Memory

Memory Devices and Interfacing (Chapter 9)
Dr. Costas Kyriacou and Dr. Konstantinos Tatas
Outline
Semiconductor Memory Basic Concepts Read Only Memory (ROM) Random Access Memory (RAM)
SRAM DRAM
Memory Interfacing
Address size expansion Word size expansion
Timing Analysis
ACOE255
Microprocessors I - Frederick University
Basic Concepts
A memory device can be viewed as a single column table.
Table index (row number) refers to the address of the memory. Table entries refer to the memory contents or data. Each table entry is referred as a memory location or as a word.
Memory Address Binary Hex 00-0000-0000 000 00-0000-0001 001 00-0000-0010 002 00-0000-0011 003 Memory Contents 10011001 00111000 11001001 00111011
Both the memory address and the memory contents are binary numbers, expressed in most cases in Hex format. The size of a memory device is specified as the number of memory locations X width or word size (in bits).
For example a 1K X 8 memory device has 1024 memory locations, with a width of 8 bits.
11-1111-1100 3FC 11-1111-1101 3FD 11-1111-1110 3FE 11-1111-1111 3FF
01101000 10111001 00110100 00011000
1024 X 8 (or 1KX8) Memory
ACOE255
Address Lines
A memory device or memory chip must have three types of lines or connections: Address, Data, and Control. Address Lines: The input lines that select a memory location within the memory device.
Decoders are used, inside the memory chip, to select a specific location The number of address pins on a memory chip specifies the number of memory locations. If a memory chip has 13 address pins (A0..A12), then it has: 213 = 23 X 210 = 8K locations. If a memory chip has 4K locations, then it should have N pins: 2N = 4K = 22 X 210 = 212 N=12 address pins (A0..A11)
A00 A01
Y00 Y01 Y02 Y03
Location 000 Location 001 Location 002 Location 003
An-2 An-1
YFC YFD YFE YFF
Location 0FC Location 0FD Location 0FE Location 0FF
ACOE255
Data Lines
Data Connections: All memory devices have a set of data output pins (for ROM devices), or input/output pins (for RAM devices).
Most RAM chips have common bi-directional I/O connections. Most memory devices have 1, 8 or 16 data lines.
Data Input Lines (DI0..DIn-1) k- address lines (A0..A m-1 )

Read (RD) Write (WR) Chip Select (CS)
2m words n-bits per word
k- address lines (A0..A m-1 )

Read/Write (R/W) Chip Select (CS)
k- address lines (A0..A m-1 )

Output Enable (OE) Chip Select (CS)
Data Output Lines (DO0..DOn-1) (2m X n) RAM with separate I/P and O/P Data lines
Data Input/Output Lines (D 0..D n-1) (2m X n) RAM with common I/P and O/P Data lines
Data Output Lines (D0..D n-1) (2m X n) ROM with only O/P Data lines
ACOE255
Control Lines
Enable Connections:
All memory devices have at least one Chip Select (CS) or Chip Enable (CE) input, used to select or enable the memory device. If a device is not selected or enabled then no data can be read from, or written into it. The CS or CE input is usually controlled by the microprocessor through the higher address lines via an address decoding circuit. Control Connections: RAM chips have two control input signals that specify the type of memory operation: the Read (RD) and the Write (WR) signals. Some RAM chips have a common Read/ Write (R/W) signal. ROM chips can perform only memory read operations, thus there is no need for a Write (WR) signal. In most real ROM devices the Read signal is called the Output Enable (OE) signal.
ACOE255 Microprocessors I - Frederick University 6
Memory Read Operations

A memory read operation is carried out in the following steps:
The processor loads on the Address bus the address of the memory location to be read (Step 1).
Some of the address lines select the memory devices that owns the memory location to be read (Step 1a), while the rest point to the required memory location within the memory device.
The processor activates the Read (RD) signal (Step 2).
The selected memory device loads on the data bus the content of the memory location specified by the address bus (Step 3).
The processor reads the data from the data bus, and resets the RD signal (Step 4).
Clock Address Bus Chip Enable Read (RD) Data Bus Step 1a Step 1 Invalid Data Step 2 Valid Data Step 3 Step 4 T1 T2 Valid Address T3
ACOE255
Memory Write Operations

A memory write operation is carried out in the following steps:
The processor loads on the Address bus the address of the memory location (Step 1).
Some of the address lines select the memory devices that owns the memory location to be written (Step 1a), while the rest point to the required memory location within the memory device.
The processor loads on the data bus the data to be written (Step 2). The processor activates the Write (WR) signal (Step 3).
The data at the data bus is stored in the memory location specified by the address bus (Step 4).
Clock Address Bus Data Bus Chip Enable Write (WR) Step 2 Step 1 Step 1a Step 3 Step 4 T1 T2 Valid Address Valid Data T3
ACOE255
Types of Semiconductor Memory Devices
Read Only Memory (ROM) A memory device that maintains its data permanently (or until the device is reprogrammed).
Non-volatile: It maintains its data even without power supply.
Random Access Memory (RAM) A memory device that can be read and written.
Volatile: It looses its data when the power supply is switched-off When the supply is switched-on it contains random data User programs that are loaded from a secondary memory (disk) Temporary data used by programs such as variables and arrays.
Used to store
Programs such as the BIOS. Data such as look tables e.g. the bit pattern of the characters in a dot matrix printer.
Used to store

A ROM device can be

1. Masked ROM (Programmed by the manufacturer) 2. Programmable ROM (can be program-erased-reprogrammed many times
A RAM device can be 1. Static 2. dynamic
ACOE255
A Read Only Memory Example

