop Bit to Set Interrupt Conditions (Note 1) 41 ns 2 Baud Out Cycles Note 1: This is an internal timing specification and is therefore not tested. TL/C/12379 – 75 FIGURE 12-7. Receiver Timing TL/C/12379 – 76 Note 1: If SCR0 e 1, then tSINT e 3 RCLKs. For a Timeout Interrupt, tSINT e 8 RCLKs. FIGURE 12-8. FIFO Mode Receiver Timing 98 12.0 Electrical Characteristics (Continued) TL/C/12379 – 77 Note 1: If SCR0 e 1, then tSINT e 3 RCLKs. For a Timeout Interrupt, tSINT e 8 RCLKs. FIGURE 12-9. Timeout Receiver Timing 12.2.7 MODEM Control Timing Max Units tMDO Symbol Delay from WR (WR MCR) to Output Parameter Conditions Min 40 ns tRIM Delay to Reset IRQ from RD (RD MSR) 78 ns tSIM Delay to Set IRQ from MODEM Input 40 ns TL/C/12379 – 78 Note 1: See Microprocessor Write cycle timing, Figure 12-3 . Note 2: See Microprocessor Read cycle timing, Figure 12-2 . FIGURE 12-10. MODEM Control Timing 99 12.0 Electrical Characteristics (Continued) 12.2.8 DMA Timing 12.2.8.1 FDC Symbol Parameter Min tKI FDACK Inactive Pulse Width 25 tKK FDACK Active Pulse Width 65 tKQ FDACK Active Edge to FDRQ Inactive tQK FDRQ to FDACK Active Edge tQP FDRQ Period (except Non-Burst DMA) (Note 3) tQQ FDRQ Inactive Non-Burst Pulse Width 300 tQR FDRQ to RD or WR Active 15 tQW FDRQ to End of RD, WR (Notes 2, 3) (FDRQ Service Time) tQT FDRQ to TC Active (Notes 2, 3) (FDRQ Service Time) tRQ RD, WR Active Edge to FDRQ Inactive (Note 1) tTQ TC Active Edge to FDRQ Inactive tTT TC Active Pulse Width Max ns ns 65 10 ns ns 8 x tDRP 50 Units ms 400 ns ns (8 x tDRP) b(16 x tICP) ms (8 x tDRP) b(16 x tICP) ms 65 ns 75 ns ns Note 1: The active edge of RD or WR and TC is recognized only when FDACK is active. Note 2: Values shown are with the FIFO disabled, or with FIFO enabled and THRESH e 0. For nonzero values of THRESH, add (THRESH x 8 x tDRP) to the values shown. Note 3: tDRP and tICP are defined in Table 12-1. TL/C/12379 – 79 FIGURE 12-11. FDC DMA Timing 100 12.0 Electrical Characteristics (Continued) 12.2.8.2 ECP Symbol Parameter Min Max Units tKIP PDACK Inactive Pulse Width 25 ns tKKP PDACK Active Pulse Width 65 tKQP PDACK Active Edge to PDRQ Inactive (Notes 2, 3) tQKP PDRQ to PDACK Active Edge 10 ns tQPP PDRQ Period 330 ms tQQP PDRQ Inactive Non-Burst Pulse Width 300 tQRP PDRQ to RD, WR Active 15 tRQP RD or WR Active Edge to PDRQ Inactive (Note 1) 65 ns tTQP TC Active Edge to PDRQ Inactive 75 ns tTT TC Active Pulse Width ns 65 a (6 x 32 x tCP) 50 400 ns ns ns ns Note 1: The active edge of RD or WR and TC is recognized only when PDACK is active. Note 2: One DMA transaction takes six clock cycles. Note 3: tCP is defined in Section 12.2.2 Clock Timing. TL/C/12379 – 80 FIGURE 12-12. ECP DMA Timing 101 12.0 Electrical Characteristics (Continued) 12.2.9 Reset Timing Symbol Parameter tRW Reset Width (Note 1) tSRC Reset to Control Inactive Min Max Units 100 ms 300 ns Note 1: The software reset pulse width is 100 ns. TL/C/12379 – 81 Note 2: DRQ and IRQ6 will be TRI-STATE after time tSRC when in the PC-AT. FIGURE 12-13. Reset Timing 12.2.10 FDC Write Data Timing Symbol Parameter Min Max Units tHDH HDSEL Hold from WGATE Inactive 