Implementation of an 8X4 ROM using (a) a decoder and OR-gates and (b) a decoder and diodes. +5V
3/8 DEC. Y0 Y1 A0 Y2 A1 Y3 Y4 A2 Y5 Y6 E Y7
3/8 DEC. Y0 Y1 A0 Y2 A1 Y3 Y4 A2 Y5 Y6 E Y7
A0 A1 A2 CS
A0 A1 A2 CS
OE
OE
D3
D2
D1
D0
D3
D2
D1
D0
Address 000 001 010 011 100 101 110 111 Data
ACOE255
0011 0010 0100 0011 1010 0000 0101 1000

Microprocessors I - Frederick University 10
A Programmable Read Only Memory Example

Implementation of an 8X4 ROM using a decoder and fused links.
+5V 3/8 DEC. Y0 Y1 A0 Y2 A1 Y3 Y4 A2 Y5 Y6 E Y7
A0 A1 A2 CS
OE
D3
ACOE255
D2
D1
D0
11
Types of semiconductor memory devices: EPROM

EPROM is a type of ROM that can be erased and re-programmed. There are two types of EPROMs: the ultra-violet (UV-EPROMs) and the electrically erasable (EEPROMs) often called the flash memory.
UV-EPROMs are erased by inserting the device in ultra violet light and programmed using a special EPROM programmer. UV-EPROMs need to be removed from the PCB in order to erased and programmed. The most common family of EPROMs is the 27XXX series, or the CMOS 27CXXX where XXX indicates the memory capacity in Kbits. Some members of this family are the following: 2716/27C16 2764/27C64 27256/27C256 (2Kx8) (8Kx8) (32Kx8) 2732/27C16 27128/27C128 27512/27C512 (4Kx8) (16Kx8) (64Kx8)
27010/27C010
27020/27C020 27040/27C040
ACOE255
(128Kx8)
(256Kx8) (512Kx8)
271024/27C1024 (64Kx16)
272048/27C2048 (128Kx16) 274096/27C4096 (256Kx16)
12
RAM Cells
Static RAM (SRAM): The basic element of a static RAM cell is the D-Latch. Data remains stored in the cell until it is intentionally modified. SRAM is fast (Access time: 1ns). SRAM needs more space on the semiconductor chip than DRAM.
SRAM more expensive than DRAM SRAM needs more space than DRAM
Dynamic RAM (DRAM): DRAM stores data in the form of electric charges in capacitors. Charges leak out, thus need to refresh data every few ms. DRAM is slow (Access time: 60ns). DRAM needs less space on the semiconductor chip than SRAM.
DRAM less expensive than SRAM DRAM needs less space than SRAM
SRAM consumes power only when accessed. DRAM needs to be refreshed DRAM is used as the main memory SRAM is used as a Cache
Bit Select
Bit Select
Data In
Data Out
Data In
Write
Data Out
En
RAM Cell
DRAM Cell
ACOE255
13
Types of semiconductor memory devices: Static RAM

Static RAM (also called SRAM)devices retain their data for as long as the DC power is applied.
The most common family of SRAM are the 61XXX, 62XXX or the CMOS 62CXXX series, where XXX indicates the memory capacity in Kbits. Some members of this family are the following: 6116/6216 61256/62256 (2Kx8) (32Kx8) 6164/6264 (8Kx8)
611024/621024 (128Kx8)
These series of SRAM devices are pin compatible with the 27XXX series of EPROMs,
with the difference that the WR signal is replaced by the programming voltage pin (Vpp) on the EPROM. This allows a single socket on the PCB hold either a SRAM, during system development, or an EPROM, after the operation of the program is verified to be the expected one. Static RAM is fast with access times much less than 100ns. SRAM chips with access times less than 10ns are often used as cache memory in computers.
DYNAMIC RAM CELL ARRAY

Asynchronous DRAM This is the basic form, from which all others are derived. An asynchronous DRAM chip has power connections, some number of address inputs (typically 12), and a few (typically 1 or 4) bidirectional data lines. There are four active low control signals: /RAS, the Row Address Strobe. The address inputs are captured on the falling edge of /RAS, and select a row to open. The row is held open as long as /RAS is low. /CAS, the Column Address Strobe. The address inputs are captured on the falling edge of /CAS, and select a column from the currently open row to read or write. /WE, Write Enable. This signal determines whether a given falling edge of /CAS is a read (if high) or write (if low). If low, the data inputs are also captured on the falling edge of /CAS. /OE, Output Enable. This is an additional signal that controls output to the data I/O pins. The data pins are driven by the DRAM chip if /RAS and /CAS are low, and /WE is high, and /OE is low. In many applications, /OE can be permanently connected low (output always enabled), but it can be useful when connecting multiple memory chips in parallel.
ACOE255
15
DRAM BLOCK DIAGRAM

Samsung Electronics
ACOE255
16
DYNAMIC RAM
DRAM requires refreshing every 2 to 4 ms . Refreshing occurs automatically during a read or write. Internal circuitry takes care of refreshing cells that are not accessed over this interval.
For a 256K X 1 DRAM with 256 rows, a refresh must occur every 15.6us (4ms/256). For the 8086, a read or write occurs every 800ns . This allows 19 memory reads/writes per refresh or 5% of the time.
DRAM technologies
EDO DRAM SDRAM DRDRAM DDR DRAM
Soft errors occur on DRAMs which often require ERROR DETECTION and/or ERROR CORRECTION A DRAM CONTROLLER is required for using DRAM
ACOE255
17
EXTENDED DATA OUTPUT (EDO) DRAM

Any memory access in an EDO memory (including a refresh) stores the 256 bits in a set of latches. Any subsequent access to bytes in this set are immediately available (without the decode time and therefore wait states). This works well because of the principle of spatial locality, and improves system performance by 15 to 25 % !
ACOE255
18
SYNCHRONOUS DYNAMIC RAM