750 ms tHDS HDSEL Setup to WGATE Active 100 ms tWDW Write Data Pulse Width Table 12-2 ns TABLE 12-2. Minimum tWDW Values Data Rate tDRP tWDW tWDW Value Units 1 Mbps 1000 2 x tICP 250 ns 500 kbps 2000 2 x tICP 250 ns 300 kbps 3333 2 x tICP 375 ns 250 kbps 4000 2 x tICP 500 ns TL/C/12379 – 82 FIGURE 12-14. Write Data Timing 102 12.0 Electrical Characteristics (Continued) 12.2.11 FDC Read Data Timing Symbol tRDW Parameter Min Read Data Pulse Width 50 Max Units ns TL/C/12379 – 83 FIGURE 12-15. Read Data Timing 12.2.12 Drive Control Timing Symbol Parameter Min tDRV DR0– 3, MTR0–3 from End of WR tDST DIR Setup to STEP Active tIW Max 100 Units ns 6 ms Index Pulse Width 100 ns tSTD DIR Hold from STEP Inactive tSTR ms tSTP STEP Active High Pulse Width 8 ms tSTR STEP Rate Time (see Table 4-15) 1 ms TL/C/12379 – 84 FIGURE 12-16. Drive Control Timing 12.2.13 IDE Timing Symbol Parameter Min Max Units tAD Delay from Address to Disable Strobe 25 ns tAE Delay from Address to Enable Strobe 25 ns TL/C/12379 – 85 FIGURE 12-17. IDE Timing 103 12.0 Electrical Characteristics (Continued) 12.2.14 Parallel Port Timing Parameter Conditions Type tPDH Symbol Port Data Hold (Note 1) 500 tPDS Port Data Setup (Note 1) 500 tPI Port Interrupt tSW STB Width Max ns ns 33 (Note 1) Units 500 ns ns Note 1: These times are system dependent and are therefore not tested. TL/C/12379 – 86 FIGURE 12-18. Compatible Mode Parallel Port Interrupt Timing TL/C/12379 – 87 FIGURE 12-19. Extended Mode Parallel Port Interrupt Timing TL/C/12379 – 88 FIGURE 12-20. Typical Parallel Port Data Exchange 104 12.0 Electrical Characteristics (Continued) 12.2.15 Enhanced Parallel Port Timing Symbol Parameter tWW WRITE Active from WR Active (Note 1) tWST DSTRB or ASTRB Active from WR Active (Notes 1, 2) tWEST Conditions DSTRB or ASTRB Active after WRITE Active Min Max Units 45 ns EEP 1.7 45 ns EPP 1.9 65 ns EPP 1.7 0 ns EPP 1.9 10 ns 50 ns tWPDh PD0–PD7 Hold after DSTRB or ASTRB Inactive tHRW IOCHRDY Active after WAIT Active (Note 3) tWPDS PD0–PD7 Valid after WRITE Active tEPDW EPP Data Width 80 ns tEPDh EPP Data Width Hold after DSTRB or ASTRB Inactive 0 ns tZWSa ZWS Valid after WR or RD Active tZWSh ZWS Hold after WR or RD Inactive EPP 1.7 40 ns D0 – 7 is stable 15 ns before WR Active 15 ns 45 0 ns ns Note 1: tWST and tWW are valid in EPP 1.9 only if WAIT is low when WR becomes active, else tWST and tWW are measured from WAIT. Note 2: The PC87306 design guarantees that WRITE will not change from low to high before DSTRB or ASTRB goes from low to high. Note 3: In EPP 1.9, IOCHRDY inactive is measured from WR or RD. TL/C/12379 – 89 FIGURE 12-21. Enhanced Parallel Port Timing 105 12.0 Electrical Characteristics (Continued) 12.2.16 Extended Capabilities Port Timing 12.2.16.1 Forward Symbol Parameter Conditions Min Max Units tecdsf Data Setup before STB Active 0 tecdhf Data Hold after BUSY 0 ns ns teclhf BUSY Setup after STB Active 75 ns techhf STB Active after BUSY 0 1 s techlf BUSY Setup after STB Inactive 0 35 ms tecllf STB Active after BUSY 0 ns TL/C/12379 – 90 FIGURE 12-22. ECP Parallel Port Forward Timing Diagram 12.2.16.2 Backward Symbol Parameter Conditions Min Max Units tecdsb Data Setup before STB Active 0 tecdhb Data Hold after BUSY 0 ns ns teclhb BUSY Setup after STB Active 75 ns techhb STB Active after BUSY 0 1 s techlb BUSY Setup after STB Inactive 0 35 ms tecllb STB Active after BUSY 0 ns TL/C/12379 – 91 FIGURE 12-23. ECP Parallel Port Backward Timing Diagram 106 12.0 Electrical Characteristics (Continued) 12.2.17 GPIO Write Timing Symbol tWGO Parameter Conditions Min Write data to GPIO update Max Units 300 (Note 1) ns Note 1: GPIO are open drain pins with 10 kX external pull-ups. TL/C/12379 – 92 FIGURE 12-24. GPIO Write Timing Note: Refer to Microprocessor Interface Timing for Read Timing. 12.2.18 RTC Symbol Parameter Conditions Min Max Units ns tRW IOR to IRQ TRI-STATE 36 tRCI MR to IRQ TRI-STATE 25 tRCL MR High Time 100 ms tVMR VCC (4.5V) to MR 10 ms ns TL/C/12379 – 93 FIGURE 12-25. IRQ Release Delay TL/C/12379 – 94 FIGURE 12-26. MR Timing 107 12.0 Electrical Characteristics (Continued) 12.2.19 Programmable Chip Select Timing Parameter Min Max Units tCE Symbol Delay from Command to Enable Chip Select 0 25 ns tCD Delay from Command to Disable Chip Select 0 25 ns TL/C/12379 – 95 FIGURE 12-27. Chip Select Timing 108 109 PC87306 SuperI/O Enhanced Sidewinder Lite Floppy Disk Controller, Keyboard Controller, Real-Time Clock, Dual UARTs, Infrared Interface, IEEE 1284 Parallel Port, and IDE Interface Physical Dimensions inches (millimeters) Plastic Quad Flatpak Order Number PC87306VUL NS Package Number VUL160A LIFE SUPPORT POLICY NATIONAL’S PRODUCTS ARE NOT AUTHORIZED FOR USE AS CRITICAL COMPONENTS IN LIFE SUPPORT DEVICES OR SYSTEMS WITHOUT THE EXPRESS WRITTEN APPROVAL OF THE PRESIDENT OF NATIONAL SEMICONDUCTOR CORPORATION. As used herein: 1. Life support devices or systems are devices or systems which, (a) are intended for surgical implant into the body, or (b) support or sustain life, and whose failure to perform, when properly used in accordance with instructions for use provided in the labeling, can be reasonably expected to result in a significant injury to the user. National Semiconductor Corporation 1111 West Bardin Road Arlington, TX 76017 Tel: 1(800) 272-9959 Fax: 1(800) 737-7018 2. A critical component is any component of a life support device or system whose failure to perform can be reasonably expected to cause the failure of the life support device or system, or to affect its safety or effectiveness. National Semiconductor Europe Fax: (a49) 0-180-530 85 86 Email: cnjwge @ tevm2.nsc.com Deutsch Tel: (a49) 0-180-530 85 85 English Tel: (a49) 0-180-532 78 32 Fran3ais Tel: (a49) 0-180-532 93 58 Italiano Tel: (a49) 0-180-534 16 80 National Semiconductor Hong Kong Ltd. 13th Floor, Straight Block, Ocean Centre, 5 Canton Rd. Tsimshatsui, Kowloon Hong Kong Tel: (852) 2737-1600 Fax: (852) 2736-9960 National Semiconductor Japan Ltd. Tel: 81-043-299-2309 Fax: 81-043-299-2408 National does not assume any responsibility for use of any circuitry described, no circuit patent licenses are implied and National reserves the right at any time without notice to change said circuitry and specifications.