In a synchronous DRAM, the control signals are synchronized with the system bus clock and therefore with the microprocessor It allows pipelined read/write operations
ACOE255
19
Double Data Rate (DDR) DRAM

An SDRAM type of memory where data are transferred on both the rising and the falling clock edge, effectively doubling the transfer rate without increasing the clock frequency DDR-200 means a transfer rate of 200 million transfers per second, at a clock rate of 100 MHz DDR1 upto 400 MHz DDR2 standard allows higher clock frequencies
ACOE255
20
Direct Rambus DRAM (DRDRAM)

A type of dual-edge SDRAM, like DDR, challenging DDR2 as the standard
ACOE255
21
ERRORS AND ERROR DETECTION AND CORRECTION

Electrical or magnetic interference inside a computer system as well as cosmic radiation can cause a single bit of DRAM to spontaneously flip to the opposite state. (soft errors) As the components on DRAM chips get smaller while operating voltages continue to fall, DRAM chips may be: affected by such radiation more frequently since lower energy particles will be able to change a memory cell's state. or since smaller cells make smaller targets individual cells may be less susceptible to such effects
A reasonable rule of thumb is to expect one bit error, per month, per gigabyte of memory Systems often use error detection and correction methods to identify and possibly correct soft errors
repetition schemes parity schemes (74AS280) cyclic redundancy checks Hamming distance based checks (74LS636)
ACOE255
22
ERROR DETECTION: PARITY

A parity bit is a bit added to a fixed number of data bits to ensure that the total number of 1s is either odd (odd parity) or even (even parity) Therefore, if the number of data bit 1s is odd in an even parity scheme, the parity bit is 1, otherwise it is 0 Likewise, if the number of data bit 1s is even in an odd parity scheme, the parity bit is 1, otherwise it is 0 The parity bit is transmitted with the data, and checked by the receiver Advantages:
Only one bit overhead Simple digital circuit implementation
Disadvantages:
Cannot correct errors, only detect them Only detects an odd number of errors
ACOE255
23
PARITY EXAMPLE
Calculate the parity bit for both even and odd parity, for the following sequence
1001 0001 1000 1000011 Assuming that the last bit is the parity bit (odd parity), determine which data transmission was successful and which unsuccessful 10001010 00111011 11011101
Design the circuit that gives the parity bit
ACOE255
24
ERROR CORRECTION: REPETITION AND MAJORITY VOTING

Data are saved (copied) in three different memory elements During a memory read, all three memories are accessed and majority voting circuitry decides the final output. Advantages: the possibility of soft errors is practically eliminated Disadvantages: Triple(!) memory space is required, and there is a performance and area overhead caused by the majority voting circuitry
Majority voter
DATA FROM MEMORY
Memory 1
Memory 2
Memory 3
DATA TO MEMORY uP
ACOE255
25
EXAMPLE
Design the majority voting circuit for one memory bit
ACOE255
26
DRAM CONTROLLER
A circuit performing address multiplexing and DRAM control signal generation
ACOE255
27
Semiconductor Memory Expansion

The size of memory devices is usually less than the memory requirements of a computer system.
In all computers, more than one memory devices are combined together to form the main memory of the system. Any computer must have at least one ROM chip and one RAM chip.
Word size memory expansion:

Most memory devices have a word size (number of data lines) of 8 or 16 bits. The word size of todays microprocessors is 32 bits (80386, 80486) or 64 bits (Pentium)
Address size memory expansion:

The size of common memory chips is usually less or in the order of 256M-byte. Most personal computers have more than 2 Gbytes of RAM. Workstations and other high throughput computers have more than 4Gbytes of RAM.
ACOE255
28
Memory Expansion on Motherboards
Memory Expansion Using 4 SIMMs on the Motherboard

Motherboard
Slot 3 SIMM Slot 4
Memory Expansion using 4 Memory Chips on a SIMM
Slot 1 SIMM
Slot 2 SIMM
Processor
ACOE255
29
Memory Address Size Expansion

More than one memory devices can be used to expand the number of memory
locations on the system. To expand the word size do the following:
Determine the number of memory chips required, by dividing the required

memory size with the size of the memory devices to be used. Connect the data lines of each memory chip in parallel on the data lines of the processor. Connect the address lines of each memory chip in parallel with the low address lines of the processor. Connect the CS lines of each memory device with the high address lines of the processor through an address decoding circuit.. Connect together all WR and all RD lines of each memory device.
Address Size Expansion: (32X4 RAM module using 8X4 RAM chips)
D0 D1 D2 D3 RAM1 D3
0 1 2
D0
RAM2 D3
0 1 2
D0
RAM3 D3
0 1 2
D0
RAM4 D3
0 1 2
D0
A0 3 A1 4 A2 5
6 7
A0 3 A1 4 A2 5
6 7
A0 3 A1 4 A2 5
6 7
A0 3 A1 4 A2 5
6 7
RD WR CS RD WR A0 A1 A2 A3 A4 A5 A6
Address Selection
RD WR CS
RD WR CS
RD WR CS
A B CS
Y0 Y1 Y2 Y3 2X4 DEC.
ACOE255
31
Memory Maps
Tables that show the addresses occupied by each memory device in a system.
In the previous example it is assumed that the processor has only 7 address line, thus it can address 128 memory locations. The size of the RAM memory module is 32 bytes, thus the module can be mapped to occupy one out of the four available memory blocks in the memory map. The memory block occupied by the memory module depends on the connection of the address selection circuit (AND gate) that enables the decoder.
A5 A6
00 - 07 RAM 1 08 - 0F RAM 2 10 - 17 RAM 3 18 - 1F RAM 4 20 - 27 RAM 1 20 - 3F Not Used 28 - 2F RAM 2 30 - 37 RAM 3 38 - 3F RAM 4 40 - 47 RAM 1 40 - 5F Not Used 40 - 5F Not Used 48 - 4F RAM 2 50 - 57 RAM 3 58 - 5F RAM 4 60 - 67 RAM 1 60 - 7F Not Used 60 - 7F Not Used 60 - 7F Not Used 68 - 6F RAM 2 70 - 77 RAM 3 78 - 7F RAM 4 40 - 5F Not Used 20 - 3F Not Used 20 - 3F Not Used 00 - 1F Not Used 00 - 1F Not Used 00 - 1F Not Used
A5 A6
A5 A6
A5 A6
ACOE255
32
Effect of the Address Selection Circuit

The memory block occupied by the memory module depends on the connection of the address selection circuit (AND gate) that enables the decoder. Two address lines are used to control the address selection circuit, thus the circuit can be configured to occupy four different areas in the address space.
Address Selection Circuit
A5 A6
A5 A6
A5 A6
A5 A6
A6 A5 A4 A3 A2 A1 A0 Mem. Map A6 A5 A4 A3 A2 A1 A0 0 0 0 0 0 0 0 00 0 0 0 0 0 0 0 RAM1 0 0 0 0 1 1 1 07 0 0 1 1 1 1 1 0 0 0 1 0 0 0 08 0 1 0 0 0 0 0 RAM2 0 0 0 1 1 1 1 0F 0 1 0 0 1 1 1 0 0 1 0 0 0 0 10 0 0 1 0 1 1 1 17 0 0 1 1 0 0 0 18 0 0 1 1 1 1 1 1F RAM3 RAM4
Mem. Map A6 A5 A4 A3 A2 A1 A0 Mem. Map A6 A5 A4 A3 A2 A1 A0 Mem. Map 0 0 0 0 0 0 0 00 Not 0 0 0 0 0 0 0 00 Not 00 Not Used Used 0 1 1 1 1 1 1 3F 1 0 1 1 1 1 1 5F Used 1F 20 RAM1 27 RAM2 RAM3 RAM4 1 0 0 0 0 0 0 40 RAM1 1 0 0 0 1 1 1 47 1 0 0 1 0 0 0 48 1 0 0 1 1 1 1 4F 1 0 1 0 0 0 0 50 1 0 1 0 1 1 1 57 RAM2 RAM3 1 1 0 0 0 0 0 60 RAM1 1 1 0 0 1 1 1 67 1 1 0 1 0 0 0 68 1 1 0 1 1 1 1 6F 1 1 1 0 0 0 0 70 1 1 1 0 1 1 1 77 RAM2 RAM3
0 1 0 1 0 0 0 28 0 1 0 1 1 1 1 2F 0 1 1 0 0 0 0 30 0 1 1 0 1 1 1 37 0 1 1 1 0 0 0 38 0 1 1 1 1 1 1 3F
0 1 0 0 0 0 0 20 Not 1 1 1 1 1 1 1 7F Used
1 0 1 1 0 0 0 58 RAM4 1 0 1 1 1 1 1 5F 1 1 0 0 0 0 0 60 Not Used 1 1 1 1 1 1 1 7F
1 1 1 1 0 0 0 78 RAM4 1 1 1 1 1 1 1 7F
1 0 0 0 0 0 0 40 Not 1 1 1 1 1 1 1 7F Used
ACOE255
33
Example: (32X4 RAM module using 8X4 RAM chips - Assume an 8-address line processor)
D3
D3 A0
D0
D3 A0 A2
D0
D3 A0 A2
D0
D3 A0 A2
D0
8x4 RAM 1
8x4 RAM 2
8x4 RAM 3
8x4 RAM 4
D0 A2 RD WR CS RD WR RD WR CS RD WR CS RD WR CS
A0 A2 A3 A4 A5 A6 A7 CS 2X4 DEC. A B Y0 Y1 Y2 Y3
ACOE255
34
Memory Map for previous example.

There are three address lines connected on the address selection circuit. Thus there can be eight different memory map configurations. Three possible memory map configurations are shown below.
A5 A6 A7 A5 A6 A7 A5 A6 A7
A7 A6 A5 A4 A3 A2 A1 A0 Mem. Map 0 0 0 0 0 0 0 0 00 0 0 0 0 0 1 1 1 07 0 0 0 0 1 0 0 0 08 0 0 0 0 1 1 1 1 0F 0 0 0 1 0 0 0 0 10 0 0 0 1 0 1 1 1 17 0 0 0 1 1 0 0 0 18 0 0 0 1 1 1 1 1 1F 0 0 1 0 0 0 0 0 20 RAM1 RAM2 RAM3 RAM4
A7 A6 A5 A4 A3 A2 A1 A0 Mem. Map 0 0 0 0 0 0 0 0 00 Not Used 1 0 0 1 1 1 1 1 9F 1 0 1 0 0 0 0 0 A0 1 0 1 0 0 1 1 1 A7 1 0 1 0 1 0 0 0 A8 1 0 1 0 1 1 1 1 AF 1 0 1 1 0 0 0 0 B0 1 0 1 1 0 1 1 1 B7 1 0 1 1 1 0 0 0 B8 1 0 1 1 1 1 1 1 BF 1 1 0 0 0 0 0 0 C0 RAM1 RAM2 RAM3 RAM4
A7 A6 A5 A4 A3 A2 A1 A0 Mem. Map 0 0 0 0 0 0 0 0 00 Not Used 1 1 0 1 1 1 1 1 DF 1 1 1 0 0 0 0 0 E0 1 1 1 0 0 1 1 1 E7 1 1 1 0 1 0 0 0 E8 1 1 1 0 1 1 1 1 EF 1 1 1 1 0 0 0 0 F0 1 1 1 1 0 1 1 1 F7 1 1 1 1 1 0 0 0 F8 1 1 1 1 1 1 1 1 FF RAM1 RAM2 RAM3 RAM4
Not Used 1 1 1 1 1 1 1 1 FF
Not Used 1 1 1 1 1 1 1 1 FF
ACOE255
35
Design Example:
Design an 8KX8 RAM module using 2KX8 RAM chips. The module should be connected on an 8-bit processor with a 16-bit address bus, and occupy the address range starting from the address A000. Show the circuit and the memory map.
Number of memory devices needed = 8K/2K =4 Decoder needed = 2X4 Number of address lines on each 2KX8 memory chip = 11 2m = 2K = 21 x 210 = 211 (A0..A10) Decoder needed = 2X4 2 address lines are needed for the decoder. (A11..A12) Number of address lines needed for the address selection circuit = 16 - 11 - 2 = 3 (A13, A14 A15)
Starting Address = A000 = 1010-0000-0000-0000 ==> A15 = 1, A14 = 0 and A13 = 1 A13 A14 Address Selection Circuit A15
A15 0 1 1 1 1 1 1 1 1 1 1 1
A14 0 0 0 0 0 0 0 0 0 0 1 1
A13 0 0 1 1 1 1 1 1 1 1 0 1
A12 0 1 0 0 0 0 1 1 1 1 0 1
A11 0 1 0 0 1 1 0 0 1 1 0 1
A10 0 1 0 1 0 1 0 1 0 1 0 1
A0 0 1 0 1 0 1 0 1 0 1 0 1
Mem. Map 0000 Not Used 9FFF A000 A7FF A800 AFFF B000 B7FF B800 BFFF C000 FFFF RAM1 RAM2 RAM3 RAM4 Not Used
ACOE255
36
Circuit Diagram
D7 D0 D7 D0 D7 D0 D7 D0 D7 2Kx8 RAM D0 A0 2Kx8 RAM A0 2Kx8 RAM A0 2Kx8 RAM A0
A10 RD WR CS RD WR A0
A10 RD WR CS
A10 RD WR CS
A10 RD WR CS
A13 A14 A15 A15
2X4 DEC. A11 A Y0 A12 B Y1 Y2 CS Y3
ACOE255
37
Address Decoding
The physical address space, or memory map, of a microprocessor refers to the range of addresses of memory location that can accessed by the microprocessor. The size of the address space depends on the number of address lines of the microprocessor.
At least two memory devices are required in a microprocessor system: one for the ROM and one for the RAM. In an 8088/8086 the high addresses in the memory map should always be occupied by a ROM, while the low addresses in the memory map should always be occupied by a RAM.
Address decoding is required in order to enable the connection of more than one memory devices on the microprocessor. Each device will occupy a unique area in the memory map.
A memory system is not fully decoded if some of the address lines are not used by the address decoding circuit or memory. In this case a memory device will occupy more than one sections in the memory map. This is referred as memory mirroring or memory imaging.
ACOE255
38
Address Decoding Circuits

A number of types of address decoding circuits can be used in a microprocessor system.The main issues related to the selection of an address decoding circuit are:
The time delays introduced by the address decoding circuit. This delays are added to the access time of the memory devices, and might yield to the insertion of wait states. The number of chips required by the address decoding circuit, as well as the complexity of the circuit (number of tracks required on the board.
An address decoding circuit must ensure that an address section is occupied by only one memory device. If two or more devices occupy the same addresses then bus contention will occur. Bus contention occurs if two of more devices drive the bus at the same time. Bus contention can be either static or dynamic. Static bus contention occurs when two or more devices drive a bus for a prolonged time period. This might damage some of the components of the system. Static bus contention might be caused by improper address decoding design, or by other faults in the system such as a short circuit of the CS of a device to the ground. Dynamic bus contention occurs when two or more devices drive a bus for a short period of time. This might change the logic levels on the bus and cause system malfunctions.Dynamic bus contention might be caused by improper address decoding design, or by wrong memory timing analysis.
Address decoding circuits using Only NAND gates

A single NAND gate is used to decode each memory device. The inputs of the NAND gate can be connected on the address lines either directly, or through inverters, according to the required memory map. This decoding circuit has the advantage that it adds a short time delay in the memory path. (td = 2 X gate delay <10ns) The disadvantage of this circuit is that too many gates (NAND and NOT) are needed for memory systems that have a few memory chips. This increases the cost of the system, adds to the complexity of the PCB board (too many chips and lines) and might create fan-out problems.
A0 A14 MEM 1 A 0 A14 CS
MEM 2
CS
A19 A18 A17 A16 A15 A14 0 0 0 1 1 1 1 0 0 0 0 0 0 1 1 0 0 0 1 0 1 0 0 0
A0 0 1 0 1
Memory Map 28000H 2FFFFH 40000H 47FFFH MEM1 MEM2
A15 A16 A17 A18 A19

ACOE255
40
Address decoding circuits using line decoders and a NAND gate

One or more line decoders such as the 74LS139 (2 x 4 decoder) or the 74LS138 (3 x 8 decoder) are used to decode(enable) one out of a number of memory device. The CS inputs of the decoders are enabled by a NAND decoding circuit, according to the required memory map. This decoding circuit has the disadvantage that it adds at least three gate delays in the memory path. The advantage of this circuit is that less gates (NAND, NOT and decoders) are needed for memory systems that have a number memory chips.
A0 A14 Mem 1 Mem 2
CS CS
Mem 8
CS
A19 A18 A17 A16 A15 A14 0 0 0 0 0 1 1 1 1 1 1 0 0 0 0 1 1 0 0 0 0 1 1 0 0 1 1 1 1 0 1 0 1 0 1
A0 0 1 0 1 0 1
Memory Map 40000H 47FFFH 48000H 4FFFFH 78000H 7FFFFH MEM1 MEM2
3X8 Dec. A15 A16 A17 A18 A19 A B C CS Y0 Y1
Y7
MEM8
ACOE255
41
Address decoding circuits using PLDs

Programmable Logic Devices (PLD) such as the Programmable Logic Array (PLA), Programmable Array Logic (PAL) or Gated Array Logic (GAL) have replaced the PROM or EPROM address decoders. These devices are easily programmed using programs such as the PALASM and EPROM/PLD programmers. This decoding circuit has the advantages of the PROM address decoding circuits, with very low delay added in the memory path. Furthermore these devices have the option of using a copy-bit, during programming, that disables reading the content of the device, thus copy protect the design.
PAL 16L8 A15 A16 A17 A18 A19

1 2 3 4 5 6
PAL Programming Mem 1

19 18
A19 A18 A17 A16 A15 A14 0 0 0 0 0 0 1 1 1 1 1 1 0 0 0 0 1 1 0 0 0 0 1 1 0 0 1 1 1 1 0 1 0 1 0 1
A 0 Memory Map 0 40000 1 47FFF 0 48000 1 4FFFF Mem1 Mem2
CHIP DECODER1 PAL16L8 ;Pin Assignment ;1 2 3 4 5 6 . . . 10 A15 A16 A17 A18 A19 NC. . . Gnd ;11 12 13 14 15 . . . . 18 19 20 NC Q8 Q7 Q6 Q5 . . . Q2 Q1 Vcc EQUATIONS /Q1 = /A15 * /A16 * /A17 * A18 * /A19 /Q2 = A15 * /A16 * /A17 * A18 * /A19 /Q8 = A15 * A16 * A17 * A18 * /A19
I1 Q1 I2 Q2 I3 I4 I5 I6 Q8 I10
CS CS
12 20 10
Mem 8
CS
Vcc Gnd
0 78000 Mem8 1 7FFFF
ACOE255
42
Address decoding circuits using comparators

Comparators, such as the 74LS85 (4-bit) or the 74LS688 (8-bit) can be used as address decoders. One set of the input lines are connected on the address bus and the other is usually connected directly to logic 0 or logic 1 according to the required memory map. This decoding circuit has the advantages of the PLD address decoding circuits, i.e. only one chip in needed and the very low delay added in the memory path. Furthermore the second set of input lines can be connected to dip switches or an O/P port so that the memory map can be easily modified. The disadvantage of this address decoding circuit is that it can enable only one device, unless if it is combined with other decoding circuits, such as line decoders.
A0 A15 MEM 1
CS
A0 A15
MEM 2
CS
X=Y X1 X2 X3 X4 +5V Gnd A16 A17 A18 A19

ACOE255
X=Y X1 X2 X3 X4 Y1 Y2 Y3 Y4
A19 A18 A17 A16 A15 1 1 1 1 0 0 1 1 0 0 0 0 1 1 0 0 0 1 0 1
A0 0 1 0 1
Memory Map 90000 9FFFF C0000 CFFFF MEM1 MEM2
Y1 Y2 Y3 Y4
43
Address decoding example

Show how a 128Kbyte RAM module can be connected on an 8088 system using 62256 SRAM chips, occupying the address range starting from the address C0000H. Use the following address decoding circuits:
1 2 3 4. Nand decoding circuits Line decoders PLD decoding circuit Comparator decoding circuit
Solution: 62256 SRAM chips: 256/8 =32 32KX8 Number of chips needed: 128K/32K = 4 Number of address lines: 32K = 25K = 25 * 210 = 215 15 address lines (A0 .. A14)
A19 A18 A17 A16 A15 A14 A13 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 0 0 1 1 0 0 1 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1
A0 0 1 0 1 0 1 0 1
Memory Map C0000 C7FFF C8000 D0000 D7FFF D8000 RAM1
RAM2 CFFFF RAM3
RAM4 DFFFF
ACOE255
44
Answer: Using NAND gates
62256 A0 D0 D0 D7 RD WR A0 A15 A16 A17 A18 A19 A14 D7
62256 A0 D0 A14 D7
62256 A0 D0 A14 D7
62256 A0 D0 A14 D7
RD WR CS
RD WR CS
RD WR CS
RD WR CS
8088 System
A19 IO/M'
ACOE255
45
Answer: Using a line decoder and a NAND gate
62256 A0 D0 D0 D7 RD WR A0 A17 A18 A19 A14 D7
62256 A0 D0 A14 D7
62256 A0 D0 A14 D7
62256 A0 D0 A14 D7
RD WR CS
RD WR CS
RD WR CS
RD WR CS
8088 System
A15 LS139 A Y0 A16 Y

B E
1
A19 IO/M'
Y2 Y3
ACOE255
46
Answer: Using a PLD decoding circuit
62256 A0 D0 D0 D7 A14 D7
62256 A0 D0 A14 D7
62256 A0 D0 A14 D7
62256 A0 D0 A14 D7
RD WR CS
RD WR CS
RD WR CS
RD WR CS
8088 System
RD WR A0
A19 IO/M'
A15 A16 A17 A18 A19 IO/M'
1 2 3 4 5 6
I1 I2 I3 I4 I5 I6 I10
Q1 Q2 Q3 Q4
19 18 17 16
PAL Programming: PAL16L8

;Pins 1 2 3 4 5 6 . . . 10 11 . . .16 17 18 19 20 A15 A16 A17 A18 A19 MEM. Gnd NC . . . Q4 Q3 Q2 Q1 Vcc EQUATIONS /Q1 = /A15 * /A16 * /A17 * A18 * A19 * /MEM /Q2 = A15 * /A16 * /A17 * A18 * A19 * /MEM /Q3 = /A15 * A16 * /A17 * A18 * A19 * /MEM /Q4 = A15 * A16 * /A17 * A18 * A19 * /MEM
10
Q8 Gnd Vcc
12 20
PAL 16L8
ACOE255
47
Answer: Using comparators
62256 A0 D0 D0 D7 RD WR A0 X A15 A16 A17 A18 A19 Gnd +5V A14 D7
62256 A0 D0 A14 D7
62256 A0 D0 A14 D7
62256 A0 D0 A14 D7
RD WR CS
RD WR CS
RD WR CS
RD WR CS
8088 System
X=Y Y X
X=Y Y X
X=Y Y X
X=Y Y
A19 IO/M'
ACOE255
48
Homework:
Show how a 32Kbyte ROM module can be connected on an 8088 system using 2764 EPROM chips, occupying the address range starting from the address E0000H. Use the following address decoding circuits:
1 2 3. 4. 5 Nand decoding circuits A line decoder and a Nand gate PLD decoding circuit Comparators only Line decoder and a comparator
Solution: Size of 2764 EPROM chips:
A19 A18 A17 A16 A15 A14 A13 A12 A11
A0
Memory Map
Number of chips needed:
Number of address lines:
ACOE255
49
Answer: Using Nand Gates only
2764 A0 D0 D0 D7 RD WR A0 A 12
OE
2764 A0 D0 A 12
OE
2764 A0 D0 A 12
OE
2764 A0 D0 A 12
OE
D7
CS
D7
CS
D7
CS
D7
CS
8088 System
A19 IO/M'
ACOE255
50
Answer: Using a line decoder and a Nand gate
2764 A0 D0 D0 D7 RD WR A0 A 12
OE
2764 A0 D0 A 12
OE
2764 A0 D0 A 12
OE
2764 A0 D0 A 12
OE
D7
CS
D7
CS
D7
CS
D7
CS
8088 System
A19 IO/M'
ACOE255
51
Answer: Using a PLD decoding circuit
2764 A0 D0 D0 D7 RD WR A0 A 12
OE
2764 A0 D0 A 12
OE
2764 A0 D0 A 12
OE
2764 A0 D0 A 12
OE
D7
CS
D7
CS
D7
CS
D7
CS
8088 System
A19 IO/M'
ACOE255
52
Answer: Using comparators only
2764 A0 D0 D0 D7 RD WR A0 A 12
OE
2764 A0 D0 A 12
OE
2764 A0 D0 A 12
OE
2764 A0 D0 A 12
OE
D7
CS
D7
CS
D7
CS
D7
CS
8088 System
A19 IO/M'
ACOE255
53
Answer: Using a line decoder and a comparator
2764 A0 D0 D0 D7 RD WR A0 A 12
OE
2764 A0 D0 A 12
OE
2764 A0 D0 A 12
OE
2764 A0 D0 A 12
OE
D7
CS
D7
CS
D7
CS
D7
CS
8088 System
A19 IO/M'
ACOE255
54
16-bit Memory Interfacing (8086, 80286, 80186, 80386SX)

The 8086 differs from the 8088 in three ways:
The data bus is 16 bits wide instead of 8 bits as on the 8088 The IO/M signal on the8088 is replaced by the M/IO on the 8086 There is a BHE (Bus High Enable) signal to enable the upper data bus lines (D8..D15). The address line A0 behaves as the BLE (Bus Low Enable) signal.
The memory is separated into the High Bank (odd addresses) and the Low Bank (even addresses). The 8086 microprocessor can access either the low bank (D0..D7), or only the high bank (D8..D15), or both banks (D0..D15). The is a need only for separate Bank Write Strobes. When the processor reads from the memory, it always reads both banks, and selects the necessary bank internally.
(BHE') FFFFF FFFFD FFFFB FFFFE FFFFC FFFFA (BLE'/A0) BHE' BLE'(A0) 0 0 1 1 0 1 0 1 Function Example
00005 00003 00001 High Bank (D15..D8) (Odd Addresses)
00004 00002 00000 Low Bank (D7..D0) (Even Addresses)
Both banks enabled (16 bit) MOV [1000H],AX High bank enabled (8 bit) MOV [1001H],AL Low bank enabled (8 bit) MOV [1000H],AL No banks enabled ---------------
ACOE255
55
16-bit Memory Interfacing using separate bank decoders

The first decoder (left side) is enabled when A0 is zero, thus it is enable with even addresses. Thus the data lines of the memory devices decoded by this decoder must be connected on the processors data lines D0..D7. The second decoder (right side) is enabled when BHE is zero, thus it is enable with odd addresses. Thus the data lines of the memory devices decoded by this decoder must be connected on the processors data lines D8..D15.
D15 D8 D7 D0 RD WR A0
A1
62256
A0 D0
A1
62256
A0 D0
A1
62256
A0 D0
A1
62256
A0 D0
A15 A14
D7
RD WR CS
A15
A14
D7
A15 A14
D7
RD WR CS
RD WR CS
A15
A14
D7
RD WR CS
8086 System
Y0 A19 A B
Y1
Y2
Y3 1E
Y0 A
Y1 B
Y2
Y3 1E
LS139
BHE' IO'/M A16 A17 A18 A19 A0 A16 A17
LS139
A18 A19
ACOE255
56
16-bit Memory Interfacing using separate bank write signals

With this method the decoder always enables both banks. On a memory read operation, the data from both banks is loaded on the data bus. The microprocessor selects internally the appropriate bank, according to the instruction being executed. On a memory write operation, only the WR signal of the appropriate bank is enabled, thus data is copied only in the appropriate memory chip.
D15 D8 D7 A15 D0 RD BHE' WR A0 A1 Y0 A A19 IO'/M A16 A17 A18 A19 B Y1 Y2 Y3 1E A1
62256
A0 D0
A1
62256
A0 D0
A1
62256
A0 D0
A1
62256
A0 D0
A14
D7
RD WR CS
A15
A14
D7
RD WR CS
A15
A14
D7
RD WR CS
A15
A14
D7
RD WR CS
8086 System
LS139
ACOE255
57
32-bit Memory Interfacing using separate bank write signals

The 80386 microprocessor has four bank enable signals to select one out of 4 memory banks. The address lines A0 and A1 are not available. On a memory read operation, the data from all banks is loaded on the data bus. The microprocessor selects internally the appropriate bank, according to the instruction being executed. On a memory write operation, only the WR signal of the appropriate bank is enabled, thus data is copied only in the appropriate memory chip.
D0 A2
62256
A0 D0
D0
A2
62256
A0 D0
D8
A2
62256
A0 D0
D16
A2
62256
A0 D0
D24
A16
A14
D7 D7
80386 Processor
D31 RD WR BE0' BE1' BE2' BE3' A2
RD WR CS
A16
A14
D7 D15
A16
A14
D7 D23
RD WR CS
RD WR CS
A16
A14
D7 D31
RD WR CS
Y0 A B
Y1
Y2
Y3 1E
LS139
A17 A18 A31 IO'/M A19 A31
ACOE255
58
Semiconductor Memory Devices:Timing Analysis

An important parameter of memory devices is the Memory Access Time(tacc). This is the time measured from the moment that a stable address appears on the address lines of the device, until the appearance of valid data at the data lines of the device.Another important parameter is the Chip Select to Output Delay (tcd). If the time allowed by the microprocessor is less than these parameters then the microprocessor will read the data bus before the memory places the data on the data bus, thus the microprocessor will read the wrong data. The time needed by the memory device to deactivate the output data buffers is also important. The parameters related to this delay are the Chip Diselect to Output Float (tdf) and the Address to Output Hold (toh) time. The output buffers must be placed in high impedance before the microprocessor starts the next memory cycle.
Address tacc CS tcd High Z Data tdf tdh
Symbol tacc tcd tdf tdh
Parameter Address to Output Delay Chip Select to Output Delay Chip Diselect to Output Float Address to Output Hold
Limit Min. Typ. Max. 250 450 120 0 0 100 100
Unit ns ns ns ns
Timing diagram of the 2764 EPROM
AC characteristics of the 2764 EPROM
ACOE255
59
Example
You are asked to interface 8Kx8 bit ROM chips with the following data to a 8088 microprocessor: Chip-select to output delay: 70ns(min) 120ns(typ) 180ns(max) Address to output delay: 230ns(min) 340ns(typ) 450ns(max) Chip deselect to output float: 80ns(typ) 100ns(max) Address to output hold: 80ns(typ) 100ns(max) Assume that buffers have a delay of 20 ns, and latches a delay of 35 ns. The delay of the wires is 20 ns A. Calculate the number of wait states (if needed) B. Draw the corresponding memory read operation timing diagram C. Calculate the number of chips required to create a 32Kbyte ROM D. Specify the memory map starting from address F8000H E. Draw the decoding circuit using NAND gates only F. Draw the decoding circuit using a decoder and NAND gates
ACOE255
60

Memory

Hochgeladen von

Dokumentinformationen

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

Memory

Hochgeladen von

Copyright:

Verfügbare Formate

Memory Devices and Interfacing (Chapter 9)

Dr. Costas Kyriacou and Dr. Konstantinos Tatas

Microprocessors I - Frederick University

11-1111-1100 3FC 11-1111-1101 3FD 11-1111-1110 3FE 11-1111-1111 3FF

01101000 10111001 00110100 00011000

1024 X 8 (or 1KX8) Memory

Microprocessors I - Frederick University

Y00 Y01 Y02 Y03

Location 000 Location 001 Location 002 Location 003

YFC YFD YFE YFF

Location 0FC Location 0FD Location 0FE Location 0FF

Microprocessors I - Frederick University

Data Input Lines (DI0..DIn-1) k- address lines (A0..A m-1 )

2m words n-bits per word

k- address lines (A0..A m-1 )

2m words n-bits per word

k- address lines (A0..A m-1 )

2m words n-bits per word

Microprocessors I - Frederick University

Memory Read Operations

Microprocessors I - Frederick University

Memory Write Operations

Microprocessors I - Frederick University

Types of Semiconductor Memory Devices

A ROM device can be

A RAM device can be 1. Static 2. dynamic

Microprocessors I - Frederick University

A Read Only Memory Example

0011 0010 0100 0011 1010 0000 0101 1000

A Programmable Read Only Memory Example

Microprocessors I - Frederick University

Types of semiconductor memory devices: EPROM

Microprocessors I - Frederick University

Microprocessors I - Frederick University

Types of semiconductor memory devices: Static RAM

DYNAMIC RAM CELL ARRAY

Microprocessors I - Frederick University

DRAM BLOCK DIAGRAM

Microprocessors I - Frederick University

Microprocessors I - Frederick University

EXTENDED DATA OUTPUT (EDO) DRAM

Microprocessors I - Frederick University

SYNCHRONOUS DYNAMIC RAM

Microprocessors I - Frederick University

Double Data Rate (DDR) DRAM

Microprocessors I - Frederick University

Direct Rambus DRAM (DRDRAM)

Microprocessors I - Frederick University

ERRORS AND ERROR DETECTION AND CORRECTION

Microprocessors I - Frederick University

ERROR DETECTION: PARITY

Microprocessors I - Frederick University

Design the circuit that gives the parity bit

Microprocessors I - Frederick University

ERROR CORRECTION: REPETITION AND MAJORITY VOTING

DATA FROM MEMORY

Microprocessors I - Frederick University

Microprocessors I - Frederick University

Microprocessors I - Frederick University

Semiconductor Memory Expansion

Word size memory expansion:

Address size memory expansion:

Microprocessors I - Frederick University

Memory Expansion on Motherboards